Michael Schwan et al., Business as Usual? Conventional Politics, Critical Junctures, and Policy Feedback in the Paycheck Protection Program (2025)

Summary

Michael Schwan and co-authors, a 2025 paper titled Business as Usual, on conventional politics, critical junctures and policy feedback in the Paycheck Protection Program. It is a 3,005,107-byte PDF whose original is on www.cambridge.org.

Summary drafted by a model from the text of the original file linked on this page 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 Schwan et al., Business as Usual? Conventional Politics, Critical Junctures, and Policy Feedback in the Paycheck Protection Program (2025).” Original at www.cambridge.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/michael-schwan-et-al-business-as-usual-conventional-politics-critical-dd2969a22f/, SHA-256 374d1cdea775125e78ea360b439ee952cafa1f832f33076edd14fdbb60f83de6.

Full text

Article

Business as Usual? Conventional Politics,
Critical Junctures, and Policy Feedback in
the Paycheck Protection Program
Michael Schwan, Marc Schneiberg and Mark K. Cassell

Conventional-politics approaches, emphasizing party ideology, electoral dynamics, committee membership, campaign donations, and
industry clout, exercise a powerful hold over assessments of public policies and their distributional eﬀects. Emerging from pluralist and
business power perspectives, such accounts see “who gets what and why” as the result of how politics and power shape policies, their
implementation, and distributional outcomes. This pervades our understanding of the Paycheck Protection Program (PPP), the US
government’s eﬀort to avert mass unemployment during the COVID-19 pandemic by lending $786 billion to small businesses to keep
employees on the payroll. Yet contrary to prior studies of the PPP, we ﬁnd that conventional-politics factors were strikingly uncorrelated
with distributional outcomes, revealing limits to such approaches to this case. Instead, we ﬁnd that an institutional politics or politics-in-
time (IP-PIT) analysis better explains the program and its trajectories. IP-PIT revises the causal sequence by emphasizing how
institutions and policies generate politics, distributional outcomes, and feedback loops. We engage both approaches via a mixed-methods
analysis of the PPP and two new datasets. We conduct a qualitative process tracing of temporal variation in policy architectures, politics,
policy revisions, and access to loans across the program’s three periods, and present quantitative analyses of loan ﬂows across congressional
districts and periods using data on the entire corpus of PPP loans. In so doing, we advance research and debate over the PPP, the
dynamics and outcomes of US policy making during crises, and the American political economy in general. Ours is the ﬁrst study of the
PPP to conduct a mixed-methods analysis of loans across congressional districts or to use conventional and institutional approaches to
address its politics, policy, and outcomes. More broadly, we document varieties of critical junctures, contribute arguments about what
might shape policy or institutional innovation in those moments, and use the PPP to identify conditions under which systems are “their
own grave diggers,” fueling negative-transformative rather than positive-reproductive feedback.

The order of author names is random. All authors contributed equally to the manuscript.

Corresponding author: Michael Schwan (m.p.schwan@rug.nl, Netherlands) is Assistant Professor of Macro-Finance and
Sustainability at the University of Groningen. He is a political economist who seeks to understand the institutional foundations
and power relations of market economies. His research focuses on ﬁnance and alternatives within the banking system, corporate
governance, the digitalization of nonﬁnancial ﬁrms and, most recently, the provision of transnational infrastructures. His work
appears in New Political Economy, Comparative European Politics, and Review of International Political Economy.

Marc Schneiberg (schneibm@reed.edu, United States) is John C. Pock Professor of Sociology at Reed College. He is an
organizational and economic sociologist whose work focuses on the rise, contemporary fates, and economic and ethnoracial
impacts of organizational diversity and alternatives to giant shareholder corporations in American capitalism. He has been elected
president of the Society for the Advancement of Socio-Economics, has served on the executive councils of that society and the
American Sociological Association and as editor of Socio-Economic Review, and has received National Science Foundation
support for his scholarship, which appears in journals like Politics and Society, American Sociological Review, American
Journal of Sociology, Organizational Studies, Economic Development Quarterly, and Socio-Economic Review.

Mark K. Cassell (mcassell@kent.edu, United States) is Professor of Political Science at Kent State University. He is a public policy
and public administration scholar focusing on the use of public-private hybrid entities to pursue public goals and on alternative
banking structures in the US and Germany. His work appears in Governance, Public Administration Review, Economic
Development Quarterly, and Social Science Quarterly. He has authored Banking on the State: The Political Economy of
Public Savings Banks (2020) and How Governments Privatize: The Politics of Divestment in the United States and
Germany (2002), among other books.

doi:10.1017/S1537592725102272 September 2026 | Vol. 24/No. 3 1271
© The Author(s), 2025. Published by Cambridge University Press on behalf of American Political Science Association. This is an Open
Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which
permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.

C
onventional-politics approaches, emphasizing party revisions, and shifting access to loans within and across
ideology, electoral dynamics, committee member- the program’s three periods. This analysis reveals a sus-
ship, campaign donations, and industry clout, exer- pension of business-as-usual political routines at the begin-
cise a powerful hold over assessments of public policies and ning of the program—a critical juncture that explains the
their distributional outcomes. Emerging from pluralist and PPP’s scale, quick passage, and initial core architectures. It
business power perspectives, such accounts see “who gets reveals further that the PPP evolved dynamically over its
what and why” as the result of how politics and power shape three lending periods, as policy feedbacks altered policies
policies, their implementation, and distributional outcomes. and outcomes. Moreover, rather than trigger positive
Indeed, this politics ! policy ! outcome framework pervades feedback in which programs foster their own supporting
our understanding of the Paycheck Protection Program coalitions, the PPP sparked negative feedback, political
(PPP), the United States government’s eﬀort to avert mass backlash, and policy revisions that transformed the pro-
unemployment during the COVID-19 pandemic by gram from one that overwhelmingly favored large busi-
channeling $786 billion over three lending periods to small nesses, aﬄuent areas, and white communities into one that
businesses to keep employees on the payroll. According to delivered millions of loans to smaller businesses, especially
Forbes, 71 public corporations received $300 million in in poorer and hard-hit Black and brown communities.
loans in the PPP’s ﬁrst week (Vardi 2020). The New York We then engage each approach further via quantitative
Times reported that the PPP “allowed big chains like Shake analyses of program lending across congressional districts
Shack, Potbelly and Ruth’s Chris Steak House to get tens of and periods using data on the entire corpus of PPP loans.
millions of dollars while many smaller restaurants walked We review and replicate prior quantitative analyses of PPP
away with nothing” (Yaﬀe-Bellany 2020). Headlines like lending dominated by conventional-politics approaches
“Banks Gave Richest Clients ‘Concierge Treatment’ for and ﬁnd that such models poorly explain loan distribu-
Pandemic Aid” (Flitter and Cowley 2020) ampliﬁed the tions. Instead, we consider and conﬁrm IP-PIT claims that
point, as did scholarly analyses (Chernenko and Scharfstein what impacts outcomes depends on timing in a develop-
2022; Fairlie 2020; Grotto, Mider, and Sam 2020; Howell mental sequence. The eﬀects on loans of business size and
et al. 2021; Sanchez-Moyano 2021; Wong, Ong, and socioeconomic factors change over the PPP’s three periods
González 2020). PPP loans ﬂowed to those with inﬂuence as its politics and policies evolve.
and access—banks’ favored clients, large and often still Through these analyses, we use one of the US govern-
proﬁtable employers, aﬄuent and white areas, districts with ment’s largest economic interventions since the New
inﬂuential representatives—rather than small businesses in Deal as a deviant case (Seawright and Gerring 2008) to
the hardest-hit Black and brown communities. advance research and debate over the dynamics and out-
Yet contrary to prior studies of the PPP, we ﬁnd that comes of US policy making during crises. This links to
conventional-politics factors were strikingly uncorrelated general debates about the American political economy,
with these outcomes, revealing limits to such approaches including the embeddedness of race and organized inter-
to this case. Instead, we ﬁnd that an institutional politics or ests, spatial variation in distributional outcomes, and
politics-in-time (IP-PIT) approach better explains the PPP the time-dependent dynamics of coalition formation
and its trajectories. IP-PIT revises the causal sequence by (Hacker et al. 2022). Ours is the ﬁrst study of the PPP
emphasizing how institutions and policies generate poli- to conduct a mixed-methods analysis of loans across
tics, distributional outcomes, and feedback loops, suggest- congressional districts or to use conventional and insti-
ing a policies ! politics/outcomes ! policies sequence tutional approaches to address its politics, policy, and
(Moynihan and Soss 2014; Pierson 2000; 2004; Sorelle outcomes. This highlights important and likely general
2023) with distinct periods. Crises, like a pandemic, spark limits on the capacities of conventional politics to explain
critical junctures and suspend conventional politics, cre- policy making during crises. It also lets us show that
ating opportunities for policy innovations. Once in place, IP-PIT arguments about critical junctures, punctuation,
“policies create new politics” (Campbell 2012; Hacker and and pathways apply not just to epochal institutional
Pierson 2019; Jacobs and Weaver 2015; Schattschneider change like the welfare state or private insurer-based
1935). They endogenously produce processes that set healthcare, but also to temporary crisis management
systems on a path, whether by generating new political episodes, which may or may not reset the system beyond
interests (Pierson 2000; Skocpol 1996), (de)mobilizing the duration of the intervention. Furthermore, our study
constituencies (Moynihan and Soss 2014; Sorelle 2023; advances speciﬁc arguments within IP-PIT approaches.
Thurston 2015), inducing new organizations (North We document varieties of critical junctures, contribute
1990), or framing calculation and debate (Dobbin 1994). arguments about what might shape policy or institutional
We engage these approaches via a mixed-methods innovation in those moments, and use the PPP to
analysis of the PPP and two new datasets. We anchor identify conditions under which systems are “their own
our study in a qualitative process-tracing analysis of tem- grave diggers,” fueling negative-transformative rather
poral variation in policy architectures, politics, policy than positive-reproductive feedback.

1272 Perspectives on Politics

Downloaded from https://www.cambridge.org/core. IP address: 73.252.170.234, on 04 Oct 2026 at 17:51:42, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S1537592725102272
The PPP and Loan Rates across rates increasing dramatically in some of the hardest-hit
Congressional Districts districts previously left out. Cumulative rates rose to 0.30
The PPP was established through the Coronavirus Aid, or more in period 2 in Atlanta, Chicago, Cleveland, Los
Relief, and Economic Security (CARES) Act on March Angeles, New York City, and southern Florida. By the end
27, 2020. It allocated $349 billion for guaranteed and of period 3 many of these places recorded loan rates up to
ultimately forgivable loans to small businesses to keep their 0.70 or more. However, striking diﬀerences remained,
employees on the payroll. Lawmakers intended the PPP as with rates in the most subscribed districts in Florida,
a one-oﬀ eﬀort to “ﬂatten the curve.” Yet they refunded Georgia, and Illinois up to three times higher than rates
the program on April 24 via the Paycheck Protection in some California, North Carolina, Pennsylvania, and
Program and Health Care Enhancement Act, allocating Texas districts, which ranged from 0.32 to 0.42. Our
an additional $320 billion, and again on December 27, via analyses explain the processes driving the evolution of
the Consolidated Appropriations Act, adding another these loan allocations across time and place.
$284 billion.1 These acts produced a program marked
by three periods of lending from April 2020 through May From Conventional to Institutional
2021, during which over 11.3 million loans totaling $786 Politics and Politics-in-Time
billion were issued.2 Academic research on the PPP builds on strong empirical
As is relatively common in US public policy, the PPP foundations when it takes understanding the program’s
was structured as a public–private collaboration (Bartik unequal and shifting distributions and failures to target the
et al. 2020) that worked through private lenders to achieve hardest-hit areas as its core analytical agenda. Moreover, in
policy goals (Prasad 2012; Quinn 2019). The CARES Act pursuing that agenda, researchers lean on conventional or
crafted the PPP around the 7(a) loan guarantee program business-as-usual politics arguments about party aﬃlia-
oﬀered by the Small Business Administration (SBA). It tions, electoral calculation, and the like to explain PPP
assigned oversight and rulemaking to the SBA, supported distributions and departures from policy goals, drawing
by the US Treasury, delegated the underwriting and selectively on a larger and well-developed scholarly tradi-
delivering of funds to the SBA’s network of lenders, and tion in policy studies. We leave aside the wider literature
relied on 7(a)’s bureaucratic machinery and electronic on conventional politics as beyond our scope, and focus
transaction (ETran) portal to authorize lenders and pro- instead on scholarly work on the PPP, which nicely covers
cess loans. Businesses needing support had to ﬁnd and key conventional-politics dynamics.
work through an authorized lender to access funds. Work on the PPP credits unequal loan allocations and
The distribution of PPP loans varied substantially over program failures to three sets of factors: (1) electoral
time and across congressional districts. Lenders issued dynamics, including party politics, party ideology, and
1.61 million loans in period 1 (April 3–16, 2020); over politicians directing funds to battleground areas or reward-
double that in period 2 (April 27–August 8, 2020; 3.48 ing their base (Birdthistle and Silver 2021; Duchin and
million); and nearly double that again in period 3 (January Hackney 2021; Ha 2023); (2) congressional committee
11–May 31, 2021; 6.28 million). The pattern drove membership or chairmanship and the inﬂuence and access
cumulative loan rates (the average number of loans divided it aﬀords for selected constituents or districts (Berger,
by total establishments and self-employed) from 0.093 in Karakaplan, and Roman 2023; Ha 2023); and (3) the
period 1 to 0.283 and 0.611 in periods 2 and 3. By the end power and lobbying of wealthy constituents, aﬄuent areas,
of the program, 45.4% of businesses had received a loan or larger employers’ banks (Berger, Karakaplan, and
(see online appendix A). Roman 2023; Duchin and Hackney 2021; Igan et al.
Moreover, congressional districts shared unequally in 2021). Analyses in this body of work span geographies and
this largesse. PPP loans ﬁrst ﬂowed to larger, often prof- generate sometimes divergent results. However, they share
itable recipients in more aﬄuent and white areas less hit by the view that these conventional politics drove the distri-
the pandemic rather than small businesses in the hardest- bution of loans in the PPP.
hit and often poorer Black and brown communities Highlighting political ideology, Berger, Karakaplan,
(Birdthistle and Silver 2021; Borawski and Schweitzer and Roman (2023) ﬁnd that greater proportions of busi-
2021; Chernenko and Scharfstein 2022; Granja et al. nesses received loans in conservative and centrist counties
2021; Howell et al. 2021; Liu and Parilla 2020; Sanchez- with high concentrations of lenders and where greater
Moyano 2021). In period 1, districts in Iowa, Kansas, percentages of businesses lobby the federal government.
Maine, Nebraska, Vermont, and Wyoming topped the list Other scholars directly target party and competitive
with loan rates between 0.14 and 0.20, exceeding by ﬁve- dynamics, and link distributions in the ﬁrst two periods
to almost tenfold loan rates in the most devastated poorer to states that voted for Donald Trump in the 2016 US
and minority districts in Arizona, California, Florida, or presidential election (Birdthistle and Silver 2021); to
New York. Distributions shifted in periods 2 and 3, with battleground states, or Republican-leaning states, districts,

September 2026 | Vol. 24/No. 3 1273

or sectors (Duchin and Hackney 2021); and to eﬀorts by 2014; Pierson 1993; 2004; Sorensen 2023). Punctua-
members of Congress or the president to enhance their tions, junctures, and innovations can ﬂow from techno-
electoral position. Sabasteanski, Brooks, and Chandler’s logical change (Tushman and Anderson 1986), wars
(2021) analysis points in the opposite direction, ﬁnding (Ikenberry 2019), depressions (Campbell and Lindberg
that Republican districts, especially those with more Afri- 1990; Fligstein 1990; Skowronek 1982), environmental
can American constituents, received less funding than disasters (Hoﬀman 1999), or pandemics (Rao and Greve
Democratic ones, which calls into question links between 2018). Once in place, new institutions become causal
party and loans. Yet Ha’s (2023) analysis of the third forces in their own right. They induce feedback—politics,
period ﬁnds that Democratic districts and pro-Biden and investments, and adaptations—that ﬁxes the system and
Democratic counties, but not swing counties, received its participants on a new trajectory leading to entrench-
more PPP funding under the Biden presidency, conﬁrm- ment and persistence (Arthur 1989; Sorensen 2023;
ing party eﬀects rooted in presidential particularism and Thurston 2015). Indeed, feedback and increasing returns
the party in power. can shift what factors shape outcomes over time by
PPP research also emphasizes how membership of rendering them increasingly dependent on the structure
powerful congressional committees lets representatives of the evolving path and increasingly impervious to func-
direct beneﬁts to their constituencies, although evidence tional economic requirements, conventional political
here is mixed. Businesses in districts with members on the inﬂuence, or interest-group demands (Pierson 2000).
House Committee on Small Business received more loans While crafted to explain epochal and long-term trans-
than otherwise in one study (Berger, Karakaplan, and formations in institutions or enduring shifts in macropo-
Roman 2023), but committee chairmanship had no sta- licy regimes like the welfare state or private insurance-
tistically signiﬁcant impact on allocations in another based healthcare systems, IP-PIT and its core elements can
(Ha 2023). In contrast, business-lobbying impacts on also apply more speciﬁcally to policy making in crises, like
PPP loans seem clearer. Businesses (and banks) that the PPP. These cases involve immediate and temporary
lobbied received (or issued) more loans than those that solutions that often do not yield long-term resets of
did not, with greater eﬀect in industries that accounted for economic, political, or policy systems.
greater shares of campaign dollars (Barrick, Olson, and Consider ﬁrst critical junctures: IP-PIT accounts view
Rajgopal 2021; Berger, Karakaplan, and Roman 2023; these as moments of profound disruption. They upend
Birdthistle and Silver 2021; Igan et. al 2021; Sabasteanski, established political and economic processes and settle-
Brooks, and Chandler 2021). ments, call into question the legitimacy of existing arrange-
These studies develop our conceptual repertoire for ments, provide leverage for challengers to advance
understanding PPP allocations, yet assume that business- alternative schemes, and induce systemic struggles over
as-usual politics carried over to a policy that emerged in a the basic rules of the game (Capoccia 2015; Capoccia and
crisis that profoundly disrupted normal political routines Kelemen 2007, 343; Fligstein and McAdam 2012;
and calculations. They also pay little attention to how Schneiberg 2005; Sine and David 2003; Sorensen 2023;
policy architectures generate their own trajectories and Thurston 2015). However, crises and wholesale disrup-
politics, shifting political calculations, distributional out- tions, like the COVID-19 pandemic or the 2008 ﬁnancial
comes, and even what shapes those outcomes over time as crisis, can suspend conventional political dynamics and
contexts evolve. This would require in-depth qualitative create windows for policy innovation without wholesale
analyses of the PPP’s policy making and implementation, struggles over the system. As Capoccia and Kelemen
and additional theoretical guidance, for which we turn to (2007) point out, short time spans in critical junctures
an alternative IP-PIT approach. come with loosened structural constraints, which may be
Emerging from historical institutionalism in compara- the case for two reasons. First, crises may involve “Lehman
tive and American politics (Brady 1991; Orren and Skow- moments,” like the 2008 collapse of the investment-
ronek 1996; Thelen and Steinmo 1992), IP-PIT revises banking giant Lehman Brothers that triggered the global
conventional analyses of politics ! policy ! outcome by ﬁnancial crisis. Such moments pose systemic threats.
theorizing policies and outcomes as developmental Under profound uncertainties, they require immediate
sequences marked by distinct periods and combinations interventions to prevent large-scale, cascading losses, and
of exogenous and endogenous forces. Sequences typically compel policy makers to suspend or bypass standard
begin with a critical juncture in which an external shock or routines for emergency processes. In addition, crises might
crisis “punctuates” institutional equilibria. This sparks also be limited in duration, or framed as such by policy
conﬂict and innovations that point the system in new makers, supporting policies advanced as short-term and
directions, followed by a reset in which new policies or temporary measures to be abandoned or revisited. In these
institutions induce new politics and other feedback pro- scenarios, uncertainties and the costs of insuﬃcient action
cesses that set the system on a new trajectory (Campbell are understood to be so high that they trump all other
2012; Hacker and Pierson 2019; Moynihan and Soss political preferences, reducing policy makers’ incentives

1274 Perspectives on Politics

Downloaded from https://www.cambridge.org/core. IP address: 73.252.170.234, on 04 Oct 2026 at 17:51:42, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S1537592725102272
and safeguards to resist emergency policy making. More- relative to alternatives. This, in turn, may induce defec-
over, with an understanding that policy architectures are tions from existing coalitions and create opportunities for
temporary solutions, policy makers may accept measures, challengers to advance new solutions (Leblebici et al.
including even quite substantial interventions, that they 1991; Schneiberg 2005). Dysfunctionalities can trigger
would not agree to under normal circumstances, and they public outcry, allowing defectors to leverage saliency.
may have little incentive to pursue broader, more system- Ultimately, this might lead to legitimacy crises. Such
atic change. To the contrary, as behavioral scholars sug- potent solvents for “iron triangles” and entrenched inter-
gest, decision makers, facing high-stakes conditions under ests (Bovens and Hart 1996; Wallner 2008) let excluded
great uncertainty, are far more likely to focus ﬁrst on what actors and entrepreneurs build public support for reform.
is already “at hand,” gravitating toward “oﬀ-the-shelf” In short, resets can trigger political backlash and coalitions
solutions with at least partly known contours and poten- suﬃciently powerful to block a policy’s operations, roll
tially predictable impacts (Cohen, March, and Olsen back implementation, and push for change (Patashnik and
1972; Kingdon 2011; Stinchcombe 1990). Zelizer 2013; Weaver 2010). Additionally, preexisting
Second, IP-PIT feedback mechanisms also apply to pol- heterogeneity in “what’s on the path” in terms of institu-
icy making during crises. Once in place, new policies or tional logics, alternative ideas, or organizational systems
institutions like the PPP, even when crafted as temporary, can amplify possibilities for negative feedback by provid-
carry independent causal weight. Their design imposes ing actors with broader perspectives on possibilities and
constraints and incentives for action, activating political organizational vehicles to redirect resources (Daugbjerg
responses and other processes that ﬁx the system on a path and Kay 2020; Schneiberg 2007; Weaver 2010; see also
or trajectory (Arthur 1989; Edelman, Uggen, and Erlanger Stryker 2000; Thornton, Occasio, and Lounsbury 2012).
1999; Pierson 2004). Path-dependence scholars initially In sum, the preceding suggests several expectations for
focused on endogenous positive feedback processes in which the PPP. First, its lending across congressional districts will
policies or institutions bias subsequent development in have been less a product of conventional politics than
reproductive directions. New arrangements entail setup outcomes of a critical juncture and responses by policy
costs, investments, and adaptation that make switching makers to the uncertainties and pressing demands associ-
paths increasingly costly, especially once actors have ated with a combined economic and public health crisis.
learned to navigate the new institutional landscape Second, the program’s distributional outcomes and evolu-
(Campbell 2012; North 1990; Pierson 2004). Institutions tion will have emanated from its architectures—from the
can also induce their own politico-cultural dynamics and feedback processes that locked-in a trajectory of policies and
support, whether by (de)mobilizing constituencies reforms that sustained and induced new organizational
(Moynihan and Soss 2014; Sorelle 2023; Thurston compositions and resource ﬂows. And ﬁnally, the eﬀects
2015), creating new vested interests (Hacker and Pierson of politics and other characteristics on loan ﬂows will have
2019; Skocpol 1996), or prompting investments in new depended on timing in the PPP’s developmental sequence.
mental maps or conceptions of governance that frame They will have shifted as the program evolved away from its
problem solving and debate (Dobbin 1994; Edelman, starting point, became a causal factor in its own right, and
Uggen, and Erlanger 1999; Fligstein 1990; Fligstein and fueled its own politics and path dependencies.
McAdam 2012). These adaptive dynamics can lock-in
new policies or institutions, rendering them and their Methods and Data
distributional outcomes less dependent on starting condi- We engage these approaches via a mixed-methods design,
tions and progressively more dependent on policy archi- combining process tracing with statistical modeling. At
tectures and the calculations, politics, and organizations root, our study analyzes temporal variations and linkages
they induce. between politics and outcomes within and across the PPP’s
Yet they can also activate negative feedback processes that three periods. We also supplement this with analyses of
undermine their basic operation and foster trajectories of PPP loan ﬂows across congressional districts. We mirror
reform and change (Daugbjerg and Kay 2020; Jacobs and our process-analytic approach of tracking politics, policy
Weaver 2015; Patashnik and Zelizer 2013; Schneiberg revisions, and shifts in loan ﬂows across periods as much as
2005; Thurston 2015; Weaver 2010). Institutional resets possible by modeling time-varying eﬀects of key factors on
produce winners and losers, including new actors with new loan distributions across program periods. In both strate-
logics and ideas and an interest in challenging the status gies, we aim not to assess conventional-politics or IP-PIT
quo (Clemens 1997; Leblebici et al. 1991; Stryker 2000; approaches generally, but rather to use the PPP as a deviant
Weaver 2010). Resets to address one set of problems can case (Seawright and Gerring 2008) to consider their
create new ones (Hollingsworth and Streeck 1994). They relevance to policy making during crises and suggest
induce patterns of competition and strategic behavior that avenues for future research.
disconnect policies and stated outcomes, generate accu- Our qualitative analysis focuses on the causal mechanism
mulating costs, and alter the payoﬀs of existing policies (Glennan 1996) linking the PPP’s policy, politics, and

September 2026 | Vol. 24/No. 3 1275

distributional outcomes. Following “theory-building pro- made from April 3, 2020, to May 31, 2021. We exclude
cess tracing” (Beach and Pedersen 2013), we do not seek to loans ﬂagged as likely fraudulent by Griﬃn, Kruger, and
test a given boilerplate mechanism but instead strive to Mahajan (2023), leaving 10.4 million. We aggregate loan
mobilize IP-PIT arguments for opening the PPP’s black numbers and value to congressional districts for each
box of policy revisions and shifting loan ﬂows. We engage lending period, creating a panel dataset and our main
counterarguments, including those grounded in conven- dependent variable: district loan counts in each period
tional politics, natural progression, or protest dynamics, to normalized by the sum of establishments and unincorpo-
probe the validity of our mechanism and its theoretical and rated, full-time self-employed individuals. We supplement
empirical “uniqueness” (Beach 2016). Finally, we reach this with (1) descriptive analyses using daily counts of
beyond the conﬁnes of within-case inference in the PPP and loans to match policy reform clusters to loan ﬂows, and
reﬂect on limitations and generalizability of IP-PIT-inspired (2) models that use weekly district loan counts and
applications to policy making during crises as components reparameterize time by compartmentalizing by clusters
of theory building. We begin with the program’s initial of reforms rather than lending rounds.
design and distribution, and then follow program changes We model conventional politics mainly using the party
and their impacts over the three program periods. At each aﬃliation of representatives in the 116th and 117th
step, we work from policy design to politics and outcomes Congresses; electoral competitiveness of House races using
and back, carefully reconstructing the timeline that links the absolute value of the Cook Political Report’s Partisan
salience, mobilization, revisions, and new loan ﬂows. This Voting Index (Cook Political Report n.d.); and member-
work detects causal process observations as elements or ship on committees involved in PPP legislation (House
building blocks of mechanisms and context (Collier, Brady, Appropriations, House Small Business, House Ways and
and Seawright 2010), letting us assess both conventional Means, and Senate Small Business). For business power,
political processes and IP-PIT arguments about path depen- we include total inside-the-district campaign donations to
dence, endogenous dynamics, and feedback. winning House candidates in the two election cycles
Qualitative data come from 82 semistructured interviews spanning the PPP’s arc using data from OpenSecrets.org
conducted from September 2022 through December 2023. (Center for Responsive Politics n.d.) and the share of large
Respondents include borrowers and lenders (41); national businesses with more than ﬁve hundred employees calcu-
and local advocacy groups (18); lender associations (eight); lated from County Business Patterns data (US Census
congressional staﬀ (seven); and local political leaders, PPP Bureau 2024). For IP-PIT arguments, we include variables
experts, and SBA oﬃcials (eight).3 We also analyze the for poverty rates, ethnoracial compositions, and period
entire body of 224 congressional hearings on the PPP interaction terms to capture time-varying eﬀects of these
between April 23, 2020, and June 30, 2021. We searched key covariates suggested by our qualitative ﬁndings.
the universe of news publications in Factiva’s database of In all models we control for district population, for
91 top news and business sources in the US and coded COVID-19 severity using Harvard Center for Population
and counted articles, using “paycheck protection” and a and Development Studies data on the number of positive
range of search terms for inclusivity and bias to measure cases per thousand residents as of July 13, 2020 (Harvard
the media’s evolving attention. Finally, we closely read Center for Population and Development Studies 2021),
key documents, including (1) reports by the Government and for economic vulnerability to the pandemic using the
Accountability Oﬃce, Congressional Research Service, share of “hardest-hit businesses” from “high contact”
and the Oﬃce of Inspector General; (2) survey reports of industries (accommodation and food, arts and entertain-
lender organizations; (3) advocacy brieﬁngs and letters; ment, wholesale and retail) calculated from County Busi-
(4) academic scholarship; and (5) periodical accounts, ness Patterns data (US Census Bureau 2024). We include
particularly from the New York Times and Washington state ﬁxed eﬀects, density ﬁxed eﬀects from CityLab’s
Post. District Density Index (Montgomery 2018), a six-step
We supplement our qualitative analysis of within-unit scale from “purely rural” to “purely urban,” and conduct
variation over time (Gerring 2004) with quantitative extensive robustness checks (online appendices B and
analyses of causal eﬀects on the distribution of PPP loans. E–F). Quantitative data used in the analysis is available
We ﬁt cluster-robust random-eﬀects models on the impact through Harvard’s Dataverse (Schwan, Schneiberg, and
of standard political factors on PPP loans, incorporating Cassell 2025).
the corpus of prior quantitative work on the PPP. We
extend this strategy to address our IP-PIT-based qualita- Qualitative Results
tive ﬁndings via models that let key eﬀects on loan ﬂows Process tracing identiﬁed a three-step causal mechanism
vary across periods as the program’s politics and policy linked to a critical juncture. We traced the PPP’s initial
architectures evolve. design backward—to a critical juncture at which policy
For these analyses, we built a dataset of PPP loans using makers crafted an “oﬀ-the-shelf” solution using the banking
SBA loan-level data for the nearly 11.4 million PPP loans system—and forward—to how that policy architecture

1276 Perspectives on Politics

produced period 1’s lopsided loan ﬂows. Working forward projections of going up to 32%. Really smart people were super
over periods 2 and 3, we traced a sequence of events from concerned about a Great Depression. If we didn’t do something,
the global economy would fall oﬀ the cliﬀ.5
the PPP’s initial architecture to its evolution and shifting
distributions, as unequal loan ﬂows generated negative Oﬃcials understood they had arrived at a critical juncture.
feedback loops and a stream of policy revisions reconﬁguring The “cash ﬂow and business reality for small businesses,
the lender system and loan ﬂows. We ﬁnd endogenous yet which comprise two-thirds of the private sector employ-
transformative dynamics in which policy generated politics ment in the US, was really scary. We knew at the time that
and reform. Figure 1 summarizes our ﬁndings. The process mass layoﬀs would be detrimental to our society,” one
began with the pandemic and PPP rollout and design explained, risking “a total breakdown of society when
(critical juncture). That design induced unequal loan distri- people are not connected to employment.”6 From this
butions (onset), resulting in criticism by media outlets and emerged a frame in which oﬃcials cast the problem as
advocacy groups (outcry) and policy revisions that induced needing to act fast to “ﬂatten the curve” of new infections.
more inclusive lending in periods 2 and 3 (outcome). They conceived the PPP as a one-shot intervention until
businesses reopened and people returned to work.7
Second, the pandemic and lockdowns profoundly dis-
Critical Juncture and Genesis
rupted the policy-making process. Oﬃcials crafted policy
Interviewees highlight the unusual circumstances leading to while socially and politically isolated, and sometimes fell
the PPP and the central role played by the Senate at the ill. No hearings, public forums, or press conferences were
outset, in particular by Senators Marco Rubio (R-FL), Susan held at this juncture. “It was chaotic,” a Republican staﬀer
Collins (R-ME), and Ben Cardin (D-MD). First, policy noted.8 With the Capitol on lockdown, virtually no one
makers in March 2020 faced unprecedented uncertainty was there except for legislators and their staﬀ, huddled
about COVID-19 and its course, profound anxiety about together in rooms, another staﬀer elaborated. This was
the economic impacts of the pandemic and closures, and an “not a normal legislative process where you have a com-
urgent need to act immediately. As a Democratic staﬀer mittee hearing, and people speak, and you see who is
explained, Congress was “trying to legislate in one of the opposed to it and who’s for it, and then it goes through.”9
most uncertain environments we’ve experienced as a coun- “We were trying to social distance doing this,” explained
try, at least in a generation.”4 Fears of massive unemploy- an SBA oﬃcial. “[I]t was like trying to run a war room, and
ment and upheaval were at the top of policy makers’ minds, you’re in separate rooms.”10
as an SBA oﬃcial related: Under intense time pressure, oﬃcials reorganized for a
Data in early March came out. We went from 4.1 or 3.9% speedy result. Meetings for the PPP were conﬁned to a
unemployment to 14% by the end of March, with credible select group of mostly Republican lawmakers; members

September 2026 | Vol. 24/No. 3 1277

and staﬀers worked virtually or in small groups; committee changing uses. … You had this existing base of about eight
chairs met weekly with then Senate majority leader Mitch hundred SBA lenders who, on day one, the thought was you
could just change via keystrokes … what they were doing with
McConnell (R-KY) to coordinate; roles were blurred. As a their existing client bases.18
staﬀer put it, “[y]ou had senators playing what in ordinary
times would be staﬀ function … because we didn’t have Senator Rubio and others recognized that retroﬁtting a
the time.”11 Further, policy making in isolation upended long-standing program addressed pressing concerns of
lobbying: urgency and scale. SBA and 7(a) were known and trusted,
mitigating uncertainty. The program also came with a
With schools closing, people shutting down … I was getting
more texts from people with questions than an organized eﬀort
network of existing, vetted, and connected lenders
from a trade association or an industry group or lobbyists or through which funds could be immediately disbursed, “a
whatever. They just didn’t move that quickly.12 coast-to-coast delivery system.”19 A key insider said this
was a decisive advantage for senators and the administra-
“We talked with the lenders after we came up with the tion:
idea,” another staﬀer related. “We knew what we wanted
to do, and the question [to banks] was, Can you handle There are banks in every city in America, just about every small
it?”13 Bankers experienced “a chaotic process,”14 but town in America. It was the best delivery mechanism that we had
at the time to get money into the hands of businesses quickly so
understood, as another staﬀer explained, that that they would hold on to their employees so that the govern-
[i]f you’re going to design something quick, quick, quick, you’re ment could pay these employees during this period of closure.20
going to keep the room pretty small. I get that speed. … [T]he
administration worked quickly and quietly, but the stakeholders
Using banks with their established assessment and distri-
were not involved.15 bution network, this public–private collaboration bypassed
practical concerns about limited administrative capacities to
Through this process, the Senate assembled a program that vet millions of transactions, while assuaging Republican
went from idea to signed legislation in under a month, an opposition to “free money” and expanding federal powers.
extraordinary achievement impossible via conventional “Treasury trusts banks,” the head of a lender association
political processes. It was “fast and furious,” a small- noted, “not this boogeyman that is the IRS.”21 Rubio’s
business association representative quipped.16 “[I]n late proposal thus triumphed quickly over proposals to provide
February,” a policy advocate related in reﬂecting on the grants to small businesses through the IRS or expand the
entire CARES package, SBA’s Economic Injury Disaster Loan relief program of
we were … talking about … hundreds of millions of dollars in direct lending and grants.
relief, hundreds of millions. Then a week later, it was half a Democrats worried that retroﬁtting SBA 7(a) and build-
trillion. And that was considered insane. … Could we even ing banks into the PPP’s core would reproduce preexisting
fathom this? And then two weeks later, we were not blinking biases and discriminatory practices. Led by Senator Ben
an eye and suggesting three trillion dollars go out the door. It was
very fast, very, very fast.17
Cardin, Democrats pushed to include a “sense of the
Senate” resolution on page 30 of the CARES Act calling
Third, these policy-making conditions led neither to for the administrator to issue guidance to lenders to
prolonged horse trading nor partisan stalemate or struggles prioritize underserved markets and businesses owned by
over the rules of the game, but rather to selecting a scalable, women and “socially and economically disadvantaged
short-term, oﬀ-the-shelf, and ready-to-go solution based individuals.” Yet even among community development
on existing policies and institutions. Senator Marco ﬁnancial institutions (CDFIs) committed to those aims,
Rubio, the SBA and Entrepreneurship Committee chair, advocates appreciated the pressures to rely on 7(a) banks.
took the lead. Rubio proposed an expanded SBA 7 As an advocacy representative noted,
(a) program in group and leadership meetings. He had
if you’re trying to move a lot of money quickly … an organization
previously crafted an industrial policy plan to scale up SBA that has scale is going to be able to do that. … Bank of America is
loan programs to bolster US innovation capacity to com- going to be able to move faster than a community-based
pete with China. He used that plan to enroll Senators lender.”22
Susan Collins and Ben Cardin and House Small Business
Chair Nydia Velázquez (D-NY) into the push for a 7 Lawmakers walked a tightrope, forging a program of
(a) option. A Rubio committee staﬀer explained, unprecedented scale built on existing institutions and
networks at a critical juncture and under extraordinary
[W]e were just using the levers available to us. … [W]e were time constraints.
building PPP oﬀ of the existing 7(a) program … making direct Launched on April 3, 2020, this architecture set into
amendments … because our goal was both as a matter of politics
for getting the thing done and as a matter of policy for having
motion a sequence, documented in ﬁgure 1, that shaped
something that could go to work quickly. Our goal was to and reshaped loan ﬂows over time. In this sequence, policy
reformat the existing program by essentially changing the num- architectures and outcomes induced, rather than merely
bers in the existing 7(a) program, changing the guarantee … reﬂected, organization and politics. In period 1, the PPP’s

1278 Perspectives on Politics

Downloaded from https://www.cambridge.org/core. IP address: 73.252.170.234, on 04 Oct 2026 at 17:51:42, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S1537592725102272
basic design prioritized subsets of lenders and borrowers, This militated against small, often informal or cash-based
yielding a one-sided distribution. This in turn activated operations in minority or poor areas that needed small
negative feedback dynamics, fueling a media storm, political loans and “a lot of hand holding and explaining [and] trust
backlash led by newly formed group the Page 30 Coalition, building”28 to connect and become eligible. “[T]he busi-
and through these, policy revisions that reconﬁgured the ness model … is very expensive,” a CDFI oﬃcer explained,
lender ecologies and loan ﬂows. “because this is a very high-touch model [and] sometimes
you have to actually go back to basics of bookkeeping
The Onset [and] tax returns.”29
Third, lawmakers, at the urging of Secretary of the
Respondents emphasized how the PPP’s design ensured
Treasury Steven Mnuchin, launched the PPP as a program
loan ﬂows in period 1 went mainly to larger, more
for employer organizations without rules or forms for solo
proﬁtable employers operating in less hard-hit, more aﬄu-
operators or Schedule C ﬁlers. When the SBA issued rules
ent, and predominantly white communities. First, using
for such ﬁlers, it required using “net proﬁts” for loan
SBA 7(a) placed the least inclusive but already approved
eligibility and amounts, disqualifying businesses with
mainstream banks ﬁrst in the queue (large bank holding
depreciation and a loss for the prior year. This excluded
corporations, larger commercial banks, community banks)
single-person operations that predominated in distressed
while closing out lenders with track records for reaching
and minority communities. One city’s Black Chamber of
borrowers in poor and minority communities (CDFIs and
Commerce president summarized these eﬀects:
ﬁnancial technology services [ﬁntechs]).23 The SBA wrote
rules to allow CDFIs and other nondepository lenders to The ﬁrst wave certainly was pretty much inaccessible to most of
participate. However, it did so only after period 1 was our Black-owned businesses. … One, they did not have estab-
lished banking relationships where their bankers could call and
underway and required at least $50 million in prior-year
say, “You need to apply for this.” Two, they did not have the
small-business loans, which “pushed out a lot of smaller, but ﬁnancials in place that qualiﬁed them for such. Third, most
high-performing CDFIs that could have made a real con- Black-owned or small businesses are solo entrepreneurs. So, they
tribution in their communities,” according to two CDFI did not have the employees that would allow them to take
observers.24 Even eligible newcomers struggled to negotiate advantage of the PPP.30
the SBA approval process, protocols, and portals, delaying Finally, respondents highlighted how quickly the program
entry. Noninclusive lenders issued 93% of period 1 loans. exhausted its ﬁrst allocation and how the ﬁrst-come, ﬁrst-
Second, the program granted lenders discretion to prior- served design sparked a competitive run on funds that
itize applicants, leaving in place mainstream banks’ preex- exacerbated inequalities. When “they opened up PPP, it
isting relationships with clients and their long history of was a stampede.”31 “[W]e were just rushing, rushing,
disconnect, discrimination, and mistrust with poor and rushing, rushing in 2020, right?” a loan oﬃcer recalled,
minority communities. This yielded “predictable results.” “’[c]ause everybody was just lending from the same or
“Many banks decided to handle only existing clients,” one borrowing from the same pool.”32 Lenders serving under-
bank’s chief loan oﬃcer reported. “Wealthy existing resourced areas could not get into the queue as large banks
clients,” its CEO added. “And they did not open those to deluged the portal with their automated systems. In
noncustomers.”25 Banks had an interest in keeping existing “period 1, we got locked out … the money was used up
clients solvent and used their ties to reach out to them to in eight days, and we only managed to get nine applica-
share information and access the PPP. Borrowers with prior tions in,” a CDFI loan oﬃcer explained.33
relationships caught banks’ attention. Additionally, “know In short, PPP’s initial design induced a deeply skewed
your customer” rules, 7(a) contingent guarantees, and the distribution of lenders and loan ﬂows, independently of
PPP’s borrower self-certiﬁcation made banks anxious and conventional political dynamics rooted in party aﬃlia-
unwilling to lend to unknown clients. As an SBA lender tions, electoral competition, or lobbying.
explained, banks eventually knew that PPP loans were
100% forgivable and 100% protected. But that was not every-
one’s ﬁrst blush of how this thing was going to roll out. The ﬁrst The Outcry
loans … were by lenders to folks that they thought they could The initial allocation of funds sparked a surge of media
count on no matter how this ended up unraveling.26
exposés and political blowback. It also facilitated the
Furthermore, the program design reinforced the deeply formation of the Page 30 Coalition in April 2020, a broad
ingrained economic calculus that small loans are too costly reform group named after the CARES Act page containing
to write. According to one observer, the “sense of the Senate” resolution, which comprised at its
peak nearly a hundred associations, advocacy organiza-
banks said, “You know what, your loan is too small for us. We
can’t justify spending time, just go to the credit union next door, tions, and lenders (see online appendix C). At the forefront
they will take care of you.” … Because look, for banks even with of PPP criticism, the Page 30 Coalition sought to raise the
the fee structure, it didn’t make sense.27 salience of the program, direct attention to its failures, and

September 2026 | Vol. 24/No. 3 1279

exploit the potential for reform. “Black ﬁrms are being worked closely with others to pressure Congress and the
shut down,” said one respondent,34 echoing observations SBA to redirect loan ﬂows to neglected businesses and
from multiple quarters (e.g., Flitter and Cowley 2020; communities. And it came forward with a stream of
Mills and Battisto 2020) that were presented in a much- concrete, expert-based proposals to (1) expand the number
cited report that found four of 10 Black businesses had and types of PPP lenders, (2) bolster lender capacities to
gone under due to COVID-19 (Fairlie 2020). The SBA deploy more resources, and (3) reduce barriers to partic-
“knew what had to be done,” the CEO of a large CDFI ipation among businesses excluded in the ﬁrst period.
expressed, “but CDFIs were not lending at that time. … Marking a clear case of policy ! politics, media outcry
We could have, but there were no regulations even and broad mobilization generated immense political pres-
there.”35 An association oﬃcer noted that this fueled sure on the Trump administration and Republicans in
“blowback to the mainstream banks.”36 Congress to revise the program. As a Senate Republican
News accounts underwrote outrage and mobilization. staﬀer mentioned,
Studies and government reports ﬁlled empirical gaps
(Chetty et al. 2020; Fairlie 2020; Government Account- when we wrote the original PPP, we were working at a breakneck
pace. … By the time you got to subsequent tranches, people were
ability Oﬃce 2021; Oﬃce of Inspector General 2020). a lot more aware and had an interest in really advocating strongly
Figure 2 plots weekly counts of articles linking the PPP to to get their piece of the pie. The challenge that continued to be
minority businesses and racial disparity (line chart) along debated was the penetration into underserved areas.38
with weekly counts of congressional hearings (bar chart)
Media coverage further ampliﬁed the situation, as a trade
and bands of policy reforms (shaded areas, detailed in
association representative related:
table 1 below).
As the article counts show, media coverage spiked There was a ton of media attention on the program. … [G]oing
dramatically as the ﬁrst period closed. This media in and trying to make the case was easier than it might be if it’s
“tsunami” was important, a coalition leader explained, as an issue that isn’t topical or isn’t part of the day-to-day
conversation.”39
focusing attention on the PPP’s lopsided distribution
enabled advocates to “get all the key groups, including This, in turn, provided leverage to organize and lobby.
some conservative-leaning groups, to say this isn’t “You cannot discount the importance of the advocates at
working.”37 Building on this, the Page 30 Coalition that point in time,” emphasized a congressional staﬀer.40

Figure 2
Media Attention and Congressional Hearings, April 2020–July 2021

1280 Perspectives on Politics

When administrative rulemaking was in play, advocates others to testify, to provide information and stories that
approached the SBA and Treasury. When legislation was deﬁned the problem for legislators, and to propose concrete,
on the table, advocates singled out individual lawmakers informed solutions. Critically, Congress actively reached
and their districts: out to these groups. As a CDFI manager related, “[t]hey
were contacting us because they wanted to understand
[We] would send the letter to all the relevant Senate oﬃces, issues that our communities were facing and that banks,
usually just those on the committees. And then you would follow CDFI, minority-owned banks were facing.”42 A trade
up by asking for a meeting to discuss it. … And then during that,
it’s uplifting stories. So, storytelling is, I think, also a piece of it, association staﬀer added, “Suddenly they were like, oh,
the person who had a Wells Fargo account but wasn’t able to we have these CDFIs, yeah, let’s focus on them, because
access the program because their loan was too small to make it otherwise, who else were they going to point to do it?”43
worth Wells Fargo’s while, pointing to the deﬁciencies with the And backing these concrete suggestions was a remarkable
current system.41 unity of focus within a coalition of diverse groups:
Heightened pressure led Congress to convene hearings. [T]hese groups don’t really play in this sandbox together often,
They addressed the PPP’s racial biases and lack of inclu- right? … In any other circumstance, it would have been incred-
sivity, creating opportunities for the Page 30 Coalition and ibly hard to get all of these groups signed on that swiftly. [But] we

September 2026 | Vol. 24/No. 3 1281

knew that we all needed to come together and lock arms. We had independent contractors in the ﬁrst reform cluster. How-
one enemy, and the enemy at the time we thought was COVID, ever, most of those borrowers were not eligible for funds
but the enemy was structural bias, racism, and the ineptitude of
the federal government.44
until after March 8, 2021, when the SBA modiﬁed
Schedule C calculations to allow gross income rather than
net proﬁts. “Schedule C for us was a particularly big push,”
The Outcome related a CDFI loan oﬃcer and Page 30 Coalition mem-
Media coverage, mobilization, and targeted advocacy ber. “[I]t was a clear way to help those smaller businesses,
yielded a stream of policy revisions that reconﬁgured the and 95 or higher percent of Black-owned businesses [that]
PPP and altered the distribution of loans. Table 1 lists key are one-person businesses.”46 This, along with period
reforms by the three program periods—seven clusters over- 3 reforms raising the minimum bank fees for loans to
all—which sought to redirect loan ﬂows in diﬀerent ways. $2,500 and removing limits on borrowers with nonfraud
First, policy makers revised rules to include and elevate felony convictions, delinquent student loans, or Individual
lenders that traditionally served smaller businesses and Taxpayer Identiﬁcation Number (ITIN) immigrant sta-
borrowers in poor and minority communities. On April tus, dramatically expanded borrower eligibility and shifted
30, the SBA expanded eligibility for CDFI and ﬁntech the cost calculation for lenders.
nonbank lenders, and the Department of the Treasury As respondents repeatedly stressed, program revisions
dropped the previous lending threshold from $50 million changed the PPP for borrowers and lenders, supporting
to $10 million. Both measures increased the number of striking shifts in lender ecologies and loan distributions.
approvals for these providers to become SBA/PPP lenders, Figures 3 and 4 overlay reform clusters with charts of daily
to lend in more places, and to issue larger amounts. In loans made in each period broken down, respectively, by
addition, Congress and the SBA dedicated funds and all loans and borrower subcategories and by lender types,
granted privileged access to the portal for CDFIs and and conﬁrm these observations. There are clear policy-
depository institutions serving low-income communities response shifts in ﬂows following ﬁve or six of the seven
or making small loans. The goal was to protect these reform clusters. Loan volumes increased in period 1 after
lenders from competitive free-for-alls and channel funds cluster I (creating the PPPLF, ﬁrst eligibility for Schedule
to communities excluded in period 1. Dedicated funds C ﬁlers) overall, for minority and high-poverty commu-
(set-asides) included $30 billion for these lenders in the nities, and especially for Schedule C borrowers and ﬁntech
Paycheck Protection Program and Health Care Enhance- lenders that concentrate on the smallest loans. Loan ﬂows
ment Act on April 24, 2020; another $10 billion for also shifted in response to four policy revisions in period
CDFIs on May 28; and $75 billion in the December 3. Starting loan volumes after reform clusters IV and V
27 Consolidated Appropriations Act 2021 for CDFIs, (revised fee structures, set-asides and exclusive access for
small depositories, ﬁrst-time borrowers, and microbusi- small depositories, small employer organizations, and
nesses. The SBA granted exclusive portal access to the CDFIs) exceeded most levels from period 2. Lending
smallest depositories in late April 2020, to community accelerated and shifted again more dramatically following
ﬁnancial institutions for the ﬁrst eight days of third-period reform cluster VI (exclusive portal access, lifting restric-
applications, and CDFIs for most of May 2021. tions for felons, immigrants, and student loan payment
Second, the Federal Reserve and the SBA adopted delays, and allowing gross income for Schedule C appli-
measures to boost the lending capacities of inclusive cants), and yet again following cluster VII (exclusive
institutions and newcomers. SBA calls, webinars, and CDFI access). Notably, increases after clusters IV
changes to the portal that included automated referral and V, but especially clusters VI and VII, were concen-
helped them to connect with borrowers and better navigate trated heavily in loans to Schedule C borrowers and
the system. Yet the “game changer,” as participants term it, businesses in minority and high-poverty communities on
was the Fed opening its PPP Liquidity Facility (PPPLF) to the borrower side, and in ﬁntechs and CDFIs that
them after 21 days on April 30, 2020, which allowed them specialize in serving those communities and borrowers
to move loans oﬀ their balance sheets. “[U]ntil the Federal on the lender side. And the dramatic upticks after cluster
Reserve window opened for us, there was no way we were VII appeared exclusively from CDFIs. The shifts in loan
going to operate at scale,” explained a CDFI loan oﬃcer. ﬂows that dominated period 3 reﬂect the impact of
“What’s 20 million in the line of credit? Literally nothing. policy reforms.
We did 200 million in PPP loans. There was no way, with Loan ﬂow overlays with period 2’s targeted reforms reveal
only our normal course of business liquidity instruments, a more complex policy-response picture. Early reforms
we will get there, guaranteed or not.”45 targeting (nonbank) CDFIs and ﬁntechs on their own had
Third and ﬁnally, revised eligibility requirements for little direct impact. Loans increased after cluster II, but for
small businesses and nonpayroll ﬁrms closed out in periods all lender types. Increases in loans after cluster III were
1 and 2 were a second “game changer.” The SBA had small, delayed, and not conﬁned to CDFIs. Respondents
authorized lending to Schedule C sole proprietors and emphasized how exploiting eligibility, getting authorized,

1282 Perspectives on Politics

making sense of SBA systems, and ﬁnding bank partners to increased nearly thirtyfold among CDFIs (from 45,300
access the PPPLF took time for newcomers to SBA 7(a), loans in period 1 to almost 1.27 million in period 3) and
suggesting delayed policy eﬀects for clusters II and III. ﬁntechs (from 59,000 to nearly 1.74 million). Growth
Moreover, reforms targeting CDFIs and ﬁntechs like autho- outpaced other lenders by an order of magnitude, driving
rization, exclusive access, liquidity support, or even revised CDFI and ﬁntech shares from 6% in period 1 to 42% in
fee structures can have little impact on loan ﬂows if there is period 3, displacing mainstream banks from their over-
no payoﬀ for the low-margin, often solo-operator clients whelming dominance and 90% share in period 1.
that these lenders serve in poor or minority areas. That is, At the same time, program revisions opened the gates for
they would have no impact unless and until they were PPP loans to the kinds of traditionally excluded businesses
combined with other policy reforms, especially the game- that predominate in underresourced and Black and brown
changing Schedule C reforms, which is precisely what we communities. From period 2 to 3, PPP loans to Schedule C
observe. Key reforms in cluster VI (March 8) dramatically borrowers skyrocketed from about 128,500 to almost 3.2
increased loan ﬂows through CDFIs and ﬁntechs and million, or almost 51% of all period 3 loans, with CDFIs
unleashed a ﬂood of Schedule C loans and loans to poor and ﬁntechs leading the way by issuing over two million
and minority communities. (or over 64%) of Schedule C loans. And with these
Individually, but especially in combination, endoge- dynamics came shifts in the socioeconomic composition
nously produced policy revisions radically recomposed of places receiving loans. Five- to sixfold increases in loan
the PPP and its loan ﬂows across lenders, borrowers, and numbers for high-poverty and majority nonwhite districts
communities. First, they pivoted the program from main- boosted loans to high-poverty and majority nonwhite
stream banks to inclusive CDFI and ﬁntech lenders. Loans districts to over 532,000 and 1.9 million, respectively, in
through all lender types doubled from period 1 to 2, and period 3. Program revisions raised high-poverty-district
nearly doubled again from period 2 to 3. Yet lending shares of PPP loans to 10% and minority-district shares

September 2026 | Vol. 24/No. 3 1283

Figure 4
Loans and Reform Clusters by Program Rounds and Select Lender Types

from 19% to 36%. Finally, they also raised loan-to-business larger ﬁrms are more informed about opportunities and
ratios in high-poverty and minority districts from 0.07 to protocols than smaller ﬁrms, possessed greater resources,
0.49 and from 0.06 to 0.39, respectively, closing the gap organizational capacities, and connections for leveraging
with more aﬄuent and majority-white ones.47 public beneﬁts, and were more attractive to banks or other
In sum, process tracing reveals that the PPP’s initial program providers. Unlike our account, the alternative
architectures induced a macrodevelopmental trajectory of suggests progression dynamics that operate independently of
politics and policy change that shaped and shifted loan the PPP’s design, subsequent revisions, and one or two
distributions across places. A new policy to support small expected loan ﬂow patterns. We might expect succession,
businesses replicated racial and class biases in banking, with better-informed large ﬁrms storming out of the gates
triggering media attention and congressional hearings that and getting loans ﬁrst, followed by medium-sized small
provided windows and leverage for group mobilization businesses, then microemploying organizations, and lastly
and targeted advocacy. Policy fueled politics, resulting in solo operators as they gather information about the pro-
administrative and legislative changes reconﬁguring the gram. Or we might expect exhaustion or crowding-out
PPP’s lending ecologies and ﬂows. processes, with better-resourced large ﬁrms muscling their
way into the program and tapping all the funds they need
before progressively smaller businesses can gain access.
Alternative Explanations Interviews and examination of PPP loans and coverage
The strength of our inference also depends on the viability by ﬁrm size do not support this account. Interviewees
of alternative explanations for shifting distributions. One stressed that banks turned away small ﬁrms that were
alternative emphasizes a natural progression by ﬁrm size in informed and actively seeking loans, indicating that infor-
which large ﬁrms precede or crowd out small ﬁrms in mation diﬀerences alone were not driving outcomes; banks
accessing program beneﬁts. Like our account, this one had to reach out to educate larger ﬁrms. Respondents also
rests on observations of the PPP and other programs that stressed how net proﬁt rules kept Schedule C solo operators

1284 Perspectives on Politics

Downloaded from https://www.cambridge.org/core. IP address: 73.252.170.234, on 04 Oct 2026 at 17:51:42, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S1537592725102272
out and that those businesses typically could not obtain online appendix E. We start with a baseline (M1) that
funding until those rules changed. Moreover, loan counts controls for population, urbanization, COVID-19 cases,
and cumulative coverage rates in ﬁgures D1 and D2 in vulnerable sectors, state ﬁxed eﬀects, and program periods.
online appendix D show similar loan ﬂow patterns for large, To that baseline, we add variables for conventional politics
medium, and small employers, with simultaneous surges and business power, separately (M2 and M3) and together
and plateaus rather than staggered sequences. They also (M5), and after including socioeconomic characteristics in
provide no evidence that small employers had to wait until a new baseline (M4), repeat the processes of adding politics
larger ones had their ﬁll. All three smaller ﬁrm classes could and business variables (M6, M7) to arrive at the full model
access funds fully despite barely a quarter of the largest ﬁrms (M8).
being covered. Moreover, while surges for Schedule C Coeﬃcients for control factors (ﬁgure 5, top left panel),
operators lagged those for larger ﬁrms, they did not occur socioeconomic conditions, and time (bottom right panel)
either right before or soon after other groups ﬂattened, but return signiﬁcant results. Loans to districts increase with
nearly a year later, following revision of net proﬁt rules, COVID-19 cases and the Black population, increase
indicating policy rather than natural progression eﬀects. dramatically in periods 2 and 3, and decrease with poverty
Size matters, but its distributional impacts depend heavily rates. In contrast, only one of the coeﬃcients for the eight
on policy architectures. conventional-politics and business power variables (top
A second argument for shifts in loans focuses on the right and bottom left panel) reaches statistical signiﬁcance
“exogenous shock” of George Floyd’s murder on May —whether the member served on the House Committee
25, 2020, and the subsequent Black Lives Matter (BLM) on Small Business that crafted the PPP’s core architecture.
movement. Protests and pressures on lenders rather than Contrary to conventional-politics expectations, that coef-
policy shifts might have produced loan ﬂows to minority ﬁcient is negative, not positive. Moreover, the array of
communities. Interviewees conﬁrm how BLM dynamics models in table E1 in online appendix E conﬁrms these
generated such pressures in the PPP: “After George Floyd, results. Four politics variables reach signiﬁcance in
it was hard for Republicans to be like … ‘The Blacks don’t reduced models. Yet donations, like committee member-
need any money,’” a staﬀer related.48 In that context, a ship, return negative eﬀects; the Partisan Voting Index
CDFI oﬃcer explained, the PPP “was one more well- shows a positive rather than the expected negative eﬀect;
documented way that Black and brown borrowers were and the party, donations, and Partisan Voting Index vari-
being left out … something that we could work together ables lose signiﬁcance when controlling for socioeconomic
on.”49 However, loan ﬂow patterns in ﬁgure 3 indicate conditions and time, leaving only one negative committee-
little if any direct impact of BLM dynamics on loans to membership eﬀect.
minority communities. Only very modest and delayed We conducted extensive robustness checks to assess
increases to those communities occurred during the sum- whether this strikingly poor performance of conventional-
mer of 2020, which were dwarfed by upticks after cluster politics models was an artifact of our modeling or data
VI reforms almost a year later. This is not to dismiss BLM choices (online appendix F). We reﬁt models using all
eﬀects. To the contrary, they powerfully impacted loan loans and just ﬁrst loans, with and without ﬂagged loans.
ﬂows, but indirectly, via policy revisions, by amplifying We use diﬀerent dependent variables (log loan numbers,
media attention and public criticism already underway, log loan values, loan and value ratios) and measures of
keeping issues of racial inequities in the PPP highly salient politics to capture battleground places, core Republican
(ﬁgure 2) and providing advocates with ammunition to districts, presidential aﬃliation, and campaign donations
push for reforms. from ﬁnancial versus nonﬁnancial industries. We even ﬁt
models separately for each loan period with each depen-
Quantitative Findings dent variable (table E4 in online appendix E). Yet these
We further engage these and conventional-politics models consistently return null or negative results, gener-
accounts via quantitative analyses and panel models of ating only occasional supportive results in a particular
the distribution of PPP loans (loan ratios) across congres- speciﬁcation or period and providing little explanation of
sional districts. We begin with models based on the distribution of PPP loans across districts.
conventional-politics approaches and then quantitatively
model some IP-PIT-inspired implications derived from
IP-PIT-Based Arguments and Findings
qualitative ﬁndings.
In contrast, the limited explanatory power of conven-
tional-politics approaches jives well with IP-PIT ideas
Conventional Politics about critical junctures and path dependence. Moreover,
Figure 5 reports ﬁndings from conventional-politics our qualitative ﬁndings about how policy generated
models by plotting coeﬃcients from the full model we politics and program trajectories imply testable expecta-
developed via the model-building process tabled in E1 in tions about how socioeconomic factors shaped PPP

September 2026 | Vol. 24/No. 3 1285

Figure 5
Coefﬁcient Plots of Full Random-Effects Model (Model 8)

distributions over time. As negative feedback induced by districts with fewer large establishments. These trends
the PPP’s initial architecture and distributional biases fuels reﬂect policy reforms favoring the smallest, solo-operator
policy changes recalibrating lenders and loan ﬂows, we businesses in some of the poorest and minority districts.
expect the impacts on ﬂows of ethnoracial composition, In a ﬁnal step, we reanalyzed these relationships and
poverty, and large establishments to shift. We expect time-varying eﬀects using a more ﬁne-grained approach
negative eﬀects of minority composition and poverty on based on our qualitative ﬁndings. We ﬁt models that use
loans in period 1 and a positive impact of larger establish- weekly loan counts as outcomes and reparametrize time
ments. Those eﬀects will diminish and even reverse in using cumulative counts of critical media articles, congres-
periods 2 and 3 as policy reforms reconﬁgure the terrain sional hearings, and reform clusters. Instead of breaking
for borrowers and lenders. down PPP lending into three program periods, these
To assess these ideas, we add interaction terms between models estimate separate interaction eﬀects for saliency,
program periods and four measures of socioeconomic hearings, and clusters as they evolve to more directly tap
conditions in districts to our entire model. Based on backlash politics and policy impacts highlighted in our
models tabled in E2 in online appendix E with estimated IP-PIT account. Tabled in E3 in online appendix E, these
eﬀects plotted in ﬁgure 6, the results in the top panel reveal models return signiﬁcant interaction coeﬃcients for all
time-varying eﬀects. Slopes in period 1 for percent Black, three moderator variables, providing clear evidence of how
Hispanic, and poor indicate negative eﬀects on loans, negative feedback, public criticism, and policy revisions
while large ﬁrms have a positive relationship. Over time, reshaped loan ﬂows. Figure 7 charts the interaction eﬀects
these eﬀects reverse. Lending in periods 2 and 3 is more for policy revisions. Before any reform, higher poverty and
inclusive, with loans ﬂowing to poorer, more diverse Black or Hispanic population shares decreased, and large-

1286 Perspectives on Politics

business prominence increased loan ﬂows. Yet as policy interviews, process tracing, and additional quantitative
reforms accumulate, these eﬀects change, switch, and modeling, we turned to institutional politics (IP-PIT) to
result—in reform cluster VII—in positive associations of better explain the PPP and its trajectories.
loan ﬂows with Black, Hispanic, and poverty covariates, IP-PIT emphasizes how policies generate politics, dis-
and negative ones for large businesses. These results tributional outcomes, and feedback loops, suggesting a
parallel those for saliency and hearings and those shown policies ! politics/outcomes ! policies sequence (Moynihan
above for the three-period analysis, and conﬁrm our key and Soss 2014; Pierson 2000; 2004; Sorelle 2023). This
ﬁnding: policy revisions reconﬁgured how the program led us to anchor our study in a qualitative analysis of
allocated funds, radically shifting the sensitivities of loan temporal variation in policy architectures, politics, and
ﬂows to district socioeconomic conditions. loan access across the program’s three periods. However,
we also engaged conventional and IP-PIT approaches via
Conclusion quantitative analyses of PPP loan ﬂows. While blending
What explains the distribution of PPP funds across con- diﬀerent approaches and methodologies is unusual and
gressional districts? While prior scholarship advances ambitious, it lets us address and contribute to debates over
conventional-politics ideas, analyzing the full corpus of the linkages between politics and policy in ways that
loans raises questions for such politics ! policy ! outcome otherwise might not be possible.
accounts. Electoral dynamics, committee membership,
campaign donations, or industry clout are largely uncor-
related with distributional outcomes, and respondents Contributions
repeatedly emphasized how the pandemic thoroughly The ﬁrst of its kind, our study contributes empirical and
undermined conventional political processes. This under- theoretical insights into not just one of the most signiﬁcant
scores the need for an alternative account. Based on public interventions in the US economy since the Great

September 2026 | Vol. 24/No. 3 1287

Figure 7
Time-Dependent Context Effects on PPP Loan Ratios by Reform

Depression, but also policy making during crises. Our induced backlash politics (Jacobs and Weaver 2015;
ﬁndings highlight likely general limits on the ability of Patashnik and Zelizer 2013; Schneiberg 2005) and policy
conventional-politics approaches to explain such issues. revisions that reconﬁgured loan ﬂows and their sensitivi-
Instead, core IP-PIT concepts of critical junctures, policy ties to socioeconomic conditions.
feedback, and path-dependent trajectories apply beyond In short, the PPP’s evolution and distributional out-
epochal institutional changes, including policies tempo- comes were rooted in a critical crisis juncture, followed by
rarily implemented as emergency measures that may or policy choices that generated feedback eﬀects and endog-
may not yield enduring resets. As shown above, the enously produced developmental trajectories. Moreover,
COVID-19 pandemic sparked a critical “crisis” juncture consistent with IP-PIT claims, the impact of socioeco-
that suspended normal politics, creating the opportunity nomics and politics on loan ﬂows depended on timing,
for a program that astonished participants for its scale, shifted across rounds, and became a causal factor in its own
speed, and scope. Once in place, PPP “policies created new right.
politics” (Hacker and Pierson 2019; Jacobs and Weaver Our work also advances theoretical understandings of
2015) that set the program on a two-year path. Its initial key IP-PIT concepts. First, it indicates varieties in critical
architectures and distributions generated new interests and junctures and their dynamics. Speciﬁcally, it highlights
organizations like the Page 30 Coalition (North 1990; junctures in crisis moments that upend business-as-usual
Pierson 2000; Skocpol 1996). They mobilized a range of politics but without systemic struggles over the rules of the
constituencies from CDFIs to Black and Hispanic cham- game or fundamental challenges to their legitimacy. The
bers of commerce (Moynihan and Soss 2014; Sorelle COVID-19 pandemic took decision makers by surprise. It
2023; Thurston 2015). They reframed public discourse, confronted them with impending existential losses and
calculations, and debate as media reports, academic stud- profound uncertainty that trumped other preferences and
ies, and hearings piled up (Dobbin 1994). They also forced them to react immediately through emergency

1288 Perspectives on Politics

Downloaded from https://www.cambridge.org/core. IP address: 73.252.170.234, on 04 Oct 2026 at 17:51:42, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S1537592725102272
instead of conventional policy making. This opened a believe that an IP-PIT approach, deployed with sensitivity
window for innovation. Yet policy makers’ beliefs were to variations across cases, can be fruitfully applied to other
also decisive: beliefs that COVID-19 was a short-term cases of policy making during crises, including the 2009
crisis to be managed in the here and now, and that policy American Recovery and Reinvestment Act (ARRA) to
solutions, however out of line with broader preferences, address the Great Recession or the 1989 Financial Institu-
were temporary measures to be set aside or revisited in the tions Reform, Recovery, and Enforcement Act (FIRREA)
near future. In these crisis junctures, decision makers are to address the savings and loan crisis. For the PPP, height-
far more likely to select short-term “oﬀ-the-shelf” solu- ened time compression and isolation during the lockdown
tions rapidly rather than pursue systemic change (Cohen, made legislators eschew conventional politics in favor of oﬀ-
March, and Olsen 1972; Kingdon 2011). the-shelf solutions. Furthermore, the pandemic exposed a
Second, our study also contributes to IP-PIT analyses of wide array of actors to unanticipated, potentially seismic
path dependence by shedding new light on dynamics and losses or shifts in payoﬀs. That created profound uncer-
conditions for negative policy feedback. As with positive tainty about how events would unfold. Framing the crisis as
feedback eﬀects, negative feedback involves political pro- a short-term challenge (“ﬂattening the curve”) required a
cesses endogenously generated by a new institution, but quick solution, which prompted oﬃcials to commit to bold
which erode support and create pressures for change but temporary interventions. If these conditions were pres-
(Clemens 1997; Daugbjerg and Kay 2020; Jacobs and ent in the ARRA (Grunwald 2012) and FIRREA (Cassell
Weaver 2015; Patashnik and Zelizer 2013; Schneiberg 2002; White 1992), then an IP-PIT framework would
2005; Thurston 2015; Weaver 2010). New institutions or apply, at least in part.
policies can foster backlash politics by creating losers and These cases are not identical to the PPP. Legislation in
opponents with new grievances who form new coalitions the ARRA and FIRREA involved longer junctures,
to unsettle settled policy and advance reform. in-person deliberation, and continued access to policy
The PPP case extends these insights. Unintended prob- makers. These diﬀerences would support more developed
lems created opportunities and leverage for reform coali- political struggles over early policy interventions and an
tions. By authorizing private lenders as program increasing reliance on conventional politics once emer-
gatekeepers, the PPP’s initial design baked the racial and gency measures were adopted. ARRA and FIRREA also
class biases of the ﬁnancial system directly into the program, diverge from the PPP with lower organizational heteroge-
fostering a spike in inequality and outrage. As the saliency of neity, which might account for their comparatively repro-
the program’s one-sided distribution increased and the ductive trajectories. Still, this comparison suggests that
broader public was drawn into policy debate, administrators IP-PIT arguments, carefully tweaked, can serve as useful
and elected oﬃcials were compelled to pay attention and starting points or foundations for future studies.
advocates for reform were able to sharpen and focus public Second, what kinds of new institutional equilibria
concerns and pressure through stories, data, and analyses. might we expect from the PPP or other cases? There is
Congressional hearings provided advocates with public some evidence of potentially lasting resets at the SBA.
venues to address the program’s legitimacy problems and During the PPP, the agency learned a great deal and built
to advance critiques and targeted solutions. administrative capacities by upgrading its portals, creating
Yet critical for policy change was the availability of digital infrastructure, and developing guidance for those
alternatives from which actors could craft viable reforms. traditionally excluded from its services. Community-based
Prior work highlights alternative models or logics (e.g., institutions and ﬁntechs also engaged in capacity building
Weaver 2010). However, within the PPP, it was organi- with potentially lasting impact. They developed new
zational heterogeneity on the path in banking and alter- underwriting protocols, adopted new technologies, and
native organizations that proved decisive for reform and established relationships with new classes of customers
change (Schneiberg 2007). CDFIs brought a long and they hoped to continue serving postpandemic. Yet these
tested history of successful relational banking in under- do not add up to enduring wholesale resets. To the
resourced and minority communities. Fintechs provided contrary, we ﬁnd here, and suggest more generally, that
borrowers with access to ﬁnance in places without banks or short-term crisis management interventions will establish
branches and the ability to scale. Both groups of organi- new equilibria or paths for the duration of the program,
zations lobbied and provided credible proposals for policy followed by reversions thereafter to a slightly modiﬁed
change. Both provided advocates and policy makers with “business as usual.” This was the case for the PPP, and we
workable platforms to alter policies and loan ﬂows. suspect for the ARRA, which modiﬁed mortgage market
strategies and volumes postcrisis and came with a new layer
of regulation, but left the basic ﬁnancial market architec-
Directions for Future Research tures and players largely in place.
Two questions suggest productive avenues for future That said, there are opportunities here for future
research. First, can we generalize from our ﬁndings? We research to address whether and how other short-term

September 2026 | Vol. 24/No. 3 1289

crisis interventions might yield more seismic resets. And 16 Interview 12.
there is also an intriguing possibility that interventions like 17 Interview 77.
the PPP might aﬀect subsequent crisis interventions by 18 Interview 39.
serving as templates for later ones or adding new tools to 19 Interview 78.
policy makers’ repertoire during crises, supporting larger 20 Interview 41.
resets and enduring path dependence in crisis management. 21 Interview 82.
22 Interview 54.
Supplementary material 23 Mainstream banks, credit unions, and farm credit
associations displayed markedly exclusive baseline
To view supplementary material for this article, please visit propensities in the PPP. They were more likely to
http://doi.org/10.1017/S1537592725102272. avoid PPP lending to high-poverty or minority com-
munities than aﬄuent or white ones and issued fewer
Data replication loans to the former on average if they did any lending
Data replication sets are available in Harvard Dataverse at: there. CDFIs and ﬁntechs, in contrast, were more
https://doi.org/10.7910/DVN/TIKAV2 likely to lend to communities that were high poverty
or majority-minority than otherwise and issued more
Acknowledgments loans on average to those communities (Cassell,
Schwan, and Schneiberg 2023; Schneiberg, Cassell,
The authors want to thank Mia Boyer-Edwards, Robin and Schwan 2024).
Hart, and Kristin Yeager for their extensive and helpful 24 Interview 54.
research assistance and technical support. The research 25 Interview 16.
reported here was supported by National Science Foun- 26 Interview 27.
dation grant no. 2218044 to principal investigators Marc 27 Interview 57.
Schneiberg and Mark Cassell, a DZ BANK Fellowship of 28 Interview 23.
the American Council on Germany to Michael Schwan, 29 Interview 48.
and a DAAD/AICGS Fellowship of the American German 30 Interview 59.
Institute to Michael Schwan. 31 Interview 53.
32 Interview 49.
Notes 33 Interview 16.
1 Congress followed up each refunding measure with 34 Interview 71.
additional legislation: The Paycheck Protection Pro- 35 Interview 23.
gram Flexibility Act of 2020, which amended program 36 Interview 54.
terms in key particulars on June 5, 2020, and the PPP 37 Interview 76.
Extension Act of 2021 on March 26, 2021. 38 Interview 41.
2 Loan ﬁgures here and below reﬂect data and authors’ 39 Interview 32.
calculations detailed in the Methods and Data 40 Interview 43.
section and online appendix A. 41 Interview 32.
3 Our human subjects review board certiﬁcation 42 Interview 16.
requires that all interviews are kept conﬁdential. 43 Interview 53.
Therefore, to preserve the conﬁdentiality of our 44 Interview 71.
respondents, we have withheld information about 45 Interview 27.
names and formal titles, as well as the locations and 46 Interview 80.
dates of the interviews. 47 These calculations substantially understate shift mag-
4 Interview 40. nitudes by ignoring reallocations within districts.
5 Interview 81. Recalculating loan ﬂows by Census tracts shows loan
6 Interview 41. numbers increasing more than ﬁvefold to high-
7 Interviews 41 and 43. poverty communities and tenfold to majority non-
8 Interview 38. white communities. This drove each of those com-
9 Interview 39. munities’ shares in PPP loans from under 10% to over
10 Interview 81. 25% and increased coverage rates ﬁve- to elevenfold,
11 Interview 38. from 10% to 51% in poor communities and from 7%
12 Interview 41. to 76% in minority communities (Schneiberg, Cassell,
13 Interview 38. and Schwan 2024).
14 Interview 53. 48 Interview 43.
15 Interview 82. 49 Interview 79.

1290 Perspectives on Politics

Downloaded from https://www.cambridge.org/core. IP address: 73.252.170.234, on 04 Oct 2026 at 17:51:42, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S1537592725102272
References Capoccia, Giovanni. 2015. “Critical Junctures and
Arthur, W. Brian. 1989. “Competing Technologies, Institutional Change.” In Advances in Comparative-
Increasing Returns, and Lock-In by Historical Events.” Historical Analysis, eds. James Mahoney and Kathleen
Economic Journal 99 (394): 116–31. DOI: 10.2307/ Thelen, 147–79. Cambridge, UK: Cambridge University
2234208. Press. DOI: 10.1017/cbo9781316273104.007.
Barrick, John, Adam J. Olson, and Shivaram Rajgopal. Capoccia, Giovanni, and R. Daniel Kelemen. 2007. “The
2021. “Returns to Seeking Political Inﬂuence: Early Study of Critical Junctures: Theory, Narrative, and
Evidence from the COVID-19 Stimulus.” Columbia Counterfactuals in Historical Institutionalism.” World
Business School Research Paper, May 14. New York: Politics 59 (3): 341–69. DOI: 10.1017/s004388710
Columbia Business School. DOI: 10.2139/ssrn. 0020852.
3845677. Cassell, Mark. 2002. How Governments Privatize: The
Bartik, Alexander W., Zoe B. Cullen, Edward L. Glaeser, Politics of Divestment in the United States and Germany.
Michael Luca, Christopher T. Stanton, and Adi Washington, DC: Georgetown University Press.
Sunderam. 2020. “The Targeting and Impact of Cassell, Mark K., Michael Schwan, and Marc Schneiberg.
Paycheck Protection Program Loans to Small 2023. “Bank Types, Inclusivity, and Paycheck
Businesses.” NBER Working Paper No. 27623, Protection Program Lending during COVID-19.”
updated June 2021. Cambridge, MA: National Bureau Economic Development Quarterly 37 (3): 277–94. DOI:
of Economic Research. DOI: 10.3386/w27623. 10.1177/08912424231163485.
Beach, Derek. 2016. “It’s All about Mechanisms—What Center for Responsive Politics. n.d. Campaign
Process-Tracing Case Studies Should Be Tracing.” New Contributions by Industry for 2018 and 2020
Political Economy 21 (5): 463–72. DOI: 10.1080/ Congressional Elections [database]. Washington, DC:
13563467.2015.1134466. OpenSecrets. https://www.opensecrets.org.
Beach, Derek, and Rasmus Brun Pedersen. 2013. Process- Chernenko, Sergey, and David S. Scharfstein. 2022.
Tracing Methods: Foundations and Guidelines. Ann “Racial Disparities in the Paycheck Protection
Arbor, MI: University of Michigan Press. DOI: Program.” NBER Working Paper No. 29748,
10.3998/mpub.2556282. February. Cambridge, MA: National Bureau of
Berger, Allen N., Mustafa U. Karakaplan, and Raluca A. Economic Research. DOI: 10.3386/w29748.
Roman. 2023. “Whose Bailout Is It Anyway? The Roles Chetty, Raj, John N. Friedman, Michael Stepner, and the
of Politics in PPP Bailouts of Small Businesses vs. Opportunity Insights Team. 2020. “The Economic
Banks.” Journal of Financial Intermediation 56: 101044. Impacts of COVID-19: Evidence from a New Public
DOI: 10.1016/j.jﬁ.2023.101044. Database Built Using Private Sector Data.” NBER
Birdthistle, William A., and Joshua Silver. 2021. Working Paper No. 27431, updated April 2023.
“Funding Crises: An Empirical Study of the Paycheck Cambridge, MA: National Bureau of Economic
Protection Program.” Buﬀalo Law Review 69 (5): 1541– Research. DOI: 10.3386/w27431.
1620. https://digitalcommons.law.buﬀalo.edu/ Clemens, Elisabeth S. 1997. The People’s Lobby:
buﬀalolawreview/vol69/iss5/5. Organizational Innovation and the Rise of Interest Group
Borawski, Garrett, and Mark E. Schweitzer. 2021. “How Politics in the United States, 1890–1925. Chicago:
Well Did PPP Loans Reach Low- and Moderate- University of Chicago Press.
Income Communities?” Economic Commentary EC Cohen, Michael D., James G. March, and Johan P. Olsen.
2021-13, May 27. Cleveland, OH: Federal Reserve 1972. “A Garbage Can Model of Organizational
Bank of Cleveland. DOI: 10.26509/frbc-ec-202113. Choice.” Administrative Science Quarterly 17 (1): 1–25.
Bovens, Mark, and Paul T Hart. 1996. Understanding DOI: 10.2307/2392088.
Policy Fiascoes. New Brunswick, NJ: Transaction. DOI: Collier, David, Henry E. Brady, and Jason Seawright.
10.4324/9781351293242. 2010. “Sources of Leverage in Causal Inference:
Brady, David W. 1991. Critical Elections and Congressional Toward an Alternative View of Methodology.” In
Policy Making. Stanford, CA: Stanford University Press. Rethinking Social Inquiry: Diverse Tools, Shared
DOI: 10.1515/9781503621442. Standards, eds. Henry E. Brady and David Collier,
Campbell, Andrea Louise. 2012. “Policy Makes Mass 161–99. Lanham, MD: Rowman & Littleﬁeld.
Politics.” Annual Review of Political Science 15 (1): 333– Cook Political Report. n.d. The Cook Partisan Voting Index
51. DOI: 10.1146/annurev-polisci-012610-135202. (Cook PVI) [database]. Arlington, VA: Cook Political
Campbell, John L., and Leon N. Lindberg. 1990. Report. https://www.cookpolitical.com/cook-pvi.
“Property Rights and the Organization of Economic Daugbjerg, Carsten, and Adrian Kay. 2020. “Policy
Activity by the State.” American Sociological Review 55 Feedback and Pathways: When Change Leads to
(5): 634–47. DOI: 10.2307/2095861. Endurance and Continuity to Change.” Policy Sciences
53 (2): 253–68. DOI: 10.1007/s11077-019-09366-y.

September 2026 | Vol. 24/No. 3 1291

Dobbin, Frank. 1994. Forging Industrial Policy: The Grunwald, Michael. 2012. The New New Deal: The
United States, Britain, and France in the Railway Age. Hidden Story of Change in the Obama Era. New York:
Cambridge, UK: Cambridge University Press. DOI: Simon & Schuster.
10.1017/CBO9781139174183. Ha, Heonuk. 2023. “Presidential Particularism, the
Duchin, Ran, and John Hackney. 2021. “Buying the Paycheck Protection Program (PPP), and the Biden
Vote? The Economics of Electoral Politics and Small- Administration.” Presidential Studies Quarterly 53 (4):
Business Loans.” Journal of Financial and Quantitative 543–69. DOI: 10.1111/psq.12850.
Analysis 56 (7): 2439–73. DOI: 10.1017/ Hacker, Jacob S., Alexander Hertel-Fernandez, Paul
S002210902100048X. Pierson, and Kathleen Thelen. 2022. “The American
Edelman, Lauren B., Christopher Uggen, and Howard S. Political Economy: Markets, Power, and the Meta
Erlanger. 1999. “The Endogeneity of Legal Regulation: Politics of US Economic Governance.” Annual Review
Grievance Procedures as Rational Myth.” American Journal of Political Science 25 (1): 197–217. DOI: 10.1146/
of Sociology 105 (2): 406–54. DOI: 10.1086/210316. annurev-polisci-051120-013916.
Fairlie, Robert W. 2020. “COVID-19, Small Business Hacker, Jacob. S., and Paul Pierson. 2019. “Policy
Owners, and Racial Inequality” NBER Reporter 4: 12– Feedback in an Age of Polarization.” Annals of the
15. https://www.nber.org/reporter/2020number4/ American Academy of Political and Social Science 685 (1):
covid-19-small-business-owners-and-racial-inequality. 8–28. DOI: 10.1177/0002716219871222.
Fligstein, Neil. 1990. The Transformation of Corporate Harvard Center for Population and Development Studies.
Control. Cambridge, MA: Harvard University Press. 2021. COVID-19 U.S. State-Level Data: Number of
Fligstein, Neil, and Doug McAdam. 2012. A Theory of Positive Cases [database]. Boston: Harvard T. H. Chan
Fields. Oxford, UK: Oxford University Press. DOI: School of Public Health. https://
10.1093/acprof:oso/9780199859948.001.0001. www.hsph.harvard.edu/population-development/data/
Flitter, Emily, and Stacy Cowley. 2020. “Banks Gave covid-19.
Richest Clients ‘Concierge Treatment’ for Pandemic Hoﬀman, Andrew J. 1999. “Institutional Evolution and
Aid.” New York Times, April 22. https:// Change: Environmentalism and the U.S. Chemical
www.nytimes.com/2020/04/22/business/sba-loans- Industry.” Academy of Management Journal 42 (4):
ppp-coronavirus.html. 351–71. DOI: 10.2307/257008.
Gerring, John. 2004. “What Is a Case Study and What Is Hollingsworth, J. Rogers, and Wolfgang Streeck. 1994.
It Good For?” American Political Science Review 98 (2): “Countries and Sectors: Concluding Remarks on
341–54. DOI: 10.1017/s0003055404001182. Performance, Convergence and Competitiveness.”
Glennan, Stuart S. 1996. “Mechanisms and the Nature of In Governing Capitalist Economies: Performance and
Causation.” Erkenntnis 44 (1): 49–71. DOI: 10.1007/ Control of Economic Sectors, eds. J. Rogers
bf00172853. Hollingsworth, Philippe C. Schmitter, and Wolfgang
Government Accountability Oﬃce. 2021. “Paycheck Streeck, 270–300. New York: Oxford University Press.
Protection Program: SBA Added Program Safeguards, Howell, Sabrina T., Theresa Kuchler, David Snitkof,
but Additional Actions Are Needed.” Report to Johannes Stroebel, and Jun Wong. 2021. “Racial
Congressional Addressees GAO-21-577, July 29. Disparities in Access to Small Business Credit: Evidence
Washington, DC: Government Accountability Oﬃce. from the Paycheck Protection Program.” CESifo
https://www.gao.gov/products/gao-21-577. Working Paper No. 9345, October. Munich: Ifo
Granja, João, Christos Makridis, Constantine Yannelis, Institute for Economic Research. https://www.ifo.de/
and Eric Zwick. 2021. “Did the Paycheck Protection en/cesifo/publications/2021/working-paper/racial-
Program Hit the Target?” BFI Working Paper disparities-access-small-business-credit-evidence-
No. 2020–52, September 21. Chicago: Becker paycheck.
Friedman Institute for Economics. https:// Igan, Deniz, Thomas Lambert, and Prachi Mishra. 2021.
bﬁ.uchicago.edu/working-paper/did-the-paycheck- “The Politics of the Paycheck Protection Program.”
protection-program-hit-the-target. CEPR Discussion Paper No. DP16842, December 27.
Griﬃn, John M., Samuel Kruger, and Prateek Mahajan. London: Centre for Economic Policy Research. https://
2023. “Did FinTech Lenders Facilitate PPP Fraud?” ssrn.com/abstract=4026723.
Journal of Finance 78 (3): 1777–1827. DOI: 10.1111/ Ikenberry, G. John. 2019. After Victory: Institutions,
joﬁ.13209. Strategic Restraint, and the Rebuilding of Order after
Grotto, Jason, Zachary R. Mider, and Cedric Sam. 2020. Major Wars, new edition. Princeton, NJ: Princeton
“White America Got a Head Start on Small-Business University Press. DOI: 10.2307/j.ctv3znx0v.
Virus Relief.” Bloomberg, July 30. https:// Jacobs, Alan M., and R. Kent Weaver. 2015. “When
www.bloomberg.com/graphics/2020-ppp-racial- Policies Undo Themselves: Self-Undermining
disparity.

1292 Perspectives on Politics

Downloaded from https://www.cambridge.org/core. IP address: 73.252.170.234, on 04 Oct 2026 at 17:51:42, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
https://doi.org/10.1017/S1537592725102272
Feedback as a Source of Policy Change.” Governance Pierson, Paul. 1993. “When Eﬀect Becomes Cause: Policy
28 (4): 441–57. DOI: 10.1111/gove.12101. Feedback and Political Change.” World Politics 45 (4):
Kingdon, John W. 2011. Agendas, Alternatives, and Public 595–628. DOI: 10.2307/2950710.
Policies, updated 2nd edition. London: Longman. ——. 2000. “Increasing Returns, Path Dependence, and
Leblebici, Huseyin, Gerald R. Salancik, Anne Copay, and the Study of Politics.” American Political Science Review
Tom King. 1991. “Institutional Change and the 94 (2): 251–67. DOI: 10.2307/2586011.
Transformation of Interorganizational Fields: An ——. 2004. Politics in Time: History, Institutions, and
Organizational History of the U.S. Radio Broadcasting Social Analysis. Princeton, NJ: Princeton University
Industry.” Administrative Science Quarterly 36 (3): 333– Press. DOI: 10.1515/9781400841080.
63. DOI: 10.2307/2393200. Prasad, Monica. 2012. The Land of Too Much: American
Liu, Sifan, and Joseph Parilla. 2020. “New Data Shows Abundance and the Paradox of Poverty. Cambridge, MA:
Small Businesses in Communities of Color Had Harvard University Press. DOI: 10.4159/
Unequal Access to Federal COVID-19 Relief.” harvard.9780674067813.
Research report, September 17. Washington, DC: Quinn, Sarah L. 2019. American Bonds: How Credit
Brookings Institution. https://www.brookings.edu/ Markets Shaped a Nation. Princeton, NJ: Princeton
articles/new-data-shows-small-businesses-in- University Press. DOI: 10.2307/j.ctvb938n3.
communities-of-color-had-unequal-access-to-federal- Rao, Hayagreeva, and Henrich R. Greve. 2018. “Disasters
covid-19-relief. and Community Resilience: Spanish Flu and the
Mills, Claire Kramer, and Jessica Battisto. 2020. “Double Formation of Retail Cooperatives in Norway.” Academy
Jeopardy: Covid-19’s Concentrated Health and Wealth of Management Journal 61 (1): 5–25. DOI: 10.5465/
Eﬀects in Black Communities.” Research report, August amj.2016.0054.
4. New York: Federal Reserve Bank of New York. Sabasteanski, Nika, Jeremy Brooks, and Thomas
Montgomery, David H. 2018. “CityLab’s Congressional Chandler. 2021. “Saving Lives and Livelihoods: The
Density Index.” Bloomberg CityLab home page, Paycheck Protection Program and Its Eﬃcacy.”
November 20. New York: Bloomberg LP. https:// EconomiA 22 (3): 278–90. DOI: 10.1016/j.
www.bloomberg.com/news/articles/2018-11-20/ econ.2021.11.004.
citylab-s-congressional-density-index. Sanchez-Moyano, Rocio. 2021. “Paycheck Protection
Moynihan, Donald P., and Joe Soss. 2014. “Policy Program Lending in the Twelfth Federal Reserve
Feedback and the Politics of Administration.” Public District.” Community Development Research Brief,
Administration Review 74 (3): 320–32. DOI: 10.1111/ March 3. San Francisco: Federal Reserve Bank of San
puar.12200. Francisco. https://www.frbsf.org/research-and-
North, Douglass C. 1990. Institutions, Institutional insights/publications/community-development-
Change, and Economic Performance. Cambridge, UK: research-briefs/2021/03/ppp-lending-12th-district.
Cambridge University Press. DOI: 10.1017/ Schattschneider, Elmer E. 1935. Politics, Pressures, and the
cbo9780511808678. Tariﬀ: A Study of Free Private Enterprise in Pressure
Oﬃce of Inspector General. 2020. “Flash Report: Small Politics, as Shown in the 1929–1930 Revision of the
Business Administration’s Implementation of the Tariﬀ. New York: Prentice-Hall.
Paycheck Protection Program Requirements.” Report Schneiberg, Marc. 2005. “Combining New
20-14, May 8. Washington, DC: US Small Business Institutionalisms: Explaining Institutional Change in
Administration. https://www.sba.gov/document/report- American Property Insurance.” Sociological Forum
20-14-ﬂash-report-small-business-administrations- 20 (1): 93–137. DOI: 10.1007/s11206-005-1899-y.
implementation-paycheck-protection-program- ——. 2007. “What’s on the Path? Path Dependence,
requirements. Organizational Diversity and the Problem of Institutional
Orren, Karen, and Stephen Skowronek. 1996. Change in the US Economy, 1900–1950.” Socio-Economic
“Institutions and Intercurrence: Theory Building in the Review 5 (1): 47–80. DOI: 10.1093/ser/mwl006.
Fullness of Time.” In Political Order, eds. Ian Shapiro Schneiberg, Marc, Mark Cassell, and Michael Schwan.
and Russell Hardin, 111–46. Nomos 38. New York: 2024. “Race, Credit and Possibilities for Inclusive
New York University Press. DOI: 10.18574/nyu/ Lending within American Banking: Lessons from
9780814788844.003.0009. Crisis.” Paper presented at the 2024 Society for the
Patashnik, Eric M., and Julian E. Zelizer. 2013. “The Advancement of Socio-Economics annual conference,
Struggle to Remake Politics: Liberal Reform and the University of Limerick, Ireland, June 29.
Limits of Policy Feedback in the Contemporary Schwan, Michael, Marc Schneiberg, and Mark K. Cassell.
American State.” Perspectives on Politics 11 (4): 1071–87. 2025. “Replication Data for: Business as Usual?
DOI: 10.1017/s1537592713002831. Conventional Politics, Critical Junctures, and Policy

September 2026 | Vol. 24/No. 3 1293

Feedback in the Paycheck Protection Program.” New Approach to Culture, Structure, and Process.
Harvard Dataverse. DOI: 10.7910/DVN/TIKAV2. New York: Oxford University Press. DOI: 10.1093/
Seawright, Jason, and John Gerring. 2008. “Case acprof:oso/9780199601936.001.0001.
Selection Techniques in Case Study Research. A Menu Thurston, Chloe N. 2015. “Policy Feedback in the
of Qualitative and Quantitative Options.” Political Public–Private Welfare State: Advocacy Groups and
Research Quarterly 61 (2): 294–308. DOI: 10.1177/ Access to Government Homeownership Programs,
1065912907313077. 1934–1954.” Studies in American Political Development
Sine, Wesley D., and Robert J. David. 2003. 29 (2): 250–67. DOI: 10.1017/s0898588x15000097.
“Environmental Jolts, Institutional Change, and the Tushman, Michael L., and Philip Anderson. 1986.
Creation of Entrepreneurial Opportunity in the US “Technological Discontinuities and Organizational
Electric Power Industry.” Research Policy 32 (2): 185– Environments.” Administrative Science Quarterly 31 (3):
207. DOI: 10.1016/s0048-7333(02)00096-3. 439–65. DOI: 10.2307/2392832.
Skocpol, Theda. 1996. Protecting Soldiers and Mothers: US Census Bureau. 2024. County Business Patterns Data.
The Political Origins of Social Policy in the United States. CPD database home page, updated July 22, 2024.
Cambridge, MA: Harvard University Press. DOI: Washington, DC: US Census Bureau. https://
10.2307/j.ctvjz81v6. www.census.gov/programs-surveys/cbp/data.html.
Skowronek, Stephen. 1982. Building a New American Vardi, Nathan. 2020. “71 Publicly Traded Companies
State: The Expansion of National Administrative Got Paycheck Protection Funding before Money Ran
Capacities, 1877–1920. Cambridge, UK: Cambridge Out.” Forbes, April 20. https://www.forbes.com/sites/
University Press. DOI: 10.1017/cbo9780511665080. nathanvardi/2020/04/20/seventy-one-publicly-traded-
Sorelle, Mallory E. 2023. “Privatizing Financial companies-got-paycheck-protection-funding-before-
Protection: Regulatory Feedback and the Politics of money-ran-out.
Financial Reform.” American Political Science Review 117 Wallner, Jennifer. 2008. “Legitimacy and Public Policy:
(3): 985–1003. DOI: 10.1017/s0003055422001071. Seeing beyond Eﬀectiveness, Eﬃciency, and
Sorensen, Andre. 2023. “Taking Critical Junctures Performance.” Policy Studies Journal 36 (3): 421–43.
Seriously: Theory and Method for Causal Analysis of DOI: 10.1111/j.1541-0072.2008.00275.x.
Rapid Institutional Change.” Planning Perspectives 38 Weaver, Kent. 2010. “Paths and Forks or Chutes and
(5): 929–47. DOI: 10.1080/02665433.2022. Ladders? Negative Feedbacks and Policy Regime
2137840. Change.” Journal of Public Policy 30 (2): 137–62. DOI:
Stinchcombe, Arthur L. 1990. Information and 10.1017/s0143814x10000061.
Organizations. Berkeley, CA: University of California White, Lawrence J. 1992. The S & L Debacle: Public Policy
Press. DOI: 10.1525/9780520909625. Lessons for Bank and Thrift Regulation. New York:
Stryker, Sheldon. 2000. “Identity Competition: Key to Oxford University Press.
Diﬀerential Social Movement Participation?” In Self, Wong, Karna, Paul M. Ong, and Silvia R. González.
Identity, and Social Movements, eds. Sheldon Stryker, 2020. “Systemic Racial Inequality and the COVID-19
Timothy J. Owens, and Robert W. White, 21–40. Homeowner Crisis.” Research report, August 27. Los
Minneapolis: University of Minnesota Press. Angeles: UCLA Center for Neighborhood Knowledge;
Thelen, Kathleen, and Sven Steinmo. 1992. “Historical Los Angeles: Ziman Center for Real Estate. https://
Institutionalism in Comparative Politics.” In escholarship.org/uc/item/2z99b26v.
Structuring Politics: Historical Institutionalism in Yaﬀe-Bellany, David. 2020. “‘The Big Guys Get Bailed
Comparative Analysis, eds. Sven Steinmo, Kathleen Out’: Restaurants Vie for Relief Funds.” New York
Thelen, and Frank Longstreth, 1–32. Cambridge: Times, April 20. https://www.nytimes.com/2020/04/
Cambridge University Press. DOI: 10.1017/ 20/business/shake-shack-returning-loan-ppp-
cbo9780511528125.002. coronavirus.html.
Thornton, Patricia H., William Ocasio, and Michael
Lounsbury. 2012. The Institutional Logics Perspective: A

1294 Perspectives on Politics

File and source

File: schwan-2025-business-as-usual-conventional-politics-critical-junctures-and-policy-feedback.pdf
Size: 3,005,107 bytes
SHA-256: 374d1cdea775125e78ea360b439ee952cafa1f832f33076edd14fdbb60f83de6
Our copy: schwan-2025-business-as-usual-conventional-politics-critical-junctures-and-policy-feedback.pdf
Original: www.cambridge.org