JLARC Report - Data Center and Manufacturing Incentives (June 17, 2019)
Summary
A report to the Governor and the General Assembly of Virginia from the Joint Legislative Audit and Review Commission (JLARC), titled Data Center and Manufacturing Incentives, dated June 17, 2019, in the Economic Development Incentives Evaluation Series. Its summary states that Virginia provides 11 incentives to promote data center and manufacturing growth, and that spending on them totaled $559 million between FY10 and FY17. The chapters cover data center incentives, semiconductor manufacturing incentives, manufacturers single sales apportionment and environmental incentives for manufacturers. The report closes with agency responses, including a letter from Stephen Moret, President & CEO, to Mr. Greer.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
Cite this document
“JLARC Report - Data Center and Manufacturing Incentives (June 17, 2019).”
Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/2020-rd148-7d9bd3e9f9/
, SHA-256 ef9b89583ee3ea40e0efafdb3e9c47b7ba632ff7386fd34f0b435e68324884d3.Full text
Commonwealth of Virginia
June 17, 2019
Report to the Governor and the General Assembly of Virginia
Data Center and Manufacturing Incentives
Economic Development Incentives Evaluation Series
JOINT LEGISLATIVE AUDIT
AND REVIEW COMMISSION
Joint Legislative Audit and Review Commission
Senator Thomas K. Norment, Jr., Chair
Delegate R. Steven Landes, Vice-Chair
Delegate Terry Austin
Delegate Betsy Carr
Delegate M. Kirkland Cox
Senator Emmett W. Hanger, Jr.
Delegate Charniele L. Herring
Senator Janet D. Howell
Delegate S. Chris Jones
Senator Ryan T. McDougle
Delegate Robert D. Orrock, Sr.
Delegate Kenneth R. Plum
Senator Frank M. Ruff, Jr.
Delegate Christopher P. Stolle
Martha S. Mavredes, Auditor of Public Accounts
JLARC staff
Hal E. Greer, Director
Kimberly Sarte, Associate Director for Ongoing Oversight and Fiscal Analysis
Ellen Miller, Chief Economic Development and Quantitative Analyst
Information graphics: Nathan Skreslet
Economic analysis
Terance J. Rephann, Regional Economist
Weldon Cooper Center for Public Service
JLARC Report 518
©2019 Joint Legislative Audit and Review Commission
http://jlarc.virginia.gov
Contents
Summary i
Recommendations and Options v
Data Center and Manufacturing Incentives 1
1. Data Center Incentives 3
2. Semiconductor Manufacturing Incentives 23
3. Manufacturers Single Sales Apportionment 39
4. Environmental Incentives for Manufacturers 51
Appendixes
A. Study mandate 65
B. Research methods and activities 67
C. State data center incentives 79
D. Corporate tax apportionment formulas by state 83
E. Semiconductor incentives by state 86
F. Semiconductor manufacturing is concentrated in six states 87
G. Virginia custom grants for semiconductor manufacturers 89
H. Semiconductor manufacturing sales and use tax exemptions by state 91
I. Pollution control equipment exemptions by state 93
J. Recycling equipment tax credits by state 99
K. Biodiesel production incentive by state 102
L. Agency responses 104
Appendixes – Online only
M. Statistical analysis and economic impact modeling
N Results of economic and revenue impact analysis
O. Findings of peer-reviewed research
P. References
Summary: Data Center and Manufacturing Incentives
Virginia provides 11 incentives to promote data center and manufacturing growth in
the state or to encourage environment-friendly practices, and spending on these in-
centives totaled $559 million between FY10 and FY17. The data center sales and use
tax exemption is by far Virginia’s largest incen-
tive in terms of forgone revenue, representing WHY WE DID THIS STUDY
more than one-fifth of Virginia’s total spending Through language in the Appropriation Act, the General
on economic development incentives during Assembly directed the Joint Legislative Audit and Review
Commission (JLARC) to review and evaluate economic
this period. development initiatives. Topics include spending on incentives
and activity generated by businesses receiving incentives; the
WHAT WE FOUND economic benefits of incentives; and the effectiveness of
Data Center Exemption appears incentives.
relatively effective and generates JLARC releases two reports each year: a high-level summary
report on overall spending and business activity and an in-
moderate economic benefits depth report on the effectiveness of individual incentives. (See
The data center exemption appears relatively ef- Appendix A: Study mandate.) JLARC contracted with the
Weldon Cooper Center for Public Service to perform the
fective. The exemption has a sizable influence
analysis for both reports.
on data center decisions to locate or expand in
This report is the third in the series of in-depth reports on the
Virginia, and it is estimated to have a moderate effectiveness of individual incentives and focuses on Virginia’s
economic benefit per $1 million in spending by data center and manufacturing incentives.
the state. It is reasonable for the state—which
has identified data centers as a targeted indus-
try—to continue the exemption. However, available information is insufficient to ac-
curately estimate the full fiscal impact and the economic benefits of the exemption,
and the exemption has not stimulated much growth in distressed areas. Some states
are taking more aggressive actions to attract data centers, which may affect Virginia’s
competitive position.
Semiconductor Custom Grants did not lead to development of
semiconductor industry but generated moderate economic benefits
Virginia provided sizable custom grants to attract two semiconductor manufacturers
to locate and expand in the state, with the hope that a semiconductor industry would
develop in Virginia. However, semiconductor activity in Virginia and the U.S. has de-
clined over the past two decades, with much of the U.S. semiconductor activity moving
overseas or remaining concentrated in a handful of states. Still, the custom grants for
both semiconductor manufacturers generated moderate benefits per $1 million in
spending by the state over the period reviewed.
i
Summary: Data Center and Manufacturing Incentives
Manufacturers Single Sales Apportionment has not resulted in
employment growth but generates moderate economic benefits
Virginia adopted provisions allowing manufacturers to use single sales apportion-
ment—a method that could reduce their income tax liability—to help stem the decline
of the industry in the state. Only a small portion of manufacturers have elected to use
single sales apportionment, and manufacturing employment in Virginia is growing
slightly less than manufacturing employment across the U.S. However, the incentive
generates a moderate economic benefit per $1 million in spending by the state. It may
be difficult to improve the effectiveness of single sales apportionment without reduc-
ing its economic benefit to the state, therefore any changes should be part of a broader
decision about the state’s apportionment policy.
Four tax incentives appear useful for other purposes even though
they generate low economic benefits
Four tax incentives—the Semiconductor Manufacturers Exemption, Semiconductor
Wafer Exemption, Pollution Control Equipment and Facilities Sales Tax Exemption,
and Recyclable Materials Processing Equipment Tax Credit—have low economic ben-
efits per $1 million spent by the state. However, they serve other purposes, such as
providing more efficient tax treatment and reducing the burden of complying with
environmental regulations. The recyclable materials tax credit could become more use-
ful if the recycling industry grows.
Two tax credits to support green industry do not achieve goals and
generate negligible economic benefits
Virginia’s Green Job Creation Tax Credit and Biodiesel and Green Diesel Fuels Pro-
ducers Tax Credit have low rates of utilization and little effect on the activity they were
designed to encourage. The Green Job Creation Tax Credit has had little to no effect
on employment in green energy jobs, and the Biodiesel and Green Diesel Fuel Pro-
ducers Tax Credit has had no effect on Virginia’s biodiesel production rate. Both cred-
its also have a negligible economic benefit to the state per $1 million spent on the
credits.
ii
Summary: Data Center and Manufacturing Incentives
Economic benefit of data center and manufacturing incentives varies
Spending Incentive Economic benefit
Program FY10–FY17 type per $1M of spending
Data Center Exemption $417.47M Exemption Moderate
Manufacturers Single Sales Apportionment 67.80 Special tax apportionment Moderate
Semiconductor Custom Grants (Micron and Qimonda) 27.00 Grant Moderate
Pollution Control Equipment and Facilities Exemption 26.59 Exemption Negligible
Recyclable Materials Processing Equipment Tax Credit 10.71 Tax credit Negligible
Semiconductor Manufacturers Exemption 7.81 Exemption Low
Semiconductor Wafers Exemption 1.14 Exemption Low
Biodiesel and Green Diesel Fuels Producers Tax Credit 0.01 Tax credit Negligible
Green Job Creation Tax Credit <0.01 Tax credit Negligible
Data Center Single Sales Apportionment 0.00 Special tax apportionment n.a.
SOURCE: Weldon Cooper Center economic impact analysis of incentives.
NOTE: Special tax apportionment that is only available to eligible corporations. Economic benefits of data center single sales apportionment
could not be estimated because the incentive has not been used to date. Micron and Qimonda received a total of $93.4 million in grant
payments but the majority of the payments were made prior to FY10.
n.a.: not available.
WHAT WE RECOMMEND
Legislative action
Further reduce or remove the minimum job creation requirement in distressed
areas or enterprise zones for the data center exemption.
Require all data centers to report employment levels, capital investment, and tax
benefit information to the Virginia Economic Development Partnership.
Require a work group to examine Virginia’s infrastructure and other policies af-
fecting data centers to identify (1) actions needed to maintain Virginia’s competi-
tive position and (2) whether opportunities exist to reduce reliance on the sizable
exemption without adversely affecting industry growth.
Eliminate the Green Job Creation Tax Credit and Biodiesel and Green Diesel
Fuel Producers Tax Credit.
Executive action
The Virginia Economic Development Partnership should report to the legisla-
ture on how potential custom grants, similar to the semiconductor grants, align
with state and regional industry clusters and strategic goals.
The Department of Environmental Quality should develop a list of pre-ap-
proved equipment and facilities to expedite certification for the pollution control
exemption.
The complete list of recommendations and options is available on page v.
iii
Summary: Data Center and Manufacturing Incentives
iv
Recommendations and Options: Data Center and
Manufacturing Incentives
RECOMMENDATION 1
The General Assembly may wish to consider amending § 58.1-609.3 of the Code of
Virginia to further reduce or remove the minimum job creation requirement of the
sales and use tax exemption for data centers locating in a distressed area or an enter-
prise zone.
RECOMMENDATION 2
The General Assembly may wish to consider including language in the Appropriation
Act directing the Secretary of Finance to convene a work group consisting of the
Secretaries of Transportation, Commerce and Trade, and Administration; the staff
directors of the House Appropriations Committee and Senate Finance Committee, or
their designee; and other relevant agency stakeholders to conduct a data center indus-
try study to examine actions that could be taken to maintain the state’s competitive
position to attract data centers and examine whether the opportunity exists to reduce
the level of the exemption without adversely affecting industry growth.
RECOMMENDATION 3
The General Assembly may wish to amend § 58.1-609.3 of the Code of Virginia to
require that (1) all data centers using the data center sales and use exemption be re-
quired to submit an annual report including their employment level, capital investment,
and tax benefit to the Virginia Economic Development Partnership (VEDP) and (2)
the Department of Taxation (TAX), in consultation with VEDP, publish an annual
report on the data center exemption which should include, at a minimum, aggregate
information on qualifying expenses that were exempt from the retail sales and use tax
and the total value of the tax benefit.
RECOMMENDATION 4
The Virginia Economic Development Partnership should report to the Major Em-
ployment and Investment Project Approval Commission how each custom grant being
considered by the commission aligns with state and regional targeted industries and
strategic plans.
RECOMMENDATION 5
The General Assembly may wish to consider amending § 30-312 of the Code of Vir-
ginia to require that information on how custom grants align with state and regional
target industries and strategic economic development plans be included in the annual
report of the Major Employment and Investment Project Approval Commission.
v
Recommendations and Options: Data Center and Manufacturing Incentives
RECOMMENDATION 6
The Department of Environmental Quality and Department of Mines, Minerals, and
Energy should develop guidance documents on (1) the types of pollution control
equipment and facilities that are exempt from the retail sales and use tax and (2) the
decision-making process for approving certification.
RECOMMENDATION 7
The Department of Environmental Quality should develop a list of pre-approved
equipment and facilities that typically meet the pollution control certification require-
ments and create an expedited certification process for equipment and facilities on that
list.
RECOMMENDATION 8
The General Assembly may wish to consider eliminating the Green Job Creation Tax
Credit and the Biodiesel and Green Diesel Fuel Producers Tax Credit.
OPTION 1
The General Assembly could amend § 58.1-609.3 or § 58.1-3660 of the Code of Vir-
ginia to clarify that the equipment or facility does not need to be constructed before
certification can be granted for purposes of claiming the Pollution Control Equipment
and Facilities Sales Tax Exemption.
Commission draft
vi
Data Center and Manufacturing Incentives
Economic Development Incentives Evaluation Series
Virginia provides economic development incentives to encourage businesses to locate
or expand their operations in the state. In order to better understand the effectiveness
of these incentives in stimulating business activity, the General Assembly directed the
Joint Legislative Audit and Review Commission (JLARC) to review and evaluate the
effectiveness and economic benefits of economic development incentives, such as
grants, tax preferences, and other assistance. (See Appendix A for the study mandate.)
This report is part of a series of annual reports that provide a comprehensive review
of individual economic development incentives offered by the state. JLARC con-
tracted with the University of Virginia’s Weldon Cooper Center for Public Service to
perform the evaluation.
This report focuses on 11 incentives designed to promote data center and manufac-
turing growth, reduce pollution from manufacturing, and to support green jobs (Ta-
ble). Two incentives target the state’s rapidly growing data center industry—a sales and
use tax exemption for computer and equipment purchases and the ability to use a For purposes of this
special apportionment (single sales factor) method to reduce corporate income tax report, spending on
liability. incentives refers to
(1) actual expenditures
Four incentives target the state’s semiconductor manufacturing industry. Two of the by the state in the form
incentives were large customized grants designed to encourage two semiconductor of grant awards and
manufacturers, Micron and Qimonda, and their predecessors, to locate in the state in (2) tax expenditures in
the form of forgone
the 1990s and expand in 2005.
revenue, through tax
The remaining five incentives target manufacturers. One of the incentives is a special credits or sales and use
tax exemptions.
apportionment formula that can lower state income taxes for some manufacturers.
The remaining incentives primarily are used by manufacturers and are designed to en-
courage environment-friendly practices.
State spending on these 11 incentives totaled $559 million between FY10 and FY17.
The data center exemption is by far the largest of the incentives evaluated in this report
and the largest offered by Virginia. State spending on the data center exemption in the
form of forgone revenue has grown with the sector’s increasing economic footprint
in the state. This exemption currently represents 22 percent of Virginia’s total spend-
ing on incentives between FY10 and FY17 ($1.9 billion). Manufacturers single sales
apportionment is also among the state’s 10 largest incentives.
1
Data Center and Manufacturing Incentives
TABLE: Eleven incentives targeting data centers and manufacturers are evaluated in this report
Purpose
Spending Data center Manufacturing Promote environment-
Program FY10–FY17 industry growth industry growth friendly practices
Data Center Exemption $417.47M
Manufacturers Single Sales
Apportionment
67.80
Semiconductor Custom Grants (two
grants for Micron and for Qimonda)
27.00
Pollution Control Equipment and
Facilities Exemption
26.59
Recyclable Materials Processing
Equipment Tax Credit
10.71
Semiconductor Manufacturers
Exemption
7.81
Semiconductor Wafers Exemption 1.14
Biodiesel and Green Diesel Fuels
Producers Tax Credit
0.01
Green Job Creation Tax Credit <0.01
Data Center Single Sales
Apportionment
0.00
All programs $558.52M
SOURCE: Weldon Cooper Center review of Code of Virginia and agency documents.
NOTE: Spending on grants includes amounts for projects that have been completed or have reached milestones and received payments;
tax credits includes amounts claimed; and exemptions are estimates of forgone revenue. Micron and Qimonda received a total of $93.4
million in grant payments, but the majority of the payments were made prior to FY10. Data centers were not able to use the single sales
apportionment until FY17, but none had met the eligibility and administrative requirements to use it by the end of FY17.
2
Data Center and Manufacturing Incentives
1. Data Center Incentives
Virginia offers two tax incentives to attract large-scale data centers: a data center retail
sales and use tax exemption and a single sales factor apportionment formula designed
to help enterprise data centers lower their state income tax liability (Table 1-1). Data
centers are facilities that house computer equipment such as servers, data storage de-
vices, and related equipment to manage, process, and distribute information. These Each state that imposes
incentives are restricted to data centers that meet minimum capital investment levels income taxes on busi-
nesses establishes an
and other criteria. Both incentives are jointly administered by the Virginia Economic
apportionment formula
Development Partnership (VEDP) and the Department of Taxation (TAX). VEDP is for businesses operating
responsible for assessing whether a data center meets the eligibility requirements and in multiple states to cal-
establishing a memorandum of understanding (MOU) with the data center. TAX is- culate the percentage of
sues exemption certificates to eligible data centers and ensures that single sales appor- their income that is taxa-
ble in the state.
tionment is used appropriately.
The General Assembly enacted Virginia’s first data center incentive in 2008, creating a Enterprise data centers
sales and use tax exemption to attract a specific data center to Virginia. This initial are in-house data cen-
exemption was limited in scope. It applied only to data centers that located in a locality ters owned and oper-
with an unemployment rate greater than 4.9 percent and required that the data center ated by a company for
its own use, or use by its
establish an MOU with VEDP during calendar year 2008. Even though the capital
customers. Examples in-
investment threshold was lower ($75 million) than the current exemption, the exemp- clude Google, Facebook,
tion had higher job creation (100 new jobs) and average wage requirements (twice the and Microsoft.
average wage in the locality). The initial exemption also applied only to purchases for Colocation data centers
data processing, storage, retrieval, or communication made on or before June 30, 2011. are owned by a “land-
lord” company that
The General Assembly adopted a broader exemption in 2010 to attract additional data leases space to busi-
centers. This was in part designed to compete with North Carolina, which had recently nesses on either a retail
been chosen for the location of two-large scale data centers—Google (a $750 million (multiple businesses or
“tenants” lease space) or
data center in 2007) and Apple (a $1 billion data center in 2009). The exemption ex- wholesale (one tenant
panded eligible purchases to include enabling hardware such as wiring, generators, leases the data center)
chillers, and other similar items used to operate data centers. Data centers with an basis.
MOU would be able to make tax exempt purchases through 2020. The General As-
sembly expanded the exemption again in 2012 and 2016.
2012: Expanded to tenants of colocation data centers and allowed jobs created
by tenants to count toward the job creation requirement, in addition to the
jobs created by the colocation data center owner.
2016: Extended the sunset date on the exemption to 2035 and allowed data
centers to relocate employees from an existing Virginia data center to meet
new job requirements for a new data center, if it makes an investment of at
least $500 million.
3
Data Center and Manufacturing Incentives
4
Data Center and Manufacturing Incentives
TABLE 1-1
Virginia provides two incentives for data centers: sales and use tax exemption and single
sales apportionment
Data Center Sales and Use Tax Exemption
Purpose Promote the establishment of large-scale data centers in Virginia
Eligible projects Enterprise or colocation data centers that make a new capital investment of at least $150 million and
create at least 50 new jobs in a Virginia locality
Minimum new job requirement is lowered to 25 if data center is located in an enterprise zone or in a
locality with an unemployment rate at least 1.5 times the average statewide unemployment rate
New jobs must pay at least 1.5 times the annual average wage in the locality where the data center
is located
Program Data centers must enter into an MOU with VEDP; colocation tenants must enter into a Participation
features Certificate Agreement with the colocation data center
Minimum capital investment, job, and wage thresholds must be met by a performance date
stipulated in the MOU, generally within 3 years after the MOU is established
Exemption can be used before meeting the minimum thresholds, but repayment is required if
thresholds are not met by the performance date
MOU is only applicable to a specific data center in a single locality; if a company builds another data
center outside of that locality, it must separately meet eligibility and file an MOU to use the
exemption
Once MOU is established for a data center and performance requirements are met, the exemption
can be used for all qualifying purchases made through 2035, unless the expiration date is changed
by the General Assembly
Use of Reduces costs of initially purchasing and upgrading or replacing eligible equipment including
exemption - servers, mainframes, network infrastructure, data storage hardware, and other computer
equipment used for data processing, storage, retrieval, or communication
- enabling hardware such as cabling, switches, chillers, generators, monitoring systems, and similar
items used to operate exempt equipment
Single Sales Apportionment for Enterprise Data Centers
Purpose Promote the establishment of large-scale enterprise data centers in Virginia
Eligible projects Enterprise data center operations (colocation centers not eligible)
Make a new capital investment of at least $150 million in Virginia
Program Data centers must enter into an MOU with VEDP
features Apportion income subject to Virginia tax based on the percentage of total sales in Virginia, rather
than using the state’s standard formula, which also accounts for the company’s proportion of total
property and payroll in Virginia. This percentage (the apportionment factor) is applied to total
taxable income to calculate income subject to Virginia’s corporate income tax
Use of incentive Allows Virginia enterprise data centers to reduce their Virginia income tax liability
SOURCE: Weldon Cooper Center review of the Code of Virginia and agency documents.
NOTE: Authorized by § 58.1-609.3 (exemption) and § 58.1-422.2 (single sales apportionment). Single sales apportionment was offered in
phases. Between July 2016 and June 2017, data centers could use a quadruple-weighted sales apportionment formula. Single sales
apportionment for data centers was fully phased in beginning July 2017.
5
Data Center and Manufacturing Incentives
In 2015, the General Assembly enacted legislation allowing enterprise data centers to
use a single sales apportionment formula to lower their income tax liability to encour-
age them to locate or expand in Virginia. Multistate corporations typically use three
factors to determine their Virginia income tax: property, payroll, and sales. Single sales
apportionment can reduce the Virginia tax income liability for data centers that have
sizable property and payroll in the state but sales that are spread out nationwide. (See
Figure 3-1 on page 39 for more detail about how each apportionment formula is cal-
culated.) To use single sales apportionment, eligible data centers must enter into an
MOU with VEDP and agree to make a new capital investment of least $150 million.
Data centers could begin using single sales apportionment after July 2017. Through
tax year 2017, no enterprise data centers had used single sales apportionment, most
likely because insufficient time had passed. Several data center operators have since
entered into MOUs with VEDP to use single sales apportionment and will likely begin
using it within the next two years. Several others have expressed interest in using the
incentive.
Virginia and many states have adopted incentives to attract data
centers—a fast-growing industry viewed as highly desirable
Since the late 2000s, states have increased efforts to attract data centers. Virginia was
the seventh state to adopt a sales and use tax exemption for data centers when it en-
acted its initial exemption in 2008. Currently, the majority of states provide favorable
tax treatment to data centers either because they currently have some form of sales
and use sales tax exemption for data centers (28) or they levy no sales and use tax (5)
(Figure 1-2). All states bordering Virginia provide some form of sales tax exemption
or other incentive to data centers. Virginia’s adoption of single sales apportionment
for enterprise data centers also makes Virginia’s corporate tax system more competi-
tive. (See Appendix C for more information about the data center exemptions offered
by other states and Appendix D for more information about corporate apportionment
by state.)
6
Data Center and Manufacturing Incentives
Figure 1-2
Majority of states provide favorable sales tax treatment to data centers
SOURCE: Weldon Cooper Center review and analysis of Mangum Economics (2018), Tarczynska (2016), and C2ER
Business Incentives Database.
NOTE: Alaska, the District of Columbia, and Hawaii also have no data center incentives. Kansas and Maryland provide
other incentives.
States view data centers as an attractive industry because they are capital intensive and
require substantial investment and reinvestment in equipment. Even though the num-
ber of jobs associated with data centers is relatively small, wages are generally high.
Construction spending for materials and labor is also substantial. Local governments
in Virginia and across the nation also find data centers attractive because they are a
valuable source of revenue. Loudoun County received more than $154 million in prop-
erty taxes from computer equipment in FY18, and Prince William County received
more than $25 million in FY17. Some localities in Virginia, such as Chesterfield, Hen-
rico, and Prince William counties, have lowered the property tax rates on equipment
used in data centers to attract data centers.
Technological advancements continue to drive growth of the data center industry. Sig-
nificant improvements in telecommunications speed, capacity, and reliability (i.e.,
broadband) have increased consumer use of mobile devices, social media, and gaming
applications and increased demand for remote storage and processing. Businesses are
increasingly outsourcing their data processing, data storage and management, and
computer applications to data center cloud operations to take advantage of low-cost
scale economies and provide more uniform and higher quality software applications.
Growth in the industry is projected to continue with the deployment of hyper-fast 5G
7
Data Center and Manufacturing Incentives
networks and increasing vehicle automation, virtual and augmented reality capabilities,
Data center absorption
and use of “smart” appliances.
is a measure of how
much of the net availa-
ble data center capacity Northern Virginia is the premier data center market in U.S. and world
is claimed via new lease
or expansion in a given
Northern Virginia is recognized as the nation’s premier data center market, with its
period, usually measured size and growth outpacing other major metropolitan markets. Northern Virginia is
in megawatts. rated first in data center space by the U.S. Business Facilities’ 13th Annual Rankings
Data centers are often Report. The Northern Virginia market absorbed 270 megawatts (MW) of power ca-
classified by their size, pacity in 2018, far outpacing London (UK) in second place with just 69MW of ab-
usually in square feet or
sorption. Loudoun County has become known as “Data Center Alley” because of its
megawatts (MW) of
power usage. massive concentration of data centers. An estimated 70 percent of international inter-
net traffic flows through its telecommunications infrastructure.
Northern Virginia has key infrastructure to attract data centers
Northern Virginia’s dominant market position was driven by a clustering of data pro-
cessing firms contracting with government agencies and high-technology government
labs. These firms located near Dulles International Airport—the “Dulles Corridor”—
Several major telecom- as early as the 1940s and 1950s. An Internet exchange point was established in Tysons
munications invest- Corner in the 1990s, and this infrastructure attracted other pioneering telecommuni-
ments in Virginia have cation firms, such as MCI, AOL, and Network Solutions. The area became known as
occurred recently. Face-
book and Microsoft
“Internet Alley” as additional telecommunication hubs were created in the region and
formed a joint venture to additional investments were made in fiber optic cable. The Northern Virginia region
build the 4,000-mile now hosts at least two dozen Internet exchange points that provide the infrastructure
long MAREA (Spanish through which regional, national, and international internet traffic flows.
for “tide”) transatlantic
cable between Spain and Industry analysts credit other favorable factors to the region’s large data center market.
Virginia Beach that was These factors include
completed in 2018.
BRUSA, another major a stable climate and low level of seismic activity,
international cable, links relatively low power rates,
Virginia Beach to Brazil
and was completed at
one of the nation’s best-educated metropolitan workforces, and
the end of 2018. proximity to major national customers (including most notably the federal gov-
SAex (South Atlantic Ex- ernment, government contractors, and tech firms that hold an enormous
press) will establish ser- amount of government and other data).
vice between South Af-
rica, Latin America, and Some also cite the importance of private-industry groups, such as the Northern Vir-
the U.S. and is scheduled ginia Technology Council (NVTC), for promoting and supporting industry investment
to be operational in and expansion.
2020.
Google announced a Most data center activity has occurred in Northern Virginia, but Richmond area
subsea cable project
called “Dunant” that will
and Southside have sizable concentrations
connect Virginia Beach While Northern Virginia continues to dominate the state’s data center market, data
with France beginning in
centers also have a sizable presence in the Richmond Metropolitan Area (15 percent
2020.
of total capital investment) and in Southside (10 percent of total capital investment)
8
Data Center and Manufacturing Incentives
(Figure 1-2). Microsoft located a data center in Mecklenburg County in 2010, and has
continued to expand its presence there with expansion announcements nearly every
year. QTS announced the location and expansion of a data center in the Richmond
Metropolitan area in 2010 and 2011, and the data center market greatly expanded with
Facebook announcing the location of large data centers (2017 and 2018). Several
smaller data centers have located in the Greater Charlottesville region (Culpeper and
Fauquier counties).
Figure 1-2
Most data center activity has occurred in Northern Virginia, followed by
Richmond and Southside (2009–2018)
This report includes a re-
vised estimate of the
forgone revenue be-
cause of the data center
exemption. Previous
JLARC reports included
estimates developed by
the Department of Taxa-
tion (TAX). Weldon
Cooper staff developed
a revised estimate for
SOURCE: Weldon Cooper Center analysis of VEDP announcements data 2009–2018, and undisclosed announcements this report. This change
by GO Virginia Region. increased estimated to-
NOTE: Does not include some recent data center announcements in Virginia Beach.
tal spending on all in-
centives between FY10
and FY17 from $1.8 bil-
Data center exemption made up more than one-fifth of Virginia’s
lion to $1.9 billion. (Eco-
incentives spending FY10–FY17 nomic Development In-
The data center exemption is by far Virginia’s largest incentive (see Economic Development centives 2018).
Incentives 2018, JLARC 2018), and estimates developed for this report suggest it repre- This method likely still
underestimates forgone
sented more than one-fifth of Virginia’s total incentive spending between FY10 and
revenue because com-
FY17. The single sales apportionment for data centers has not been used to date. The panies with an existing
amount of forgone revenue because of the special apportionment formula is expected MOU may make major
to be minimal, at least initially. data center campus ex-
pansions at an existing
Data centers saved an estimated $417 million because of data center exemption location without needing
to enter into another
FY10–FY17
MOU.
The estimated value of the data center exemption was approximately $417 million
between FY10 and FY17. As of FY2017, 159 data centers (including enterprise data
9
Data Center and Manufacturing Incentives
centers and landlords and tenants of colocation centers) have signed MOUs with
VEDP to use the exemption. The state revenue impact of the current sales and use
tax exemption has increased since its introduction in 2010, from an estimated $3 mil-
lion in FY10 to an estimated high of $93 million in FY15 (Figure 1-4). The revenue
impact was highest between FY13 and FY15 when substantial data center investments
were made. Much of the investment that was made in FY16 and FY17 was for replac-
ing and upgrading equipment.
Figure 1-4
Data centers saved an estimated $417 million from data center equipment
purchases because of exemption (FY10–FY17)
SOURCE: Weldon Cooper Center analysis of information provided by VEDP.
NOTE: Estimates of spending on exemption are underestimates because some data centers report only the minimum
threshold amounts of capital expenditures. See Appendix B for more information about how the estimate was de-
rived. The estimated exempt amount for FY18 is $86.5 million but is not included to maintain consistency with the
reporting of information for other incentives in this report.
While the exemption was originally adopted to benefit enterprise data centers, coloca-
tion data centers now are the largest group of beneficiaries. Many colocation centers
began to qualify for the exemption after changes were implemented in FY13. Of the
data centers using the sales and use tax exemption as of June 2017, 10 were enterprise
or single use data centers, and the remainder were landlords of colocation data centers
(14) or their tenants (135). Capital investments reported by data centers to VEDP, as
required by their MOUs, show that investment for enterprise data centers slowed after
FY13 and that most of the investment growth since then has been for colocation data
centers (Figure 1-5). By FY17, cumulative investments by colocation landlord and ten-
ants had reached $8 billion (69 percent of total data center investment).
10
Data Center and Manufacturing Incentives
Figure 1-5
Colocation data centers are increasingly benefiting from the exemption as
indicated by capital expenditures (FY10–FY17)
SOURCE: Weldon Cooper Center analysis of MOU information provided by VEDP.
NOTE: Colocation includes landlord and tenant. Capital expenditures for some data centers exceed the announced
capital investments in Figure 1-2. Capital expenditures are assigned to the year in which MOU was signed.
No data centers have used single sales apportionment, and future forgone
revenue is expected to be minimal
Single sales apportionment for enterprise data centers has not been used yet on tax
Because data centers
returns, although some data centers may begin using it soon. It was not used on income have not used single
tax returns for tax year 2017—the first year that firms were eligible to begin using it— sales apportionment, it
according to TAX staff. Several data centers since have entered into MOUs with could not be evaluated
at this time. The remain-
VEDP. It is estimated that corporate tax liability could be reduced by $100,000 in FY19
der of the section fo-
and $240,000 by FY24 if data centers and current data center prospects begin using cuses on the data center
single sales apportionment (Table 1-2). (See Appendix B for the methodology for com- exemption.
puting these revenue reductions.)
The projected revenue reduction from single sales apportionment is much less than
for the data center exemption because it can only be used by enterprise data centers.
The revenue reduction is also dependent on data center profits rather than expendi-
tures.
11
Data Center and Manufacturing Incentives
Table 1-2
Single sales apportionment for data centers is projected to
lead to minimal reductions in state revenue
Year Reduction in state revenue
FY19 $0.10M
FY20 0.13
FY21 0.15
FY22 0.17
FY23 0.19
FY24 0.22
FY25 0.24
7-year total $1.20M
SOURCE: Weldon Cooper Center estimates based on VEDP MOU data
for enterprise data centers and secondary public data.
Exemption is influential in data center growth
Tracking growth in data
center employment and
Virginia has several advantages to attract data centers, but the exemption has been a
other activity over time is key factor in the industry’s growth. As previously mentioned, the state offers the nec-
difficult. Data centers, un- essary fiber and power infrastructure, lower utility rates, and a skilled workforce for
like many industries, do building and operating data centers. Data center representatives acknowledge these
not have a predefined in-
dustry or NAICS code.
advantages have strongly contributed to the high level of industry clustering in Vir-
The NAICS code most ginia. They also indicated that incentives have increasingly become more important
likely to include data cen- since the late 2000s, when additional states began offering incentives to attract data
ters is 518210—data pro- centers.
cessing, hosting, and re-
lated services—but a re-
view of data from the Vir-
Data center employment and investment increased as Virginia broadened its
ginia Employment Com- incentives
mission suggests the ma-
Data center job creation and capital investment, in particular, have increased as Vir-
jority of data center em-
ginia expanded eligibility of its data center incentives (Figure 1-6). These increases also
ployment in Virginia is in-
cluded in other sectors. correspond with steady growth of cloud computing. Job creation and investment in-
See Appendix B for the creased following the extension of the data center exemption to colocation data center
methodology for tracking tenants after July 2012.
data center employment.
Job creation and capital investment increased substantially after the exemption sunset
date was extended to 2035 in 2016 and again when enterprise data centers could use
single sales apportionment in 2017. Even though single sales apportionment has not
been used through tax year 2017, its adoption was instrumental in attracting large en-
terprise data center investments that plan to use it in the future, according to data
center representatives. The majority (58 percent) of the job creation and most (72
percent) of the investment appear to have been by data centers using the exemption
based on analysis of VEDP announcement and MOU data. This information alone,
however, is only suggestive of cause and effect and does not separate the influence of
these tax policy changes from pre-existing growth trends in the industry and other
factors.
12
Data Center and Manufacturing Incentives
Figure 1-6
Data center employment and investment increased in Virginia as data center
incentives broadened
SOURCE: Weldon Cooper Center analysis of VEDP Announcements and Closings Database.
NOTE: SSF, single sales factor. Job creation and investment are assigned to the year they were announced.
Virginia data center representatives rated exemption among top location
factors
Data center representatives in Virginia rated incentives among the top four location
factors they use to select a site for location or expansion. Although the importance of
specific site selection factors differed among companies, they generally indicated that
top factors included infrastructure and utility costs, availability of skilled labor for data
center operation and construction, the business-friendliness of the state and local ju-
risdiction, and state and local business taxes and tax incentives. All data center repre-
sentatives said the exemption was instrumental in tipping the balance for a site when
site development and operating costs were similar.
The sales and use tax exemption is regarded as important because of the capital inten-
sity of the industry. The initial investment to establish the data center is substantial.
Additional substantial investments are necessary to sustain data centers because they
must replace or upgrade their equipment frequently, usually every three to five years.
Local property taxes, corporate income taxes, and availability of discretionary eco-
nomic development incentives were also identified as factors that played a role in lo-
cation decisions but were mentioned less frequently than the sales and use tax exemp-
tion.
13
Data Center and Manufacturing Incentives
Exemption size makes it an influential factor in data center growth
The scale estimating the
It is estimated that the majority of the economic activity generated by the data centers
amount of economic ac- using the exemption is attributable to the incentive. The exempted amount represents
tivity attributed to an in- a sizable fraction of data centers’ estimated operating costs for the new activity. Data
centive is based on the centers using the exemption saved an estimated $417 million in taxes from FY10–
incentive amount as a
FY17, representing 21 percent of operating costs for this new activity over a 20-year
percentage of the busi-
ness’s operating costs period. Based solely on this cost analysis, the exemption is estimated to induce up to
for the related activity 90 percent of the economic activity of these data centers, according to a scale devel-
over a 20-year period. oped by a leading researcher of incentives from analysis of incentive activity across
The estimate is based on
the nation (Bartik 2018b). However, it is difficult to precisely estimate the impact in-
costs and does not ac-
count for other factors centives have on business location decisions, including the impact of Virginia’s data
that may influence a center incentives. Business executives consider many site location factors, and it is im-
business’s location or ex- possible to verify their motivations. This report estimates incentives’ impacts using a
pansion decisions. See
cost analysis based on best available research. However, it does not include other site
Appendix M [online
only] for more detail on location factors or incentive features that may affect data center location decisions
the difficulty of precisely (sidebar).
estimating incentives’ ef-
fects and the methodol- Virginia’s data centers also received incentives from several other grant and tax credit
ogy used in this report. programs, but it is unlikely that much data center activity can be directly attributed to
them. The amount data centers received from these other programs is small ($17 mil-
lion) relative to the total exemption amount ($417 million). Most of the $17 million
went to one company (Microsoft for its data centers in Mecklenburg County).
Exemption has moderate economic benefit and moderate return in
state revenue
The data center exemption is estimated to have generated additional economic activity
for the state between FY10 and FY17. Estimates show that each year private sector
employment increased by 7,665 jobs, Virginia GDP increased by $1.3 billion, and
statewide personal income increased by $724.9 million, on average, because of the data
center exemption (Table 1-4). These estimates assume that 90 percent of total eco-
nomic activity generated by the incentivized data centers is attributed to the data center
exemption.
Much of the economic benefit occurs during construction of the data center and from
initial and ongoing expenditures for equipment, rather than the daily operation of the
data center. Capital investments to build and equip the facility and to update or replace
equipment every three to five years account for approximately two-thirds of the addi-
tional employment (67 percent), Virginia GDP (58 percent), and personal income (64
percent) that was generated by the exemption between FY10 and FY17.
14
Data Center and Manufacturing Incentives
TABLE 1-4
Data center sales tax exemption has moderate economic benefit and moderate
return in revenue to the state (FY10–FY17)
Annual average (FY10–FY17)
Net impact to Virginia economy
Economic impact
Private employment 7,665 jobs analysis of expenditures
Virginia GDP $1.3 billion
by exemption recipients
Personal income $724.9 million
between FY10 and FY17
Impact to Virginia economy per $1 million of exemption
was conducted using
Private employment 155 jobs
economic modeling
Virginia GDP $26.5 million
software developed by
Personal income $14.6 million
REMI, Inc.
Impact to state revenue
Total revenue $37.7 million (See Appendix M [online
Exempted amount $52.2 million only] for the economic
Revenue net of exempted amount ($14.5 million) impact analysis used in
Return in revenue 72¢ per $1 spent this study.)
SOURCE: Weldon Cooper Center economic analysis of estimated exempted amounts between FY10 and FY17.
NOTE: Includes direct, indirect, and induced impacts. Assumes that 90 percent of the economic activity is attributable
to the exemption. The gross impact on Virginia’s economy is used to calculate the impact per $1 million in exempted
amount. This is consistent with how the economic development research literature typically calculates these impacts.
(See Appendix N [online only] for detailed results on total impact of the exemption, impact of raising income taxes Net impact is the
by the amount exempted [opportunity cost], and revenue generated by source.) increase in economic
activity induced by the
exemption after adjust-
Overall, this economic benefit is moderate compared to the economic benefits of ing for the opportunity
other incentives that have been assessed so far in this series. The data center exemption cost of increasing taxes
is estimated to generate $27 million in Virginia GDP per $1 million in spending on the to pay for the exemp-
exemption. This amount is higher than the Virginia GDP per $1 million in spending tion.
generated by most other incentives reviewed in this report and slightly less than the (See Appendix N [online
economic benefits of several workforce grant programs reviewed in 2018. (See Work- only] for information on
the total economic im-
force and Small Business Incentives, JLARC, 2018.) Although the economic benefit is mod- pact and the opportunity
erate compared to incentives overall, it is particularly large for a tax incentive. In fact, cost of increasing taxes.)
it is substantially larger than the estimated $1.6 million in Virginia GDP generated per
$1 million spent on all exemptions and the estimated $1 million in Virginia GDP gen-
erated per $1 million spent on all tax credits. (See Economic Development Incentives 2018,
JLARC 2018.)
The data center exemption also has a moderate return in revenue of 72¢ for every $1
spent on the exemption. Again, this return is higher than the return in revenue for
most other incentives reviewed in this report and slightly less than the return in state
revenue generated by several workforce grant programs reviewed in 2018. This return
is also substantially higher than the return in revenue for other tax incentives reviewed
to date, on average. (See Economic Development Incentives 2018, JLARC 2018.)
Even though the program likely influences more economic development projects than
the typical Virginia incentive, the sizable tax revenue generated ($38 million) by the
exemption does not cover its costs. This exemption is much more expensive in terms
of direct job creation and capital investment than most other incentives. The cost is
15
Data Center and Manufacturing Incentives
$223,125 per data center job compared to $4,189 per job for the average Virginia in-
centive. It is not surprising that the exemption has a higher cost per job because the
data center industry is not job-intensive. Instead, it is heavily capital intensive. How-
ever, the cost is $33 per $1,000 in capital investment, which is higher than the cost of
the average Virginia incentive of $24 per $1,000 in capital investment. (See Economic
Development Incentives 2018, JLARC 2018.)
Representatives reported data centers have other economic and social
benefits
Data center representatives reported that data centers provide other economic and
social benefits to a region that are not captured in the economic analysis described
above. Several representatives indicated that data center investments, particularly if
they are clustered within a region like Northern Virginia, have indirect or “knock-on”
effects that result in other economic development. The clustering of data centers, re-
lated businesses, and skilled workers can further improve the attraction of the region
and lead to additional telecommunication investments. The location of large enterprise
data centers—such as the location of Facebook in Henrico County—may also have
“follow the leader” effects by attracting smaller data centers to locate nearby. A smaller
data center may not be eligible for the exemption, but it may benefit from using infra-
structure that serves the enterprise data center at a lower cost than if it had to install
the infrastructure itself. Data centers may also have “add on” effects for a region.
These effects may occur if a company with a data center presence in Virginia makes a
broader commitment to the state by moving other company operations, such as a cor-
porate office, to the state. In this case, the presence of the data center may have served
as a recruitment tool.
Several data center representatives indicated that data centers have had a positive effect
on the development of the clean and renewable energy industry in the state. Despite
data center commitments to renewable energy procurement policies, there is signifi-
cant controversy over the actual impact of the state’s data center industry to Virginia’s
carbon footprint and renewable energy supply, particularly because data centers are a
primary source of growing electricity demand in the state. Concerns have also been
raised about the impact increased data center demand will have on utility prices, par-
ticularly as costs for some infrastructure projects may be paid for by all electricity rate
payers.
Continuation of data center exemption appears reasonable but
several actions should be considered
The data center exemption is the costliest of Virginia’s incentives in terms of forgone
revenue, but it appears to be relatively effective. It has a sizable influence on data center
behavior—even in Northern Virginia which has multiple other advantages—because
interstate competition to attract data centers has increased over the past decade. The
data center exemption also generates moderate economic benefits to the state relative
16
Data Center and Manufacturing Incentives
to other incentives and generates substantially higher economic benefits than other tax
incentives. Data centers also provide sizable tax revenue to the localities in which they
locate.
Substantial changes to Virginia’s data center exemption do not appear necessary to
keep Virginia competitive. However, actions can be taken to attract data centers to
distressed areas that could benefit from the additional local tax revenue. In addition,
the state should consider actions that could reduce the level of the incentive offered
without adversely affecting industry growth.
Virginia’s data center exemption is considered very competitive
Some data center representatives characterized Virginia’s exemption as one of the
more aggressive incentives to attract data centers. Nearly all thought Virginia’s data
center exemption was competitive with incentives provided by other states. Colocation
data centers find it particularly attractive because few other states offer a special ex-
emption for this type of data center.
Eligibility requirements were generally characterized by data center representatives as
favorable. Representatives said the exemption’s three-year investment timetable, set in
the MOU with VEDP, is reasonable. In addition, Virginia’s capital expenditure thresh-
old ($150 million) is appropriate because the industry is so capital intensive. Virginia’s
lack of a minimum square footage requirement also was viewed positively. One data
center representative stated that square footage requirements in other states limit flex-
ibility in choosing the optimal architecture and layout for the data center. The current
2035 sunset date is also viewed positively because it creates long-term certainty for
business planning, but representatives indicated removal of the sunset would provide
even greater certainty.
A lower job creation threshold could encourage data center growth outside of
Northern Virginia
Multiple data center representatives said that the exemption’s new jobs threshold was
a barrier to data center growth, particularly in areas outside of Northern Virginia. Be-
cause the Northern Virginia region faces rising real estate and congestion costs, a de-
creasing supply of developable land, and mounting difficulties building power distri-
bution and other infrastructure, data centers are likely to seek lower cost locations
elsewhere, either inside or outside of the state.
Data center representatives had two main concerns with the exemption’s job require-
ments. One concern is that the 50-job requirement is out of sync with the capital
investment requirement. One representative said: “Data centers are not job creators;
they are revenue producers.” Data centers continue to become more efficient through
automation, which means fewer jobs are necessary. Based on an analysis of capital
investment and job creation figures from data centers’ MOUs with the state, one job
is associated with $6.3 million in capital investment. Thus, a $150 million investment
would be expected to create 24 jobs, on average. Virginia and Mississippi are the only
17
Data Center and Manufacturing Incentives
For purposes of the data states that have a 50-job requirement. Most data center sales and use incentives from
center exemption, a dis- other states do not have employment creation requirements or they are lower. (See
tressed area is a locality Appendix C for minimum requirements in other states.)
that had an unemploy-
ment rate for the previ- The other main concern is that the job creation threshold is a barrier for locating data
ous year of at least 150 centers in more distressed areas of the state, an area where data centers could have a
percent of the average
big impact on the regional economy, according to data center representatives. Even
statewide unemploy-
ment rate as determined though the job threshold is reduced to 25 jobs for distressed areas and enterprise
by the Virginia Economic zones, it has had little to no impact on encouraging data centers to locate in these areas.
Development Partner- So far, only one data center (Microsoft) using the exemption has located in a distressed
ship.
locality (Mecklenburg County). Representatives indicated that smaller data centers
meeting the capital investment threshold may be more likely to locate in distressed
areas if the exemption did not have a minimum job requirement. Distressed regions
generally do not already have the skilled workforce necessary, and it is often difficult
to relocate workers from elsewhere. Savings from the exemption can provide resources
VEDP requires that all
to address these challenges.
jobs must be within the
same locality. The statu- Data center representatives also noted that all jobs must be within the same locality to
tory language states the
qualify for the exemption, which is a barrier for distressed areas. Jobs created at a data
minimum requirements
are for a “data center” center owned by the company in another locality do not count toward meeting the
but does not define minimum threshold. This requirement encourages a campus approach of clustering
what constitutes a data data center buildings in select localities, such as in Loudoun and Prince William coun-
center. VEDP requires
ties, when some companies may prefer to spread their operations out statewide.
the job creation and
capital investment all be Data center executives suggested several changes to the job creation requirement:
within the same locality,
because the statute remove requirement that all jobs be in the same locality to encourage data cen-
makes multiple refer- ter growth statewide;
ences to “in the locality.”
further reduce or remove the minimum job creation threshold in distressed ar-
eas and enterprise zones to encourage data center growth outside of Northern
Virginia; and
reduce or eliminate the job creation thresholds (both the standard 50 job re-
Chapter 512 of the 2019
quirement and the 25 job requirement in distressed areas or enterprise zones.)
Acts of Assembly (§ 2.2- The best approach at this time may be to reduce or remove the minimum job creation
621) specifies that the
creation of off-site or
threshold in distressed areas and enterprise zones (55 localities met at least one of the
teleworking jobs for Vir- criteria in 2017) to encourage data center growth in these areas. The General Assembly
ginia residents by a re- could also allow the provisions of new legislation adopted in 2019 (Chapter 512) to
cipient company or its apply to the exemption, for data centers in distressed areas or enterprise zones. Local-
affiliates may be in-
cluded in assessing com-
ities with data centers reported data centers are important sources of tax revenue, and
pliance with a job crea- they do not require substantial local government services. Therefore, data centers
tion requirement for a could provide substantial benefits to distressed areas. The other two suggested changes
grant or incentive issued could increase the amount of forgone revenue to the state by allowing smaller data
by a state agency.
centers owned by the same company, that would have located in Virginia regardless of
the incentive, to qualify for the exemption. They also do not directly incentivize data
center growth outside Northern Virginia.
18
Data Center and Manufacturing Incentives
RECOMMENDATION 1
The General Assembly may wish to consider amending § 58.1-609.3 of the Code of
Virginia to further reduce or remove the minimum job creation requirement of the
sales and use tax exemption for data centers locating in a distressed area or an enter-
prise zone.
Virginia should examine how to best maintain the state’s competitive position
The state has indicated that attracting data centers is important by offering multiple
incentives targeted specifically to the industry. VEDP has also identified data centers
as one of its targeted industries. However, data center representatives indicated that
the state risks losing its competitive position because other states are adopting more
aggressive infrastructure policies.
The state should commission a comprehensive data center industry study to assess
opportunities to maintain Virginia’s competitiveness. Although Virginia has developed
an excellent and diverse fiber network, particularly in Northern Virginia, a study could
address the challenges of attracting data center development in distressed or rural re-
gions, such as lack of redundant telecommunications fiber, difficulties in accessing
utilities, and workforce readiness. For example, Utah has encouraged the expansion of
broadband infrastructure into rural areas through public-private partnerships and by
leveraging rights-of-way along interstate and state highways. In addition, Utah lowered
installation costs for broadband service providers by facilitating cooperative fiber and
conduit trades. Virginia could similarly improve its utility and telecommunications in-
frastructure through how it addresses co-development projects and easements. The
study should also assess how Virginia’s renewable energy policies, local personal prop-
erty taxation policies, and regulatory framework affect data centers.
Given the size of the exemption, the study should examine whether the opportunity
exists to reduce the level of the incentive and still maintain Virginia’s competitive po-
sition. Economic theory also suggests incentives should be used to help emerging mar-
kets and industry clusters develop in a region. However, the current exemption is
mostly used by data centers in Northern Virginia where a strong industry cluster is
already well-developed.
RECOMMENDATION 2
The General Assembly may wish to consider including language in the Appropriation
Act directing the Secretary of Finance to convene a work group consisting of the
Secretaries of Transportation, Commerce and Trade, and Administration; the staff
directors of the House Appropriations Committee and Senate Finance Committee, or
their designee; and other relevant agency stakeholders to conduct a data center indus-
try study to examine actions that could be taken to maintain the state’s competitive
position to attract data centers and examine whether the opportunity exists to reduce
the level of the exemption without adversely affecting industry growth.
19
Data Center and Manufacturing Incentives
Better track data center investments to improve estimates of fiscal impact and
economic benefit
Current information collected from data centers using the sales and use tax exemption
is limited in scope. During a three-year performance period, VEDP requires data cen-
ters using the exemption to provide progress reports and a final performance report
that provides information on job creation and capital investment. TAX, which is fre-
quently asked to report on the fiscal impact of the data center exemption by legislative
committees, uses the capital investment information in these reports to develop an
estimate of the forgone revenue from the exemption. VEDP recently began asking
data centers to include the tax benefit of the exemption in these reports.
The reports, however, do not have enough information to accurately estimate the ex-
emption’s fiscal impact and economic benefits. Some data centers report only jobs and
spending up to the minimum threshold, so the reports underestimate data centers’
employment and capital investment. In addition, data centers are required to submit
these reports only until the end of the performance period, even though they can
continue to use the exemption until 2035. Therefore, no new information on employ-
ment, investments, and tax benefits is available and must be estimated from past levels.
The General Assembly could require that all data centers using the exemption provide
an annual report to VEDP that includes their current employment levels, capital in-
vestment, and tax benefits, even after their performance period. The information re-
ported by data centers could be considered proprietary confidential information and
exempt from public disclosure under the Virginia Freedom of Information Act, which
is the current practice. The new reporting requirement would allow TAX and JLARC
to develop more accurate estimates of the exemption’s fiscal impact and economic
benefit. Improving the accuracy of tax benefit estimates for previous years also would
be beneficial, and the General Assembly could require data centers to report historical
information for past years in which they used the exemption, at least for the past three
years. It may be difficult for data centers that have been using the exemption for many
years to provide accurate information beyond three years. The General Assembly may
wish to require TAX to publish an annual report on the data center exemption, similar
to the report it is required to publish on the Motion Picture Production Tax Credit.
Information could be reported in aggregate form only, statewide and by region, to
protect the confidentiality of proprietary information.
VEDP should work with TAX and JLARC staff to ensure that the annual reports
from data centers include the necessary information so that (1) VEDP can ensure data
centers achieve their minimum requirements pursuant to the MOU; (2) TAX can de-
velop a more accurate estimate of the cost of the data center exemption; and (3)
JLARC staff can evaluate more precisely the impact of the data center exemption.
20
Data Center and Manufacturing Incentives
RECOMMENDATION 3
The General Assembly may wish to amend § 58.1-609.3 of the Code of Virginia to
require that (1) all data centers using the data center sales and use exemption be re-
quired to submit an annual report including their employment level, capital investment,
and tax benefit to the Virginia Economic Development Partnership (VEDP) and (2)
the Department of Taxation (TAX), in consultation with VEDP, publish an annual
report on the data center exemption that should include, at a minimum, aggregate
information on qualifying expenses that were exempt from the retail sales and use tax
and the total value of the tax benefit.
21
Data Center and Manufacturing Incentives
22
Data Center and Manufacturing Incentives
2. Semiconductor Manufacturing Incentives
Virginia has offered custom incentive grants and special sales tax exemptions to attract
semiconductor manufacturing to the state. The custom grants by far are the largest of
these incentives and were created to attract three semiconductor plants to Virginia in
Semiconductors are
the late 1990s when the state aspired to establish a “Silicon Dominion.” Custom grants electronic chips or cir-
have periodically been offered to attract specific large manufacturers or other compa- cuits made out of sili-
nies to the state, because the state’s typical package of incentives (grants, tax credits, con—a material that has
exemptions, etc.) provides relatively small to modest awards (awards of $205,000, on both conductive and in-
sulative properties—and
average, between FY10 and FY17). Custom grants often provide large award amounts are the size of a postage
ranging from $30 million to $550 million. stamp or smaller. The
two main types of semi-
Multiple states offer “megadeal” incentives to attract semiconductor conductors are logical
process chips (micropro-
manufacturers cessors) and memory
Semiconductor manufacturers are viewed as valuable to states: they require a large chips for storage and re-
trieval. Anything that is
workforce and regularly purchase, replace, and upgrade expensive equipment. Some
computerized depends
of the jobs require highly skilled workers to design and fabricate the electronic chips on semiconductors.
from a silicon-based wafer. The technology used to produce semiconductor chips
quickly becomes obsolete requiring equipment to be replaced about every five to seven
years. Semiconductor manufacturing also relies on substantial investments in research
and development to design increasingly smaller and more powerful semiconductors.
Like Virginia, other states have frequently used large incentives packages or “mega-
deals” (i.e., $75 million or more in total incentives) to incentivize the location or ex-
pansion of semiconductor manufacturers (Figure 2-1). These projects would also qual-
ify for most available job creation tax credits, investment tax credits, R&D tax credits,
or similar incentives, but these incentives may not be sufficient to sway the location
decision of semiconductor manufacturers because they would only amount to a rela-
tively small part of the relocation and expansion expenses.
23
Data Center and Manufacturing Incentives
Figure 2‐1
24
Data Center and Manufacturing Incentives
Multiple states have offered “megadeal” incentives to semiconductor
manufacturers, with several exceeding $1 billion
SOURCE: Weldon Cooper Center review of information available from Good Jobs First.
NOTE: See Appendix E for more detail about the incentive packages awarded by state.
Semiconductor industry in U.S. remains concentrated in six states
The U.S. semiconductor industry has not moved much beyond its origins. The industry
was established by Texas Instruments in Dallas, Texas, and Fairchild Semiconductor
in San Jose, California, in the late 1950s and 1960s. Corporate spinoffs (AMD, Intel,
and National Semiconductors) by former Fairchild employees remained in the area
around San Jose resulting in the creation of “Silicon Valley.” Much of today’s semi-
conductor activity in the U.S. can be traced to these companies, with later manufactur-
ing plants mostly in the western and southwestern U.S. To remain competitive, most
U.S. semiconductor companies now manufacture semiconductor chips overseas or
outsource fabrication to other businesses that specialize in manufacturing. These “fa-
bless firms” focus on design and engineering instead, the bulk of which continues to
occur in California.
Currently, most U.S. semiconductor activity is located in six states: Arizona, California,
Upstream industries
Idaho, Oregon, Texas, and Massachusetts (Figure 2-2). These six states account for 74 process basic or raw ma-
percent of the nation’s employment in semiconductor manufacturing and 85 percent terial into an intermedi-
of the nation’s employment in semiconductor machinery manufacturing. These states ary product that is con-
verted into a finished
also represent nearly half (43 percent) of the employment in the electronics manufac-
product by other down-
turers upstream industries—like computer storage device manufacturing and elec- stream industries.
tronic computer manufacturing—that use semiconductor chips. Only three of these
states offer “megadeal” incentives to these firms.
25
Data Center and Manufacturing Incentives
Figure 2-2
Semiconductor manufacturing and related supply chain industries are
concentrated in a handful of states
SOURCE: Weldon Cooper Center analysis of EMSI employment data.
NOTE: See Appendix F for more detail of employment and location quotient by state.
Location quotient indi-
These six states also have high concentrations of semiconductor and related employ-
cates how concentrated ment, with sizable location quotients particularly for semiconductor manufacturing
an industry or occupa- and semiconductor machinery manufacturing. In contrast, Virginia has a low rate of
tion is in a region com- concentration for all semiconductor-related industries, with location quotients less
pared to the national av-
than 1.0. The absence of these industries and other semiconductor R&D activities
erage.
point to substantial challenges for growing a semiconductor industry cluster within the
A location quotient
above 1.0 indicates the state.
industry or occupation in
a region is more concen- Virginia approved five grants totaling $195 million but paid out less
trated than the national
than half that amount (FY96–FY17)
average. A location quo-
tient below 1.0 indicates Between FY96 and FY17, the General Assembly authorized five grants totaling $195
it is less concentrated. million to attract semiconductor manufacturers (Table 2-1). The first three were au-
thorized in the mid-1990s to attract three semiconductor manufacturers to locate pro-
duction facilities in Virginia. They included grants for
Motorola to locate a plant in Goochland County (1996),
An industry cluster is a
Dominion Semiconductor LLC to locate a plant in the City of Manassas
regional concentration
of related businesses in- (1996), and
cluding suppliers, service White Oak Semiconductor Partnership to locate a plant in Henrico County
providers, and other (1997).
agencies and institutions
to support the industry. In 2001, the General Assembly authorized additional grants for the expansion of sem-
iconductor production facilities in Manassas and Henrico. The expansion plans,
26
Data Center and Manufacturing Incentives
though, were delayed because of a downturn in the semiconductor chip market and
were never awarded. In 2005, the General Assembly revised the statutes authorizing
the grants to offer a custom grant for
Micron, which purchased Dominion Semiconductors in 2002, to expand the
Manassas plant and
Infineon, formerly White Oak Semiconductors Partnership, to expand the
Henrico plant.
In 2019, the General Assembly authorized another $70 million custom grant for the
expansion of Micron in Manassas. Because it is newly authorized, this grant is not
reviewed in this report.
Table 2-1
Virginia approved $195 million for semiconductor manufacturers but paid out
only $93 million in awards (FY96–FY17)
Capital
Year Job creation investment Maximum Amount
Custom grant project authorized requirement requirement award paid
Motorola (Goochland) 1996 -- $1.0B $60.0M $0.0M
Dominion Semiconductor LLC
1996 4,000 FTEs 4.0 38.4 18.6
(Manassas)
White Oak Semiconductor Partnership
1997 1,500 1.5 15.0 15.0
(Henrico)
Micron (Manassas) 2005 860 1.2 27.0 27.0
Infineon (Henrico) 2005 1,200 1.2 55.0 32.8
All custom projects 7,560 FTEs $8.9B $195.4M $93.4M
SOURCE: Weldon Cooper Center review of Chapters 642, 645, 651, and 652, 1996 Acts of Assembly; Chapters 789, 1997 Acts of Assembly;
Chapter 392, 2005 Acts of Assembly; and documentation provided by VEDP.
NOTE: Reflects year grants were authorized in statute by the General Assembly. FTEs, full time equivalents. Job creation requirements for
Dominion Semiconductor (1996) and White Oak Semiconductor reflect the requirements stated in the memorandum of understanding
between the company and VEDP. See Appendix G for more detail on the requirements and the payment schedule for the grants.
Virginia spent $93.4 million on the custom grants for semiconductor manufacturers
between FY96 and FY17, which is less than half of the total approved ($195 million).
No funding was paid out for the Motorola facility in Goochland County, which was
approved in 1995. The project was just underway when Motorola halted construction
because of declining market conditions, and the grant was canceled in 1998. Dominion
Semiconductors LLC in Manassas, which later was purchased by Micron, only met the
first milestone of its 1996 grant, receiving $18.6 million out of a potential $38.4 mil-
lion. The grant for Micron’s expansion in 2005 was paid in full. The Henrico County
plant, which underwent several different owners before shutting down, received full
payment for its first grant approved in 1997 but only partial payment for its 2005 grant.
Virginia only awarded $32.8 million of the $55 million maximum award for the Hen-
rico plant because the company (Qimonda) shut down the plant in 2009 because of
corporate insolvency. The grant’s final milestone was not met.
27
Data Center and Manufacturing Incentives
In addition to receiving the custom grants, the state also provided other incentives to
Qimonda and Micron. Qimonda received a $3 million Governor’s Development Op-
portunity Fund (now Commonwealth’s Opportunity Fund) grant. Micron received
$4.3 million in additional grants since 2005 including
a $1 million grant from the Virginia Investment Partnership;
two grants totaling $2.5 million from the Governor’s Development Oppor-
tunity Fund; and
two grants totaling $790,486 from the Virginia Jobs Investment Partnership
program.
Virginia adopted two sales and use tax exemptions for semiconductor
manufacturers resulting in $9 million in forgone revenue (FY10–FY17)
In 2006, the General Assembly also adopted two retail sales and use tax exemptions
for semiconductor manufacturers to extend the retail sales and use tax exemption for
purchases of manufacturing inputs to purchases of inputs used indirectly in the semi-
conductor manufacturing process (Table 2-2). Semiconductor manufacturers—like
any manufacturer—can claim the manufacturing exemption for purchases of goods
that are used directly in the manufacturing of semiconductor chips.
TABLE 2-2
Virginia offers two sales and use tax exemptions for semiconductor manufacturers
Exemption Purpose
Semiconductor Exempts equipment, fuel, power, energy, and supplies used primarily in the integrated
Manufacturers Exemption process or sub-process of designing, developing, manufacturing, or testing semiconductor
chips
Extends existing exempt status to items used before or after production, or in the cleanroom
(lab designed to maintain low levels of particulates and necessary to ensure quality and
integrity of product)
Semiconductor Wafer Allows all semiconductor wafers used by semiconductor manufacturers to be tax exempt, not
Exemption just those used directly in the manufacturing process
Extends exempt status to wafers used indirectly in production for cleaning, process control,
and testing
SOURCE: Weldon Cooper Center review of the Code of Virginia and agency documents.
NOTE: Authorized by § 58.1.-609.3(14)(15). Semiconductor and other manufacturers can exempt purchases of equipment, supplies, and
other tangible property used directly in the manufacturing process through the state’s manufacturing exemption.
It is estimated that the cost of the two sales and use tax exemptions in terms of for-
gone revenue has been relatively small, totaling only $9 million between FY10 and
FY17. Most of the forgone revenue comes from the semiconductor manufacturing
exemption, which is estimated to have resulted in $7.8 million in forgone revenue dur-
ing the period, or nearly $1 million per year, on average. The semiconductor wafer
exemption is estimated to have resulted in $1.1 million in forgone revenue during the
28
Data Center and Manufacturing Incentives
period, or $142,000 per year, on average. The estimates are small because very few
businesses in Virginia are in semiconductor manufacturing and are eligible for the ex-
emptions. Industry representatives reported that only a small portion of their total
purchases are related to indirect activities.
Virginia incentives for semiconductor manufacturers have had limited
effects on employment
The semiconductor manufacturing industry and related jobs—like all manufacturing
jobs—have been declining across the nation. Semiconductor production in the U.S.
made up 42 percent of global production in 1980 but fell to 30 percent in 1990 and to
13 percent in 2015. The decline comes from several challenges, including the rise of
global competitors, increasing costs of design and fabrication, technological change,
and difficulties in finding technical talent. U.S. production of semiconductors has off-
shored to countries such as Taiwan, Korea, and China, where producers benefit from
proximity to other electronic industries and from active government-sponsored indus-
trial policies that promote industry growth, including incentives and research sponsor-
ship.
Semiconductor manufacturing employment in Virginia declined more than
industry employment nationally
Virginia semiconductor manufacturing employment declined 70 percent between 2001
and 2017 (Figure 2-3). This decline is at a substantially greater rate than the 38 percent
decline nationwide, despite Virginia’s custom incentive grants. The decline in Virginia
was temporarily halted between 2003 and 2008 by the expansion of the plant in Hen-
rico but then sharply declined again after Qimonda closed.
The greater employment declines in Virginia partly reflect company specialization in
memory chips rather than microprocessors (like U.S. companies such as Intel). The
memory chip market is a more “commodified” market because chips between one
manufacturer and another can be easily substituted. The U.S. market for memory chips
also faces fierce international competition with Asian producers. While Virginia has
other businesses classified as semiconductor and related devices, they include only a
handful of companies with few employees.
29
Data Center and Manufacturing Incentives
Figure 2-3
Virginia semiconductor manufacturing employment has declined more than
industry employment nationally (2001–2017)
A synthetic control
group to represent Vir-
ginia without semicon-
ductor incentives was
constructed from states SOURCE: Weldon Cooper Center analysis of EMSI employment data for the semiconductor and related device man-
that did not have signifi- ufacturing sector for Virginia and U.S. states.
NOTE: Excludes employment in fabless design and engineering semiconductor firms.
cant semiconductor in-
centives during the time
period studied. Select Grants may have provided a short-term boost to semiconductor manufacturing
characteristics of states
employment
that made up the syn-
thetic control group The custom incentive grants may have provided a substantial short-term boost to sem-
were weighted so that iconductor manufacturing employment in Virginia. Between 2005 and 2009, Virginia
once combined, it
semiconductor manufacturing employment was higher than what it likely would have
closely resembled Vir-
ginia on those character- been without the incentives (Figure 2-4). This time period reflects the five-year per-
istics. (See Appendix M formance period agreed to in the MOU between the companies and VEDP. After
for research methods 2009, semiconductor manufacturing employment in Virginia resumed its decline in a
used for this analysis.)
pattern similar to states with comparable characteristics that did not award incentives.
30
Data Center and Manufacturing Incentives
Figure 2-4
Virginia semiconductor custom grants had short-term positive impacts on
semiconductor employment (2001–2017)
SOURCE: Weldon Cooper Center analysis of EMSI data for the semiconductor and related device manufacturing sector.
NOTE: Excludes employment in fabless design and engineering semiconductor firms.
While incentives are important to manufacturers, Virginia’s custom
grants may not have been sufficient to sway behavior
Semiconductor manufacturers have reported that financial incentives and subsidies
rate highly among site location factors. Tax advantages ranked first in international
location decisions, followed by the supply of an engineering and technical workforce
and the availability and reliability of utilities, according to a survey of 10 executives at
major semiconductor companies (Leachman and Leachman 2004). A study of the
semiconductor industry in the Kumamoto Prefecture of Japan indicates that financial
incentives and subsidies ranked fourth, preceded by workforce (first), proximity to
customers (second), and input costs (third). Representatives of Global Foundries in-
dicated that the large incentive package was a crucial component of the decision to
locate in upper New York, although workforce/education, existing semiconductor in-
dustry operations, R&D infrastructure, and stable geology were also important. Vir-
ginia’s custom incentives for semiconductor manufacturers, which have been substan-
tially less than incentives provided by many other states, represent a very small portion
of operating costs for the new semiconductor manufacturing activity and, alone, may
not be enough to sway decisions.
31
Data Center and Manufacturing Incentives
The scale estimating the Estimates indicate a relatively small portion of economic activity by
amount of economic ac- semiconductor manufacturers in Virginia is likely attributed to custom grants
tivity attributed to an in-
The custom grants—while sizable compared to the average Virginia incentive—rep-
centive is based on the
incentive amount as a resented only a small fraction of the total operating costs for the Micron (0.8 percent)
percentage of the busi- and Qimonda (0.7 percent) semiconductor plants over a 20-year period. These low
ness’s operating costs percentages are expected to induce less than 10 percent of Micron’s (8 percent) and
for the related activity
Qimonda’s (6 percent) economic activity, according to a scale developed by a leading
over a 20-year period.
The estimate is based on researcher of incentives (Bartik 2018b).
costs and does not ac-
The custom grants may not be big enough to compete with less expensive labor and
count for other factors
that may influence a capital overseas. The development of new large semiconductor fabrication plants has
business’s location or ex- become increasingly rare in the U.S., with most investment now occurring at existing
pansion decisions. See facilities. One industry source reports that building, equipping, and operating a fabri-
Appendix M [online
cation facility in the U.S. costs $1 billion more than a similar facility overseas. Virginia’s
only] for more detail on
the difficulty of precisely
custom grant payments to Micron and Qimonda totaled $45.6 million and $47.8 mil-
estimating incentives’ ef- lion, respectively.
fects and the methodol-
ogy used in this report. Exemptions for semiconductor manufacturers likely have had limited effect
The sales tax exemptions for semiconductor manufacturers also likely had a limited
effect on economic activity in the state. Studies of sales and use tax exemptions and
tax rates on manufacturing and other industry economic activity are limited and gen-
erally have not found a strong relationship between sales and use taxes and industry
employment. Two recent studies have attempted to measure the economic effects of
similar exemptions and have found little to no impact. One study examined the effect
of the state sales and use tax burden on machinery and materials and found they are
negatively associated with manufacturing capital expenditures and employment, but
the effect is relatively small (Hageman, Bobek, and Luna 2013). Another study using
county data across state borders finds that retail sales taxation levels for manufacturing
machinery and equipment has little or no effect on employment levels (Mikesell and
Ross 2017).
Economic benefits of semiconductor custom grants are moderate and
Economic impact
analysis of expenditures
are larger for Micron than for Qimonda
by custom grant recipi- The custom grants for Micron and Qimonda generated economic activity for Virginia.
ents between FY98 and
Each year employment increased by 287 jobs, Virginia GDP increased by $100 million,
FY17 was conducted us-
ing economic modeling and personal income increased by $58 million, on average, because of the grants be-
software developed by tween FY98 and FY17 (Table 2-3). These estimates assume that 8 percent of the ac-
REMI, Inc. tivity generated by Micron and 6 percent of the economic activity generated by
(See Appendix M [online Qimonda are attributable to the custom grants.
only] for the economic
impact analysis used in The economic benefits of both Micron and Qimonda were highest during site devel-
this study.) opment and construction. Once both companies were operational, the highest impacts
32
Data Center and Manufacturing Incentives
occurred when the companies were at their highest levels of employment (1,800 work-
Net impact is the in-
ers for Micron and 2,500 for Qimonda). The economic benefits were larger for Micron
crease in economic ac-
even though Qimonda was the larger project in terms of jobs created and capital in- tivity induced by the in-
vestment. Economic benefits generated by Qimonda over the 20-year period are lower, centive after adjusting
because the Qimonda plant shut down in 2009. for the opportunity cost
of increasing taxes to
The economic benefits generated by the custom grants are moderate compared to pay for the loan assis-
those generated by other incentives. The custom grant for Micron generated an esti- tance.
mated $30 million in Virginia GDP per $1 million spent on the grant, and the grant (See Appendix N [online
for Qimonda generated an estimated $14 million in Virginia GDP per $1 million spent. only] for information on
the total economic im-
These amounts are higher than the estimated Virginia GDP per $1 million spent for pact and the opportunity
tax incentives (with the exception of the data center exemption) reviewed in this report cost of increasing taxes.)
and in other reports in this series. These amounts are similar to the estimated benefits
generated collectively by all Virginia’s incentive grants per $1 million spent on grants.
(See Economic Incentive Grants 2018, JLARC, 2018.)
Table 2-3
Economic benefits of semiconductor custom grants are moderate and are
larger for Micron (FY96–FY17)
Annual average FY98–FY17
Total semiconductor
Micron Qimonda custom grants
Net impact to Virginia economy
Private employment 154 jobs 134 jobs 287 jobs
Virginia GDP $66.9 million $31.6 million $98.5 million
Personal income $38.9 million $19.2 million $58.1 million
Impact to Virginia economy per $1 million of grants
Private employment 73 jobs 62 jobs 67 jobs
Virginia GDP $30.1 million $14.0 million $21.9 million
Personal income $17.8 million $8.9 million $13.3 million
Impact to state revenue
Total revenue $2.2 million $1.2 million $3.4 million
Grant awards $2.3 million $2.4 million $4.7 million
Revenue net of awards ($0.1 million) ($1.2 million) ($1.3 million)
Return in revenue 97¢ for every $1 spent 49¢ for every $1 spent 72¢ for every $1 spent
SOURCE: Weldon Cooper Center economic impact analysis of amount of credits claimed FY98–FY17.
NOTE: Includes direct, indirect, and induced impacts. Gross impact on Virginia’s economy is used to calculate impact per $1 million in incentive
awards. This is consistent with how the economic development research literature typically calculates these impacts. (See Appendix N [online
only] for detailed results on total impact of the custom grants, impact of raising income taxes by the amount of the grants [opportunity cost],
and revenue generated by source.)
The return in revenue from both custom grants is also moderate, with Micron again
yielding a higher return. The return in revenue for every $1 spent on the Micron cus-
tom grants was 97¢ annually, on average, and the return in revenue for the Qimonda
custom grants was 49¢. These returns in revenue are similar to the returns (55¢) for all
grants collectively per $1 dollar in total grant spending. (See Economic Incentive Grants
2018, JLARC, 2018.)
33
Data Center and Manufacturing Incentives
Because the custom grants are well-targeted to projects expected to have large eco-
nomic benefits, these economic benefits and the returns in revenue should be higher
than those across all incentive grants. However, the custom grants are substantially
costlier on a per-job basis (approximately $21,457 per job for both Qimonda and Mi-
cron) than the cost per job ($4,189) for the typical incentive grant provided by Vir-
ginia’s programs.
Economic benefits of semiconductor exemptions are low and returns
in revenue are negligible
The semiconductor manufacturers and wafers exemptions generate relatively low eco-
nomic benefits and returns in revenue to the state, according to estimates. These ex-
emptions on average generated an additional 43 jobs, $7 million in Virginia GDP, and
$4.3 million in personal income annually, on average, between FY10 and FY17 (Table
2-4). These economic benefits are low relative to the economic benefits generated by
other incentives. The exemptions are estimated to generate an additional $7.1 million
in Virginia GDP per $1 million spent on the exemptions. This amount is less than the
$14.6 million in Virginia GDP generated across all incentive grants per $1 million in
grant spending. (See Economic Development Incentive Grants 2018, JLARC, 2018.) How-
ever, this amount is higher than the additional Virginia GDP generated per $1 million
spent by most other tax exemptions and credits that have been reviewed to date in this
series of reports. One factor likely limiting the economic benefits of the sales tax ex-
emptions for semiconductors is that, unlike the data center exemption and many in-
centive grants, eligibility is not contingent on the companies achieving certain levels
of job creation and capital investment.
The return in state revenue for the exemptions, like most other tax incentives reviewed
in this series, is negligible. For every $1 spent on the semiconductor manufacturing
exemption, the return in state revenue is 5¢. The return in state revenue for the semi-
conductor wafers exemption is slightly higher at 12¢ per $1 dollar spent.
34
Data Center and Manufacturing Incentives
Table 2-4
Economic benefits of semiconductor manufacturing exemption and wafers exemption are low
(FY10–FY17)
Annual average FY10–FY17
Semiconductor
Manufacturing Semiconductor Wafer Total semiconductor
Exemption Exemption exemptions
Net impact to Virginia economy
Private employment 36 jobs 7 jobs 43 jobs
Virginia GDP $5.8 million $1.2 million $7.0 million
Personal income $3.6 million $0.7 million $4.3 million
Impact to Virginia economy per $1 million of exemptions
Private employment 45 jobs 57 jobs 47 jobs
Virginia GDP $6.8 million $9.2 million $7.1 million
Personal income $4.5 million $5.9 million $4.7 million
Impact to state revenue
Total revenue $47,359 $16,878 $64,237
Spending on exemption $975,760 $142,357 $1,118,117
Revenue net of awards ($928,402) ($125,479) ($1,053,881)
Return in revenue 5¢ for every $1 spent 12¢ for every $1 spent 6¢ for every $1 spent
SOURCE: Weldon Cooper Center economic impact analysis of amount of exemption spending FY10–FY17.
NOTE: Includes direct, indirect, and induced impacts. Gross impact on Virginia’s economy is used to calculate impact per $1 million in ex-
emption spending. This is consistent with how the economic development research literature typically calculates these impacts. (See Appen-
dix N [online only] for detailed results on total impact of the exemptions, impact of raising income taxes by the amount of the exemptions
[opportunity cost], and revenue generated by source.)
Exemptions have little effect on semiconductor activity but may
create more efficient tax
Even though the semiconductor manufacturers and wafer exemptions likely have little
effect on semiconductor activity in the state and have low economic benefits and re-
turns in state revenue, the exemptions appear warranted to create a more efficient tax.
Purchases of machinery and equipment directly used in manufacturing industries and
for research and design have been exempted since the sales tax was adopted in Virginia.
These exemptions expand this tax treatment to purchases of equipment and supplies
that are not in the final product, but are arguably important for development of quality
final products. From a tax policy perspective, sales and use taxes are generally consid-
ered by economists as most efficient when taxing only final sales for personal con-
sumption. These exemptions, however, should be considered for elimination if its
main beneficiary, Micron, ceases operation in the state.
Virginia should develop principles for providing custom grants
Virginia provided semiconductor manufacturers with custom grants in the 1990s and
in 2005 with hopes of establishing a semiconductor industry cluster—the “Silicon
Dominion”—in the state. At one time, three semiconductor manufacturers and mul-
tiple suppliers had a presence in the state, but ultimately an industry cluster was never
developed and only one major company, Micron, remains. Based on reviews of the
35
Data Center and Manufacturing Incentives
process for awarding the custom grants to Micron and Qimonda (or their predeces-
sors) in the 1990s and 2005 and the economic activity generated by the companies on
a long-term basis, several principles were identified that should be used for awarding
future custom grants. These principles could improve the effectiveness (short- and
long-term) of future custom grants offered by Virginia.
Ensure custom grants generally align to state or regional target industry
clusters and strategic goals
Economic incentives are likely to be most effective when they leverage investment in
an industry targeted by a region in its economic development strategy. However, it is
unclear that there is deliberate alignment between custom grants and targeted indus-
tries during the decision-making process in Virginia, based on attendance of legislative
meetings and discussions with legislative staff.
The Major Employment and Investment (MEI) Project Approval Commission, which
reviews and approves custom grants, could assess the extent to which proposed grants
align with state and regional targeted industry sectors and strategic plans. VEDP pro-
vides one or more presentations about the project and its economic impact to the MEI
commission. VEDP’s internal project review process considers whether projects align
with strategic industry sectors and state and local strategies, according to agency doc-
uments. The MEI Commission could direct VEDP to include in its presentation a
discussion of how the custom grant aligns with state and regional targeted industries
and strategic plans, or provide other justification for the necessity of providing custom
incentives. The MEI Commission could then include this information, at least at a
basic level to avoid revealing company proprietary information or confidential strate-
gic plan details, in its annual report.
RECOMMENDATION 4
The Virginia Economic Development Partnership should report to the Major Em-
ployment and Investment Project Approval Commission how each custom grant being
considered by the commission aligns with state and regional targeted industries and
strategic plans.
RECOMMENDATION 5
The General Assembly may wish to consider amending § 30-312 of the Code of Vir-
ginia to require that information on how custom grants align with state and regional
target industries and strategic economic development plans be included in the annual
report of the Major Employment and Investment Project Approval Commission.
36
Data Center and Manufacturing Incentives
Ensure custom grants are matched with local incentives
With the escalating capital investments required to open new facilities or expand ex-
isting ones and the attractiveness of high-tech development, “megadeal” packages of-
ten are accompanied by local incentives. Local incentives serve as a signal to the state
that a locality is a willing economic development partner.
The local incentives provided to Micron and Qimonda made up a large part of the
total incentive package from Virginia, and are estimated to have doubled Virginia’s
ability to encourage the companies to locate and expand in the state. Both companies
received large local incentives. Micron received an estimated $93 million in local incen-
tives over the 20-year period, two-thirds of the total incentive value, and Qimonda
received an estimated $27 million, one-third of the total incentive value. These local
incentives increase the percentage of operating costs covered to 2.3 percent for Mi-
cron and 1.0 percent for Qimonda, which is estimated to increase the amount of eco-
nomic activity that can be attributed to the incentives to 20.7 percent for Micron and
to 9.6 percent for Qimonda. (The economic activity that can be attributed to the state
incentives alone is 8 percent for Micron and 6 percent for Qimonda.)
The economic benefits and returns in revenue of the custom grants for Micron and
Qimonda to the state are estimated to be substantially larger when accounting for the
fact that the local communities made a sizeable contribution to the total incentive
packages that attracted the companies. The local incentives are estimated to have more
than doubled the chances of attracting the businesses, thereby doubling the estimated
economic benefits and returns in revenue to the state that can be attributed to the
grants (Table 2-5). Thus, local contributions play an important part in determining the
impact on state revenue of the grants at current levels of support ($19,176 per job for
Qimonda and $24,756 per job for Micron compared to a Virginia average incentive of
$4,189).
For these custom grant projects, ensuring localities’ participation seems warranted to
increase the chances of securing a location or expansion decision, increase the eco-
nomic benefits to the state, and help offset the relatively high cost of providing incen-
tives. Many other grant-funded projects in Virginia already receive local incentives. The
Commonwealth’s Opportunity Fund grant requires a 100 percent local match. Several
projects in Virginia that received custom grants have also received local incentives,
including Micron, Amazon, Newport News Shipbuilding, and Rolls-Royce. By nature
of the scale of the custom projects, it is likely that the local government will also
provide monetary or in-kind incentives to help attract the project to its area. To ensure
local participation, the MEI Commission could consider establishing a policy that re-
quires custom grants to be accompanied by local incentives, either monetary or in-
kind.
37
Data Center and Manufacturing Incentives
Table 2-5
Economic benefits of semiconductor custom grants are higher when local incentives are
included
Annual average FY98–FY17
Total semiconductor
Micron Qimonda custom grants
Net impact to Virginia economy
Private employment 650 jobs 209 jobs 859 jobs
Virginia GDP $66.9 million $31.6 million $235.7 million
Personal income $38.9 million $19.2 million $139.8 million
Impact to Virginia economy per $1 million of grants
Private employment 290 jobs 93 jobs 190 jobs
Virginia GDP $82.9 million $21.2 million $51.4 million
Personal income $49.0 million $13.4 million $30.9 million
Impact to state revenue
Total revenue $6.1 million $1.8 million $7.9 million
Grant awards $2.3 million $2.4 million $4.7 million
Revenue net of awards $3.8 million ($0.6 million) ($3.2 million)
Return in revenue $2.66 for every $1 spent 74¢ for every $1 spent $1.68 for every $1 spent
SOURCE: Weldon Cooper Center economic impact analysis of amount of credits claimed FY98–FY17.
NOTE: Includes direct, indirect, and induced impacts. Gross impact on Virginia’s economy is used to calculate impact per $1 million in incentive
awards. This is consistent with how the economic development research typically calculates these impacts. (See Appendix N [online only] for
detailed results on total impact of the custom grants, impact of raising income taxes by the amount of the grants [opportunity cost], and
revenue generated by source.)
38
Data Center and Manufacturing Incentives
3. Manufacturing Single Sales Apportionment A uniform apportion-
ment formula was devel-
Following the lead of many states, the General Assembly adopted legislation in 2009 oped by a group of state
to allow manufacturers to use a single sales income tax apportionment formula that tax professionals to en-
could reduce their Virginia income tax liability. Multistate manufacturers are typically sure that states do not
tax more than their share
required to use a three-factor apportionment formula based on the proportion of their of income, and thus com-
total property, payroll, and sales in Virginia. This formula double-weights the sales ply with the U.S. Com-
factor, and corporations must use this “double-weighted sales” formula unless other- merce Clause require-
wise specified in statute. The single sales formula allows manufacturers to calculate ment for fair apportion-
ment. The formula appor-
their taxable income in Virginia based solely on the proportion of total sales that are
tions income to states ac-
in Virginia. This formula allows manufacturers that have a sizable proportion of prop- cording to where prod-
erty and payroll in Virginia and a smaller proportion of sales to lower their tax liability ucts are made (where
(Figure 3-1). Thus, it is considered a tool to attract businesses to locate and expand in property and payroll are
located) and where prod-
the state. Single sales apportionment was phased in from July 2011 to July 2014 with
ucts are sold, with each of
increasing weights for the sales factor each year until July 2014, when companies could the three factors (prop-
use just sales (Table 3-1). erty, payroll, and sales)
equally weighted. Nearly
all states adopted the for-
Figure 3-1 mula. States began to
Single-sales apportionment allows manufacturers to potentially reduce their double-weight the sales
Virginia tax income factor to equalize the
weight given to supply
(property and payroll)
and demand (sales). Later,
states began to use only
sales in the formula.
SOURCE: Weldon Cooper Center review of the Code of Virginia, Virginia Administrative Code, and TAX forms and
guidance documents.
39
NOTE: Beneficiary savings for 2017 are estimated by inflating 2016 savings.
40
Virginia adopted a single sales apportionment option for
manufacturers to maintain industry employment
Virginia’s manufacturing industry, which is relatively small, has declined at a much faster
Virginia’s manufacturing
rate—34 percent—than the 26.5 percent decline nationwide over the past two decades industry is relatively
(1998 to 2017). Virginia’s manufacturing employment declined for reasons similar to the small, representing 4.8
decline in national employment. Many overseas manufacturers, particularly in China and percent of total employ-
other Asian countries, have low labor costs. In addition, technological improvements ment in the state com-
pared with 6.8 percent
have reduced the demand for U.S. labor. Cyclical forces (a lack of available jobs because nationwide.
of decreased demand) contributed to a sharper reduction in Virginia manufacturing em-
ployment during and immediately following the 2007–09 recession. Much of the em-
ployment loss was in lower-wage industries—textiles, apparel, furniture, and wood prod-
ucts—that were more concentrated in Virginia than in other states. The textile, apparel,
and furniture industries faced increasing competition from lower price imports. The
wood products industry was affected by the downturn in the housing market and con-
struction activity during the recession and the slow recovery afterward.
TABLE 3-1
Manufacturers in Virginia may elect to use a single sales apportionment formula
Purpose Incentivize the state’s manufacturing industry and slow the decline of manufacturing jobs
Eligible Corporations classified under the traditional manufacturing NAICS codes of 31, 32, and 33
businesses and the agriculture, forestry, fishing and hunting NAICS code of 11
Maintain employment levels at 90% of the base year (year elected to use single sales
apportionment) for 3 years
Pay employees at least as much as the state or local average weekly wage for the company’s
industry, whichever is higher, for 3 years
Program features Must use the single sales apportionment for 3 years after electing to use it
- Must maintain employment and wages at levels described or face recapture of the
difference between tax liability calculated under single sales apportionment and tax
liability under double-weighted sales apportionment plus interest
- To avoid recapture and interest in a tax year, company can notify the Department of
Taxation (TAX) that it expects to not meet the wage or employment requirements and
instead use double-weighted sales apportionment
After the 3 years, not required to maintain the employment and wage requirements; can
continue to use the single sales apportionment or switch back to double-weighted sales
Applies to corporations that pay corporate income taxes and businesses organized as
pass-through entities whose owners pay individual income taxes
Use of incentive Allow manufacturers to reduce tax liability by choosing to use a more beneficial
apportionment method, particularly if they have high proportions of property and payroll
in the state relative to their sales in the state
SOURCE: Weldon Cooper Center review of the Code of Virginia and agency documents.
NOTE: NAICS, North American Industrial Classification System. The elective single sales apportionment statute in-
cluded a 10 percent penalty for not meeting the requirements and required businesses to maintain base year em-
ployment (at 100 percent). The penalty was removed and the maintenance level reduced to 90 percent during the
2012 legislative session.
41
Data Center and Manufacturing Incentives
Virginia adopted the elective single sales apportionment for manufacturers to slow the
Estimates of forgone decline in manufacturing employment. The goals were to (1) incentivize Virginia manu-
revenue begin in tax
year 2014 because it is
facturers to remain in the state rather than moving operations to a state with a more
the first year that single favorable tax system and (2) attract additional manufacturers to the state. Many other
sales apportionment was states were beginning to use a single sales factor formula at the time of Virginia’s adop-
used. In previous years, tion, usually for all corporations rather than for certain industries, such as manufacturers.
sales were triple and
quadruple weighted.
Elective single sales apportionment for manufacturers resulted in $68
Estimates do not include
forgone revenue from million in forgone revenue 2014–2017
pass-through entities
The elective single sales apportionment resulted in an estimated $68 million in forgone
whose owners pay indi-
vidual income taxes. For- revenue between tax year 2014 and tax year 2017. Manufacturers electing to use single
gone revenue for these sales apportionment increased substantially between 2014 and 2015, from 16 to 76
businesses is expected to taxpayers. As a result, forgone revenue increased from $1.9 million to $20.9 million.
be small, based on a pre- Forgone revenue in tax year 2016 was slightly higher at $22.3 million, but the number
vious JLARC report (Re-
view of Virginia’s Corpo-
of taxpayers was lower (70).
rate Tax System, JLARC, Manufacturers electing to use single sales apportionment typically included larger em-
2010).
ployers from certain sectors. More than 50 percent of the forgone revenue is from
manufacturers with 500 or more employees. They are typically concentrated in certain
manufacturing sectors: more than 80 percent of the forgone revenue is from manu-
facturers in the machinery, food, and plastics and rubber products manufacturing sub-
sectors (Figure 3-2).
Figure 3-2
Majority of forgone revenue was for large manufacturers and manufacturers
producing machinery, food, and plastics and rubber products (2016)
SOURCE: Weldon Cooper Center analysis of corporate income tax returns for tax year 2016 and Virginia Employment
Commission ES202 data.
Forgone revenue could increase in future years. Several manufacturers that elected to
use single sales apportionment appear to have reverted back to using the standard
“double-weighted” apportionment in some years within the three-year period after the
42
Data Center and Manufacturing Incentives
election. This reversion is allowed to avoid the process of recapture, in which the
manufacturer would have to pay the difference between tax liability using single sales
and double-weighted sales plus interest, if the company does not expect to meet the
wage or job requirements. Once this three-year period is up, these manufacturers can
use single sales apportionment without needing to meet job or wage requirements. It
would take only the addition of a small number of highly profitable manufacturers
using single sales apportionment to have a large effect on forgone revenue.
Single sales apportionment for manufacturers has had mixed impacts
Employment levels of
Some manufacturers interviewed for this report said the elective single sales appor- manufacturers that
tionment formula played a role in their decisions to locate in or to remain in Virginia elected to use single
and preserve jobs. The single sales formula also offsets some negative features of state sales apportionment in
and local government taxation of capital. This includes local machinery and tools taxes tax year 2014 and used
single sales apportion-
in Virginia, which make the state less competitive according to state rankings on busi- ment for tax years 2014–
ness taxation (e.g., Tax Foundation State Business Tax Climate Index and Location Matters). 2016 (2014 cohort) were
Further review, however, indicates that the impact of elective single sales apportion- tracked from 2013 (the
ment is mixed. base year) to 2017. Em-
ployment records for 11
of the 13 firms in the co-
Manufacturers using single sales apportionment had more employment growth, hort were identified.
but some growth likely would have occurred without incentive
Collectively, the employment levels of Virginia manufacturers using single sales appor-
tionment in tax year 2014 grew by 16.3 percent (or 316 jobs) between 2013 and 2017.
This growth rate is substantially greater than the 1.5 percent growth rate of total man- The scale estimating the
ufacturing employment in Virginia over the same period (Figure 3-3). Nine of the 11 amount of economic ac-
manufacturers added jobs, and all others maintained employment of at least 90 percent tivity attributed to an in-
centive is based on the
of the baseline (2013) employment levels. Only one firm dropped below the baseline
incentive amount as a
by 2017. Further statistical analysis indicates that early single sales apportionment elec- percentage of the busi-
tors added significantly more jobs than other multistate manufacturers. (See Appendix ness’s operating costs
M, available online, for more information on the regression analysis.) for the related activity
over a 20-year period.
However, manufacturers that elected to use single sales apportionment may have The estimate is based on
grown anyway. Therefore, not all of the job growth can be attributed to single sales costs and does not ac-
apportionment. Manufacturers must maintain employment levels for at least three count for other factors
that may influence a
years, so manufacturers that expected to grow were more likely to use it than those business’s location or ex-
that are contracting. Estimates indicate that elective single sales apportionment may pansion decisions. See
induce a modest portion of the economic activity generated by manufacturers making Appendix M [online
the election. Tax savings for electors represent 1.7 percent of their estimated costs for only] for more detail on
precisely estimating in-
the new economic activity, which is expected to induce 15 percent of the economic centives’ effects and the
activity generated (Bartik 2018b). (See Appendix M for the methodology used in esti- methodology used in
mating the percentage of economic activity influenced by incentives.) this report.
43
Data Center and Manufacturing Incentives
Figure 3-3
Employment of manufacturers electing single sales apportionment grew more
than Virginia manufacturers’ overall employment (2013–2017)
SOURCE: Weldon Cooper Center analysis of corporate income tax returns and Virginia Employment Commission
ES202 employment records.
NOTE: SSF, single sales factor. SSF electors are those in the 2014 cohort.
Virginia manufacturing growth is slightly less than national growth despite
adoption of single sales apportionment
Even though elective single sales apportionment has had a positive effect on manufac-
turing employment in the state, Virginia’s manufacturing growth is still slightly less
than the nation. After manufacturing single sales apportionment was fully imple-
mented in 2014, Virginia’s manufacturing industry grew 2.6 percent between 2013 and
2016 compared with 4.3 percent nationally (Figure 3-4).
Only a small portion of manufacturers in Virginia are using single sales apportion-
ment. Electors account for only 70 tax returns out of approximately 6,000 manufac-
turers in 2016, though several returns contained multiple businesses. In addition, man-
ufacturers using single sales apportionment in tax year 2016 employed 12,641 workers,
or just 5 percent of the state’s manufacturing jobs. The majority of the benefit is also
used by a few large taxpayers, with five taxpayers, out of the 70 returns, in tax year
2016 receiving 61 percent of the total savings. These five taxpayers employed approx-
imately 28 percent of the total employment of all manufacturers that elected to use
single sales apportionment.
44
Data Center and Manufacturing Incentives
Figure 3-4
Virginia manufacturing employment grew slightly less than the nation after the
recession, even after adoption of single sales apportionment
SOURCE: Weldon Cooper Center analysis of Bureau of Economic Analysis employment data for Virginia and the
United States.
It is unlikely that the single sales apportionment for manufacturers will encourage em-
ployment growth in the future. Most existing Virginia manufacturers that are likely to
use the single sales formula are probably already taking advantage of it, according to
stakeholders and several companies using it. Manufacturers may have difficulty main-
taining employment because of technology changes, difficulties recruiting skilled
workers in tight labor markets, the fading appeal of manufacturing jobs to younger
workers, and retirements. Some manufacturers in rural regions may face difficulty
meeting the wage requirement, which is based on industry average wages, even if they
pay substantially more than the average wage for the region. Virginia manufacturers
that have elected not to use single sales apportionment also have lower proportions of
property (4 percent) and payroll (6 percent) than manufacturers that have elected to
use it (32 percent and 24 percent, respectively). These manufacturers would not bene-
fit, or would achieve only minimal benefits, from using single sales apportionment.
However, some manufacturers relocating to Virginia may choose to use single sales
apportionment.
Research suggests increased sales factor weighting has less effect than
previously indicated
In theory, a heavier weighting of the sales factor in the apportionment formula should
stimulate state economic activity in many circumstances. It should decrease the tax
burden of multistate businesses that export a high share of their product out-of-state
45
Data Center and Manufacturing Incentives
Empirical studies of the
but have a strong payroll and property presence in the state. It may also encourage
effects of single sales ap-
portionment have gener- businesses that sell a large number of goods in Virginia but that produce their goods
ally focused on four types primarily outside the state to increase their state production footprint.
of economic impacts: (1)
employment, (2) capital Advocates of single sales apportionment frequently touted two early studies from 2000
expenditures, (3) sales or and 2002 that showed a substantial positive effect of sales reapportionment on man-
shipments, and (4) tax ufacturing employment. These studies found that states that moved from an equally
revenues.
weighted formula to a single sales factor formula would see a 1.1 percent increase in
manufacturing employment (Goolsbee and Maydew 2000a) and a 2.0 percent increase
in firm property and payroll (Edmiston 2002). Subsequent studies indicate that states
would see significant manufacturing job growth from switching to single sales factor,
with New York experiencing a 3.5 percent increase in manufacturing jobs as a result
The Goolsbee and
of moving from double-weighted sales to single sales apportionment (Goolsbee and
Maydew (2000a) and Ed-
miston (2002) studies Maydew 2000b).
were presented as evi-
More recent studies have generally found smaller or no effects. Later replications of
dence for adopting single
sales apportionment to the earlier studies have not corroborated their findings. One study replicated the anal-
Virginia’s Joint Subcom- ysis for the same time period (1978–1993) and found that the relationship between
mittee Studying Benefits sales apportionment and manufacturing employment was not statistically significant
of Adopting a Single
(Bernthal et al 2012). The study also used a newer data set (1978–2010) and found a
Sales Factor for Corporate
Income Tax Purposes in statistically significant relationship, but the magnitude of the effect was just one-fourth
2008. of the size previously reported. Another replication study (Merriman 2014) suggests
that the previous results are not reflected if new sample data or improved economet-
rics methods are used. That study and several others have failed to find empirical sup-
port for greater sales weighting. (See Appendix O, available online, for more detail of
each study.)
There are at least two possible explanations for the differences between the earlier and
more recent studies. First, many earlier studies relied on a relatively basic representa-
tion of the corporate tax code and used older econometric methods. Since then, many
researchers have concluded that other factors affect corporate tax liability, such as tax
rates, combined reporting rules, depreciation rules, and tax base definition (Serrato
and Zidar 2017; Moore and Bruce 2014). After accounting for these factors and cor-
recting for recognized econometric problems, the effects of sales factor weighting ap-
pear smaller in magnitude or statistically insignificant. Second, many of the earlier
studies were conducted when single sales apportionment was less prevalent, and as
with any incentive, first adopters experience the largest impacts. As more states adopt
the incentive, net gains to later adopting states dissipate (Edmiston 2002).
Single sales apportionment for manufacturers has moderate
economic benefits and moderate return in state revenue
The elective single sales factor apportionment for manufacturers has generated eco-
nomic activity for the state economy. It is estimated that each year private employment
increased by 126 jobs, Virginia GDP increased by $23.9 million, and personal income
increased by $8.8 million because of elective single sales apportionment (Table 3-2).
46
Data Center and Manufacturing Incentives
These findings assume that 15 percent of the economic activity can be attributed to
elective single sales apportionment.
TABLE 3-2
Elective single sales apportionment for manufacturers has moderate benefit
Economic impact
and moderate return in revenue to the state
analysis of taxpayer ex-
Annual average penditures that elected
FY14–FY17 to use single sales ap-
Net impact to Virginia economy portionment in tax year
Private employment 126 jobs 2014 (2014 cohort) be-
Virginia GDP $23.9 million tween FY14 and FY17
Personal income $8.8 million was conducted using
Impact to Virginia economy per $1 million of forgone revenue economic modeling soft-
Private employment 98 jobs ware developed by REMI,
Virginia GDP $17.6 million Inc.
Personal income $6.9 million (See Appendix M [online
Impact to state revenue only] for the economic
Total revenue $0.8 million impact analysis used in
Forgone revenue because of single sales election $1.4 million this study.)
Revenue net of awards ($0.7 million)
Return in revenue 54¢ for every $1 spent
SOURCE: Weldon Cooper Center economic impact analysis of taxpayers that elected to use the single sales appor-
tionment for manufacturers in tax year 2014 (2014 cohort) over the time period FY14–FY17.
NOTE: Includes direct, indirect, and induced impacts. Estimates assume that 15 percent of the 316 jobs created by
the 2014 cohort can be attributed to elective single sales apportionment. Gross impact on Virginia’s economy is used
to calculate impact per $1 million in incentive awards. This is consistent with how the economic development research
literature typically calculates these impacts. (See Appendix N [online only] for detailed results on total impact of the
incentive, impact of raising income taxes by the amount of the incentive [opportunity cost], and revenue generated
by source.)
These economic benefits are moderate relative to the economic benefits generated by
other incentives. Elective single sales apportionment for manufacturers is estimated to
generate $17.6 million in additional Virginia GDP for every $1 million in forgone rev- Net impact is the
increase in economic
enue (Table 3-2). This is higher than estimates of additional Virginia GDP per $1 activity induced by the
million spent for many other incentives included in this report and similar to estimates incentive, adjusted for
of additional Virginia GDP per $1 million in spending generated by all grant programs. the opportunity cost of
(See Economic Development Incentives 2018, JLARC 2018.) Like the data center exemption, increasing taxes to pay
for the incentive.
this estimate is also higher than estimates for other tax incentives.
(See Appendix N [online
The return in revenue to the state is also moderate at 54¢ for every $1 in forgone only] for information on
revenue because of manufacturers using the elective apportionment. This return is the total economic im-
substantially higher than the return for other tax incentives and similar to the return pact and the opportunity
cost of increasing taxes.)
(55¢) for all grant programs. (See Economic Development Incentives 2018 JLARC, 2018.)
The economic benefits are moderate because elective single sales apportionment is
relatively well-targeted to businesses that should generate higher economic benefits.
Only multistate businesses apportion their income among states, therefore the tax in-
47
Data Center and Manufacturing Incentives
centive is targeted to businesses that likely have greater opportunities to locate else-
where. The elective apportionment also effectively targets manufacturers that export
a majority of their product outside of the state. The higher the sales outside the state,
the higher the amounts of new revenue brought into the state economy. Therefore,
the elective single sales apportionment for manufacturers is probably one of the most
effective incentive programs at targeting exporters. Many of Virginia’s grant programs
benefit exporters, but the elective single sales apportionment is the only one where the
benefit increases as the export percentage increases.
Several factors limit the benefits of the elective single sales apportionment. It does not
incentivize new investment or job creation like incentives for other industries, such as
Virginia’s single sales apportionment for data centers. Some states, such as Florida and
Oklahoma, incentivize new activity in their single factor apportionment formula re-
quirements. Virginia’s elective single sales apportionment for manufacturers instead
incentivizes job maintenance, but only during the three years after manufacturers elect
to use single sales apportionment.
The estimated cost of single sales apportionment per job ($17,939 per job) is more
than four times the cost per job across all incentive programs ($4,189 per job). (See
Economic Development Incentives 2018, JLARC 2018.) Only the Transportation Partner-
ship Opportunity Fund ($48,348), Advanced Shipbuilding Training Facility Grant
($32,778), and the semiconductor grants and data center exemption reviewed in this
report have higher costs per job. Manufacturers that have elected to use single sales
apportionment have also tended to be in lower-wage sectors, paying an average wage
of $67,362 compared to an average wage of $70,244 for all manufacturers, based on
average wages reported by the Bureau of Economic Analysis for 2017.
Changes to elective single sales apportionment should be part of
broader tax policy decision
The elective single sales factor apportionment has been somewhat effective. Manufac-
turers using it have experienced higher rates of employment growth than manufactur-
ers using standard apportionment. However, their employment growth has not been
large enough to bring Virginia more in line with manufacturing growth nationally
(based on the 50-state average), and it seems unlikely that many more manufacturers
will choose to use it. Still, it has generated moderate economic benefits and returns in
state revenue relative to other incentives.
It may be difficult to improve its effectiveness without negatively affecting its eco-
nomic benefits and return in revenue to the state. In order to incentivize more manu-
facturers to use it, the job and wage requirements would likely need to be removed.
This change would allow manufacturers with declining employment levels and lower
wages to use it, which would reduce the economic benefits and return in revenue gen-
erated.
48
Data Center and Manufacturing Incentives
Instead, changes to Virginia’s elective single sales apportionment would be better made
as part of changes to the state’s broader income tax apportionment policy. Today, most
states are using single sales apportionment, or phasing it in, and almost all apply it
uniformly across industries and make it mandatory. (See Appendix D for more detail
about apportionment factors by state.) Adopting single sales apportionment across all
industries and making it mandatory would provide more equitable treatment across
taxpayers and simplify administrative and compliance activities on behalf of both the
Department of Taxation and the taxpayer. By using methods consistent with other
states, Virginia may reduce the potential for income shifting to states that use more
favorable apportionment rules and also ensure that these rules are not viewed as a
detriment to businesses location decisions.
These changes would reduce state revenue generated by the corporate income tax.
However, keeping it optional would reduce revenue by a substantially larger amount.
In 2008, when Virginia originally considered adopting single sales apportionment more
broadly, the Department of Taxation estimated that uniformly applying mandatory
single sales apportionment to all industries would reduce state revenue by $47 million.
Making it optional would reduce revenue by $123 million because taxpayers would
choose the method that would most reduce their tax liability.
49
Data Center and Manufacturing Incentives
NOTE: Beneficiaries for the tax credits are the number of tax returns claiming the credit and not the count of the businesses claiming it.
50
Data Center and Manufacturing Incentives
4. Environmental Incentives for Manufacturers
Virginia offers four incentives to encourage economic activity that is environmentally
friendly or to compensate facilities required to control or abate pollution. These in-
centives include the Pollution Control Equipment and Facilities Tax Exemption and
three tax credits—the Green Jobs Creation Tax Credit, Biodiesel and Green Diesel
Fuels Producers Tax Credit, and Recyclable Materials Processing Equipment Tax
Credit. These incentives are all administered by the Department of Taxation (TAX),
and three require certification by the Department of Environmental Quality (DEQ)
or another agency to verify eligibility. These incentives are relatively small compared
to other incentives reviewed in this report. In total, they represent less than 2 percent
of Virginia’s total spending on incentives in FY17.
Pollution Control Equipment and Facilities Sales Tax Exemption
reduces costs of regulatory compliance
Virginia adopted a retail sales and use exemption for pollution control equipment and The 1972 legislation ex-
facilities in 1972 (Table 4-1). Initially manufacturers and other businesses were ex- empted certified pollu-
tion control equipment
empted from paying sales taxes on purchases of property and equipment used primar-
from local taxation, in-
ily to abate or prevent air or water pollution. The property or equipment had to be cluding local property
certified by the appropriate state agency to be eligible for the exemption. The exemp- taxes. A provision of this
tion has been expanded several times to additional industries: legislation also ex-
empted pollution control
1995: coal, oil, and gas production and waste disposal equipment from all state
taxes, including sales
2003: forestry and other vegetative waste products
taxes. This report focuses
2006: landfill gas and recovery on the sales tax exemp-
2016: property used to generate solar and wind energy. tion.
Table 4-1
Virginia provides a Pollution Control Equipment and Facilities Sales Tax Exemption
Purpose Reduce financial burden of complying with federal environmental regulations and encourage use of
pollution control measures
Eligibility Purchase of real or personal property, equipment, facilities, or devices used primarily for the purpose
of abating or preventing pollution
Purchasers may include either the manufacturer or a contractor or subcontractor that installs the
equipment for use by the manufacturer
Use of Lowers the capital costs associated with installing required pollution control equipment
exemption
SOURCE: Weldon Cooper Center analysis of the Code of Virginia and agency documents.
NOTE: Section § 58.1-609.3 exempts pollution control equipment as defined in § 58.1-3660 (local property tax exemption) from the retail
sales and use tax. Before adoption of the exemption, TAX administratively exempted pollution control measures from the sales tax under
the manufacturing exemption if the equipment and facility was used directly in the manufacturing process.
51
Data Center and Manufacturing Incentives
Certifying authorities include DEQ on behalf of the State Water Control Board, State
Air Pollution Control Board, and Virginia Waste Management Board; the Department
of Mines, Minerals, and Energy; and “any interstate agency authorized to act in place
of a certifying authority of the Commonwealth.”
The financial impact of
Every state currently offers incentives to offset the costs, which can be substantial, for
environmental regula- installing pollution control equipment. Pollution from U.S. manufacturing has declined
tions per manufacturer considerably in recent decades, and most of the decline is attributed to strict environ-
is eight times more than mental regulations, such as the federal Clean Air Act. However, these regulations also
the impact on the aver-
have the side effect of dampening manufacturing employment, productivity, and for-
age U.S. company (Na-
tional Association of eign investment in the U.S.
Manufacturers 2014).
Sales and use exemptions are the most common state-level pollution control equip-
ment incentive. Only five states with sales and use taxes do not offer an exemption for
pollution control equipment. Instead, these states offer other incentives, such as local
property tax exemptions. Some states, such as Kansas and Louisiana, exempt pollution
control equipment as part of their general manufacturing sales and use exemption if
the equipment is used in the manufacturing process. Some states offer incentives in
addition to sales and use tax exemptions, such as local property tax exemptions, tax
credits, and loan assistance programs. (See Appendix I for more detail on pollution
Prior to 1972, TAX ad- control equipment incentives by state.)
ministratively exempted
purchases of pollution
control equipment un-
Pollution control equipment sales tax exemption saved businesses an estimated
der the state’s manufac- $27 million between FY10 and FY17
turing exemption. How-
Manufacturers and contractors received an estimated $26.6 million in tax savings from
ever, the equipment had
to be used directly in the the pollution control equipment sales tax exemption between FY10 and FY17. Savings
manufacturing process. are estimated to be approximately $3.3 million per year, on average. These savings were
estimated based on a Weldon Cooper Center survey of 230 businesses deemed eligible
to use the exemption from a review of certification records maintained by TAX. (See
Appendix B for more detail about the methodology.)
Businesses reported administrative burdens to obtain certification
Manufacturers reported that pollution control equipment makes up a large portion of
their capital expenses, but they often forgo using the sales tax exemption because of
Section 58.1-3660 re- the administrative burden of achieving certification, particularly from DEQ. Manufac-
quires that pollution turers indicated that the certifying agency reviews each request on a case-by-case basis,
control equipment be sometimes resulting in a lengthy approval process. The certification process can also
certified as “constructed,
reconstructed, erected,
be complicated because third parties (contractors and subcontractors installing equip-
or acquired.” ment and engineers that designed the equipment) often must be involved. As required
by statute, DEQ requires that the equipment be installed or the facility built before
granting certification, meaning manufacturers must pay the sales tax upfront before
later seeking a refund. Construction times are often lengthy, sometimes resulting in
decisions to forgo using the exemption entirely.
52
Data Center and Manufacturing Incentives
Pollution control equipment exemption may be an ineffective tool for reducing
pollution
While the pollution control equipment exemption helps businesses defray part of the
costs of complying with federal environmental regulations, it may not be an efficient Evaluating the effective-
ness of the exemption
and effective tool for reducing pollution. Research suggests these subsidies violate the on industrial activity or
“pollution payer principle,” which requires the pollution emitter to bear the cost of pollution reduction was
managing or reducing pollution. The exemption instead passes the cost on to society not performed because
in the form of reduced tax revenues. Exemptions and other subsidies may also en- industries are required
by federal regulation to
courage the use of pollution control equipment instead of the adoption of newer, less reduce pollutants, often
polluting production technologies and practices. Exemptions and other subsidies, through pollution con-
which reduce costs of production, may encourage businesses to produce more, which trol devices.
results in more pollution than would occur without the subsidy (Jenkins and Lamech
1994; Goulder and Parry 2008).
Recyclable Materials Processing Equipment Tax Credit was adopted to
encourage businesses to invest in equipment to recycle waste
The Recyclable Materials Processing Equipment Tax Credit was adopted in 1990 to
reimburse businesses for purchasing equipment to produce new products from recy-
cled materials (Table 4-2). The intent of the credit appears to be twofold. It aims to
encourage businesses to recycle industrial and consumer solid waste and encourage
additional employment and economic activity associated with recycling. The credit has
been modified several times since its adoption, including increasing the amount of the
credit from 10 percent to 20 percent of eligible purchases, imposing an annual $2
million cap per fiscal year, and extending the carryover period from five to 10 years.
Virginia is one of only 13 states that offers a recycling equipment tax credit. (See Ap-
pendix J for recycling equipment tax credits by state.)
Table 4-2
Virginia offers a Recyclable Materials Processing Equipment Tax Credit
Purpose Promote recycling of industrial and consumer solid waste and expand recycling equipment
investment
Eligibility Businesses that purchase machinery and equipment predominantly used to manufacture, process,
compound, or produce items of tangible goods from recyclable materials to be sold
Machinery and equipment must be certified by DEQ as integral to the recycling process
Credit features Nonrefundable credit equal to 20 percent of purchase price of eligible machinery and equipment
Credit can be claimed against individual and corporate income taxes but cannot exceed 40 percent of
taxpayer’s total tax liability; can be carried over for 10 years
Capped at $2 million in total credits per year
SOURCE: Weldon Cooper Center review of Code of Virginia and agency documents.
NOTE: Authorized by § 58.1-439.7 and expires in 2020.
53
Data Center and Manufacturing Incentives
Recycling materials tax credit saved businesses $10.7 million FY10–FY17
The total number of
Businesses producing goods from recyclable materials saved $10.7 million because of
businesses using the the recyclable materials tax credit between FY10 and FY17. Beneficiary savings each
credit is unknown. Half year have been fairly stable, averaging $1.3 million per year. Manufacturers accounted
(54 percent) of the credit for over 80 percent of the amount claimed on corporate returns between FY10 and
amount was claimed on
FY17, mostly by chemical product and preparation manufacturers (36 percent) and
corporate tax returns,
and the remaining iron and steel mills and ferroalloy manufacturers (29 percent). These manufacturers
amount was claimed by recycle the waste generated during the production of their primary products. Other
business owners on indi- notable non-manufacturers that have claimed the credit include miscellaneous durable
vidual tax returns be-
goods merchant wholesalers (17 percent), which include businesses involved in recy-
cause the business was
structured as a pass- cling and wholesale distribution of automotive scrap, industrial scrap, and other con-
through entity. sumer and commercial waste.
Manufacturers expect the credit may be used more frequently in the future, but the
current annual cap of $2 million could be a barrier. Some industries, such as pulp and
paper companies, are required to use a certain amount of recycled materials in their
products (e.g., corrugated paper uses post-consumer materials), and the U.S. Plastics
Resin Producers has established a national goal to recycle 100 percent of plastics. In
addition, U.S. companies have historically exported recyclable plastic, but China has
banned imports of certain grades of recycling plastic making growth of the domestic
recycling industry more likely.
Effect of recyclable materials tax credit on industrial recycling activity is
unknown
National sources for in-
Limited data makes it difficult to assess the effect of the recyclable materials tax credit
dustrial recycling are on industrial recycling activity. Businesses using the tax credit may recycle municipal
limited. The U.S. Environ- solid waste (residential and commercial trash) or recycle their own industrial solid
mental Protection waste. Statewide and local statistics for municipal solid waste recycling in Virginia are
Agency Toxics Release
available from DEQ, but statistics of industrial solid waste recycling are not. Municipal
Inventory provides a na-
tional database of toxic solid waste represents only a small share of total waste generation. National sources
chemical and other ma- for industrial recycling are either limited or have been discontinued. Currently, it is
terial recycling by state estimated that 236 million tons of municipal solid waste and 7.6 billion tons of non-
but covers only selected
hazardous industrial solid waste are generated per year nationally.
industries.
The U.S. Census Pollu- It is also difficult to assess industry activity, such as employment and production
tion Abatement Costs changes. There are several industry sectors for recycling processors, but these exclude
and Expenditures survey scrap materials recycled by other sectors, such as pulp, paper, and paperboard mills,
was discontinued in
2008. It provided costs
and plastics manufacturers. Most of these manufacturing industries, however, use orig-
of industrial pollution inal materials rather than recyclable scrap materials. Therefore, their industrial activity
abatement related recy- cannot be used to infer recycling activity levels by these industries. These sectors also
cling by state. tend to not report recycled material.
54
Data Center and Manufacturing Incentives
Green Job Creation Tax Credit was adopted to encourage creation of
Virginia is one of the few
green energy jobs but use has been low states with green job
The Virginia Green Job Creation Tax Credit provides a nonrefundable tax credit to tax incentives. Massa-
chusetts established a
Virginia businesses that create and retain new jobs that promote renewable and alter-
“green jobs” program in
native energy (Table 4-3). The credit targets green jobs, which include employment in 2008, which created a
the renewable and alternative energy industry, and other jobs as defined by the Secre- Clean Energy Technol-
tary of Commerce and Trade. Virginia’s green job tax credit is set to expire January 1, ogy Center that provides
grant funding to
2021.
startups. Pennsylvania
The green job tax credit was adopted in 2010 as part of Governor McDonnell’s job created the Alternative
and Clean Energy Pro-
creation and green energy agendas. Virginia and many states were increasingly pro-
gram to provide grants
moting green industry activity, corresponding with the clean energy and green industry of $10,000 per clean and
funding initiatives in the American Recovery and Reinvestment Act of 2009. These alternative energy job.
programs were motivated in part by a desire to stimulate job growth as well to enhance Tax credits adopted by
energy security, promote conservation, and expand investment in research and devel- New Mexico and Arizona
also target industry clean
opment on new environmental technologies. However, Virginia is one of the few energy investments and
states that created a job creation incentive targeting green jobs. job creation.
TABLE 4-3
Virginia offers a Green Job Creation Tax Credit
Purpose Encourage the creation and retention of green energy jobs
Eligibility Businesses that create new green jobs that are full-time and pay at least $50,000 annually
Green jobs include jobs in
- industries relating to the field of renewable, alternative energies, including the manufacture of
products used to generate electricity and other forms of energy from alternative sources
- businesses that produce goods or provide services that benefit the environment or conserve
natural resources
- which workers’ duties involve making their establishment’s production processes more
environmentally friendly or using fewer natural resources
Credit features Nonrefundable tax credit of $500 per job annually, up to 350 green jobs per business; can be
claimed against corporate and individual income tax
Claimed after the job is filled for 1 year; can be claimed for up to 5 years (for a total of $2,500 per
job) as long as job is continuously filled
5-year carryover period
Cannot claim Major Business Facility Tax Credit or federal credit for investing in clean energy
technologies for the same job
SOURCE: Weldon Cooper Center review of the Code of Virginia and agency documents.
NOTE: Authorized by § 58.1-439.12:05 and sunsets in 2021.
55
Data Center and Manufacturing Incentives
Green job tax credit use is low likely because other incentives offer higher
values and fewer restrictions
Only $4,362 in green job tax credits have been claimed on taxpayer returns between
FY10 and FY17. Thus, it has had the lowest level of utilization of any economic de-
velopment incentive program reviewed by JLARC to date. However, $63,501 in green
job credits were claimed in FY18 by taxpayers who earned credits in previous years
but did not have sufficient tax liability to claim them.
This credit is likely only attractive to businesses that do not meet eligibility criteria for
more valuable Virginia economic development incentives. For example, a business cre-
ating a substantial number of jobs that is eligible for both the Green Job Creation and
the Major Business Facility Jobs tax credits would likely claim the Major Business Fa-
cility Jobs Tax Credit. Even though the green jobs tax credit offers higher potential tax
savings ($2,500) per job, a taxpayer can only claim the credit for up to 350 jobs with
annual salaries of $50,000. While the Major Business Facility Jobs Tax Credit pays only
$1,000 per job, it has no maximum number of jobs, no minimum wage requirement,
and no restrictions on claiming federal credits for the same jobs.
Many blue-collar, green job occupations pay less than $50,000 annually, which likely
prevents more taxpayers from claiming the green job creation credit. Median average
wages were $39,490 for solar photovoltaic installers, $41,400 for hazardous materials
removal workers, and $45,490 for environmental science and protection technicians in
2017, according to the Bureau of Labor Statistics. The $50,000 wage requirement is
also substantially higher than many other grant and tax credit programs.
Research on effect of green job incentives on employment is limited and
inconclusive
Existing literature on environmental incentives’ impacts on state and local green in-
dustry growth is limited and provides conflicting results. (See Appendix O, available
online, for findings of peer-reviewed research.) Even if green job incentives create
new green jobs, the net effect on total employment is unclear. Although there is some
evidence that clean energy employment is more likely to be found in export-based
industries than general employment (Muro, Rothwell, and Saha 2011), not all green
jobs are export-oriented. Virginia’s green jobs tax credit does not specifically target
export-based sectors as do many other Virginia economic development incentive pro-
grams. Thus green incentives, in general and in Virginia, could fund jobs that substitute
for existing jobs and have no net employment creation effects. Some critics contend
that green jobs programs fund lower productivity employment that displaces tradi-
tional fossil fuel related jobs, which tend to be export-oriented and pay high wages.
56
Data Center and Manufacturing Incentives
Biodiesel and Green Diesel Fuels Producers Tax Credit was designed
Biodiesel is fuel made
to promote clean energy production but use has been low from new and used veg-
etable oils or animal fats
The Biodiesel and Green Diesel Fuels Producers Tax Credit was adopted to promote
or similar materials. Bio-
production of biodiesel and green diesel in the Commonwealth (Table 4-4). The pro- diesel is commonly
gram appears to have multiple goals, including supporting farm and forestry commod- made from recycled
ity production, promoting waste recycling, stimulating value-added refinery produc- cooking oils and fats de-
rived from restaurant
tion, and improving the environment by promoting clean energy substitutes for fossil
waste.
fuels. Since the state imports the vast majority of its energy needs, including gasoline
Green diesel is fuel
and diesel, the program encourages the substitution of Virginia-produced fuel for im- made from nonfossil re-
ported fuels and the creation of new employment opportunities. The biodiesel tax newable resources in-
credit was adopted in the 2008 session and provides a nonrefundable but transferable cluding cultivated plants
credit of 1¢ per gallon of fuel produced that can be claimed for up to three years, with and trees, their byprod-
ucts, animal fats, and
a taxpayer cap of $5,000 per year. Thus, it applies mostly to smaller producers. The other renewable re-
Department of Mines, Minerals, and Energy certifies whether production require- sources.
ments have been met for receiving the credit.
Table 4-4
Virginia offers a Biodiesel and Green Diesel Producers Tax Credit
Purpose Promote production of biodiesel and green diesel products
Eligibility Producers of biodiesel or green diesel (1) making up to 2 million gallons of biodiesel or green diesel
per year and (2) within the first 3 years of production (cannot be claimed after the first three years of
production)
Certified by Virginia Department of Mines, Minerals, and Energy that meet requirements
Credit features Nonrefundable credit equal to 1ȼ per gallon of fuel produced; can be claimed against individual and
corporate income taxes
Maximum credit allowed is $5,000; can be carried over for 3 years; unused credits can be transferred for
use by another taxpayer
SOURCE: Weldon Cooper Center review of the Code of Virginia and agency documents.
NOTE: Authorized by § 58.1-439.12:02.
All U.S. states except for Delaware had at least one type of renewable fuel tax incentive
when Virginia adopted the biodiesel credit. These incentives were popular following
the federal enactment of the biodiesel tax credit in 2004 (equal to $1 per gallon for
agricultural-based biodiesel and 50¢ per gallon for recycled feedstock) and later Amer-
ican Recovery and Revitalization Act fiscal stimulus programs that included temporary
federal programs for environmental and green job creation. As these federal programs
ended, state support for these incentives also declined. A number of states that had
biodiesel production incentives have repealed them in the last decade, including Ar-
kansas (eliminated a grant but still has a tax credit), Florida, Kansas, Maine, Mississippi,
Nebraska (eliminated one of two tax credits), North Carolina, and Oklahoma. Cur-
rently, only 10 states have biofuels tax incentives. (See Appendix K for information on
biodiesel credits by state.)
57
Data Center and Manufacturing Incentives
Use of biodiesel tax credits was much lower than expected
Only six taxpayers claimed a total of $11,212 in biofuel tax credits between FY10 and
FY17, according to annual reports by TAX. This credit amount—valued at 1¢ per
gallon—represents only 2 percent (1,121,200 gallons) of total biodiesel capacity in the
state during the same time period. This amount is significantly lower than estimated
claims of $15,000 per year that TAX projected based on potential usage by three bio-
refineries in Virginia at the time. In addition, TAX also found that some companies
were claiming the credit without certification. TAX issued fewer than four biodiesel
tax credit approvals since the credit was adopted. A total of $4,309 in properly issued
credits have been claimed in total, all of which was claimed in FY11. Some of the
companies that erroneously claimed the credit may be involved in biodiesel blending
and distribution rather than production. Therefore, eligible credits totaled less than the
$11,212 claimed between FY10 and FY17.
Several factors could explain low usage of the credits. Some of Virginia’s biodiesel
refineries may be too large to claim the credit. The Environmental Protection Agency
and Biodiesel Magazine report there were five registered biodiesel refineries in Virginia
in 2018 (Ditzel et al. 2018) and all but one had the capacity to exceed the tax credit’s
production limit of two million gallons per year:
Reco Biodiesel LLC in Richmond (3.6 million gallons),
Red Birch Energy Inc. in Bassett (3 million) (now closed),
Shenandoah Agricultural Products in Frederick (300,000 gallons),
The federal tax credit
and fuel standard regu- Synergy Biofuels LLC in Pennington Gap (3 million gallons) (now closed), and
lations (blending bio- Virginia Biodiesel Refinery LLC in West Point (3 million).
diesel with standard pe-
troleum-based diesel The refineries also all existed prior to the adoption of the tax credit, and the credit
fuel products) have may not have been available during the first three years of production. In addition,
played a critical role in companies may not have had sufficient tax liability, the certification process and pa-
U.S. biodiesel produc-
tion. Biodiesel produc-
perwork may be too onerous, or the incentive amount too meager to make applying
tion in the U.S. was less for it worthwhile.
than one million gallons
in 1999. Federal subsi- Virginia’s biodiesel credit is likely too small to influence production decisions
dies and fuel standards
were adopted (federal Virginia’s biodiesel tax credit has not had an impact on biodiesel production in the
Energy Policy Act of state, likely because it is too small to make a difference in production decisions. Vir-
2005) and later ex- ginia’s biodiesel production capacity has dropped from approximately 15 million gal-
panded (Energy Inde-
pendence and Security
lons in 2009, when the tax credit was first available, to nine million gallons in 2018, a
Act of 2017). By 2006, bi- decrease of 36 percent (Figure 4-1). In contrast, national production capacity has in-
odiesel production had creased by 14 percent. Virginia’s credit is valued at 1¢ per gallon of biodiesel produced
risen to 250 million gal- and is lower than credits in other states (ranging from 5¢ to 30¢ per gallon) and sub-
lons, and by 2017, it had
risen to 1.6 billion gal-
stantially lower than the federal credit ($1 per gallon for agricultural based biodiesel
lons. and 50¢ per gallon for recycled feedstock). It is also smaller than other Virginia incen-
tives that were unsuccessful at attracting biodiesel producers to the state and have since
been repealed.
58
Data Center and Manufacturing Incentives
Clean-Fuel Vehicle and Vehicle Emissions Testing Equipment Tax
Credit (1996–2014)—tax credit of $700 per job for a variety of biofuel-related
manufacturing and energy conversion activities, including the production of
biodiesel fuel manufactured from biomass or algae.
Biofuels Production Grant Incentive Program (2006–2011; 2014–2017)—
grant valued at 10¢ per gallon to new producers who produce at least 10 mil-
lion gallons per year of unblended biodiesel or ethanol fuels and existing pro-
ducers that increased their production by an additional 10 million gallons. The
program was repealed, then reenacted later but with reduced value per gallon,
and then repealed again.
Clean Energy Manufacturing Incentive Grant Program (2011–2017)—
custom grant to encourage eligible clean energy manufacturers, including bio-
fuel producers, who made a capital investment of at least $50 million and cre-
ated 200 new full-time jobs. This custom grant was adopted to attract a
Hopewell facility.
Biodiesel production is concentrated in a handful of states, with approximately half
of U.S. biodiesel production capacity in 2017 occurring in the 10 states that produce
the majority of the U.S. soybean crop—the principal oilseed for making green diesel.
Virginia, however, ranks 19th for soybean production despite it being the state’s largest
cash crop.
Figure 4-1
Virginia biodiesel production decreased after biodiesel tax credit was adopted
SOURCE: Weldon Cooper Center analysis of Energy Information Administration Monthly Biodiesel Production Re-
ports for Virginia and the U.S.
59
Data Center and Manufacturing Incentives
Increased biodiesel production likely has some negative environmental effects
While one of the motivations for encouraging biodiesel production subsides is to im-
prove the environment, it may have some negative environmental effects. Biofuel pro-
duction may not lead to reduced greenhouse emissions because crop cultivation re-
quires fertilizers that release greenhouse gases during production and consumption.
In addition, converting forest and grassland to farmland may release greenhouse gas-
ses (Ng, Ng, and Gan 2010). Increased crop production for biodiesel may have other
negative environmental consequences, including water pollution from increased nutri-
ent runoff, soil erosion, and decreased biodiversity (Janda, Kristoufek, and Zilberman:
Searchinger 2008). Substituting biodiesel for fossil fuels should lower carbon emis-
sions, but production subsidies also lower the consumer price of fuel, which may result
in increased consumption of fuel and increased consumer emissions (de Gorter and
Just 2010; Yoder 2008).
Biofuel and biodiesel tax credits may not be as cost effective as other market-based
policies for reducing greenhouse gas emissions. The cost of reducing emissions using
biodiesel tax credits is substantially higher than the cost of reducing emissions through
a cap-and-trade program that imposes a price on emissions and then allows the market
Economic impact
analysis of expenditures
to determine how to reduce emissions throughout the economy (CBO 2010).
by incentives recipients
between FY10 and FY17 Environmental incentives generate negligible economic benefits and
was conducted using returns in state revenue
economic modeling
software developed by The manufacturing environmental incentives generate relatively low economic activity
REMI, Inc. for the Virginia economy. Of the four incentives, the pollution control equipment ex-
(See Appendix M [online emption generated the highest levels of additional economic activity (Table 4-5). It is
only] for the economic also the largest of the programs. It was the only incentive estimated to increase private
impact analysis used in
this study.)
employment (24 new jobs annually, on average) between FY10 and FY17. The pollu-
tion control exemption annually generated nearly $700,000 in additional Virginia GDP
and $1.75 million in additional personal income between FY10 and FY17, but in-
creases to Virginia GDP and personal income for the other three incentives were far
Net impact is the less. Estimated economic activity is low because the programs provide relatively low
increase in economic levels of financial assistance (less than $10,000 in total for the green jobs and biodiesel
activity caused by the in- tax credits) to a few businesses.
centives after adjusting
for the opportunity cost When measured per $1 million in spending, the economic benefits of the manufactur-
of increasing taxes to ing environmental incentives are negligible relative to other incentives. They are the
pay for the incentives. lowest of the 23 incentives that JLARC has evaluated to date. The estimated Virginia
(See Appendix N [online GDP generated for each of the four incentives is $1 million or less for every $1 million
only] for information on
spent on the incentives.
the total economic im-
pact and the opportunity The returns in state revenue are also negligible relative to other incentives. Returns
cost of increasing taxes.)
range from a low of 3¢ (recyclable materials tax credit) to a high of 13¢ (biodiesel tax
60
Data Center and Manufacturing Incentives
credit). This amount is less than the estimated 19¢ return across all economic devel-
opment incentives. (See Economic Development Incentives, JLARC 2018.) This amount,
however, is similar to the returns for most other tax incentives.
TABLE 4-5
Manufacturing environmental incentives generate negligible economic
benefits and returns in state revenue (FY10–FY17)
Annual average (FY10–FY17)
Pollution Recyclable Biodiesel and
Control Materials Green Diesel
Equipment Processing Green Job Fuels
and Facilities Equipment Creation Tax Producers
Exemption Tax Credit Credit Tax Credit All programs
Net impact to Virginia economy
Private employment 24 jobs 0 jobs 0 jobs 0 jobs 24 jobs
Virginia GDP $0.71 M $0.06 M <$0.01 M <$0.01 M $0.77 M
Personal income $1.75 M ($0.07 M) <$0.01 M <$0.01 M $1.68 M
Impact to Virginia economy per $1 million of spending on incentive
Private employment 15 jobs 8 jobs 13 jobs 4 jobs 13 jobs
Virginia GDP $1.11 M $0.94 M $0.56 M $0.74 M $1.1 M
Personal income $1.28 M $0.69 M $1.04 M $0.43 M $1.1 M
Impact to state revenue
Total revenue $198,295 $40,943 $27 $181 $239,446
Cost of incentive $3,323,469 $1,338,166 $545 $1,401 $4,663,581
Net revenue ($3,125,174) ($1,297,223) ($518) ($1,220) $1,302,086
Return in revenue
6¢ 3¢ 5¢ 13¢ 5¢
for every $1 spent
SOURCE: Weldon Cooper Center economic impact analysis of economic activity induced by manufacturing environ-
mental incentives between FY10 and FY17.
NOTE: Includes direct, indirect, and induced impacts. The gross impact on Virginia’s economy is used to calculate the
impact per $1 million in incentive awards. This is consistent with how the economic development research literature
typically calculates these impacts. (See Appendix N [online only] for detailed results on total impact of the loans,
impact of raising income taxes by the amount of the incentives [opportunity cost], and revenue generated by source.)
Changes should be considered to improve the exemption, and green
job and biodiesel tax credits should be eliminated
None of the four environmental incentives appears to be effective at increasing indus-
try productivity or improving the environment. Further, they all have a negligible eco-
nomic benefit and return in revenue to the state. Two of the incentives—the pollution
control equipment exemption and recyclable materials tax credit—should be consid-
ered for continuation. The pollution control equipment sales tax exemption was pri-
marily designed to reduce the burden of regulatory compliance rather than to promote
economic activity. Repealing it could also place Virginia at a competitive disadvantage
61
Data Center and Manufacturing Incentives
with other states, all of which provide some type of pollution control equipment in-
centive. The recyclable materials tax credit could be continued—and the expiration
date continued—to see if its use increases because of additional need to recycle in-
dustrial and consumer waste. If more businesses use it, and the amount businesses
seek to claim exceeds the $2 million annual cap, the General Assembly may want to
increase the credit cap or adopt minimum job creation or capital investment require-
ments. Two tax credits—the green job and biodiesel tax credits—should be eliminated.
Changes to certification process for pollution control equipment exemption
could increase transparency and reduce compliance burden
The certification process for the pollution control equipment exemption could be im-
proved to reduce compliance burden. To use the exemption, businesses generally must
receive certification from either DEQ—regional offices rather than the central of-
fice—or DMME. Each agency has its own process and uses its own forms to collect
information. Neither agency provides much information about the process on its web-
site, with exception of the certification for solar energy equipment or facilities. The
certification process, therefore, is not transparent and likely not consistent, adding to
the compliance burdens reported by exemption users.
Several changes could be made to increase transparency and reduce compliance bur-
den. At a minimum, DEQ and DMME should develop guidance documents to pro-
vide greater assistance to businesses and contractors as to the types of equipment and
facilities that are exempt and the decision-making process used to approve certifica-
tion. Guidance documents should be available on the agency website.
RECOMMENDATION 6
The Department of Environmental Quality and Department of Mines, Minerals, and
Energy should develop guidance documents on (1) the types of pollution control
equipment and facilities that are exempt from the retail sales and use tax and (2) the
decision-making process for approving certification.
DEQ should also develop a list of pre-approved equipment or facilities that are typi-
cally exempt and develop an expedited certification process for certifying property or
facilities on this list. It does not appear necessary for DMME to take similar action at
this time. DMME staff indicated they approve solar equipment and facility certifica-
tion requests within 48 hours and coal, oil, and gas equipment certification requests,
for which few have been made, within 30 days. Several states—Iowa, Michigan, Mis-
sissippi, and Texas—have developed lists of pre-approved equipment. Texas has a
three-tiered application process for its pollution control equipment property tax ex-
emption. A Tier 1 application is used for certification of equipment that is on a pre-
approved list and that will be used solely for pollution control purposes. A certification
decision must be made within 30 days. A Tier 2 application is used for property that
will be used solely for pollution control but is not on the pre-approved list, and a Tier
3 application is used for property that is partially used for pollution control. Tier 2 and
62
Data Center and Manufacturing Incentives
Tier 3 applications are limited to a 60-day review period as long as the application is
complete. A tiered process may be particularly beneficial for smaller manufacturers or
contractors that may not have the resources to devote substantial effort to applying
for the exemption.
RECOMMENDATION 7
The Department of Environmental Quality should develop a list of pre-approved
equipment and facilities that typically meet the pollution control certification require-
ments and create an expedited certification process for equipment and facilities on that
list.
Several stakeholders said requiring that pollution control equipment be installed or the
facility built before certification is burdensome. The statutory language requiring the
state certifying authority to certify that the pollution control equipment has been “con-
structed, reconstructed, erected, or acquired” is within the Code section for local prop-
erty taxation. It is unclear whether it was the General Assembly’s intent to require that
pollution control equipment be in place before companies could receive certification
to use the sales tax exemption. Other sales tax exemptions in Virginia are self-enforc-
ing and do not have a certification requirement, but this requirement for pollution
control equipment and other equipment or property is not uncommon for other states.
Statute also allows certain pollution control equipment or property to be exempt
whether or not it has been certified, allowing the sales tax exemption to be claimed
before equipment is in place. The General Assembly could amend the Code if mem-
bers think certification should be allowed before the pollution control equipment is in
place or the facility built.
OPTION 1
The General Assembly could amend § 58.1-609.3 or § 58.1-3660 of the Code of Vir-
ginia to clarify that the equipment or facility does not need to be constructed before
certification can be granted for purposes of claiming the Pollution Control Equipment
and Facilities Sales Tax Exemption.
DEQ, DMME, and TAX could collaborate and develop a web-based application sys-
tem for businesses to apply for certification and then be issued exemption certificates.
This system would streamline the process for obtaining certification and the exemp-
tion certificate. This web-based system could also be used to uniformly collect esti-
mates about the costs of the equipment and facilities to be purchased that would pro-
vide more accurate estimates of forgone revenue and economic benefits because of
the exemption. According to DEQ staff, some applicants already provide this infor-
mation even though it is not required.
63
Data Center and Manufacturing Incentives
Green jobs and biodiesel tax credits should be eliminated
The Green Jobs Creation Tax Credit and the Biodiesel and Green Diesel Fuels Pro-
ducers Tax Credit should be eliminated. Alternatively, the green jobs tax credit could
be allowed to expire on its expiration date in 2021. Both tax credits have low rates of
utilization and appear to have little effect on the activity they were intended to incen-
tivize. The tax credits are also estimated to provide negligible economic benefits to the
state.
RECOMMENDATION 8
The General Assembly may wish to consider eliminating the Green Job Creation Tax
Credit and the Biodiesel and Green Diesel Fuel Producers Tax Credit.
64
Appendixes
Appendix A: Study mandate
2018-2020 Appropriation Act
Passed as Chapter 2 of the Acts Assembly, May 20, 2018
§ 1-11 Item 31 F
F.1. The General Assembly hereby designates the Joint Legislative Audit and Review Commission
(JLARC) to conduct, on a continuing basis, a review and evaluation of economic development initia-
tives and policies and to make such special studies and reports as may be requested by the General
Assembly, the House Appropriations Committee, or the Senate Finance Committee.
2. The areas of review and evaluation to be conducted by the Commission shall include, but are not
limited to, the following: (i) spending on and performance of individual economic development in-
centives, including grants, tax preferences, and other assistance; (ii) economic benefits to Virginia of
total spending on economic development initiatives at least biennially; (iii) effectiveness, value to tax-
payers, and economic benefits to Virginia of individual economic development initiatives on a cycle
approved by the Commission; and (iv) design, oversight, and accountability of economic development
entities, initiatives, and policies as needed.
3. For the purpose of carrying out its duties under this authority and notwithstanding any contrary
provision of law, JLARC shall have the legal authority to access the facilities, employees, information,
and records, including confidential information, and the public and executive session meetings and
records of the board of VEDP, involved in economic development initiatives and policies for the
purpose of carrying out such duties in accordance with the established standards, processes, and prac-
tices exercised by JLARC pursuant to its statutory authority. Access shall include the right to attend
such meetings for the purpose of carrying out such duties. Any non-disclosure agreement that VEDP
enters into on or after July 1, 2016, for the provision of confidential and proprietary information to
VEDP by a third party shall require that JLARC also be allowed access to such information for the
purposes of carrying out its duties.
4. Notwithstanding the provisions of subsection A or B of § 58.1-3 or any other provision of law,
unless prohibited by federal law, an agreement with a federal entity, or a court decree, the Tax Com-
missioner is authorized to provide to JLARC such tax information as may be necessary to conduct
oversight of economic development initiatives and policies.
5. The following records shall be excluded from the provisions of the Virginia Freedom of Infor-
mation Act (§ 2.2-3700 et seq.), and shall not be disclosed by JLARC:
(a) records provided by a public body as defined in § 2.2-3701, Code of Virginia, to JLARC in con-
nection with its oversight of economic development initiatives and policies, where the records would
not be subject to disclosure by the public body providing the records. The public body providing the
records to JLARC shall identify the specific portion of the records to be protected and the applicable
provision of the Freedom of Information Act or other provision of law that excludes the record or
portions thereof from mandatory disclosure.
65
Appendixes
(b) confidential proprietary records provided by private entities pursuant to a promise of confidenti-
ality from JLARC, used by JLARC in connection with its oversight of economic
development initiatives and policies where, if such records are made public, the financial interest of
the private entity would be adversely affected.
6. By August 15 of each year, the Secretary of Commerce and Trade shall provide to JLARC all
information collected pursuant to § 2.2-206.2, Code of Virginia, in a format and manner specified by
JLARC to ensure that the final report to be submitted by the Secretary fulfills the intent of the General
Assembly and provides the data and evaluation in a meaningful manner for decision-makers.
7. JLARC shall assist the agencies submitting information to the Secretary of Commerce and Trade
pursuant to the provisions of § 2.2-206.2, Code of Virginia, to ensure that the agencies work together
to effectively develop standard definitions and measures for the data required to be reported and
facilitate the development of appropriate unique project identifiers to be used by the impacted agen-
cies.
8. The Chairman of JLARC may appoint a permanent subcommittee to provide guidance and direc-
tion for ongoing review and evaluation activities, subject to the full Commission's supervision and
such guidelines as the Commission itself may provide.
9. JLARC may employ on a consulting basis such professional or technical experts as may be reason-
ably necessary for the Commission to fulfill its responsibilities under this authority.
10. All agencies of the Commonwealth shall cooperate as requested by JLARC in the performance of
its duties under this authority.
66
Appendixes
Appendix B: Research activities and methods
JLARC contracted with the University of Virginia’s Weldon Cooper Center for Public Service (Weldon
Cooper Center) for this review. Key research activities performed by Weldon Cooper Center staff for
this study included
collection and analysis of national- and state-level financial and economic data and state
agency incentive program data;
estimation of business savings and tax revenue impacts from incentives;
program employment performance tracking;
statistical analysis of incentive program effects and quantitative analysis of the economic and
fiscal impacts of Virginia incentives using a dynamic economic model (See Appendix M, avail-
able online, for more detail on the analyses);
interviews with agencies and stakeholders;
review of other states’ manufacturing and data center incentive programs; and
review of documents and literature.
Collection and analysis of national- and state-level financial and economic data
and state agency incentive program data
This report drew on several federal, state, and private industry sources of economic data. Some of
this data was used primarily for descriptive purposes, including to highlight trends in state economic
activity, such as manufacturing employment and biodiesel production (Table B-1).
Information from state agencies, including the Virginia Economic Development Partnership, Virginia
Employment Commission, and Department of Taxation was used for both descriptive and analytical
purposes. First, project-level information was aggregated to show characteristics of program users
and features of the programs, including industry and employment size. Second, agency data was used
in conjunction with other data, such as confidential Virginia Employment Commission (VEC) ES202
payroll employment records, to track employment outcomes and conduct economic analyses. Third,
Department of Taxation data on form ST-11A pollution control equipment and facility users and
VEDP data center MOU records were used to estimate program revenue impacts. These analyses are
described further in the sections that follow.
Obtaining reliable public data for empirical analysis of data center activity was a challenge. Data cen-
ters do not have a dedicated industry code, such as semiconductor manufacturing; instead, data centers
can be found in a variety of industries. Most companies have strict privacy/confidentiality require-
ments to protect their clients and ensure facility and data security. Thus, less is known about the in-
dustry than others. In addition, the data center market changes rapidly with addition of many new
companies and frequent mergers and acquisitions, making it difficult to track company employment
activity over time using unit record data.
67
Appendixes
TABLE B-1
Multiple data sources were collected and used for a variety of analyses
Data source Description of data Analysis
Financial and economic data
State Personal Income and
Compute state and national employment
Bureau of Economic Analysis Employment—Data by two-digit NAICS
change for manufacturing industry
industry
Identify large state semiconductor
Good Jobs First Subsidy tracker
manufacturing incentives
Employment data by six-digit NAICS Compute state and national employment
EMSI
industry change for semiconductor industry
Information on educational achievement
levels of residents 25 years and older by
U.S. Census Bureau, American state and county; information on Use in statistical analyses (synthetic
Community Survey average wage levels by state for control analysis and panel data regression)
computer and electronic product
manufacturing
Compute state and national biofuel
Monthly Biodiesel Production Report
U.S. Energy Information production and capacity, use in statistical
(2009-2018); Electric Power Monthly
Administration analysis of semiconductor manufacturing
(2001-2017)
employment
Announcements and Closings Database
Virginia Economic Development Analysis of data center responsiveness to
(including data not disclosed to the
Partnership state and local economic incentives
public)
Track employment performance and
Virginia Employment Employment Security (ES 202) payroll
conduct quasi-experimental statistical
Commission employment records
analyses
Information on local tangible personal Estimate size of local machinery tools and
property rates for data center tax incentives for semiconductor plants;
Weldon Cooper Center equipment and machinery and tool conduct statistical analysis of tangible
rates for semiconductor industry from personal property tax rates on data center
Virginia Local Tax Rates location
Virginia incentive programs
Tax credit utilization for the Green Job
Creation Tax Credit, Biodiesel and Green
Computation of tax credit usage by fiscal
Department of Taxation Diesel Fuel Producers Tax Credit, and
year
Recyclable Materials Processing
Equipment Tax Credit
Corporate income tax Form 500 and Tabulate revenue impact of single sales
Department of Taxation
Form 500A data for tax years 2013–2016 apportionment factor for manufacturing
Conduct survey of pollution control
Records on firms in ST-11A certification
Department of Taxation equipment and facility sales and use tax
database
exemption
Data on Data Center Sales and Use Tax
Virginia Economic Development Estimate economic and revenue impact of
Exemption MOU employment and
Partnership data center sales and use tax exemption
capital investment attainment
68
Appendixes
Other
Census of Government, Annual
State tax revenue by tax category and
Survey of State Government Tax revenue impact analysis
fiscal year
Finances
Regional SAM Balances, institution Estimation of tax revenue impacts of
IMPLAN industry demand, regional employment semiconductor manufacturers and
multipliers, study area industry data semiconductor wafers exemptions
Tax revenue impact analysis. Computation
Demand by industry, GDP, personal of “but for” effect of incentives
income, and transfer receipts by year; (semiconductor grants, data center tax
REMI PI+
value added and employment by exemption); computation of value-added
industry per employee by industry for "but for"
calculations
Returns of active corporations: income
Urban-Brookings Tax Policy by industry 2009–2013 based on Internal Calculate fiscal impact of data center
Center Revenue Service, Statistics of Income, single sales factor apportionment
corporation income tax returns
SOURCE: Weldon Cooper Center.
Some industry studies have used the industry “data processing, hosting, and related services” (NAICS
518210) as a proxy for data center industry activity. This is likely to be a poor measure for several
reasons. First, data centers were identified using a variety of sources (DataHawk, VEDP announce-
ments data, and search engine searches of company data, and other sources) and combined with Vir-
ginia Employment Commission establishment employment data for 2016. The result was that the bulk
of “data processing, hosting, and related services” were found to be various types of data processing,
IT work, document scanning, online business and educational services, and software development.
Much of it is connected to federal IT contracting in Northern Virginia. Approximately 15 percent of
the total employment in the sector is data center employment with other data center, cloud computing,
and cybersecurity-related support services making up perhaps 2 percent to 5 percent more.
Second, data centers are more likely to be found outside of “data processing, hosting, and related
services” than inside it. An examination of data center employment indicated that less than half (41
percent) of data center employment was in the data processing, hosting, and related services sector
(Figure B-1). Three telecommunications sectors (i.e., wired telecommunications carriers, telecommu-
nications resellers, and all other telecommunications carriers), collectively made up an even higher
share (44 percent). These figures do not include employment for certain enterprise data centers located
adjacent to other larger company operations, where it was not possible to segregate data center em-
ployment from other principal operations at the location, and almost all collocating firms that make
up the bulk of employment in the sector. Finding colocation center tenants is difficult if not impos-
sible because their employment may be reported in other branch locations. However, if these jobs
were included, a significant portion would likely be found outside of the data processing, hosting, and
related services sector.
69
Appendixes
Figure B-1
Majority of Virginia data center employment is found outside the data processing, hosting,
and related services industry
SOURCE: Weldon Cooper Center analysis of Virginia Employment Commission data.
Third, actual employment data for the data processing, hosting, and related services sector does not
appear to represent growth in the data center industry. VEDP announcement data suggests that em-
ployment in the sector has grown rapidly in the last 15 to 20 years. Yet EMSI employment data for
Virginia shows that the data processing, hosting, and related services sector has experienced little net
growth over the 2001–2017 period.
Estimating business savings or forgone taxes because of incentives
Information to estimate business savings, or forgone taxes because of incentive programs, came from
several sources, including agency records and other data, secondary source data, and a Weldon Cooper
Center survey. The amounts of funds awarded and disbursed to semiconductor customized grant
recipients (Micron and Qimonda) were obtained from Virginia Economic Development Partnership
MOUs and correspondence with the firms. Firm tax credit data for the Biodiesel and Green Diesel
Fuels Production, Green Job Creation, and Recyclable Materials Processing Equipment Tax credits
was obtained from the Virginia Department of Taxation Annual Reports. Business savings and asso-
ciated state forgone revenue for the other incentive programs were estimated using both primary and
secondary data as explained below.
Manufacturers Single Sales Apportionment
The Weldon Cooper Center estimated forgone revenue from single sales apportionment for manufac-
turers for tax years 2014–2017 using Virginia Department of Taxation corporate tax form submissions
from tax years 2013–2016. Information from Form 500 (Virginia Corporation Income Tax Return)
and Schedule 500A (Multistate Corporation Allocation and Apportionment of Income) were used.
Schedule 500A identifies multistate firms using the manufacturer’s modified apportionment method
70
Appendixes
for the corresponding year. Together, both forms are used to compute corporate income tax liability.
Businesses using single sales apportionment are required to provide only information on the sales
factor and not information on all three factors (property, payroll, and sales). The property, payroll, and
sales factor information was obtained from the closest year before each firm elected the single sales
apportionment method to compute the baseline double-weighted sales factor the company would use
otherwise. For most firms, this was either tax year 2013 or tax year 2014. Additional information was
obtained from individual hardcopy corporate tax records.
Two estimates were calculated: one estimate apportioned taxable income and calculated tax liability
using the traditional double-weighted sales factor (the default apportionment formula) to determine
baseline data. The second estimate apportioned taxable income and calculated tax liability with single
sales factor apportionment using the baseline data. The two estimates (double-weighted sales estimate
and single-factor sales estimate) were compared using the same baseline data. The implicit assumption
in both calculations is that the property and payroll factor proportions did not change markedly over
the period; there is no way of validating this assumption. However, a comparison of the sales factor
proportions indicated that the sales factor did not vary much. In order to provide the most conserva-
tive estimate of fiscal impact, the minimum of these two estimates is shown as the final estimate. Since
no information was available from tax year 2017 returns to estimate tax returns, a consumer price
index cost escalation factor is applied to the tax year 2016 estimate to obtain the revenue impact for
that year.
The estimates made here did not take into account any recapture for single sales factor electors who
did not meet the wage and employment requirements for using the single sales factor manufacturing
apportionment. There is no way to determine the status of such recaptures using electronic returns
data.
Semiconductor Manufacturing and Wafers exemptions
The Weldon Cooper Center estimated the semiconductor sales and use tax exemptions based on IM-
PLAN data and assumptions about the percentage of semiconductor company purchases likely to be
affected by the exemption. IMPLAN is a commercial economic impact model produced by MIG Inc.
It is based on input-output analysis, which requires estimates of the value of intermediate input pur-
chase for each industry. The intermediate input purchase estimates for Virginia formed the basis of
the relevant sales tax base for sales and use tax revenue impact calculations. In the case of the semi-
conductor manufacturing exemptions, spending on eligible inputs (i.e., equipment, fuel, power, en-
ginery, supplies, or other tangible personal property) is estimated by multiplying semiconductor and
related device manufacturing (IMPLAN sector 309) industry output by gross absorption coefficients
for IMPLAN commodity sectors 3040-3051 (utilities) and 3112-3394 (durable and some nondurable
manufactured goods) using the most recent data for 536 industries and commodities. These coeffi-
cients represent the input purchases for various commodities per dollar of output. For example, the
semiconductor and related device manufacturing industry spent $0.0180 per dollar of output on com-
modity 49 (electricity transmission and distribution). This absorption coefficient was multiplied by the
output of the semiconductor and related device manufacturing for 2016 ($973,343,567) to obtain the
estimated expenditure on this input ($17,520,184). Since only a small portion of this input was eligible
for the manufacturing exemption for purposes such as design, development, and testing purposes,
71
Appendixes
only a portion of these purchases (obtained from a semiconductor industry source) was used for the
base calculations. The eligible purchases base estimate was multiplied by the state sales and use tax to
obtain the revenue impact. The same methodology was used for estimating the semiconductor wafer
exemption. However, in this instance, only purchases of semiconductor and related device inputs (a
commodity category that includes wafers) were counted. Once again, an industry source provided an
estimate of the percentage of wafer inputs that were used as filler, cleaning, engineering, or process
control purposes.
The semiconductor manufacturing and wafers exemptions estimates provided in this report are sub-
stantially lower than those published in previous JLARC reports (Economic Development Incentives, 2017
and 2018). For those reports, the wafer exemption was based on an assumption that 3.3 percent of
wafers would be scrapped using IBM estimates. However, this estimate includes loss due to defects
and production losses in addition to the eligible exemption uses defined by the statute. An industry
source provided a smaller estimate of the percentage of wafers that would be eligible for the exemp-
tion. For the manufacturing exemption, it was assumed in previous JLARC reports that 33.5 percent
of purchases in eligible commodities would be eligible for the semiconductor manufacturing exemp-
tion. This estimate was based on the percentage of total semiconductor and other electronic compo-
nent manufacturing (NAICS 3344) employment in related scientific and engineering occupations
(computer and mathematical occupations, architecture and engineering occupations, and life, physical,
and social science occupations) from U.S. Bureau of Labor Statistics data. An industry source indicated
that the actual proportion of eligible activities is significantly lower than the percentage of personnel
in related occupations.
Data Center Exemption
For most sales and use tax exemptions, no agency records are available to estimate tax revenue impacts.
However, for the data center exemption, users are statutorily required to enter into a memorandum
of understanding (MOU) with the Virginia Economic Development Partnership (VEDP). Firms must
report their capital investment to determine whether they have complied with terms of the MOU.
JLARC and Weldon Cooper Center obtained data from VEDP regarding all MOUs since enactment
of the exemption. Under the terms of the program, data center firms (including colocating firms)
must make a total capital investment of at least $150 million. Capital investment includes new real
property purchases and eligible IT and related machinery and equipment used to establish the facility.
Only the portion of the capital investment made in machinery and equipment is eligible for the ex-
emption. The assumption is made that 77.72 percent of total capital investment was used for IT and
other eligible equipment. This percentage is based on capital costs of construction breakdown ob-
tained from a publication by the U.S. Chamber of Commerce Technology Engagement Center, Data
centers: Jobs and opportunities in communities nationwide (Day and Pham 2017).
The data file from VEDP included date of MOU execution, targeted capital investment amount, and,
for firms that qualified for the exemption, actual capital investment amounts for enterprise data cen-
ters and for both landlords and tenants of colocation centers. For most qualified firms, firms report
actual capital investment levels; however, some report only that they met the minimum statutory re-
quirement ($150 million). For firms that had entered into a MOU but had not yet completed their
performance period (or those that received extensions to comply with the MOU, usually employment
72
Appendixes
requirement levels), no information is available on capital investment to date. However, these firms
use the exemption during the interim period, which results in forgone state tax revenue. To adjust for
these data centers, exemption use on all estimated eligible equipment purchases are imputed. Based
on qualified exemption firms, capital investment targets underestimated actual investment by a factor
of 1.74. Therefore, for firms that did not report actual capital investment, actual levels were estimated
to be 1.74 times as large as the targeted (usually statutory minimum) level. In addition, capital invest-
ment was assumed to be spread out over a three-year period.
Finally, based on data center firm interviews, the IT equipment refresh cycle (three to five years) forms
the basis for future sales and use tax savings after the facility is established and equipped. Weldon
Cooper assumes that new equipment equivalent in value to the original equipment investment occurs
five years later. For example, a firm that signed an MOU in 2010 and completed the facility in 2013
would begin a new equipment refresh cycle in year 2015 and again in 2020.
The revenue impact estimates provided here are significantly higher than previous estimates developed
by the Department of Taxation and used in previous JLARC reports (Economic Development Incentives,
2017 and 2018). The 2018 JLARC report indicates forgone revenue totaled $284.9 million between
FY10 and FY17 compared to $417.47 million in this report, a 40 percent increase. The primary reason
for the discrepancy is this report uses information on actual investment for qualified projects rather
than minimum qualifying equipment targets. It also extrapolates the investment performance of qual-
ified projects to other pending projects. This method likely still underestimates the actual forgone
revenue because companies with an existing MOU do not need to apply for an additional MOU when
they undertake major expansions. For example, Microsoft has expanded its Mecklenburg data center
campus several times since entering into its MOU in 2010. The documented capital investment would
reflect only re-equipping the original buildings and not the current use of the exemption.
Data Center Single Sales Apportionment
Ordinarily, corporate tax returns could be used to determine the impact of the data center single sales
apportionment on state revenues. However, most companies operating enterprise data centers in the
state (e.g. Amazon, Microsoft, Visa) also have other unrelated operations in the state (i.e., multiproduct
firms), and it is difficult to disentangle these operations and employment from data centers. Further-
more, no information is available about future data center expansion. Therefore, revenue impacts are
estimated using Virginia Economic Development Partnership MOU and Department of Taxation
information in conjunction with secondary data and certain assumptions to arrive at plausible esti-
mates and projections of future revenue impact. These are described below.
Revenue estimates depend on data from the Internal Revenue Service Corporation Income Tax Re-
turns by Industry reported by the Brookings-Urban Tax Policy Center for five years (2009–2013).
Income subject to tax for the data processing, hosting, and related service and telecommunications
(including paging, cellular, satellite, cable, and internet service providers) industries (50-50 split) is
assumed to be representative of the data center industry based on information presented earlier. Em-
ployment data for the same sectors and corresponding years was obtained from Bureau of Economic
Analysis State Personal Income and Employment. Federal taxable income per employee was calculated
for each tax year and inflated to 2018 using the consumer price index. The average of five years was
assumed to be representative of the typical year and would persist into the future with only a consumer
73
Appendixes
price index inflation adjustment (assumed to be 2.2 percent per year over the period 2019–2024). It is
assumed that federal taxable income per employee nationwide is an adequate gauge of Virginia taxable
income per employee (i.e., there are no other tax additions and subtractions).
Weldon Cooper Center collected information on enterprise data centers that are currently eligible for
single sale apportionment by reviewing data that VEDP collects from companies using the data center
sales tax exemption. Qualifying companies must make capital investments in data center operations to
be eligible for the exemption; the same capital investment requirement ($150 million) applies to using
single sales apportionment. The MOU data for the sales tax exemption also contains information on
actual job creation associated with each data center investment. Seven enterprise data centers qualified
for a sales and tax use exemption over the 2011–2018 period. Over that period, the companies created
an estimated 812 jobs (102 per year). This annual rate of employment gain was projected to persist
over the 2019–2024 period for eligible firms and incrementally increase Virginia taxable income. To
qualify to use single sales apportionment, an enterprise data center would have to make an investment
and enter into an MOU with VEDP after July 1, 2015, the effective date of the incentive. For taxable
years 2018 and beyond, two firms that qualified for the exemption have also qualified to use single
sales apportionment. Single sales apportionment for data centers may have a tax revenue impact for
FY18 and beyond depending on the profitability of these and other firms that may qualify in the
future.
Calculating the effect of switching from standard double sales apportionment to single sales requires
estimates of companies’ payroll, property, and sales. Estimates of these factors came from two
sources. Actual apportionment factor data was obtained from corporate returns filed with the Depart-
ment of Taxation. When this data was not available, estimates were made by assuming that the Virginia
percentage of total data centers was representative of the property and payroll factors. For example,
if a company had 52 U.S. data centers and 37 were located in Virginia, it was assumed that 71 percent
(37/52) of property and payroll are located in Virginia. For these same firms, it was assumed that 5
percent of company sales originated in Virginia. This percentage is roughly double Virginia’s percent-
age of total U.S. population (2.6 percent). Using this approach, it was estimated that the typical enter-
prise data center has 5 percent of payroll, 20 percent of property, and less than 1 percent of sales in
Virginia. Thus, under standard double sales factor approximately 6 percent of income would be taxa-
ble. Under single sales apportionment less than 1 percent would be taxable. The difference of the two
was applied to estimated income subject to Virginia tax. This difference then was multiplied by the 6
percent corporate tax rate.
This method may slightly undercount or over count the impact on revenue for several reasons. First,
it is possible that data centers are more profitable (and have greater federal tax liability per employee)
than the composite telecommunications and data processing, hosting, and related service industry
firm. Second, data centers that invested at least $150 million but did not meet job creation thresholds
for the program would not be captured in the MOU data used to estimate revenue impact. Third, the
projects do not take into consideration expansion of existing data center campuses that had already
qualified for the tax exemption and whose subsequent expansions investments are not captured in the
MOU data. Fourth, failure to account for tax subtractions to arrive at Virginia taxable income may
lead to over counting of state tax liability per employee.
74
Appendixes
Pollution Control Equipment and Facilities Exemption
Limited data is available to estimate the forgone revenue of the pollution control equipment and
facilities exemption. To be eligible for the exemption, manufacturing firms must first obtain certifica-
tion from the relevant state agency. For most water, air, and solid waste mitigation equipment, the
Department of Environmental Quality provides the certification based on federal and state regulatory
requirements. The Department of Mines, Minerals, and Energy provides the certification for solar
projects and coal, oil, and gas production related pollution mitigation equipment. Manufacturers who
make the improvements must complete form ST-11 (Sales and Use Tax Certificate of Exemption)
and provide this form to suppliers to receive the exemption. Construction and other firms who install
and equip pollution control facilities for manufacturers and other firms must first obtain approval
from the Department of Taxation and file instead form ST-11A (Sales and Use Tax Certificate of
Exemption for construction contractors and non-manufacturers).
The Department of Taxation archives information obtained from regional DEQ offices on candidates
for obtaining the ST-11A form, including contractors and other firms. In some instances, the files
contain information regarding the entity receiving certification, including public entities like local gov-
ernments that did not need to use the exemption since they use the governmental exemption instead.
The files contain information on claimants, including the construction contractor or other firm pur-
chasing and installing the equipment, a description of the project, a contact address that is either the
installation site or the contracting firm address, date of application, and date of approval. This infor-
mation was compiled into a database, cleaned, and aggregated to identify firms potentially using the
exemption in years 2010–2017 and the number of pollution mitigation projects for which they might
use the exemption. In total, 280 firms were identified as potentially using the exemption during the
2015–2017 period for an estimated 1,284 projects. In addition, since the contact information was often
imprecise or incorrect, supplemental contact information was gathered from a variety of industry
sources, including points of contact within the firm such as CEO/president/owner or CFO/director
of finance/comptroller.
Using this contact database, the Weldon Cooper Center conducted a survey of the 280 firms. Firms
were asked to provide the total amounts spent on exemption eligible equipment and number of pol-
lution control projects for each year over the period 2015–2017. In addition, firms were asked to
identify their 2-digit NAICS industry. Finally, firms were given an opportunity to offer comments
about the program and their use or non-use of the exemption during the period. Firms were contacted
multiple times by email, regular mail, and telephone to encourage participation in the survey. The
survey was offered in an online format using Qualtrics software with the option of mailing a com-
pleted survey form and sending by mail in a postage-paid, self-addressed envelope.
Seventy-four firms responded out of an adjusted sample size of 230 firms (accounting for duplicates,
firm closures, and other usual circumstances) for an adjusted response rate of 32.2 percent. Estimates
of the exemption-eligible purchases were made by extrapolating the average spending on a per-project
basis ($208,310), weighted by the original number of projects in the contact database. These estimates
were extrapolated backwards in time (2010–2014) based on the number of projects and an inflation
adjustment factor based on the consumer price index. State tax revenue impact was obtained by mul-
tiplying the effective state sales and use tax rates for each fiscal year (3.92 in 2010; 3.98 in 2011 and
2012; and 4.28 in 2014–2017) by estimated eligible sales.
75
Appendixes
The results (an average revenue impact of $3.815 million in FY15–FY17) are similar to Department
of Taxation estimates based on an earlier (pre-2010) survey. Using this base information, TAX pro-
jected forward by inflating the baseline number by the consumer price index to adjust to nominal
values, with FY17 estimated at $3.64 million. Another comparison was made to estimates based on
information from the U.S. Census Bureau’s Pollution Abatement Costs and Expenditures (PACE):
2005 and the Annual Survey of Manufacturers. The Annual Survey of Manufacturers reports total
capital expenditures of $2.731 billion, while PACE reports depreciation of $46.1 million on pollution
control equipment used by manufacturers for 2005. To determine the tax impact, the Weldon Cooper
Center used depreciation as a proxy for replacement expenditures needed to maintain pollution con-
trol efforts and extrapolated Virginia manufacturing capital expenditures of $2.963 million in 2016.
Therefore, Weldon Cooper estimated that Virginia manufacturers in 2016 spent $50 million to gener-
ate a $2.65 million sales and use tax impact.
Employment performance tracking
Four of the incentive programs examined in this report establish employment performance bench-
marks as a condition for receiving the incentive. The MOUs between VEDP and Micron and
Qimonda specify employment levels before grant funds will be disbursed. The data center tax exemp-
tion also establishes minimum job requirements. Manufacturers single sales apportionment requires
firms to maintain employment for three years to retain eligibility. Employment must be kept at least
90 percent of the year before electing single sales apportionment. Information on compliance with
the employment requirements for the Micron and Qimonda grants and data center exemption was
obtained from VEDP based on company reporting and VEDP follow-up.
The Weldon Cooper Center evaluated electors’ compliance with employment performance require-
ments by examining Virginia Employment Commission ES202 annual employment data for the 2013–
2017 period. Program project records for tax year 2014 manufacturers singles sales apportionment
electors were matched with 2013–2017 quarterly VEC ES202 payroll employment data using FEIN
(i.e., taxpayer identification number) and company name from Department of Taxation Corporate
Income tax records (forms 500 and 500A). The FEIN (Federal Employer Identification Number) is a
unique nine-digit number that identifies a firm for federal tax purposes. VEC employment data was
aggregated to the FEIN level for each firm.
Interviews with agencies and stakeholders
JLARC and Weldon Center Cooper staff held meetings and phone conference calls with staff from
agencies administering the incentives evaluated for this report and include the
Department of Environmental Quality;
Department of Mines, Minerals, and Energy;
Department of Taxation;
Virginia Economic Development Partnership; and
Virginia Clean Energy Program.
In addition, meetings were held with interested members of the Virginia Manufacturers Association
Taxation Committee to discuss the importance of manufacturing incentives and consultants for the
76
Appendixes
Northern Virginia Technology Council. Representatives of semiconductor manufacturers and data
centers were also interviewed.
In December 2018 and January 2019, Weldon Cooper Center and JLARC analysts interviewed seven
data center companies with a presence in Virginia. The names and contacts for the data centers were
provided by the Northern Virginia Technology Council as representatives of the industry. They in-
cluded firms operating enterprise and colocation data centers, with some interviewees operating facil-
ities outside of the Northern Virginia region. Those interviewed included company staff involved in
finance and accounting, state and local taxation, state relations, public policy, technology, economic
development, and site selection. Each had significant experience with their firms and experience
needed to assess site selection criteria and the importance of economic development incentives. All
of the firms interviewed are currently using the data center sales and use tax exemption. Some of the
enterprise data centers interviewed might also be eligible to use the newly enacted single sales appor-
tionment for data centers in the future. Most had centers elsewhere in the U.S. and some internation-
ally. They represent collectively over one quarter of the estimated number of data centers in Virginia.
During the interviews, data center representatives were asked about what location and expansion fac-
tors are most important for their firm; the role of incentives in location decisions; Virginia’s infra-
structure and other factors, including incentives, considered in site location decisions compared with
other states; potential improvements of Virginia’s data center incentive programs; data centers’ impact
on local and state economic development; and the technological, economic, and social trends affecting
data center growth in Virginia and elsewhere.
Review of data center and manufacturing incentives in other states
Weldon Cooper Center staff reviewed several sources to obtain information on comparable manu-
facturing and data center incentives offered by other states. Sources often varied by the type of in-
centive, since there is no authoritative comprehensive source on all state incentives. The Council for
Community and Economic Research (C2ER) online State Business Incentives Database, which re-
flects incentive programs in place as of 2015 (the last time the database was updated), was used to
confirm and supplement information for several programs (e.g., data centers, semiconductor incen-
tives, biodiesel tax credits, recycling tax credits). For corporate manufacturing apportionment factors,
Commerce Clearing House or CCH (2018) and Federation of Tax Administrators (2018) survey in-
formation was used. CCH (2018) was also used to identify recycling equipment tax credits. CCH
(2017) and Potter, Steward and Kessler (2017) were used to assemble information on pollution con-
trol equipment and facilities sales and use exemptions and other similar types of pollution control
investment incentives (e.g., tax credits, property tax abatements, loans). Keve, Verdi, and English
(2010) served as the starting point for state biodiesel incentives, although many of the programs
listed by that publication have since expired. Search engine keyword searches and information culled
from a report by Washington JLARC (2016) were the primary sources of information to identify
comparable semiconductor manufacturing sales and use tax exemptions, while Good Jobs First pro-
vided information on large semiconductor industry customized incentive programs to attract new
plants or encourage plant expansion. Information on state data center incentives was assembled with
assistance of Tarczynska (2016) and Mangum Economics (2018). In each instance, these sources
were supplemented with reviews of state departments of economic development and tax/revenue
77
Appendixes
websites to confirm that the information was still current or to obtain a better understanding of par-
ticular features of incentives.
Review of documents and literature
During this study, several sources of information, including documents, reports, and published or
unpublished research, were examined. The purpose of this literature review was to understand the
purpose and goals of Virginia incentive programs, industry site location factors, role and importance
of economic incentives, market imperfection rationales for programs, and methodological ap-
proaches for quantifying the economic and tax revenue impacts of economic incentives. Sources
consulted included
materials describing the programs, Virginia agency reports describing program usage, and leg-
islative statutes authorizing the programs;
state evaluations and economic impact studies published by state agencies or their consultants
in other states;
scholarly books and articles that examine the (1) features and locational characteristics of the
manufacturing industry, semiconductor manufacturing industry, and data centers and (2) effect
of corporate tax apportionment factors and other types of economic incentives on firm
growth, employment, and other outcomes (e.g., environmental emissions); and
studies that attempt to quantify the economic impact of economic development incentives us-
ing forecasting and retrospective modeling methods.
78
Appendixes
Appendix C: State data center incentives
A majority of states offer data center sales and use tax exemptions to attract data centers. This ex-
emption has greater value for data centers than for other industries, including manufacturers, because
of the high cost of capital investment in plant and equipment and a rapid depreciation cycle of three
to five years. Twenty-eight states currently have some form of sales and use tax exemption for data
centers (Table C-1).
Table C-1
Data center sales and use tax exemptions by state
Year
Other data center
State established/ Requirements Eligible purchases
incentives
last enhanced
Create at least 20 jobs and average
compensation of $40,000. Sales and Construction and Property Tax
Alabama 2012
use exemption period varies with equipment Abatement
amount of investment
Alaska
Capital investment $25 million to $50
Arizona 2013 Equipment
million within 5 years
State offers energy
efficiency program
Connecticut to enhance efficiency
for energy-intensive
businesses
Capital investment of at least $150
million dollars over 5 years and a
“critical IT load” of 15 megawatts or Infrastructure,
Enterprise zone
Florida 2017 higher and a critical IT load of 1 equipment, personal
program eligible
megawatt or higher dedicated to property, and electricity
each individual owner or tenant
within the data center.
Tax credit available
for
''telecommunication
Sales and use exemption for $15 Computer hardware support companies”;
Georgia 2005
million capital investment and software 8% of qualified
investments for
investment of $50k
or more
79
Appendixes
Year
Other data center
State established/ Requirements Eligible purchases
incentives
last enhanced
Create at least 20 "quality jobs" (30-
hour regular workweek, pays at least
110% of average wage) for single
user and colocation and qualified Equipment, materials,
Georgia 2018 customers. Minimum investment components used in
thresholds for 7 year period are $100 data centers
million for counties with populations
less than 30k, $150 million for 30–
50k), and $250 million (50k+)
Computing, networking,
$10 million investment in qualifying
data storage
Indiana 2012 high technology districts and pays at
equipment, and power
least 125% of average county wage
supply equipment
Exemption ($200M investment) and
refund (50%) with items and duration
Construction,
depending on investment level ($1
computers and Personal property tax
Iowa 2007 / 2009 million and $200 million). Facility
equipment, fuel, and abatement
must adhere to sustainable design
electricity
standards and be at least 5,000
square feet
State does not have
Kansas new equipment
property tax
Communications
Tax refund for capital equipment equipment, computer
Kentucky 2009
investment of at least $100 million equipment, and
computer software
Data energy
Maryland
efficiency grant
Exemption for equipment purchases
for enterprises that receives 75% or Construction materials
Michigan 2015 / 2016
more of its sales from colocated and equipment
businesses
Exemption for data centers with at
Equipment, software, Property tax
Minnesota 2012 least 25,000 square feet and
and energy abatement
investment of $30 million
Exemption for investment of at least
Construction and
Mississippi 2010 $50 million, creates at least 50 jobs,
equipment
and pays 150% of average state wage
New centers must have investment of
at least $25 million and create 10
jobs. Existing centers must invest $5
Missouri 2010 / 2015
million and create 5 jobs. Jobs must
pay at least 150% of county average
wage
Property tax
Montana No state sales tax
abatement
80
Appendixes
Year
Other data center
State established/ Requirements Eligible purchases
incentives
last enhanced
General economic development
Job creation tax
program has multitiered system of
credit. Investment
sales tax refunds based on
tax credit and
Nebraska 2012 investment and employment levels
personal property
with specific thresholds. Data center
tax exemption also
eligibility starts at $3 million in
are awarded.
investment and 30 new full-time jobs
Partial (75%) sales and use exemption Partial personal
Nevada 2015 for $25 million investment and 10 property tax
jobs at 100% of state average wage abatement
New
No state sales tax
Hampshire
Equipment, software,
New York 2000 Internet web hosting data centers and other personal
property
North $75 million investment and meet Equipment, software,
2006 / 2015
Carolina county wage and benefit standards and electricity
Data centers at least 16,000 square
North Dakota 2015
feet
$100 million investment and $1.5
Construction materials
Ohio 2011 / 2013 million payroll. Colocation centers
and equipment
eligible
Computer system sales tax
Oklahoma 1993 exemption (not data center specific) Computer systems
if 80% of sales from out-of-state
Property tax
Oregon No state sales tax
abatements
Sales and tax refund. Minimal
investment of $50 million in counties
with 250,000+ population or $25
Equipment, software,
million with less than 250,000
Pennsylvania 2016 and other personal
population and pays at least $1
property
million in total employee
compensation. Program cap of $5
million.
Rhode Island
$50 million minimum investment and
South
2012 25 job creation minimum. Jobs must
Carolina
pay at least 150% of average wage.
Computer systems,
$100 million investment and 15 jobs
communication
created that pay 150% of state
Tennessee 2007 / 2016 systems, software,
average occupational wage and offer
backup power and
health-care benefits.
HVAC equipment
Computers, equipment,
$200 million investment and create at
HVAC and power
Texas 2013 least 20 jobs that pay 120% of
equipment, electricity
average county wage.
and fuel
81
Appendixes
Year
Other data center
State established/ Requirements Eligible purchases
incentives
last enhanced
Machinery, equipment
and replacement parts,
Utah 2010 / 2016 At least 150,000 square feet
with economic life of at
least one year
$150 million investment, create 50
jobs that pay 150% of average Property tax
Virginia 2009 / 2016 county wage. Job creation threshold abatement (local
is 25 jobs in rural or economically option)
distressed counties.
Rural enterprise zone. Must be at Server equipment and
Washington 2010
least 100,000 square feet power infrastructure
Construction materials,
No specific minimum investment or Personal property tax
West Virginia 2009 equipment, and
job creation levels abatement
software
Lower tier investment
creates exemption for
Two-tier structure. $5 million and $50
computer equipment.
million infrastructure investment tiers
Wyoming 2011 Higher tier qualifies
combined with $2 million in
power supply and
equipment and software expenditure.
generation and HVAC
equipment also.
SOURCE: Weldon Cooper Center analysis based on Mangum Economic (2018), Tarczynska (2016), and C2ER Business Incentives Data-
base.
82
Appendixes
Appendix D: Corporate tax apportionment formulas by state
Despite the original recommendation by state tax professionals, many states have questioned the ra-
tionale and continued relevancy of the standard equally weighted, three-factor (property, payroll, and
sales) apportionment formula. Currently, only a few states still use it. Today, most states use single
sales factor apportionment (Table D-1). Unlike Virginia, most states require all industries to use single
sales apportionment rather than allowing only certain industries to use it.
TABLE D-1
Corporate tax apportionment rules by state
State Apportionment Formula
Alabama 3 factors with sales double-weighted
Alaska Sales, property, and payroll equally weighted
Arizona Elective sales or 3 factors with sales double-weighted
Arkansas 3 factors with sales double-weighted
California Sales
Colorado Sales
Connecticut Sales
Delaware 3 factors with sales triple-weighted; Single-sales factor phased in by tax year 2020
District of Columbia Sales
3 factors with sales double-weighted. Corporations making capital investment of at least $250
Florida
million may choose one-factor sales
Georgia Sales
Hawaii Sales, property, and payroll equally weighted
Idaho 3 factors with sales double-weighted
Illinois Sales
Indiana Sales
Iowa Sales
Sales, property, and payroll equally weighted. Corporations where payroll exceeds 200% of
Kansas
average of property and sales factor may choose two-factor formula of property and sales.
Kentucky Sales
Louisiana Sales
83
Appendixes
State Apportionment Formula
Maine Sales
Maryland Sales/3 factors with sales triple-weighted; Single-sales factor phased in by tax year 2022
Massachusetts Sales
Michigan Sales
Minnesota Sales
Factor formula varies by manufacturing category with sales, property, and payroll equally weighted
Mississippi for wholesale sellers, three factors with sales double-weighted for retail sellers, and single sales for
selected other incentivized manufacturers
Missouri Sales, property, and payroll equally weighted
Montana Sales, property, and payroll equally weighted
Nebraska Sales
Nevada No state corporate income tax
New Hampshire 3 factors with sales double-weighted
New Jersey Sales
New Mexico Elective sales factor or 3 factors equally weighted
New York Sales
North Carolina Sales
Sales, property, and payroll equally weighted. Recently underwent gradual phase in of incentive
North Dakota that since tax year 2018 allows manufacturers to elect to use sales as a single-factor. Election is
binding for five years.
Ohio N/A*
Sales, property, and payroll equally weighted. Corporations making capital investment of at least
Oklahoma
$200 million may choose three factors with sales double-weighted.
Oregon Sales
Pennsylvania Sales
Rhode Island Sales
South Carolina Sales
South Dakota No state corporate income tax
Tennessee 3 factors with sales triple-weighted
Texas Sales
Elective 3 factor with sales double-weighted or 3 factors equally weighted. Single sales option will
Utah
be phased in by tax year 2021 with optional opt-out category
Vermont 3 factors with sales double-weighted
Virginia Elective sales or 3 factors with sales double-weighted for manufacturing and data centers
Washington No state corporate income tax
West Virginia 3 factors with sales double-weighted
84
Appendixes
State Apportionment Formula
Wisconsin Sales
Wyoming No state corporate income tax
SOURCE: Weldon Cooper Center review of 2019 U.S. Master Multistate Corporate Tax Guide and Federation of Tax Administrators (2018).
*Ohio has a gross receipts tax instead of corporate income tax.
85
Appendixes
Appendix E: Semiconductor incentives by state
States want to recruit and retain semiconductor manufacturers because they require large capital in-
vestments and create a large number of jobs. States often offer large “megadeal” ($75 million or more)
incentives or customized incentive packages (Table E-1).
Table E-1
Multiple states have offered “megadeal” incentives to attract semiconductor manufacturers
State Incentive package
$84 million in tax abatements, exemptions, and credits for Intel to locate $1.3 billion plant employing
Arizona 1,500 workers (1994)
$17.5 million in tax credits for Intel plant expansion (2014)
$116 million in tax abatements, tax increment financing, and Business Equipment Tax
Reimbursements for National Semiconductor (later acquired by Texas Instruments) to establish $500
Maine million plant and create 450 jobs (1997)
$17 million from the Business Equipment and Tax Reimbursement program to Texas Instruments
(2009–2016)
$645 million in property and other tax abatements, tax credits, and corporate income tax reductions
New Mexico to Intel for expansion (1993)
$2 billion in financial assistance through an industrial revenue bond for Intel expansion (2004)
$1.5 billion in grants, loans, property tax abatements, sales tax exemptions, tax credits, and state
investment in R&D and education facilities for IBM (now owned by Global Foundries) (2008)
New York
$1.4 billion in capital grant and tax credits to Global Foundries for establishing $4.6 billion facility in
Saratoga County (2010-2012)
North $14.1 million in grants and tax credits to RF Micro Devices and Triquint Semiconductors, and their
Carolina successor Qorvo, for plant expansions and other activities (2002–2012)
$600 million in local property tax abatements over 24 years and a $50 million grant to Texas
Instruments for establishing a $3 billion plant employing 1,000 workers (2003)
Texas $233.4 million in local property abatements and a $10.8 million grant to Samsung to locate a $3.5
billion plant to employ 900 workers (2006)
$83.6 million in property tax abatements to Samsung for expansion of plant (2012)
$121.5 million in incentives to establish Intel plant (1996)
$200 million in incentives for Intel expansion (1999)
Oregon
$579 million in tax incentives for retention and expansion of Intel plant (2005)
$2 billion in property tax abatements for additional investment by Intel (2014)
$125 million in property tax abatements, sales tax exemptions, and other assistance to Micron to
establish plant (1995)
Utah
$45.9 million in tax increment financing to Micron for a $1.5 billion plant expansion to employ 200
additional workers in joint venture with Intel (2011)
$18.6 million custom grant for Dominion Semiconductor for $4 billion plant creating 4,000 jobs (1996)
$15 million custom grant for White Oak Semiconductor Partnership for $1.5 billion plant creating
Virginia 1,500 jobs (1997)
$59.8 million in custom grants for expansion of Micron and Qimonda plants to collectively employ
2,060 additional workers and invest $2.4 million (2005)
SOURCE: Weldon Cooper Center review of Good Jobs First.
86
Appendixes
Appendix F: Semiconductor manufacturing is concentrated in
six states
Most (74 percent) semiconductor and related device manufacturing is concentrated in six states with
location quotients well over 1.0: Arizona, California, Idaho, Massachusetts, Oregon, and Texas (Table
F-1). Three of these states—California, Massachusetts, and Oregon—also have most (77 percent) of
the nation’s employment in semiconductor machinery manufacturing.
Table F-1
Semiconductor manufacturing and related supply chain industries are highly concentrated in
a few states
Semiconductor and related Semiconductor machinery
device manufacturing manufacturing Upstream industries
Location Location Location
Jobs quotient Jobs quotient Jobs quotient
Arizona 17,443 5.02 837 2.20 5,049 0.69
California 46,568 2.12 10,179 4.23 91,996 2.00
Colorado 2,388 0.71 81 0.22 6,938 0.98
Florida 7,602 0.71 188 0.16 9,747 0.43
Idaho 8,619 9.39 18 0.18 838 0.43
Maine 796 1.01 0 0.00 472 0.28
Massachusetts 8,302 1.86 2,403 4.92 15,887 1.70
Minnesota 2,170 0.61 78 0.20 10,316 1.37
Missouri 1,102 0.31 2 0.00 3,355 0.45
New Hampshire 500 0.60 50 0.55 3,046 1.75
New Jersey 1,828 0.37 30 0.06 3,982 0.38
New Mexico 1,943 1.86 0 0.00 3,596 1.64
New York 7,731 0.67 545 0.43 16,582 0.68
North Carolina 4,646 0.84 25 0.04 8,169 0.70
Ohio 1,732 0.26 88 0.12 11,028 0.79
Oregon 25,279 10.52 2,672 10.15 2,583 0.51
Pennsylvania 1,936 0.27 440 0.55 8,800 0.58
South Dakota 673 1.21 0 0.00 831 0.71
Texas 27,869 1.82 763 0.46 47,537 1.48
Utah 2,272 1.26 14 0.07 964 0.25
Vermont 3,030 7.47 27 0.62 182 0.21
Virginia 1,602 0.33 7 0.01 4,452 0.43
87
Appendixes
Semiconductor and related Semiconductor machinery
device manufacturing manufacturing Upstream industries
Location Location Location
Jobs quotient Jobs quotient Jobs quotient
Washington 3,090 0.73 0 0.00 12,525 1.41
All Others 2,901 0.05 1,502 0.22 113,594 0.87
Total 182,022 19,952 382,469
SOURCE: Weldon Cooper Center analysis of EMSI employment data..
NOTE: Upstream electronic industries using approximately 20% of semiconductor and related device output according to IMPLAN in-
clude printed circuit assembly manufacturing (334418), computer storage device manufacturing (334112), wireless telecommunications
carriers (517312), electronic computer manufacturing (334111), telephone apparatus manufacturing (334210), and motor vehicle electri-
cal and electronic equipment manufacturing (336320).
88
Appendixes
Appendix G: Virginia custom grants for semiconductor
manufacturers
TABLE G-1
Virginia approved five custom grants to attract semiconductor production facilities FY96–
FY17
Custom grant: Motorola
Purpose Encourage White Oak Semiconductor to construct and operate a 200mm semiconductor production
facility in Henrico County
Performance Enter into MOU with VEDP
requirements Create 1,500 FTEs and $1.5 billion in capital investment
Payment In the amount of $100 per memory wafer and $250 per logic wafer produced and sold by the Henrico
structure and facility
schedule Authorized between 2003 and 2009 if performance achieved and VEDP notified pursuant to MOU.
Maximum aggregate payment was $15 million to be paid in annual installments that may vary
depending on attainment of performance and number of chips sold but are not to exceed
- $3 million in 2003
- $6 million in 2004 (less any amounts paid in 2003)
- $9 million in 2005 (less any amounts paid since 2003)
- $12 million in 2006 (less any amounts paid since 2003) and
- $15 million for 2007 through 2009 (less any amounts paid since 2003)
Custom grant: Dominion Semiconductor LLC (became Micron)
Purpose Encourage Dominion Semiconductor—a joint venture of Toshiba and IBM—to occupy a shutdown
facility in the City of Manassas and produce DRAM (dynamic random-access memory) memory chips
Performance Enter into MOU with VEDP
requirements Create 4,000 FTEs and $4 billion in capital investment within 3 performance milestones (cumulative):
- Milestone 1: 1,200 FTEs and $1 billion in capital investment by December 1998
- Milestone 2: 2,500 FTEs and $2.5 billion in capital investment by December 2002
- Milestone 3: 4,000 FTEs and $4 billion in capital investment by December 2005
Payment Begin 5 years after milestone is reached and VEDP is notified
structure and In the amount of $100 per memory wafer and $250 per logic wafer produced and sold by the
schedule Manassas facility with annual and cumulative maximums depending on the milestone achieved.
- Milestone 1: $3,720,000 annual and $18.6 million cumulative
- Milestone 2: $6,080,000 annual and $30.4 million cumulative
- Milestone 3: $7,680,000 annual and $38.4 million cumulative
89
Appendixes
Custom grant: White Oak Semiconductor Partnership (became Infineon)
Purpose Encourage White Oak Semiconductor to construct and operate a 200mm semiconductor production
facility in Henrico County
Performance Enter into MOU with VEDP
requirements Create 1,500 FTEs and $1.5 billion in capital investment
Payment In the amount of $100 per memory wafer and $250 per logic wafer produced and sold by the Henrico
structure and facility
schedule Authorized between 2003 and 2009 if performance achieved and VEDP notified pursuant to MOU.
Maximum aggregate payment was $15 million to be paid in annual installments that may vary
depending on attainment of performance and number of chips sold but are not to exceed
- $3 million in 2003
- $6 million in 2004 (less any amounts paid in 2003)
- $9 million in 2005 (less any amounts paid since 2003)
- $12 million in 2006 (less any amounts paid since 2003) and
- $15 million for 2007 through 2009 (less any amounts paid since 2003)
Custom grant: Micron
Purpose Encourage Micron to expand its 300mm semiconductor wafer plant in Manassas
Performance Enter into MOU with VEDP
requirements Create 860 FTEs and $1.2 billion in capital investment within 3 performance tiers (cumulative)
- Tier 1: 300 FTEs and $400 million in capital investment by 2007
- Tier 2: 640 FTEs and $800 million in capital investment by 2009
- Tier 3: 860 FTEs and $1.2 billion in capital investment by 2011
Payment Maximum of $27 million
structure and Begin 5 years after performance for each tier reached and VEDP notified
schedule
- Tier 1: $8 million to be paid in 5 annual installments of $1.6 million
- Tier 2: $11 million to be paid in 5 annual installments of $2.2 million
- Tier 3: $8 million to be paid in 5 annual installments of $1.6 million
Custom grant: Infineon (became Qimonda)
Purpose Encourage Infineon to expand its semiconductor plant in Henrico
Performance Enter into MOU with VEDP
requirements Create 1,200 FTEs with an average annual salary of at least $65,000 and make $1.2 billion in capital
investment within 3 performance tiers (cumulative)
- Tier 1: $550 million in capital investment by 2006
- Tier 2: 1,000 FTEs by 2008
- Tier 3: 200 additional FTEs and $1.2 billion in capital investment by 2011
Payment Maximum of $55 million
structure and Begin 5 years after performance for each tier reached and VEDP notified
schedule
- Tier 1: $15 million to be paid in 2 annual installments of $7.5 million
- Tier 2: $35 million to be paid in 4 annual installments of $8.75 million
- Tier 3: $5 million to be paid in 2 annual installments of $2.5 million
SOURCE: Weldon Cooper Center review of Chapters 642, 645, 651, and 652, 1996 Acts of Assembly; Chapter 789, 1997 Acts of Assembly;
Chapter 392, 2005 Acts of Assembly; and documentation provided by VEDP.
90
Appendixes
Appendix H: Semiconductor manufacturing sales and use tax
exemptions by state
Only nine states, including Virginia, offer exemptions from the sales and use tax to semiconductor
manufacturers (Table H-1).
TABLE H-1
Semiconductor manufacturing sales and use tax exemptions by state
State Name Description
Exempts semiconductor clean rooms used for
Clean room* sales and use tax
Arizona manufacturing, processing, fabrication, or R&D
exemption
from sales and use tax.
Sales and use tax exclusions for qualifying
manufacturing enterprises, including
Manufacturing and R&D partial sales
California semiconductor industry. Includes tangible
and use exemption
personal property used in manufacturing, R&D,
testing, and special purposes.
Sales and use tax exemption on semiconductor,
Semiconductor, defense, or space defense, and space technology-based purchases
Florida
technology sales and use tax exemption of production and/or research and development
equipment.
Exempts machinery and equipment and materials
used in the construction or operation of a clean
Georgia Clean room sales and use tax exemption room of Class 100 or less when the clean room is
used directly in the manufacture of tangible
personal property.
Sales and use tax exemption on tangible personal
Clean room environment in
property used in, or to maintain, clean rooms and
Idaho semiconductor manufacturing sales and
for research and development or production of
use tax exemption
semiconductors.
Additional sales exemption for semiconductor
manufacturers beyond what is available to other
Semiconductor manufacturing manufacturers, including tangible personal
Texas
exemption property outside the cleanroom such as
integrated systems, airflow systems, fixtures and
piping, and other items
Sales and use tax exemption for fabricating,
Semiconductor fabricating, processing, processing, or R&D materials used for research or
Utah
or R&D material exemption development, manufacturing, or fabricating of
semiconductors.
Semiconductor manufacturing Sales and use tax exemptions for equipment, fuel,
Virginia exemption; Semiconductor wafer supplies, wafers used in semiconductor
exemption manufacturing
91
Appendixes
State Name Description
The Reduced B&O tax rate for manufacturers of
Reduced business and occupation semiconductor materials is available to
(B&O) tax rate for manufacturers of manufacturers or processors for hire of
Washington semiconductor materials; sales/use tax semiconductor materials; the sales and use tax
exemption for purchases of exemption for purchases of semiconductor gasses
semiconductor gasses & chemicals and chemicals is available to manufacturers or
processors for hire of semiconductor materials.
SOURCE: Weldon Cooper Center analysis based on Washington JLARC (2016) and C2ER Business Incentives Database.
NOTE: *Clean rooms are laboratories that contain extremely low levels of air contaminants. They are used by specialty manufacturers, such
as semiconductor and biotechnology companies.
92
Appendix I: Pollution control equipment exemptions by state
Sales and use tax exemptions are the most common incentive for the purchase of pollution control equipment and facilities (Table I-1). Only
five states with sales and use taxes do not offer the exemption, though they typically offer other incentives such as property tax exemptions.
TABLE I-1
Pollution control equipment sales and use tax exemptions by state
Have Industry or other Other pollution control
State exemption Air Water Land/solid Other Eligible purchases restrictions incentives
Facilities, machinery, Gross receipts tax deduction,
Alabama Yes X X
equipment and materials property tax exemption
No sales
Alaska
tax
93 Manufacturing,
processing, power Appendixes
Arizona Yes X X X Machinery and equipment generation, mining, Business Income tax credit
and selected other
industries
Tangible personal property
installed and used by
Arkansas Yes X X
manufacturing and processing
industries
Have Industry or other Other pollution control
State exemption Air Water Land/solid Other Eligible purchases restrictions incentives
Exemption
restricted to
transfers between
California Yes California Pollution Loan programs
Control Financing
Authority and
participating party
Exemption
availability depends
Colorado Yes X Equipment
on amount of state
fiscal surplus
X (Solid
Tangible personal property Business income tax credit,
Connecticut Yes X X X waste to
installed or used property tax exemption
energy)
No sales
94 Delaware
tax Appendixes
District of
No
Columbia
Manufacturing and
Facility, equipment, machinery other goods
Florida Yes X X X Property tax
and chemicals. producing
industries
Machinery, equipment, and Business income tax credit (part of
Georgia Yes X X X
parts investment tax credit)
Facility, machinery, equipment,
Hawaii Yes X Property tax exemption
and supplies
Tangible personal property Business income tax credit (part of
Idaho Yes X X
used for pollution control investment tax credit)
Have Industry or other Other pollution control
State exemption Air Water Land/solid Other Eligible purchases restrictions incentives
Restricted to low
sulfur dioxide
Illinois Yes X Equipment Loan program
emission coal fueled
devices
Manufacturing,
Tangible personal property
processing, refining,
Indiana Yes X X used in construction and Property tax exemption
mining, or
operation
agriculture
Iowa Yes X Equipment Manufacturing Property tax exemption
Manufacturing or Property tax exemption (power
Kansas Yes X X X Machinery and equipment
processing generation only)
Facility, machinery, and
Kentucky Yes X X X X (noise) Reduced personal property tax rate
equipment
80% property tax abatement for
Louisiana Yes X X X X (noise) Machinery and equipment
equipment
95
Machinery, equipment, and
Maine Yes X X Property tax exemption
parts
Appendixes
Construction and items
Property tax exemption for coal
Maryland Yes X X considered part of real
industry
property
Business income tax credit (part of
Massachusetts No investment tax credit), property tax
exemption
Facilities, machinery, and
Michigan Yes X X Property tax exemption
equipment
Steel reprocessing
firm, resource
X (solid
Equipment; facilities for waste- recovery facility, and
Minnesota Yes X X X waste to Property tax exemption
to-energy waste-to energy
energy)
resource recovery
facility
Have Industry or other Other pollution control
State exemption Air Water Land/solid Other Eligible purchases restrictions incentives
Manufacturing and
Mississippi Yes X X X Equipment
processing
Materials and supplies for
Missouri Yes X X construction of facility;
machinery and equipment
No sales
Montana Property tax exemption
tax
Tangible personal property
Nebraska Yes X X used for pollution control
facility
Nevada No Property tax exemption
New No sales
96 Property tax exemption
Hampshire tax
Appendixes
X (zero Wastewater
New Jersey Yes X X emission Equipment treatment facilities, Property tax exemption
vehicles) other
New Mexico No Property tax exemption
Manufacturing,
Business income tax credit (part of
must result in over
New York Yes X X Machinery and equipment investment tax credit), property tax
50% waste
exemption
reduction
North
Yes X X X Equipment and chemicals Manufacturing Property tax exemption
Carolina
Power plants, oil
refineries, gas
North Dakota Yes X X X Machinery and equipment Property tax exemption
processing plants,
recycling facility
Have Industry or other Other pollution control
State exemption Air Water Land/solid Other Eligible purchases restrictions incentives
X Materials or parts used in
Ohio Yes X X Property tax exemption
(noise) facility
Machinery, equipment, fuels,
Oklahoma Yes X
and chemicals
No sales
Oregon Property tax exemption
tax
Manufacturing,
Machinery, equipment, and
Pennsylvania Yes X X X (noise) processing, mining,
supplies
and public utilities
Tangible personal property Business income tax credit (part of
Rhode Island Yes X X X used in construction or investment tax credit), property tax
operation exemption
Manufacturing,
South processing,
Yes X X X (noise) Machinery Property tax exemption
97 Carolina recycling, and
mining Appendixes
South Dakota No Property tax exemption
Systems, chemicals, and
Tennessee Yes X X X Property tax exemption
supplies
Offshore spill
response
containment
property and repair,
remodeling, and
Texas Yes X X X Property tax exemption
restoration services
performed on
personal property
required to protect
environment
Have Industry or other Other pollution control
State exemption Air Water Land/solid Other Eligible purchases restrictions incentives
Property, materials, or services
Utah Yes X X used in construction or
installation
Vermont No Property tax exemption
X Manufacturing and
(landfill other goods
Personal property, equipment,
Virginia Yes X X X gas, producing Property tax exemption
facilities, and devices
solar, industries, mining,
storage) and public utilities
Manufacturing,
testing, research
Equipment or tangible
and development Business and occupation tax
Washington Yes X personal property, labor, and
operation; thermal exemption, utility tax exemption
services depending on industry
electric generation
98
facility Appendixes
Manufacturing,
transportation,
Tangible personal property or communication,
West Virginia Yes X X X Property tax exemption
services transmission, or
production of
natural resources
Vehicles, waste
treatment facility,
recycling, biomass
Wisconsin Yes X X X to residential fuel; Property tax exemption
wind, solar, and
gas-powered
products
Wyoming No Property tax exemption
Source: Weldon Cooper Center analysis based on 2018 State Tax Handbook, Potter, Stewart and Kessler (2017), and online search of state tax/revenue department websites
Appendixes
Appendix J: Recycling equipment tax credits by state
Thirteen states offer some form of recycling tax credit (Table J-1). One state—Nebraska—offers a
recycling grant program to businesses for recycling and waste reduction. Direct comparisons of the
tax credits are difficult because eligibility requirements vary widely, including eligible industries, types
of waste targeted, and eligible purchases. Several states had recycling equipment tax credits but allowed
them to expire: New Jersey (1996), Oklahoma (Recycling Reuse, and Source Reduction Incentive ex-
pired in 2014), and Oregon (2002 and 2008).
TABLE J-1
Recycling equipment tax credits by state
State Tax credit Eligibility Incentive amount
Available to manufacturers for cost of
10% of qualifying
Environmental Technology building or improving facilities where
Arizona construction and
Facility Tax Credit products are made out of recycled
equipment costs
materials.
Taxpayers engaged in business of
reducing, reusing, or recycling waste
for commercial purposes may claim
credit for purchase of related
equipment. Waste reduction, reuse,
or recycling equipment must be used
30% of cost, including
exclusively in the collection,
Arkansas Recycling Equipment installation, of
separation, processing, modification,
equipment purchased
conversion, treatment, or
manufacturing of products containing
at least 50% recovered materials, of
which at least 10% of the recovered
materials shall be post-consumer
waste.
99
Appendixes
State Tax credit Eligibility Incentive amount
(1) Manufacturer that derives at least
25% of weight of raw materials from $650 per new
recycled materials or materials employee plus $650
removed from solid waste stream. (2) per $100,000 of
(1) Recycled Materials and (2) Engaged in business of processing investment. In
Delaware Waste Materials Processing Tax materials from solid waste stream for targeted areas, credit
credits resale as raw materials to increased to $900 per
manufacturers. Facility must employ employee and $900
five or more employees and have per $100,000
qualified investment of $200,000 or investment
more in facility.
Manufacturing or
Investment Tax Credit can be
Georgia telecommunications facility. Can be 3–8% of investment
used for recycling equipment
used for construction or equipment.
Purchased equipment used to
manufacture a product from
postconsumer or postindustrial waste
(i.e., glass, paper, or plastic). Ninety
percent of equipment’s production
Idaho Postconsumer Waste Equipment must result in products using 20% of investment
postconsumer or postindustrial
waste. Product must comprise at
least 50% of postconsumer or
postindustrial waste ($30,000 tax
credit limit).
(1) credit equals 50%
of installed cost of
equipment. Credit
claimed in tax year
limited to 10% of
total credit allowable
(1) Purchasing of Recycling or
Purchase of recycling or composting and 25% of tax
Kentucky Composting Equipment (2)
equipment liability for year.
Major Recycling Project Credit
Amount is increased
for large projects:
limit is 50% of tax
liability in year and
maximum of $2.5
million.
14% of costs of
equipment or service
contract. Amount
New Recycling Manufacturing Purchases of qualified recycling
claimed may not
Louisiana or Process Equipment and/or manufacturing or process equipment
exceed 20% of total
Service Contracts Tax Credit or qualified service contracts.
credit allowable. May
not exceed 50% of tax
liability for tax period.
100
Appendixes
State Tax credit Eligibility Incentive amount
25% of cost of
property for first
Investments in depreciable property
$250,000 invested;
Investment in Property for used primarily to collect or process
Montana 15% for next
Recycling reclaimable material or manufacture a
$250,000 invested,
product from reclaimed material.
and 5% for next
$500,000 invested
Nevada No corporate income tax
Facility must invest at least $300
million and create at least 250 new
North 50% of purchase or
Recycling Facility Investments full-time jobs within four years of
Carolina lease costs
construction. Must be located in tier-
one development area.
20% of net
Businesses engaged in recycling,
investment in
Oklahoma Hazardous Waste Control reuse, or source reduction of
equipment and
hazardous waste
installation
Business operating a qualified
South recycling facility that makes minimum
Recycling Facility Investment 30% of investment
Carolina investment of $300 million by five
years after beginning construction
South Dakota No corporate income tax
Credit of up to $2,000 annually for
expenditures paid to third parties for 20% of net
Recycling Market Development
Utah expenses of establishing and expenditures up to
Zone Credits
operating recycling or composting $10,000
technology in state
Recycling machinery and equipment
20% of purchase
used predominantly in manufacturing
Purchases of Recycling price. May not exceed
Virginia facility or plan that manufactures,
Machinery and Equipment 40% of business tax
processes, or produces products from
liability in year.
recyclable materials for sale.
Washington No corporate income tax
Wyoming No corporate income tax
Source: Weldon Cooper Center Analysis based on 2019 U.S. Master Multistate Corporate Tax Guide, C2ER Business Incentives Database,
and online search of state taxation/revenue and economic development agency websites
101
Appendixes
Appendix K: Biodiesel production incentive by state
All states except for Delaware had at least one type of renewable energy tax incentive when Virginia’s
biodiesel tax credit was created. The high point for state incentives seems to be around the time of
the enactment of the federal credit and the temporary federal programs for environmental and green
job creation that were part of the American Recovery and Revitalization Act fiscal stimulus program.
Now, only nine states have biodiesel production incentives (Table K-1).
TABLE K-1
Biodiesel and green diesel production incentives by state
State Tax credit Requirements Other biofuel/biodiesel incentives
Income tax credit to biodiesel
suppliers equal to 5% of the cost of
Arkansas Biodiesel Incentive Act facilities and equipment used directly
in the wholesale or retail distribution
of biodiesel fuels.
Tax credit of 5 cents per gallon of low-
Ultra-Low Sulfur
California sulphur diesel fuel produced by a
Diesel Fuel Credit
small refiner.
Credit for biofuel (including biodiesel).
The Alternative Energy Loan Program
Company must produce at least 15
was established to help full-time
billion British thermal units (BTU) of
farmers, ranchers, and
Renewable Fuels renewable fuels per year. The credit is
Hawaii aquaculturalists produce renewable
Production Tax Credit equal to 20 cents per 76,000 BTU of
energy through sources such as
renewable fuels. The credit is capped
photovoltaic, hydroelectric, wind,
annually at $3 million per taxpayer
methane, biodiesel, and ethanol.
and $3 million in the aggregate.
Tax credit of $1 per gallon for
Kentucky Biodiesel Tax Credit production and blending of biodiesel
(credit cap of $10 million).
Missouri Qualified Biodiesel Producer
Incentive Fund, biodiesel produced in
the state by a facility that is at least
51% owned by Missouri agricultural
producers or which uses feedstock
that is at least 80% of Missouri origin,
Missouri
is eligible for a grant in any fiscal year
equal to 30 cents per gallon for the
first 15 million gallons produced from
Missouri agricultural products and 10
cents per gallon for the next 15
million gallons.
102
Appendixes
State Tax credit Requirements Other biofuel/biodiesel incentives
Credit for up to 15% of the cost of
Biodiesel/Biolubricant
investments in depreciable property
Montana Production Facilities
for constructing or equipping a
Credit
facility.
Biodiesel Facility Credit (30% of
Nebraska
amount invested in biodiesel facility)
Biofuel tax credit (15 cents for each
Biofuel Production Tax gallon of biofuel produced after
New York
Credit production of first 40,000 gallons;
capped at $2.5 million)
Biofuels PACE Program provides
interest buydown on loans to
biodiesel, ethanol, or green diesel
production facilities and livestock
Credit for investment in biodiesel or operations. The Envest Loan is to
Biodiesel or Green
green diesel production facility equal purchase shares in an agriculture
North Dakota Diesel Production
to 10% of equipment cost per year for processing plant intended to process
Equipment
five years (maximum of $250,000). North Dakota products or for the
purchase of equity shares in a North
Dakota feedlot or dairy operation
that feeds a byproduct of an ethanol
or biodiesel facility.
South Carolina provides several
income tax credits for a facility that
South Ethanol or Biodiesel produces ethanol or biodiesel at a
Carolina Production Credits plant in South Carolina at which all
fermentation, distillation, and
dehydration takes place.
A credit is available for Virginia
biodiesel and green diesel fuel
Virginia Biodiesel fuels credit
producers who produce up to two
million gallons of fuel per year.
Source: Weldon Cooper Center Analysis based on 2019 U.S. Master Multistate Corporate Tax Guide, C2ER Business Incentives Database,
eve, Verdi, and English (2010) and online search of state taxation/revenue and economic development agency websites.
103
Appendixes
Appendix L: Agency responses
As part of an extensive validation process, the state agencies and other entities that are subject to a
JLARC assessment are given the opportunity to comment on an exposure draft of the report. JLARC
staff sent an exposure draft of this report to the Virginia Economic Development Partnership; De-
partment of Taxation; Department of Environmental Quality; Department of Mines, Minerals, and
Energy; Secretary of Commerce and Trade; and Secretary of Finance.
Appropriate corrections resulting from technical and substantive comments are incorporated in this
version of the report. This appendix includes response letters from the
Department of Taxation,
Department of Environmental Quality, and
Virginia Economic Development Partnership.
104
COMMONWEALTH of VIRGINIA
I)c’jirtiiic’nt of JLUUtiOIl
June 10, 2019
Mr. Hal E. Greet, Director
Joint Legislative Audit and Review Commission
919 East Main Street, Suite 2101
Richmond, irginia 23219
Dear . reer:
Thank you for the opportunity to review and comment on the exposure draft report:
Data Center and Manufacturing Incentives. We believe the report is very well done and will
be useful to the members of the General Assembly going forward. As requested, we provided
our comments and clarification Linder separate cover, for your consideration in the report.
Thank you again for the opportunity to review the draft report. Should you have any
additional questions, please feel tree to contact me.
Sinc
Tax missioner
C: The Honorable Aubrey L. Layne, Jr., Secretary of Finance
Commonwealth of Virginia
VIRGINIA DEPARTMENT OF ENVIRONMENTAL QUALITY
1111 E. Main Street, Suite 1400, Richmond, Virginia 23219
P.O. Box 1105, Richmond, Virginia 23218
(800) 592-5482
www.deq.virginia.gov
Matthew J. Strickler David K. Paylor
Secretary of Natural Resources Director
(804) 698-4000
June 5, 2019
VIA EMAIL
Ms. Ellen J. Miller
Chief Economic Development and Quantitative Analyst
Joint Legislative Audit and Review Commission
ejmiller@jlarc.virginia.gov
RE: DEQ Comments on Provided Portions of JLARC Exposure Draft for
Data Centers and Manufacturing Incentives
Dear Ms. Miller:
DEQ appreciates the opportunity to provide comments regarding tax exemption certification for
pollution control equipment and facilities included in shared portions of JLARC’s Exposure
Draft for “Data Center and Manufacturing Incentives” (pp 49-62). In addition to specific
comments regarding the language provided, it is important to add some context and general
discussion regarding DEQ’s implementation of §58.1-3660.
DEQ currently has individual guidance available for each division (i.e., air, land, water) to
provide clarity for DEQ’s required certification of pollution control equipment and facilities
pursuant to §58.1-3660. DEQ’s role is clear; as the State Certifying Authority, DEQ certifies
that certain property meets the definition in the statute. Specifically, this section of the Code of
Virginia falls within Chapter 36 “Tax Exempt Property” under TAX’s Subtitle III “Local Taxes.”
§58.1-609.3 is a separate section within Chapter 6 “Retail Sales and Use Tax” under TAX’s
Subtitle I “Taxes Administered by the Department of Taxation” that provides commercial and
industrial exemptions for certified pollution control equipment and facilities as defined in
§58.1-3660. So, briefly stated, DEQ certifies that certain property meets the definition for a
local property tax exemption, and TAX provides a separate (piggybacked) sales and use tax
exemption for the property that DEQ has certified.
Ms. Ellen J. Miller
June 5, 2019
Page Two
DEQ has an additional responsibility under the Constitution of Virginia. Specifically Section 6 f
of Article X states in part that
“Exemptions of property from taxation as established or authorized hereby shall be
strictly construed...”
As a result, DEQ has attempted to meet this requirement by considering whether the property
requiring certification meets all parts of the definition. §58.1-3660 B defines “certified pollution
control equipment and facilities” as
“…any property, including real or personal property, equipment, facilities, or devices,
used primarily for the purpose of abating or preventing pollution of the atmosphere or
waters of the Commonwealth and which the state certifying authority having jurisdiction
with respect to such property has certified to the Department of Taxation as having been
constructed, reconstructed, erected, or acquired in conformity with the state program or
requirements for abatement or control of water or atmospheric pollution or
contamination.” [The definition goes on to include other property that is exempt
“…whether or not such property has been certified…” and will therefore not be
discussed with these comments.]
To certify (and strictly construe as required by the state constitution) whether property meets this
lengthy multipart definition, DEQ must consider each clause individually and in the context
provided. More specifically, DEQ is tasked by §58.1-3660 to certify the following:
“…any property, including real or personal property, equipment, facilities, or devices…”
(Is this property as opposed to a service or intangible?)
“…used primarily for the purpose of abating or preventing pollution of the atmosphere
or waters of the Commonwealth…”
(What is its primary purpose?)
“…and which [DEQ] having the jurisdiction with respect to the property…”
(Is DEQ the delegated authority for the program or requirements?)
“…having been constructed, reconstructed, erected or acquired…”
(Is it already onsite and installed?)
“…in conformity with the state program or requirements for abatement or control of
water or atmospheric pollution or contamination…”
(Is it required and installed in compliance with those requirements?)
2
Ms. Ellen J. Miller
June 5, 2019
Page Three
On the other hand, §58.1-609.3 is not prescriptive and simply requires TAX to exempt property
that DEQ has already certified under §58.1-3660. Considering a sales tax exemption at the time
of purchase instead of after DEQ certifies the property has been constructed, reconstructed,
erected, or acquired could result in purchasers not paying sales tax on a prospective property
improvement, and then owing sales tax on all or portions of the purchase that was not used to
install it. (Example: An owner/contractor/subcontractor prospectively purchases concrete and
related materials to build the foundation for an air pollution control device, but not all of it is
used and/or the contract is not awarded. How is the exempted sales tax recovered?)
DEQ provides the aforementioned information as general comments to JLARC’s draft report to
provide context and background for the following specific comments regarding pp 49-62.
Page 49, Paragraph 1, Sentence 1: The exemption for pollution control equipment and
facilities does not encourage or incentivize voluntary installation of the property; rather,
the exemption is provided only for equipment that is required for abatement or control of
pollution. Add “…or to compensate facilities required to control or abate pollution.”
Page 49, Paragraph 2: Per differences discussed above regarding §58.1-3660 and §58.1-
609.3, this paragraph appears to be incorrect (i.e., conflates the two statues).
Page 49, Paragraph 2, Sentence 1: Change “retail sales and use” to “property”
Page 49, Paragraph 2, Sentence 2: Change “sales taxes on purchases of” to “taxes on”
Page 49, Table 4-1, Title: Change “a” to “two” and “Exemption” to “Exemptions”
Page 49, Table 4-1, Eligibility: Change “Purchase of” to “Possession (for property tax
exemption) and purchase (for sales tax exemption) of”
Page 49, Table 4-1, Use of Exemption: Change “installing” to “required”
Page 50, Under Subtitle Achievement of Purpose: How is “high administrative costs”
defined? Does this consider an hourly rate for request preparers and DEQ reviewers
compared to the exemption received? Per the JLARC report, $26.59M over 8 years for 86
beneficiaries averages $309K per beneficiary. Many tax exemption certification requests
are only a few pages, and if the definition is clearly met, then the request and DEQ’s
response may be simple and brief. Other requests may not provide enough information to
clearly meet each of the definition’s requirements, and the iterative process may continue
until all pertinent information is submitted, received, and reviewed.
3
Ms. Ellen J. Miller
June 5, 2019
Page Four
Page 51, Paragraph 3: This discussion omits the fact that Virginia offers local property
tax exemption (the purpose of §58.1-3660, on which DEQ guidance is available), as well
as sales and use tax exemption (i.e., §58.1-609.3) provided by TAX after DEQ certifies
property under §58.1-3660.
Page 51, Margin Comment 2: §58.1-609.3 appears to have been enacted in 1993. §58.1-
3660 was enacted in 1972 and provides a property tax exemption, not a sales and use
exemption.
Page 51, Paragraph 5, Sentence 1: See discussion above on “high administrative costs”.
Is there any information available indicating that the costs to request certification is
actually more than the one-time sales tax exemption and/or the continuous property tax
exemptions received?
Page 51, Paragraph 5, Sentence 2: A case-by-case basis allows requests and DEQ staff to
spend an appropriate amount of resources based on the complexity of the request.
Anecdotally, there are examples of short as well as lengthy processing times.
Page 51, Paragraph 5, Sentence 4: Pursuant to the definition in §58.1-3660, DEQ cannot
certify the property prospectively. The language of the statute is in the past tense (i.e.,
“used,” “constructed,” “reconstructed,” “erected,” “acquired”) and thereby requires
DEQ’s approach. §58.1-609.3.9 requires that before the sales tax exemption can apply,
the pollution control equipment and facilities be “certified,” which DEQ cannot do before
it is “constructed, reconstructed, erected, or acquired.”
Page 51, Paragraph 5, Sentence 5: It is important to note that the property tax exemption
is continually provided every year it continues to meet the definition in §58.1-3660, and
only a single initial certification is required by the statute. Sales tax by its nature is a
very small percentage of project costs, and the timing of sales tax exemption under §58.1-
609.3.9 (i.e., post certification) does not change the actual cost of the project.
Page 52, Paragraph 1, Sentence 3: While “passes the cost on to society” may be correct
on a macro level, the impact of the lost revenue resulting from the exemptions is borne by
the people of the Commonwealth and its localities. The sales tax exemption in §58.1-
609.3.9 impacts the Virginia revenue directly once per project, and property tax
exemption takes money from the coffers of Virginia localities continually, every year the
property continues to meet the definition in §58.1-3660. The funds that are lost by state
and local government to the exemptions are either made up from other sources (i.e., other
taxpayers), or result in fewer government services being provided to the citizens of the
state and localities.
4
Ms. Ellen J. Miller
June 5, 2019
Page Five
Page 59, Paragraph 3 and Table 4-5: Because the pollution control equipment eligible for
the exemption are required by law and regulation, it seems counterintuitive to conclude
that all of the one-time sales taxes exempted, and continual property taxes exempted
generate positive economic activity. Based on total project costs for locating,
constructing, and operating a facility in Virginia, it seems unlikely the relatively small
amount of monies recouped by the exemptions contribute significantly to a decision to
locate and operate a facility in Virginia.
Page 61, Paragraph 1 and 2: DEQ maintains agency guidance for its staff in each media
that process exemption requests, and all DEQ guidance is required to be located on the
Virginia Regulatory Town Hall. However, availability of guidance is not the only
measure of transparency. DEQ staff work with requesters on a case-by-case basis to
address projects that range over a full spectrum of complexity. Off-the-shelf automatic
certification would be difficult to address in concise guidance or lists when
considerations must be made for each of the provisions within the §58.1-3660 definition
(See bulleted list in general comments). However, it is important to note (as JLARC does
on Page 49) that the statute itself does exactly that in several instances, providing
property tax exemption for the following…
Coal, oil and gas production and waste disposal
Forestry and other vegetative waste products
Landfill gas recovery [and related energy production]
Property used to generate solar and wind energy
…”whether or not such property has been certified to the Department of Taxation by a
state certifying authority”. (Note that in general these categories of equipment and
facilities arguably might not meet §58.1-3660’s primary purpose test for tax exemption.)
Page 62, Paragraph 1, Sentence 5: DEQ appreciates that JLARC recognizes the limits of
DEQ’s statutory authority for certification in its comment to the General Assembly that
certification prior to construction would require an amendment to the Code.
5
Ms. Ellen J. Miller
June 5, 2019
Page Six
Again, DEQ appreciates the opportunity to comment on this and all drafts prepared by JLARC
for reporting to the Governor and the General Assembly of Virginia. Please contact Michael
Dowd if you have any questions.
Sincerely,
David K. Paylor
cc: Hal Greer
Mike Dowd
Jeff Steers
6
June 10, 2019
Mr. Hal E. Greer, Director
Joint Legislative Audit & Review Commission
919 East Main Street, Suite 2101
Richmond, VA 23219
Re: VEDP response to sections 1 (data center incentives) and 2 (semiconductor manufacturing
incentives) of the draft JLARC report, Data Center and Manufacturing Incentives
Dear Mr. Greer:
Thank you for providing an opportunity for us to comment on sections 1 (data center incentives) and 2
(semiconductor manufacturing incentives) of the Joint Legislative Audit & Review Commission’s
(JLARC’s) draft report, Data Center and Manufacturing Incentives.
We agree with the draft report’s finding that the availability of data center sales and use tax exemptions is
among the top few site-selection considerations for data centers. Indeed, we believe that Virginia’s data
center sales and use tax exemption has been a major contributor to the Commonwealth’s success in
becoming one of the largest, most attractive markets in the world for data center investment. The General
Assembly’s relatively recent action to extend the sunset of the data center sales and use tax exemption
program to 2035 further solidified Virginia’s competitive position in the marketplace.
We further agree with the draft report’s finding that states have offered very large incentive packages to
attract semiconductor manufacturing facilities, as such projects are among the most coveted economic
development opportunities due to their large job counts, high wages, technology-intensive processes, and
large capital investment levels.
We generally are comfortable with all of the recommendations in the draft report that pertain to data
centers and the Major Employment and Investment Project Approval Commission (i.e., recommendation
nos. 1-5). With respect to recommendation no. 3, it would be important to design the disclosure
requirement to protect confidential location and investment information associated with individual data
centers (e.g., the General Assembly could require site-specific data submitted to VEDP and TAX to be
aggregated at the GO Virginia region level for any public reporting).
I want to thank JLARC for the work it has done to markedly increase the sophistication of its return-on-
investment (ROI) analyses (including in particular “but-for” estimates) for incentives to better reflect
varying levels of impact that incentives can have based on their size, design, and eligibility standards
(e.g., whether a competitive situation is required). Even so, I think it is important to acknowledge that such
ROI analyses with estimated “but-for” assumptions rarely can be performed with precision (by JLARC,
VEDP, or any third party) because the inputs into the full decision framework used by each firm (e.g., how
they weight various factors) typically are unknown. Accordingly, the actual ROI levels of the various
incentive programs could be materially higher (or lower) than what is depicted in the report.
901 E. Cary Street, Suite 900
Richmond, Virginia 23219
VEDP.org
Mr. Greer
Page 2 of 2
As usual, we appreciated the professionalism of JLARC staff during the project, as well as compliment
your team on its thoughtful work and recommendations. We also appreciated how well the draft report
captured important nuances about both of the referenced industry sectors.
Sincerely,
Stephen Moret
President & CEO
JLARC.VIRGINIA.GOV
919 East Main St. Suite 2101
Richmond, VA 23219
File and source
- File
- 2020-RD148.pdf
- Size
- 4,011,815 bytes
- SHA-256
- ef9b89583ee3ea40e0efafdb3e9c47b7ba632ff7386fd34f0b435e68324884d3
- Our copy
- 2020-RD148.pdf
- Original
- No public link identified.