Michael J. Boskin, Tully M. Friedman Professor of Economics, Stanford University — Testimony

Summary

Testimony of Michael J. Boskin, Tully M. Friedman Professor of Economics at Stanford University and Hoover Institution Senior Fellow, for the House Transportation Committee on July 29, 2021, titled Federal Infrastructure Policy in the Current Macroeconomic Context. Boskin states that the $5 trillion cumulative Covid response spending by the federal government thus far and related FED policies helped the economy recover more quickly and cushioned hardship. He says he supports policies to mitigate short-run economic pain from a crisis like the COVID-19 pandemic as long as the long-run cost is reasonable. He argues that the potential short-run macroeconomic benefits of additional spending are much lower now than in March and April 2020.

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

“Michael J. Boskin, Tully M. Friedman Professor of Economics, Stanford University — Testimony.” Original at docs.house.gov. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/hhrg-117-pw00-wstate-boskinm-20210729-c87c8161e5a6-4760c64c68/, SHA-256 c87c8161e5a6a023aa302e06212926dd5cad2a1a414c280d8745946d6b2aa536.

Full text

Michael J. Boskin
Tully M. Friedman Professor of Economics, Stanford University
Wohlford Family Hoover Institution Senior Fellow
former Chair, President’s Council of Advisers
July 29, 2021 11am (ET)
Assessing the Federal Government’s COVID-19 Relief and Response Efforts and Its Impact

1
Federal Infrastructure Policy in the Current Macroeconomic Context
Michael J. Boskin

Testimony for the House Transportation Committee
July 29, 2021

The “Covid-recession” was unprecedented in its cause and in its depth and speed as

dated by my colleagues on the NBER Business Cycle Dating Committee. The $5 trillion dollar

cumulative Covid response spending by the federal government thus far and related FED

policies certainly both helped the economy recovery more quickly and cushioned the hardship

experienced by many. Given the immense ex ante uncertainty that accompanied the rapid

lockdown of much of the economy and the early massive unemployment, it should be expected

that the policy response likewise was unprecedented.

I support policies to mitigate short-run economic pain caused by a crisis like the COVID-19

pandemic and help spur recovery, as long as the long-run cost is reasonable. As the economy has

recovered considerably since those horrible days of March and April 2020, the potential short-run

macroeconomic benefits of additional spending are much lower now than then and any additional

spending is better focused on long-run societal benefits with spending levels, allocations among

projects and financing methods designed to pass rigorous national cost-benefit tests.

It is early days in the detailed evaluations of the economic effects of the several

responses to the Covid crisis and recession, and their many components, by independent

scholars, I will address my comments to the desirability of additional spending, and its methods

of finance under consideration for traditional infrastructure. I present what I believe are the

2
best estimates based on leading academic research on the short-run impacts of government

spending, especially on infrastructure (as opposed to, say, transfer payments, for which

whatever the desirability may be on other grounds, the macroeconomic benefits are far less).

Much of the research I cite below is based on evaluations of the 2009 ARRA, but some focus as

well as other data and periods.

But first, America certainly has infrastructure needs. The American Society of Civil

Engineers, serious if somewhat self-interested, rates the nation’s infrastructure a C-. Some

claim there is a multi-trillion dollar “infrastructure deficit” and others have long blamed

inadequate public investment in infrastructure for holding back U.S. economic productivity

(e.g., Aschauer, 1991). Yet others argue that a closer analysis shows U.S. infrastructure in much

better shape, and advocate for improving the allocation of funding over massive new

expenditures (Duranton, Nagpal, and Turner, 2020). In a similar vein, the World Economic

Forum rates U.S. infrastructure 13th out of 141, behind top rated Singapore and Hong Kong, but

ahead of countries like Sweden and Denmark.

At even a fraction of some of the infrastructure spending being discussed today, there is

ample opportunity to do considerable productive long-run infrastructure investment (the

attached Table is my attempt to give a sense of the scale of the nation’s infrastructure). But

only some of that is appropriately a governmental, and only a part of that is appropriately a

federal, responsibility. And the long-run net economic effects of a new federal infrastructure

program, following the expiration of the 2015 FAST ACT and the impending exhaustion of the

Highway Trust Fund will depend not just on the level of spending, but also the ex -ante quality

of the projects funded, their ex post execution and their financing method(s).

3
Done well, the program can produce substantial societal benefits; but done to excess or

with poor design incentives, a plethora of poor return projects, even boondoggles, would likely

result. As a general guide, the larger the appropriated spending, the greater the likelihood of

the laws of diminishing returns and of unintended consequences creating a large set of sub-

standard projects. Ditto the further the financing deviates from use of already appropriated

funds and user fees or their gas and vehicle miles driven tax cousins. In short, the federal

infrastructure program should fund projects that pass rigorous national cost benefit tests and

the better aligned the incentives for state and local officials, federal authorities and private

citizens, the more likely that result becomes.

It should be noted that the economy is now back to its pre-pandemic level and is

growing solidly. While risks remain and we should keep a close eye on job growth to make sure

unemployment continues its downward movement to full employment, it does not appear

likely to need considerable additional short-run stimulus on top of that already provided and in

process. And some argue the risk of entrenching longer-term some of the considerable recent

inflationary pressures are a greater risk. Certainly the larger the program and the more front-

loaded, the bigger that risk.

Some suggest that huge additional infrastructure spending will dramatically spur growth

and employment. With government borrowing rates low, the argument goes, deficit finance

amounts to a cheap way to increase employment. In fact, existing research suggests that is a

misguided conclusion. First, while infrastructure spending may have made for good short-run

stimulus in the 1930s, that is not the case today (Glaeser 2016). The best evidence (Ramey

4
(2019, 2020) is that each dollar of infrastructure spending would increase GDP only 60 cents—

even in a soft economy, there are no magically large multipliers.

And, of course, when the dollar is financed with taxes (now, or in the event debt-finance

is used, later), the tax hikes exact a cost on the economy. Students learn in Economics 1 that

cost rises with the square of tax rates; doubling rates, quadruples the harm from the distortion

of private decisions to work, hire, save, invest, innovate, etc. Thus, the incremental cost rises

with the tax rate. This is not a doctrinal issue; it has to do with the area under supply and

demand curves. A rough estimate would be that each dollar of taxes (now or in discounted

present value terms later to pay interest on the debt) costs the economy about $1.25-$1.30. So

the projects chosen for funding really do need to be prospectively high societal return, yet the

CBO estimated the return on public infrastructure investment at 5%, just half of the return on

the private investment likely to be crowded out by taxes or debt. The best way to minimize

these distortionary costs is to finance the spending with user fees or, where applicable, their

tax cousins, the gas tax and vehicle miles driven tax, that tie the responsibility for payment

closely to the benefits received.

Debt finance of the Covid relief funding was certainly justifiable in a deep recession and

early in an uncertain recovery, but is unwise, even risky (Boskin, 2020) in normal times. Historically,

huge debt buildups have usually been followed by serious problems: sluggish growth, an uptick in

inflation, a financial crisis, or all of them. We cannot be certain which problems will occur or what

debt-to-GDP ratio will signal trouble for which countries. And the US does have the advantage of

issuing the world’s leading reserve currency, at least for time being. But inflation risks are rising – a

trend that more deficit-financed spending will only accelerate. To be sure, I support policies to

5
mitigate the short-run economic pain caused by a crisis like the COVID-19 pandemic and help spur

recovery, as long as the long-run cost is reasonable. But be careful not to run huge deficits that

persist long after the economy is back to full employment (the Administration projects the

unemployment rate to be down to 4.1% next year and a low of 3.8% thereafter).

It is also important to realize that only a very small fraction of those unemployed today

have the skills and experience for the kind of work required by today’s infrastructure

challenges; we cannot instantly train the unemployed to safely and effectively operate tower

cranes or giant excavators, for example. There are few public infrastructure projects that

require only a shovel. Additionally, planning and approval hurdles that were absent in the

1930s are omnipresent today, slowing the speed with which funds can be disbursed and

infrastructure built. As a result, research has found that large increases in infrastructure

spending within a short window of time may lead not to increases in employment, but to

backlogs that result in higher profits for a relatively small set of contractors (Balat, 2017).

Worse yet, when federal funding is too abundant and not closely tied to national, as

opposed to local, benefits, political incentives exacerbate the tendency to fund too many low

return projects. A prime example is the boondoggle of California’s High-Speed Rail project,

which originally used a grant from the 2009 ARRA to pay, six years later, for a tiny initial rail

line. The entire project is mired in a tripling of cost estimates, technical problems and epic

mismanagement for what is now prospectively blended speed rail amid widespread lack of

support and outright hostility in the Central Valley where billboards clamor to “Build Dams Not

Trains.”

6
Second, large public infrastructure projects—highways, dams, etc.—are designed to last

many decades, and eventually interest rates on government debt will rise and render rolling

over the larger debt much more expensive. CBO (2021) estimates net interest payments as a

share of (rising) GDP will increase sharply beginning in the middle of this decade and more than

triple in the following two decades, with interest costs exceeding even rapidly growing

spending on Social Security, and dwarfing all discretionary spending, including on defense.

Thus, the view that infrastructure spending in today’s low-interest rate environment is

essentially a free lunch is misguided.

While large changes in interest rates are unlikely in the near term, the is fact financial

markets and government and private forecasters have often failed to anticipate them – for

example, during the inflation of the 1970s and the disinflation of the early 1980s. After 2008, all

grossly underestimated how long the Fed would keep its target interest rate at zero.

Infrastructure spending seems to be one of a few areas of potential bipartisan

agreement. As mentioned above, there is ample room for a considerable well-crafted

infrastructure spending program that is economically beneficial. Some policymakers, interest

groups, and constituents still view infrastructure spending as shovel-ready work that is both

desperately needed and great at creating new jobs. As President Obama eventually stated

“there’s no such thing as shovel-ready projects.” The New Deal did not end the Great

Depression, nor did Japan’s massive ongoing infrastructure expenditures spare it from its “lost

decades.” To repeat, most of the unemployed do not have the skills or experience to operate

modern equipment such as giant excavators and tower cranes.

7
Recent academic evidence on the matter, however, does suggest that better allocation

of infrastructure spending is more important for long-run productivity than increased spending

(Duranton, Nagpal, and Turner, 2020), and casts doubt on whether a large allocation of federal

funds for infrastructure will work as an effective stimulus (e.g., Balat, 2017; Gallen and Winston,

2019; Ramey, 2020).

Garin (2019) studies how funding allocated by the federal government for road

construction projects through the 2009 American Recovery and Reinvestment Act (ARRA)

affected local employment. He finds that every dollar of ARRA spending increased local

construction payrolls by thirty cents, but had virtually no effect on employment. Balat (2017)

analyzes the effect of ARRA spending on highway-related procurement in California, finding

that the sudden infusion of cash into an industry that was already working near capacity did not

grow the number of construction firms or construction employment, but resulted in higher

procurement prices. This capacity constraint is directly at odds with a 1930s vision of what

infrastructure spending can accomplish. The highly specialized and technologically advanced

nature of the work now requires skills, experience, and certifications that make it difficult to

quickly expand the number of firms and workers. In California, Balat (2017) finds that the

government paid 6.2% more on stimulus projects and 4.8% more on other projects as a result

of ARRA, stimulating the economy by increasing construction company revenues, but forgoing

about $335 million that could have been spent on other road projects.

Additional work, such as Ramey (2020), demonstrates that infrastructure spending is

usually slow to move from appropriation to implementation to actual use, making even the

most productive and most shovel-ready projects poor candidates for short-run economic

8
stimulus. In fact, as Gallen and Winston (2019) argue, disruptions that come from a slew of

highway infrastructure projects can even result in negative short-run effects on total

employment. Studies of the ARRA also provide cautionary tales on the ability of infrastructure

spending to create jobs in the short-run and on the cost of doing so. Leduc and Wilson (2017),

for example, find a cost of $500,000 per job in 2010, considerably costlier than the roughly

$125,000-$200,000 per job that other papers have attributed to ARRA spending overall (Wilson,

2012; Conley and Dupor, 2013).

Long-run productivity is a different story, but the devil is in the details. As discussed

above, the research literature generally stresses that quality and rigor behind fund allocation

and incentive preserving funding mechanisms are key to large enough long-run returns to

justify the spending, much more so than the sheer volume of spending. In particular, repairs

and maintenance seem to have consistently higher returns than new construction.

In conclusion, the evaluation of government spending in response to previous

recessions suggests quite limited short-run macroeconomic benefits. The ex-post estimates of

spending “multipliers” for ARRA was one-third that of Administration economists is 2009.The

response to the unprecedented Covid pandemic and recession likely did somewhat better, and

the humanitarian case given the widespread suffering was compelling. But as the economy

approaches full employment, the case for additional spending as “stimulus” is far weaker and

tax and deficit financed spending likely will be quite costly in the longer-term. The most

compelling case for additional federal spending is simply to find good projects and methods of

financing that pass rigorous national (not local) cost-benefit tests based on sensible estimates

of the factors affecting benefits and costs. That should conceptually govern the size of the

9
spending. But given the budgetary process, the headline number will undoubtedly be

determined by other factors. But then the federal, state and local agencies involved need to

hew as closely as possible to the dictates of cost-benefit tests to maximize the opportunity for

reasonable returns on the (current and/or future) taxpayers’ investment.

References:

Aschauer, David A. 1991. “Infrastructure: America’s Third Deficit,” Challenge, 34(2), pp.39-45.

Balat, J., 2017. “Highway Procurement and the Stimulus Package: Identification and Estimation
of Dynamic Auctions with Unobserved Heterogeneity.” Working Paper.

Boskin, Michael J. 2020. “Are Large Deficits and Debt Dangerous?” NBER Working Paper No.
26727 (condensed version published in American Economic Association Papers and
Proceedings, Vol. 110, May, pp145-48).

Brooks, Leah and Liscow, Zachary. 2019. “Infrastructure Costs,” Hutchins Center Working Paper
No. 54, Brookings Institution.

Brooks, Leah and Liscow, Zachary. 2020. “Can America Reduce Highway Spending? Evidence
from the States” in Economic Analysis and Infrastructure Investment. University of Chicago
Press.

Congressional Budget Office, The 2021 Long-Term Budget Outlook, March.

Conley, Timothy G. and Dupor, Bill. 2013. “The American Recovery and Reinvestment Act: Solely
a Government Jobs Program?” Journal of Monetary Economics, Vol. 60, Iss. 5, pp. 535-549.

Duranton, Gilles, Nagpal, Geetika, and Turner, Matthew A. 2020. “Transportation Infrastructure
in the US,” in Economic Analysis and Infrastructure Investment. University of Chicago Press.

10
Gallen, Trevor and Clifford Winston, 2019. “Transportation and its Effects on the U.S. Economy:
A General Equilibrium Approach.” Technical report, Purdue University.

Garin, A. 2019. “Putting America to Work, Where? Evidence on the Effectiveness of
Infrastructure Construction as a Locally Targeted Employment Policy,” Journal of Urban
Economics, 111, 108-131.

Glaeser, Edward L. 2016. “If You Build It…Myths and Realities About America’s Infrastructure
Spending,” City Journal.

Leduc, Sylvain, and Daniel Wilson. 2017. "Are State Governments Roadblocks to Federal
Stimulus? Evidence on the Flypaper Effect of Highway Grants in the 2009 Recovery
Act." American Economic Journal: Economic Policy, 9 (2): 253-92.

Ramey, Valerie A.2019. “Ten Years After the Financial Crisis: what Have We Learned From the
Renaissance in Fiscal Research?” Journal of Economic Perspectives, Vol. 33, No. 2.

Ramey, Valerie A. 2020. “The Macroeconomic Consequences of Infrastructure Investment” in
Economic Analysis and Infrastructure Investment. University of Chicago Press.
http://www.nber.org/chapters/c14366

Wilson, Daniel J. 2012. “Fiscal Spending Jobs Multipliers: Evidence from the 2009 American
Recovery and Reinvestment Act,” American Economic Journal: Economic Policy, Vol. 4, Iss. 3, pp.
251-282.

11
America’s Vast and Varied Infrastructure Enterprise

Boskin and Bolotnyy Infrastructure Policy 3 / 20

File and source

File: hhrg-117-pw00-wstate-boskinm-20210729-c87c8161e5a6.pdf
Size: 376,734 bytes
SHA-256: c87c8161e5a6a023aa302e06212926dd5cad2a1a414c280d8745946d6b2aa536
Our copy: hhrg-117-pw00-wstate-boskinm-20210729-c87c8161e5a6.pdf
Original: docs.house.gov