880013

research-article2019
UARXXX10.1177/1078087419880013Urban Affairs ReviewDonegan et al.

Original Manuscript
Urban Affairs Review
1­–26
Striking a Balance: A © The Author(s) 2019
Article reuse guidelines:
National Assessment of sagepub.com/journals-permissions
DOI: 10.1177/1078087419880013
https://doi.org/10.1177/1078087419880013
Economic Development journals.sagepub.com/home/uar

Incentives

Mary Donegan1, T. William Lester2 ,

and Nichola Lowe2

Abstract
The use of incentive packages has intensified as local governments compete
for new plants and corporate relocations and as private firms increasingly
demand a deal. While incentives promise jobs and tax revenue, scholars
and practitioners criticize their high cost and limited accountability.
Through a comparison of matched establishments, this article explores how
governmental incentive-granting strategy impacts incentive performance.
We examine the overall impact of incentives and whether incentives granted
to smaller firms perform better. Using economic development budget data,
we also assess the state’s overall approach to economic development to
determine which strategies are prioritized through funding. By showing
that incentivized firms fail to create more jobs than matched controls, our
analysis casts doubt on claims that “but for” incentives job creation would
not occur. Still, our findings suggest that states are smarter in their incentive
use when they strike a balance between recruiting industry and supporting
“homegrown” businesses and technology.

Keywords
incentives, mediating policies, employment, equity, economic development

1University of Connecticut, Storrs, CT, USA

2The University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Corresponding Author:
T. William Lester, Department of City and Regional Planning, The University of North
Carolina at Chapel Hill, 320 New East (MC #3140), Chapel Hill, NC 27599-3140, USA.
Email: [email protected]
2 Urban Affairs Review 00(0)

Incentive packages to private businesses have become the modus operandi

among state and local governments competing for large-scale corporate relo-
cations and new branch plants, bringing the promise of increased employ-
ment and tax revenue. But incentives are also widely criticized, with scholars
and practitioners equally frustrated by their limitations and costs. As private
businesses increasingly demand incentive payments to locate or remain
within a region, economic developers have responded with efforts to make
incentive payments more effective by using a mix of more sophisticated ana-
lytical tools, integrating incentives within broader sectoral or industrial tar-
geting strategies, or adding additional checks to curb excessive incentive use.
New data sources create the possibility to study some of these moderating
steps and explore their varying effect on incentive performance. In this arti-
cle, we present the results of a national, time-series, establishment-level
study of state and local economic development incentive deals, comparing
establishments that received an incentive package to a control group of simi-
lar, nonincentivized establishments. While incentives can be used to stimu-
late a range of economic outcomes, they are most frequently used with job
creation and job retention goals in mind. Our analysis focuses on employ-
ment effects of incentivized economic development deals by using three
national databases: the Subsidy Tracker from Good Jobs First (GJF), the
National Establishment Time-Series (NETS) database, and the State
Economic Development Expenditure Database from the Council for
Community and Economic Research (C2ER).
Combining these data sources, we seek to answer three interrelated ques-
tions: First, do firms that receive incentive details outperform nonincentiv-
ized firms? Second, does prioritizing incentives to establishments within a
region’s targeted or specialized industries enhance the efficacy of the incen-
tives? Finally, do states with more balanced economic development “portfo-
lios” (i.e., similar public investment in recruitment and entrepreneurship)
make more effective use of their incentive dollars with respect to employ-
ment ends, when compared to those states with less balanced portfolios?
Our findings will inform critical policy debates surrounding economic
development incentives. By conducting the first nationwide study with a
causal research design, we provide new empirical evidence showing that eco-
nomic development incentives, on average, fail to produce new employment
opportunities. By comparing incentivized establishments to a carefully
selected control group, we cast doubt on the biggest claim made by incentive
proponents that “but for” the incentive payment, job creation would not
occur. This simple but direct finding—that incentives do not create jobs—
should prove critical to policymakers. However, we also show how incen-
tives can be more effective by examining the disparate impacts by firm size.
Donegan et al. 3

Here, we find that incentives granted to smaller establishments have better

performance in terms of job creation compared to very large establishments,
which we find to have starkly negative employment effects. Finally, we show
that these large negative impacts for larger establishments are mitigated in
states that “balance” their economic development spending portfolios
between traditional business attraction activities and support for entrepre-
neurial development. Before turning to our analysis, we provide an overview
of the literature on incentive-granting and common strategies to enhance
incentive performance.

Motivation and Literature Review

Incentives have been widely criticized by economic development scholars,
and for good reason (Eisinger 1988; Hanley and Douglass 2014; Markusen
and Nesse 2007). Record-breaking mega-deals given to large multinationals
involving millions of dollars are particularly concerning. Not only do these
deals subsidize already wealthy corporations, they often fail to perform as
hoped by issuing municipal or state authorities. Much is known about these
high-stakes incentive “failures,” including questionable practices by some
footloose corporations that jump from location to location to extract addi-
tional rounds of public funding and support (Freyer 2017; LeRoy 2005). The
result is not only increased interjurisdictional competition, but at best a “zero-
sum” outcome for national economic development (Chirinko and Wilson
2008; Malizia 1994).
Many economic development practitioners would agree that, all things
equal, they would much rather retire tax incentive programs and related cor-
porate “giveaways” (Rubin 1988). These programs often draw down essen-
tial yet scarce sources of public funding that could be better spent on other
kinds of economic and community-enhancing activities, including public
education, vocational training, or income-stabilizing social services (Bartik
2014). Yet, empirical studies show that the practice of offering economic
development incentives is deeply entrenched, having survived decades of
legislative and lay challenges (Bartik 2017; Pollard 2015). In fact, state and
local incentive use in the United States has seen a notable uptick in the wake
of the 2008–2009 Great Recession (Osgood, Opp, and Bernotsky 2012;
Warner and Zheng 2013), reflecting increased desperation for additional
jobs. We need look no further than Wisconsin’s $3 billion incentive package
to lure Foxconn in 2017, or more recent frenzy around Amazon and Google’s
decisions to establish a second U.S. headquarters to see that communities
throughout the country and decision makers within them are willing to pull
out all the stops.
4 Urban Affairs Review 00(0)

But this rise in incentive use is also happening at a time when many U.S.
cities and communities are facing hardening budget constraints (Donald et al.
2014; Pollard 2015) and thus need to be smarter with their public investments
and avoid costly deals that fail to deliver on their promises. With this in mind,
we add our voices to a growing chorus of economic development scholars
and incentive reform advocates that are pushing to refocus debate over incen-
tives from a simple yes or no proposition to a more nuanced consideration of
how to improve incentive-granting implementation and practice—that is,
how to ensure state and local governments more effectively allocate and
manage incentives to support their intended economic development goals?
Scholarship on better incentive-granting points to three reinforcing strate-
gies, starting with actions taken to improve the inner workings of any incen-
tive deal. Research in this area recognizes the transactional nature of incentive
granting and thus opportunities for economic development practitioners to
better negotiate contractual terms and conditions when extending a publicly
funded incentive or subsidy to a private company (Weber 2007). At a basic
level, practitioners can reduce community risk by including clawbacks and
related performance mechanisms in their written contracts, thus ensuring
they have legal recourse to recapture public money from companies that fail
to meet or honor these terms (Ledebur and Woodward 1990; Markusen and
Nesse 2007; Sullivan 2002; Weber 2007). But equally, practitioners can also
secure encompassing, high-impact “community benefits” by linking incen-
tive offers to explicit requirements for living wages, job training guarantees,
and targeted local hiring provisions (Bartik 2005; Dewar 2013; Fitzgerald
2004; Lester 2011; Lowe and Morton 2008; Weber 2002; Wolf-Powers 2010).
Incentives can also produce stronger economic and employment returns
when channeled toward smaller-sized establishments (Bartik 2018). That
said, the general literature on economic development incentives has been
rather agnostic to questions of firm size and thus has overlooked the potential
for differential employment outcomes when incentives are given to smaller
rather than larger firms. Important exceptions are studies that have focused
on the use of geographically bound incentives to revitalize blighted or under-
developed urban neighborhoods and as part of state urban enterprise zone
programs (see Wilder and Rubin 1996, for a review). While these studies find
considerable variation in incentive performance at this microgeographic
scale, a key take-away is that job growth is greatest and most consistent when
incentives are used to support smaller-sized establishments, including expan-
sion of incumbent businesses already located within the zone. Interestingly,
employment outcomes also intensify in enterprise zones that layer on multi-
ple incentives and target programmatic support to smaller firms (Harrison
and Glasmeler 1997; Wilder and Rubin 1996).
Donegan et al. 5

At one level, this outcome reflects the fact that smaller, especially younger
firms have greater propensity to add more jobs relative to larger, more mature
establishments—a distinction noted in numerous studies of small business growth
trends in the United States (Birch 1979; Davis, Haltiwanger, and Schuh 1996). But
there is a moderating element at play here as well, insofar as economic develop-
ment practitioners are more likely to gain the upper hand when negotiating incen-
tives with smaller, resource-constrained firms (Weber 2007). In addition, as they
engage these firms, state and local economic developers can strengthen funding
impacts by coupling incentives with other forms of small-business-oriented assis-
tance (Wilder and Rubin 1996). As this implies, incentive outcomes are not just
affected by the actions of a skillful dealmaker—they can be furthered magnified
through interventions that simultaneously resolve constraints that certain busi-
nesses face because of their smaller size or related organizational limits.
But there are additional gains to be leveraged when incentive granting is
treated as one element of a much larger and more sophisticated strategy of
industrial recruitment (Lowe 2014). The focus is not the incentive, per se, but
the larger industrial recruitment strategy—more specifically, the selective use
of recruitment to build-out existing industrial strengths and the promotion of
nonincentive advantages in order to attract, anchor, and retain firms within
specialized sectors (Feser and Luger 2003; Goetz, Deller, and Harris 2009;
Lowe and Freyer 2015). Strategies of targeted recruitment can also help
reduce demand for cash-equivalent incentives by shifting the focus toward
other regional advantages that offer greater value to firms as well as their host
communities, including access to quality infrastructure, transferable work-
force skills, and specialized industry support services or institutions (Bartik
2005, 2018; Lester, Lowe, and Freyer 2014; Schunk and Woodward 2003). In
this respect, targeting reinforces what most scholars and practitioners already
know: incentives are rarely a crucial factor in firm location. They may tip the
balance but only once other factors are in play (Bartik 2005; LeRoy 1997).
These two approaches—better incentive dealmaking and targeted recruit-
ment with selective incentive use—have received considerable attention
from economic development scholars in recent years. However, most incen-
tivized deals entail funding tools that are themselves embedded within an
even larger state economic development system (Lowe and Feldman 2018).
Few states focus their entire economic development effort on incentive grant-
ing for business attraction or retention purposes; rather, most have other eco-
nomic development programs in place, including robust funding streams to
boost entrepreneurship and small business development; assist business and
industry expansion through market development, including international
trade; and even strengthen local economies through recurring investments in
infrastructure and community development (Hanley and Douglass 2014).
6 Urban Affairs Review 00(0)

Scholars tend to study each of these strategy areas in isolation, at times

casting the others as competing policy priorities (Ross and Friedman 1990).
Yet practitioners, for their part, take a more integrated approach. To illustrate
this point, a senior fellow for research and policy at North Carolina’s Rural
Economic Development Center recently used the analogy of a three-legged
stool to describe economic development practice in that state (Gray 2017). In
public comments before a university economic development working group,
he noted that one leg represents strategies that focus on retaining incumbent
firms and supporting their continued growth, the second leg is industrial
recruitment or attraction of new businesses to the state, and the third focuses
on creating and nurturing homegrown entrepreneurial firms, technology, and
talent. By default, if one leg is removed the stool would fall over because the
weight of the economy would no longer be equally distributed.
But there is a stabilizing element in play with implications for increasing
the effectiveness of incentive granting. The three elements of the economic
development stool help strike a balance between exogenous and endogenous
development strategies—that is, between recruiting business to a region ver-
sus supporting homegrown development. Each leg adds to the soundness of
the state’s overall economic development program, reducing the risk that any
one leg will dominate and risk destabilizing the whole. Extending the anal-
ogy further, it could be that this mix of multiple budget priorities creates a
form of “productive friction” (Stark 2011), in turn helping to moderate exces-
sive or inappropriate incentive use. In addition, this balanced portfolio could
encourage new, creative forms of cooperation and knowledge sharing among
practitioners, including the possibility for traditionally distinct areas of eco-
nomic development practices to be better sequenced or intentionally copro-
duced and combined (Lowe and Feldman 2018; Lowe and Wolf-Powers
2018). With this in mind, and including the aforementioned moderating steps
of industry targeting and enhanced dealmaking with smaller firms, we turn to
new national data on state incentive deals to explore questions about incen-
tive deal performance and efficacy.

Data and Method

Data Sources and Sampling
We use three sources of data to build a national, time-series, establishment-
level database: the Subsidy Tracker from GJF, the NETS database, and the
State Economic Development Expenditure Database from the C2ER. First,
we use GJF to identify deals issued by either a state or local (county or
municipal) government. While there is no single public source of incentive
Donegan et al. 7

deals in the United States, the GJF data set is the most complete and compre-
hensive listing of deals and includes information on firm name, location
(state, with some records including city, county, and street address), and deal
year, and incentives offered, with some deals also including information on
the number of jobs promised.1 We downloaded a snapshot in May 2016,
when the database contained 500,593 subsidy deals. To create our sample, we
exclude federal and Washington, DC, deals (184,688 and 342, respectively),
leaving us with 315,563 deals. Since a key purpose of this article is to assess
the relationship between state budgeting patterns and deal efficacy, we do not
consider federal and DC deals to be a part of our study population. We next
remove deals issued either before 1995 or after 2010 (28,902 and 132,071,
respectively), as those deals will not allow for pre- or posttest employment
trends. We exclude 199 deals GJF codes as “megadeals,” as those rare deals
worth $75 million or more would limit the external validity of findings.
Given that a primary interest is post deal employment changes, we exclude
deals not promising to either retain or increase employment by 100 employ-
ees or more, bringing the sample down to 13,324 deals. While this may sound
like a restrictive sample limitation, deals that announce at least 100 expected
jobs make up the majority of state and local government incentive dollars.
When we breakout our analysis by firm size, we provide additional discus-
sion of this limitation. Finally, to aid in matching these deals to their records
in other data sources, we removed 2,048 deals that did not include city or
county locations. This brought the sample down to 11,276 deals.
Next, we gather data on our GJF sample deals from the NETS database, a
national, private, longitudinal database comprising 25 snapshots of Dun &
Bradstreet (D&B) data (1990–2014). Each establishment in the NETS data-
base has a D-U-N-S Number (the unique D&B identifier for each establish-
ment) and allows us to gather data on firm name, location (state, county/
municipality, and address), subsidiary status (i.e., whether the establishment
operates as a subsidiary of another establishments), SIC code, employment,
and sales. For a discussion of the use of NETS data to assess the efficacy of
subsidies, see Lester, Lowe, and Freyer (2014). We rely on a text-matching
algorithm that minimizes differences in company names to locate GJF records
in the NETS database.2 We were able to find information in the NETS data-
base for 2,604 subsidy deals. These deals correspond to 3,321 records in the
NETS database. The numerical discrepancy between deals and records reflect
the D-U-N-S filing practice of a single physical establishment choosing to
report to D&B with multiple D-U-N-S numbers.3
Our third data source, C2ER, uses consistent funding categories to aggre-
gate state budget data for economic development activity across 15 func-
tional areas. We choose the data collected from the 2007 economic
8 Urban Affairs Review 00(0)

development survey, rather than a more recent year (2013–2017), as it

approximates the midpoint of the period in which we observe subsidy deals
and thus more accurately reflects the policy orientation of states within the
study period. We focus on two distinct encompassing categories of economic
development funding, drawn C2ER’s 15 functional areas.4 Within the first
group, which we call entrepreneurial, we combine four functional areas:
business assistance, which includes funding for small business development
centers, counseling for small businesses and technical assistance in support
of small business procurement; entrepreneurial development, which includes
all forms of state assistance to help new business ventures develop and thrive;
minority business development, which supports underserved needs for minor-
ity-owned businesses, including women-owned establishments; technology
transfer, which supports local innovation and technology development
through process and product improvement, research and development and
manufacturing modernization and extension. In our second group, which we
call recruitment, we include four other categories that are primarily oriented
toward business recruitment, attraction, and retention. The first and most
obvious two in this group are the strategic business attraction fund, which
captures state resources allocated for business incentive grants and the
domestic recruitment category which includes funding for advertising and
marketing for business recruitment and site location preparation. Within this
second group, we also include two additional categories from the C2ER data-
base, workforce preparation and development and special industry assis-
tance. The C2ER survey is designed to only capture workforce funding
within a narrower economic development framework and more specifically,
funding for customized vocational training that can be included within a busi-
ness recruitment and retention package. While some of this funding could
potentially flow to homegrown firms, there is a strong likelihood that it used
as a form of incentive in support of customized training for new employees at
recruited establishments. It is also important to note that this category does
not represent all state spending on workforce development and vocational
education, most of which will undoubtedly be captured in other, noneco-
nomic development areas of the overall state budget. With special industry
assistance, our fourth functional area in this group, the link is a bit more tenu-
ous. Still, this category captures the share of the economic development bud-
get that is going to the development of strategic industrial sectors—sectors
that the state has deemed important to expand, including through strategically
targeted business recruitment activities.
We then construct a sample-limiting dummy variable called “balanced,”
which is coded at the state level when their budgeted economic development
spending is within one quartile of the median (i.e., between the 25th and
Donegan et al. 9

Table 1. Control Establishment per Treatment Establishment.

Number of Controls Establishments

1 167
2 136
3 126
4 129
5 2,615
Total 3,173

75th percentile) for both the entrepreneurial development and recruitment

funding groups.

Research Design and Identifying Control Establishments

For each of the 3,321 incentive-receiving treatment establishments, we aim to
identify five control establishments. Control criteria include perfect matches on
four treatment establishment characteristics: three-digit SIC code, state, subsid-
iary status (e.g., whether the establishment is its own headquarters), and employ-
ment category.5 We also require that the first year of recorded NETS data be
within three years. We gather these data for the treatment record’s year of analy-
sis. For treatment records with NETS data recorded in the year of the GJF sub-
sidy (i.e., the establishment was born before or concurrent to the subsidy deal),
the year of analysis is the year of the subsidy deal. For treatment records that do
not have NETS data recorded until up to three years after the subsidy deal, the
year of analysis is the establishment’s first year in the NETS data.
To identify control records, we rely on a matching-without-replacement
program in SAS so that each control establishment is used just once. We ran
this program five times and were unable to find any control establishments
for 147 of our 3,321 treatment records; these treatment records were dropped
from the analysis. For 2,615 of our treatment establishments we have five
controls, and the remaining treatment records had between one and four con-
trol matches (see Table 1). This brought our final GJF sample to 2,486 firms
linked to 3,173 establishment records, drawn from 35 states (see Table 2).
This means that we ultimately pull NETS records for 17,581 establishments:
3,173 treatment establishments and 14,408 control establishments.
We next gathered data from the NETS database for both the treatment and
control records on yearly in-state sales and employment, our two dependent
variables. For deals with multiple D-U-N-S, we create a single record with
combined data.
10 Urban Affairs Review 00(0)

Table 2. Incentive-Receiving Sample Firms, by State.

State Number of Deals

Alabama 20
Arizona 80
California 130
Colorado 14
Delaware 16
Florida 226
Georgia 24
Iowa 88
Indiana 86
Kansas 8
Kentucky 195
Louisiana 63
Massachusetts 57
Maryland 36
Maine 18
Michigan 180
Minnesota 1
Missouri 21
Mississippi 3
Montana 5
North Carolina 66
Nebraska 60
New 19
Hampshire
New Jersey 73
New Mexico 3
Nevada 5
New York 722
Ohio 128
South 29
Carolina
South Dakota 4
Texas 38
Utah 12
Virginia 30
Wisconsin 6
West Virginia 20
Total 2,486
Donegan et al. 11

Identification Strategy
Our basic research design consists of testing the impact of an economic
development incentive given to a particular establishment on subsequent
changes in employment and sales at that establishment relative to a set of
control establishments. To test this basic question, we conduct a simple dif-
ference-in-differences analysis using the panel data set described above. The
panel consists of observations of establishment’s characteristics for busi-
nesses which received an incentive and the same information for each of the
control establishments matched to each incentivized establishment. The
panel uses data on employment and sales from 1990 to 2014. Equation (1)
summarizes our basic empirical strategy:
ln Eit = a + β1 INCENTIVEit + δi + γ ct + eit (1)

The dependent variable is the natural log of employment or sales at the

establishment level. Logging the dependent variable helps smooth out scale
differences in employment changes at small and large firms and enables us to
interpret the value of β1 as a semielasticity, or the percentage change in
employment/sales resulting from an incentive. The key independent variable
INCENTIVEit is coded 0 for each year before the incentive is granted and 1
for the year of the event and each subsequent year.
The model includes an establishment fixed effect δi , which controls for
idiosyncratic differences across establishments that do not vary over time.
The final fixed effect γ ct is a time-fixed effect that is specific to each “control
set,” which is a dummy variable that represents each unique set of an indi-
vidual treatment observation and the three to five control establishments
matched to it. In this way, each “control set” has its own unique time trend
that not only controls for common macroeconomic trends over time but also
allows for the identification of β1 to be limited to comparisons between the
treated case and its own controls. Thus, we interpret β1 as a difference-in-
differences estimator in that it is created by comparing employment (or sales)
changes in establishments that received an incentive (i.e., in which the incen-
tive dummy variable changes from zero to one) with those that never received
an incentive (i.e., matched control group of establishments). So, a statistically
significant positive coefficient means that, on average, treatment establish-
ments created more jobs after receiving an incentive compared to the set of
matched control establishments. We do not include other controls for indus-
try, firm size, or geography since these are the core variables used to select
the matched pairs for each treatment case.
While the identification strategy presented above and summarized by
equation (1) is used to assess the overall effectiveness of all incentives and to
12 Urban Affairs Review 00(0)

answer the simple question of, “Do incentives work?” we move beyond this
question and ask, How do incentives work? First, we seek to understand
whether incentives granted in industries in which the state granting the incen-
tive has an established industry concentration—measured as having an
employment location quotient above 1.1 in the same three-digit industry as
the incentivized (treatment) establishment—perform better than incentives in
nonspecialized sectors. While we do not specifically observe whether the
industry in question is a strategic target industry of each state, we use the
location quotient as a proxy for industry targeting.
Next, we break down our analysis by firm size to see whether incentives
granted to small- and medium-sized establishments are more efficient than
those granted to large and very large companies. As discussed above, there
are several reasons to suggest that incentives offered to smaller companies
may produce better outcomes in terms of job growth, including the fact that
dealmakers may have more leverage compared to bargaining with very large
firms that may be more likely to relocate after the grant is made. Smaller
firms may also be more likely to be in the earlier stages of a product life
cycle, therefore offering more of an upside to local policymakers.
Finally, we seek to understand whether incentives perform better in states
that have the most balanced economic development portfolios, especially
when they treat industry recruitment and entrepreneurship as equally impor-
tant strategy areas. As described above, we use data from C2ER to divide our
national sample of states into two groups that balance economic development
expenditures between traditional business attraction activities versus invest-
ments in local technologies and businesses. We again split our sample and
rerun equation (1) separately for the sample of most balanced states and all
other states. Table 3 provides the share of each state’s total economic devel-
opment expenditures that are allocated to entrepreneurial activities versus
business attraction and recruitment. It also captures the portion of expendi-
tures that are passed from the state to the local jurisdiction and as a result are
difficult to classify as either entrepreneurial or recruitment-focused.

Results
When we examine the overall effectiveness of state incentive grants on firm-
level performance, we find little evidence that they generate new jobs or other
direct economic benefits to the states that employ them. As described above,
our difference-in-differences research design identifies the impact of an incen-
tive treatment on establishments that received a grant relative to their set of
matched control establishments. Since the control establishments were located
in the same state and operate in the same industry (three-digit SIC), we can
interpret the results summarized in Table 4 as what would have happened to
Donegan et al. 13

Table 3. State Share of Expenditure Allocated to Entrepreneurship, Recruitment,

and Local Economic Development.
Entrepreneurial Balanced Portfolioa: Decentralizedb as
and Small Strategic Business Entrepreneurial Local Economic
State Business Support Attraction and and Industrial Development
Abbreviation (%) Recruitment (%) Recruitment Resources (%)

OH 36 45 1 13
MT 42 34 1 24
GA 30 22 1 33
MD 38 27 1 33
SC 22 34 1 41
WI 20 36 1 44
KY 9 29 1 58
VA 24 16 1 60
NM 21 77 0 0
LA 59 41 0 0
DE 5 69 0 0
NJ 79 17 0 2
NV 1 75 0 4
WV 3 1 0 4
NH 18 2 0 4
UT 85 7 0 8
NY 33 13 0 8
IN 9 76 0 9
TX 59 28 0 13
MS 66 18 0 16
KS 64 14 0 17
IA 53 29 0 18
MO 13 67 0 18
FL 38 11 0 21
AZ 0 70 0 25
ME 44 25 0 28
MA 8 58 0 31
MI 16 47 0 37
NC 7 31 0 38
AL 9 52 0 40
SD 59 0 0 41
NE 24 11 0 65
CA 1 19 0 80
MN 4 12 0 83
CO 2 14 0 84

Source. C2ER 2007 State Economic Development Expenditure Survey.

a. Balanced Portfolio states indicate those where their budgeted economic development spending is within
one quartile of the median (i.e., between the 25th and 75th percentile) for both the entrepreneurial
development and recruitment funding groups.
b. Decentralized refers to the share of the state expenditure for economic development that is allocated to
local jurisdiction to support city- and county-level economic development activities.
14 Urban Affairs Review 00(0)

Table 4. Impact of State-Level Incentives on Employment and Sales, by Industry

Specialization and Establishment Size.

(1) (2)

Specification Ln(Emp) Ln(Sales)

Full sample −0.0372*** −0.0137
N = 253,462, R2 = .923 (0.00775) (0.00838)
Industry targeting
State LQ >1.1 in 1990 −0.0639*** −0.0498***
N = 77,916, R2 = .917 (0.0155) (0.0167)
State LQ >1.1 in 2010 −0.108*** −0.0840***
N = 84,958, R2 = .916 (0.0149) (0.0160)
Establishment size
Less than 50 employees 0.0759*** 0.0892***
N = 47,889, R2 = .931 (0.0122) (0.0141)
50–250 employees 0.0731*** 0.0864***
N = 90,895, R2 = .867 (0.0117) (0.0129)
250–1,000 employees −0.0497*** −0.0142
N = 76,102, R2 = .879 (0.0147) (0.0158)
1,000 or more employees −0.588*** −0.529***
N = 32,785, R2 = .893 (0.0320) (0.0334)

Note. All specifications include establishment fixed effects as well as a control-set-specific

time-fixed effect. Robust standard errors listed below each estimate (of beta from equation
(1)). LQ = location quotient.
Significance levels are indicated as follows: *p < .1. **p < .05. ***p < .01.

employment “but for” the incentive. Thus, our analysis is one of the few
empirical studies to evaluate incentives based on a plausible counterfactual of
economic trends common to similar nonincentivized establishments.
The overall effect of incentives in our full national sample is listed in the
first row of Table 4. Note that each coefficient listed in Table 4 represents a
different regression and we do not report the point estimates on any fixed
effects for the sake of brevity. The estimated impact on employment is
slightly negative and statistically significant. The point estimate of −.037
implies that establishments that received an incentive experienced employ-
ment growth that was 3.7% slower than nonincentivized establishments. The
impact on sales (column 2) growth is close to zero and not significant. There
are several reasons why this negative affect is identified. First, states could be
offering incentives to establishments that were already declining before the
incentive, thus hoping unsuccessfully that growth will occur. Alternatively,
the incentivized establishment may simply grow slower than promised or
Donegan et al. 15

slower than the control establishments. Finally, it could be that the establish-
ment goes out of business entirely after receiving an incentive, leading
employment and sales to go to zero. While these outcomes are of unique
interest to policymakers, we cannot distinguish between these competing
interpretations without data and research design.
While understanding that the overall impact of incentives is critical for
policymakers, we also explore whether incentives are more or less effective
in different industrial contexts and for certain types of establishments. First,
we limit our sample to analyze only those incentive deals made where the
state offering the incentive has a preexisting specialization in the establish-
ment’s industry. Specifically, we define an “existing specialization” as hav-
ing an employment location quotient above 1.1 in the same three-digit NAICS
(North American Industry Classification System) code as the incentivized
establishment. This sample limitation focuses the analysis on deals that may
have the potential to complement a state’s previously targeted industrial sec-
tors or competitive clusters. While we cannot directly observe targeting
activity, focusing on states with a high location quotient can test whether
incentives are more effective if they occur alongside existing economic
strengths. The results indicate otherwise, however, as the point estimates for
both logged employment and sales are negative and statistically significant.
The results are similar when we define state location quotients in either 1990
or 2010. There are a number of factors that could explain this counterintuitive
result, including the presence of legacy industries that are declining in total
state employment but still indicate high levels of regional concentration.
While we recognize that location quotients are a blunt analytical tool, else-
where, we do show that strategic industry targeting and related institution
building at the state level results in robust employment growth from busi-
nesses recruitment in those targeted industries (see Lester, Lowe, and Freyer
2014). This suggests an opportunity to collect more detailed data on industry
targeting across all states, including identifying the specific industries and
industry-supporting institutions that each state promotes and develops to sus-
tain and increase employment growth.
Next, we segment our analysis by the size of the establishment receiving
the incentive. Here, we use four different employment size categories ranging
from small (less than 50 employees) to very large (over 1,000). Interestingly,
our findings are quite different across establishment size categories. The esti-
mated employment and sales impacts are positive and significant for both the
small- and medium-sized categories (50–250). This implies that incentives
seem to be more effective for smaller enterprises. While policymakers could
interpret this finding as a signal to shift incentive activities to smaller firms,
it could also be the case the states and local governments that offer incentives
16 Urban Affairs Review 00(0)

to smaller firms may have more information about the future growth pros-
pects of incentivized firms. Alternatively, these smaller firms that seek out
incentives may be in a growth cycle—and may have already been intending
to expand production or enter new markets. Job growth among small firms is
concentrated in a tiny minority of firms (Henrekson and Johansson 2010), but
researchers have struggled to separate a priori those firms that will grow from
those that will not—meaning that our control firms may not be true compari-
sons and thus a potential threat to study validity. While we cannot detect the
precise reason why incentives granted to smaller firms are more successful,
when compared to the large negative coefficients for the largest category,
incentive use for smaller firms appears to be less risky.
The large negative estimates observed for very large establishments appear
to be driving the overall negative result. There are many possible interpreta-
tions as to why this figure is so negative for very large establishments. It
could be that large firms that seek incentives are moving activity from one
state to another in a process of consolidation that is proximate in time with an
overall downward trend in business activity. Alternatively, large establish-
ments may simply be more likely to play the incentive game and are less
likely to be experiencing a positive growth cycle or inventing new goods and
services. Whether these large incentives are retention or recruitment deals
(which we cannot directly observe in the GJF database), our findings could
also reflect the fact that large establishments have greater bargaining power
and thus are able to extract incentives, even though they are not producing or
promising higher-than-average employment gains. Regardless of the reasons
why, our findings should be a caution to policymakers interested in chasing
very large establishments.
Leaving the story here might suggest the lesson for states is to avoid
incentivizing large establishments and instead put incentive dollars behind
smaller establishment deals. However, there might be mediating effects in
play that complicate this simplified small- versus large-firm narrative. For
this reason, we split the sample again based on the orientation of each state’s
overall economic development strategy. As described above, we use the
C2ER database to construct a variable to represent states that have the most
balanced budget portfolios. We use the term balanced portfolio to categorize
a subset of states that divide their economic development budget more
equally across entrepreneurial support programs and more traditional recruit-
ment activities, including incentive-based recruitment efforts and firm-spe-
cific customized training. In other words, these states strike a better balance
between endogenous and exogenous strategies—recognizing these not just as
competing interests but also potentially complementary approaches to eco-
nomic development.
Donegan et al. 17

Table 5. Impact of State-Level Incentives on Employment, by Budget Type and

Establishment Size.
Less than 50 1,000 or more
Employees 50–250 Employees 250–1,000 Employees Employees

(1) (2) (3) (4) (5) (6) (7) (8)

lnemp 0.0619* 0.0780*** 0.0683** 0.0739*** −0.0696** −0.0445*** −0.0943 −0.675***

(0.0344) (0.0131) (0.0338) (0.0123) (0.0302) (0.0168) (0.0962) (0.0338)
Observations 6,481 41,408 14,258 76,637 15,819 60,283 4,413 28,372
R2 .927 .932 .832 .874 .875 .879 .847 .900
Balanced Yes No Yes No Yes No Yes No

Note. All specifications include establishment fixed effects as well as a control-set-specific time-fixed effect.
Robust standard errors listed below each estimate (of beta from equation (1)).
Significance levels are indicated as follows: *p < .1. **p < .05. ***p < .01.

After taking balanced portfolios into account, we find a rather striking

divergence in results among these largest establishments (see Table 5). In
states that have a more balanced economic development expenditure portfo-
lio, establishments that receive incentives no longer stand out for their
consistent negative employment effects when compared to their non-incen-
tive-receiving controls—there is no statistically significant difference between
the two. Those results do not hold in other, less balanced states, where incen-
tivized large firms underperform their control group peers.
Our initial interpretation is that a more balanced economic development
budget could mean that states recognize that employment gains can come
from a wide range of economic development activities and enterprise types—
some endogenous, some exogenous. This knowledge might help shape deci-
sion-making with respect to incentive granting, especially when large
establishments are involved. One possibility may be that states with espe-
cially balanced economic development budgets employ more careful selec-
tion criteria when evaluating potential large firm deals, and as a result, pick
only the more promising deals among multiple options to incentivize. These
states might also be better at negotiating with larger establishments, pushing
for better contractual terms that are tied to better labor market outcomes (and
related to that they might also be better at enforcing these contracts, so they
meet their intended employment targets). But equally, we recognize that there
could be other intervening factors that affect not only incentive granting, but
also the decision by a state to take a more balanced approach to budget allo-
cation. In this regard, we are cautious not to oversell a causal argument here,
but rather seek to draw attention to this strong correlation in order to encour-
age additional research on the performance of public economic development
18 Urban Affairs Review 00(0)

investments. Furthermore, the fact that six states in our analysis reallocate
more than 50% of their state-approved economic development expenditures
to local jurisdictions reinforces the need for future scholarship to more deeply
explore the factors that shape economic development decision-making.

Summary and Broader Implications

Economic development incentives are widely used, yet remain highly contro-
versial. Policymakers and political leaders often claim incentives are needed
to generate job growth and private investment. However, others present a
compelling counterclaim, noting that scarce public tax dollars would be bet-
ter spent on infrastructure, education, or other state programs. Businesses
undoubtedly gain from these indirect institutional investments as well, but so
do other members of the wider community. Still, attempts to use this logic to
push an outright ban on incentive use have failed to gain traction at the fed-
eral, state, or local level, creating the need for timely research on a national
scale to understand moderating steps that can be taken to ensure economic
development incentives generate anticipated gains in employment and job
quality.
This article offers some insights from analysis of a national study of incen-
tive granting that integrates three innovative data sets—the GJF Subsidy
Tracker, the NETS database, and the C2ER State Economic Development
Expenditure Database. When combined, these data sources allow us to com-
pare employment growth at companies that receive a government-funded
incentive to a carefully selected group of nonincentivized control establish-
ments. Overall, we find that incentivized establishments lead to lower
employment gains than their nonincentivized peers. However, we also
observe important sources of variation when we factor in differences in firm
size. As our results indicate, incentives offered to firms that employ 1,000 or
more workers produce far fewer jobs compared to their nonincentivized
counterparts. By contrast, incentivized firms that are smaller-sized generate
positive and significant employment effects when compared to nonincentiv-
ized control establishments.
At first glance, these results could support an argument for states and
localities to redirect their incentive dollars to small- and medium-sized
firms—those that, on average, face greater resource constraints relative to
their larger counterparts and thus may be better poised to benefit from public
assistance. But this conclusion, in isolation, overlooks the role that incentives
can also play in boosting employment effects of larger firms. Recognizing
this, we take the additional step of factoring in the breakdown of state eco-
nomic development budgets, which allows us to examine whether incentives
Donegan et al. 19

offered to larger firms perform better in states that strike a balance in their
economic development spending. In particular, we look at the relative share
of the state economic development budget going to industrial recruitment
(where incentive granting is most common) versus spending on small busi-
ness assistance, including entrepreneurial-supporting activities and support
programs.
We find that there is a potential tempering effect in states that more equally
distribute funding streams so as to simultaneously support major economic
development functions, business recruitment, and homegrown business
development. While both academics and practitioners alike agree that incen-
tives should be scaled back if not eliminated entirely, doing so in practice is
difficult if not impossible as long as neighboring states continue to incentiv-
ize. Yet, by nesting incentives within a balanced economic development port-
folio, states continue to compete while removing the most pernicious negative
effects that incentives to larger firms can lead to. In more balanced states,
larger firms that receive a public incentive display similar employment
effects to their nonincentivized peers. In other words, the labor market pen-
alty that is initially associated with large-firm incentive deals is greatly
diminished.
Additional research is needed to unpack the underlying mechanisms that
are in play. Admittedly, there is some potential that this cross-state variation
simply reflects a spurious economic effect, meaning that states with already
strong economic conditions can better support budgetary balance—and
where those same underlying economic strengths are also reflected in stron-
ger labor market outcomes, irrespective of the incentive. However, we believe
our findings speak to a more nuanced interpretation, namely, the need to situ-
ate incentive granting within a broader economic development framework—
one that recognizes the potential for incentive performance to be further
enhanced through the act of balancing multiple economic development
objectives.
Viewed through that lens, it could also be the case that incentive deals are
more highly scrutinized by state agencies that are juggling other competing
uses of public spending. And by extension, economic development practitio-
ners in that same context might be less inclined to extend incentive offers to
job-shedding companies. Furthermore, state agencies might also feel more
emboldened to make strong demands of large firms, requiring those that
receive an incentive to commit to retaining or adding jobs, rather than using
public money to subsidize labor-displacing technologies or restructuring.6
These possibilities speak to an expanded research opportunity to explore the
dynamic relationship between budget allocations in economic development
and economic performance of incentivized firms. In their study of
20 Urban Affairs Review 00(0)

county- and city-level economic development, Zhang, Warner, and Homsy

(2017) found increased support for homegrown business development in
communities that also had broad public representation in local economic
development decision-making. This suggests an opening for further research
to more critically examine the combination of actors and organizations that
participate in state-level economic development planning. Equally, our find-
ings also align well with recent calls to action in support of greater incentive
transparency (Markusen and Glasmeier 2008)—most recently, visible in
coordinated campaigns that call into question excessive incentive offers by
cities hoping to lure Amazon’s second headquarters (Donegan 2019).
Beyond this, we also recognize the need for additional inquiry on the ques-
tion of firm size, and especially to consider whether incentives offered to
smaller firms might be further enhanced through better economic develop-
ment oversight. Research on smaller firms may want to lower the threshold for
jobs growth, as our 100-employee cutoff may have limited our sample to
“gazelles.” In addition, while our findings suggest that incentives generate
greater employment effects when channeled to smaller-sized firms, they say
nothing yet about the quality of those additional jobs. In his seminal 1994
article titled “The Small Firm Myth,” Bennett Harrison (1994) raised concerns
about diminished job quality with shrinking firm size. After 25 years later, that
concern still resonates (see Berger 2013; Osterman and Weaver 2014). A
recent paper even notes that wage inequality is heavily influenced by variation
in firm-specific pay and with interfirm variance greatest among small- and
medium-sized firms (Song et al. 2015). In other words, the gap between busi-
nesses that pay high versus low wages is greatest among smaller firms.
This clearly has implications for the types of specific performance criteria
that are tied to economic development incentives. At a fundamental level, eco-
nomic development practitioners should develop mediating strategies that use
incentives as a tool for improving wage standards among smaller firms. This
includes layering on performance standards that more intentionally tie those
incentives to living wages and enhanced working conditions. But, more
broadly, this also creates the potential for worker advocacy campaigns—
including community groups involved in the Fight for $15—to engage in the
micropolitics of incentive accountability, critically assess contemporary
incentive use, and in the process, push greater demands that public funding for
economic development be used to support ethical and equity commitments.

Acknowledgments
We are grateful to Caroline Hanley, Greg LeRoy, Ned Hill, Mereb Hagos, Daniel
Hartley, Allison Hewitt, Yasukui Motoyama, James Umbanhowar and Rachel Weber
for their insightful comments, as well as those shared by others in attendance at the
Donegan et al. 21

ACSP conference in Denver in 2017. We also wish to thank three anonymous review-
ers for their helpful suggestions. Research assistance was provided by Griffen Rice.
Authorship is listed alphabetically.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research,
authorship, and/or publication of this article: This project received funding support
from the Ewing Marion Kauffman Foundation.

ORCID iD
T. William Lester https://orcid.org/0000-0003-2132-8665

Notes
1. Good Jobs First (GJF), the developer of the searchable Subsidy Tracker data-
base, is a nonpartisan, nonprofit national policy resource center. Incentives in
the Subsidy Tracker include those issued for property tax abatements, corpo-
rate income tax credits, sales tax exemptions, megadeal, and workforce training,
among others. The Tracker codes each deal with just one of these subsidy types,
meaning we cannot include information about the full subsidy package in our
analysis.
2. The program relies on the COMPGED function in SAS, which calculates the
Levenshtein distance between two string variables (Staum 2007). The algo-
rithm calculates the distance between the GJF firm name and each firm name
the National Establishment Time-Series (NETS) database, sorts the numerical
distances from smallest to largest, and returns the 20 pairs with the smallest
numerical distances (see Donegan 2016, for full discussion). We use the name of
the deal-receiving firm from GJF and the establishment’s firm name in the NETS
database. Matches must meet four criteria. First, the GJF firm name and the
NETS firm name must be recognizably similar. While an exact match in names
would be ideal, firms occasionally change names or spellings; databases (in our
case, both GJF and NETS) may shorten or misspell a firm’s name. Second, the
NETS record must be in the city or county recorded in the GJF database either
at the time of the subsidy deal or in the following three years. Third, the subsidy
had to be the first deal the establishment had received—and not a repetition of
an earlier deal. In other words if, through the matching process, we discover
that multiple GJF deals match to the same NETS record, we retain only the first
match and discard the rest. Fourth, if multiple NETS establishments in the city
or county meet the first three criteria, no match is made and the deal is discarded
22 Urban Affairs Review 00(0)

as we cannot accurately determine which establishment received the deal. Two

researchers reviewed algorithm results to ensure records marked as matches
met all four criteria.
3. This often happens when multiple business activities occur at a location—for
example, a single physical location with both an engineering department and
a manufacturing plant may file with two D-U-N-S numbers linked to two SIC
codes. In still other instances, establishment addresses overlapped at some
point in the study period before splitting into multiple locations or merging into
one; we concluded that this was a single establishment for study purposes and
included multiple records.
4. The 15 functional areas contained in the C2ER state economic development
expenditure database include the following: (1) business finance, (2) strategic
business attraction fund, (3) business assistance, (4) international trade and
investment, (5) domestic recruitment, (6) workforce preparation and develop-
ment, (7) technology transfer, (8) entrepreneurial development, (9) minority
business development, (10) community assistance, (11) tourism and film, (12)
special industry assistance, (13) program support, (14) administration, and (15)
other program areas.
5. We use four employment categories. The first includes establishments with 1–19
employees, the second includes those with 20–99 employees, the third with 100–
499 employees, and the fourth with 500 and more employees.
6. We thank Ned Hill for raising this possibility at American Collegiate Schools of
Planning 2017 Annual Conference.

References
Bartik, Timothy J. 2005. “Solving the Problems of Economic Development
Incentives.” Growth and Change 36 (2): 139–66.
Bartik, Timothy J. 2014. From Preschool to Prosperity: The Economic Payoff to
Early Childhood Education. Kalamazoo: W.E. Upjohn Institute for Employment
Research.
Bartik, Timothy J. 2017. A New Panel Database on Business Incentives for Economic
Development Offered by State and Local Governments in the United States.
Prepared for the Pew Charitable Trusts. Kalamazoo: W.E. Upjohn Institute for
Employment Research.
Bartik, Timothy J. 2018. “Who Benefits from Economic Development Incentives?
How Incentive Effects on Local Incomes and the Income Distribution Vary with
Different Assumptions About Incentive Policy and the Local Economy.” Upjohn
Institute Technical Report No. 18-034, W.E. Upjohn Institute for Employment
Research, Kalamazoo.
Berger, Suzanne. 2013. Making in America: From Innovation to Market. Cambridge:
MIT Press.
Birch, David L. 1979. The Job Generation Process: Final Report to Economic
Development Administration. Cambridge: MIT Press.
Donegan et al. 23

Chirinko, Robert S., and Daniel J. Wilson. 2008. “State Investment Tax Incentives: A
Zero-Sum Game?” Journal of Public Economics 92 (12): 2362–84.
Davis, Steven J., John C. Haltiwanger, and Scott Schuh. 1996. Job Creation and
Destruction. Cambridge: MIT Press.
Dewar, Margaret. 2013. “Paying Employers to Hire Local Workers in Distressed
Places.” Economic Development Quarterly 27 (4): 284–300.
Donald, Betsy, Amy Glasmeier, Mia Gray, and Linda Lobao. 2014. “Austerity in
the City: Economic Crisis and Urban Service Decline?” Cambridge Journal of
Regions, Economy and Society 7 (1): 3–15.
Donegan, Mary. 2016. “Inside the Triangle: Advancing Research on Entrepreneurship
and Firm Locations.” Unpublished PhD dissertation, Univ. of North Carolina at
Chapel Hill.
Donegan, Mary. 2019. “Post–Amazon HQ2 Incentive Reform: We Can Have Good
Deals—But Not Without Transparency and Meaningful Public Participation.”
Metropolitics, May 21. Accessed September 21, 2019. https://www.metropoli-
tiques.eu/Post-Amazon-HQ2-Incentive-Reform-We-can-have-good-deals-but-
not-without.html.
Eisinger, Peter K. 1988. The Rise of the Entrepreneurial State: State and Local Economic
Development Policy in the United States. Madison: Univ. of Wisconsin Press.
Feser, Edward J., and Michael I. Luger. 2003. “Cluster Analysis as a Mode of Inquiry:
Its Use in Science and Technology Policymaking in North Carolina.” European
Planning Studies 11 (1): 11–24.
Fitzgerald, Joan. 2004. “Moving the Workforce Intermediary Agenda Forward.”
Economic Development Quarterly 18 (1): 3–9.
Freyer, Allan M. 2017. “Incentives in a New Light: Enhancing the Effectiveness
of State Economic Development Through Policy Embeddedness.” PhD thesis,
Univ. of North Carolina at Chapel Hill.
Goetz, Stephan J., Steven C. Deller, and Thomas R. Harris. 2009. Targeting Regional
Economic Development. New York: Routledge.
Gray, Jason. 2017. “Rural North Carolina: Connections and Threats.” Presentation,
UNC Working Group on Economic Development, Chapel Hill, NC, October 26.
Hanley, Caroline, and Michael T. Douglass. 2014. “High Road, Low Road, or Off
Road? Economic Development Strategies in the American States.” Economic
Development Quarterly 28 (3): 220–29.
Harrison, Bennett. 1994. “The Small Firms Myth.” California Management Review
36 (3): 142–57.
Harrison, Bennett, and Amy K. Glasmeler. 1997. “Response: Why Business Alone
Won’t Redevelop the Inner City: A Friendly Critique of Michael Porter’s Approach
to Urban Revitalization.” Economic Development Quarterly 11 (1): 28–38.
Henrekson, Magnus, and Dan Johansson. 2010. “Gazelles as Job Creators: A Survey
and Interpretation of the Evidence.” Small Business Economics 35 (2): 227–44.
Ledebur, Larry C., and Douglas Woodward. 1990. “Adding a Stick to the Carrot:
Location Incentives with Clawbacks, Recisions, and Recalibrations.” Economic
Development Quarterly 4 (3): 221–37.
24 Urban Affairs Review 00(0)

LeRoy, Greg. 1997. No More Candy Store: States and Cities Making Job Subsidies
Accountable. Washington, DC: Good Jobs First.
LeRoy, Greg. 2005. The Great American Jobs Scam: Corporate Tax Dodging and the
Myth of Job Creation. San Francisco: Berrett-Koehler.
Lester, T. William. 2011. “The Impact of Living Wage Laws on Urban Economic
Development Patterns and the Local Business Climate: Evidence from California
Cities.” Economic Development Quarterly 25 (3): 237–54.
Lester, T. William, Nichola Lowe, and Allan Freyer. 2014. “Mediating Incentive Use:
A Time-Series Assessment of Economic Development Deals in North Carolina.”
Economic Development Quarterly 28 (2): 132–46.
Lowe, Nichola. 2014. “Beyond the Deal: Using Industrial Recruitment as a Strategic
Tool for Manufacturing Development.” Economic Development Quarterly 28
(4): 287–99.
Lowe, Nichola, and Maryann Feldman. 2018. “Breaking the Waves: Innovation at
the Intersections of Economic Development Policy.” Economic Development
Quarterly 323 (3): 183–94.
Lowe, Nichola, and Allan Freyer. 2015. “A Moving Target: Rethinking Industrial
Recruitment in an Era of Growing Economic Uncertainty.” Environment and
Planning C: Government and Policy 33 (5): 1284–1300.
Lowe, Nichola, and Brian J. Morton. 2008. “Developing Standards: The Role of
Community Benefits Agreements in Enhancing Job Quality.” Community
Development 39 (2): 23–35.
Lowe, Nichola, and Laura Wolf-Powers. 2018. “Who Works in a Working Region?
Inclusive Innovation in the New Manufacturing Economy.” Regional Studies 52
(6): 828–39.
Malizia, Emil E. 1994. “A Redefinition of Economic Development.” Economic
Development Review 12 (2): 83–84.
Markusen, Ann, and Amy Glasmeier. 2008. “Overhauling and Revitalizing Federal
Economic Development Programs.” Economic Development Quarterly 22 (2):
83–91.
Markusen, Ann, and Katherine Nesse. 2007. “Institutional and Political Determinants
of Incentive Competition.” In Reining in the Competition for Capital, edited
by Ann Markusen, 1–42. Kalamazoo: W.E. Upjohn Institute for Employment
Research.
Osgood, Jeffery L., Jr., Susan M. Opp, and R. Lorraine Bernotsky. 2012. “Yesterday’s
Gains Versus Today’s Realties: Lessons from 10 Years of Economic Development
Practice.” Economic Development Quarterly 26 (4): 334–50.
Osterman, Paul, and Andrew Weaver. 2014. “Skills and Skill Gaps in Manufacturing.”
In Production in the Innovation Economy, edited by Richard Locke and Rachel
L. Wellhausen, 17–50. Cambridge: MIT Press.
Pollard, Randall B. 2015. “‘Was the Deal Worth It?’: The Dilemma of States with
Ineffective Economic Incentives Programs.” Hastings Business Law Journal 11
(1): 1–28.
Donegan et al. 25

Ross, Doug, and Robert E. Friedman. 1990. “The Emerging Third Wave: New
Economic Development Strategies.” Entrepreneurial Economy Review 90 (1):
3–11.
Rubin, Herbert J. 1988. “Shoot Anything that Flies; Claim Anything that Falls:
Conversations with Economic Development Practitioners.” Economic
Development Quarterly 2 (3): 236–51.
Schunk, Donald L., and Douglas P. Woodward. 2003. “The Case of BMW in South
Carolina.” In Financing Economic Development in the Twenty-First Century,
edited by Sammis B. White, Richard D. Bingham, and Edward W. Hill, 145–69.
Armonk: Routledge.
Song, Jae, David J. Price, Fatih Guvenen, Nicholas Bloom, and Till von Wachter.
2015. “Firming Up Inequality.” Working Paper No. 21199, National Bureau of
Economic Research, Cambridge.
Stark, David C. 2011. The Sense of Dissonance: Accounts of Worth in Economic Life.
Princeton: Princeton Univ. Press.
Staum, Paulette. 2007. “Fuzzy Matching Using the COMPGED Function.”
Proceedings of the 2007 Northeastern SAS Users Group (November 2007), Paul
Waldron Consulting, West Nyack, NY.
Sullivan, Daniel Monroe. 2002. “Local Governments as Risk Takers and Risk
Reducers: An Examination of Business Subsidies and Subsidy Controls.”
Economic Development Quarterly 16 (2): 115–26.
Warner, Mildred E., and Lingwen Zheng. 2013. “Business Incentive Adoption in the
Recession.” Economic Development Quarterly 27 (2): 90–101.
Weber, Rachel. 2002. “Do Better Contracts Make Better Economic Development
Incentives?” Journal of the American Planning Association 68 (1): 43–55.
Weber, Rachel. 2007. “Which Local Governments Have the Most Bargaining
Leverage?” In Reining in the Competition for Capital, edited by Ann Markusen,
141–59. Kalamazoo: W.E. Upjohn Institute for Employment Research.
Wilder, Margaret G., and Barry M. Rubin. 1996. “Rhetoric Versus Reality: A Review
of Studies on State Enterprise Zone Programs.” Journal of the American Planning
Association 62 (4): 473–91.
Wolf-Powers, Laura. 2010. “Community Benefits Agreements and Local Government:
A Review of Recent Evidence.” Journal of the American Planning Association
76 (2): 141–59.
Zhang, Xue, Mildred E. Warner, and George C. Homsy. 2017. “Environment,
Equity, and Economic Development Goals: Understanding Differences in Local
Economic Development Strategies.” Economic Development Quarterly 31 (3):
196–209.

Author Biographies
Mary Donegan is assistant professor in Residence in Urban and Community Studies
at the University of Connecticut. She earned her MRP and PhD from the Department
of City and Regional Planning at The University of North Carolina at Chapel Hill.
26 Urban Affairs Review 00(0)

T. William Lester is associate professor in the Department of City and Regional

Planning at The University of North Carolina at Chapel Hill. His research focuses on
quantitative analysis of economic development policy tools and their impact on equi-
table development at the neighborhood and regional scales. He earned a PhD from UC
Berkeley.
Nichola Lowe is a professor in the Department of City and Regional Planning at The
University of North Carolina at Chapel Hill. Her research primarily focuses on the
institutional arrangements that lead to more inclusive forms of economic development
and specifically, the role that practitioners can play in aligning growth, equity and
innovation goals through their engagement with private business. She earned her PhD
in economic development and planning from MIT in 2003.