How do nonprofits respond to negative wealth shocks? The impact of the 2008 stock market collapse on hospitals

• Author: David Dranove,Craig Garthwaite,Christopher Ody
• DOI: http://doi.org/10.1111/1756-2171.12184
• Date: 01 May 2017
• Published date: 01 May 2017

RAND Journal of Economics

Vol.48, No. 2, Summer 2017

pp. 485–525

How do nonprofits respond to negative

wealth shocks? The impact of the 2008

stock market collapse on hospitals

David Dranove∗

Craig Garthwaite∗

and

Christopher Ody∗

The theory of cost shifting posits that nonprofit firms “share the pain” of negative financial

shocks with their stakeholders, for example, by raisingprices. We examine hownonprofit hospitals

responded to the sharp reductions in their assets caused by the 2008 stock market collapse. The

average hospital did not raise prices, but hospitals with substantial market power did cost shift

in this way. We find no evidence that hospitals reduced treatment costs. Hospitals eliminated but

left unchanged their offerings of profitable services. Taken together, our results provide mixed

evidence on whether nonprofits behave differently from for-profits.

1. Introduction

There is growing interest in the behavior of nonprofit firms in the United States that

reflects the increasing prevalence of this organizational form. From 1985 to 2004, nonprofit

organizations saw an inflation-adjusted increase in assets and annual revenue of 222% and 171%

percent respectively (IRS, 2008), and by 2010, nonprofits accounted for approximately 5.5% of

US Gross Domestic Product (GDP) (Roeger, Blackwood, and Pettijohn, 2012). This growing

source of economic activity is exempt from all taxes.In 2009, just the state and local property tax

component of these exemptions was estimated to be worthbetween $17 and $32 billion (Sherlock

and Gravelle, 2009). To place this in context, the tax breaks represent 4% to 8% of overall property

tax receipts for these governments (Kenyon and Langley, 2011).

In exchange for these valuable tax benefits, nonprofits are expectedto behave differently than

profit-maximizing organizations. Forexample, universities are expected to subsidize research and

∗Northwestern University; d-dranove@kellogg.northwestern.edu, c-garthwaite@kellogg.northwestern.edu, c-

ody@kellogg.northwestern.edu.

The authors wouldlike to thank Jen Brown and Jon Skinner for valuable comments and guidance, as well as Katie Johnson

from the National Research Corporation and Laura Johnson from the Kaiser Family Foundationfor providing useful data.

Stephanie Holmes, Matthew Schmitt, and TongtongShi provided valuable research assistance. All errors remain our own.

C2017, The RAND Corporation. 485

486 / THE RAND JOURNAL OF ECONOMICS

hospitals are expected to provide “community benefits” such as uncompensated care. However,

despite the growing economic importance of this sector and the corresponding costs of the tax

benefits to the government, prior research has failed to convincingly demonstrate consistent,

meaningful differences between nonprofit and for-profit behavior.1

As federal and state budgets grow tighter, it is important to understand whether nonprofits

are merely “for-profits in disguise,” that is, firms that are simply receiving inframarginal transfers

from tax payers. We examine this question by estimating the reaction of nonprofit firms to large

financial shocks. Firms of all types often experience negative lump-sum financial shocks. For

example, a firm may lose a lawsuit, experience an accident, or see its investment portfolio decline

in value. Economists usually presume that a profit-maximizing firm has previouslyfully exploited

all opportunities to reduce costs or raise revenues, so absent a fundamental rethinking of the firm’s

strategy, it would have to absorb the loss.2However, theory and evidence suggest that nonprofits

should respond differently (Dranove,1988; Brown et al., 2014). In particular, a nonprofit that was

previously sacrificing some profits in order to provide other benefits to the community should

mitigate the effects of a negative financial shock by“sharing the pain” with its stakeholders. This

could involve raising prices, reducing quality, or some other action that decreases the size of the

profit-sacrificing community benefits previously offered by the firm. When accomplished using

price changes, such “sharing of the pain” is traditionally described as cost shifting and should not

be observed in profit-maximizing firms.

Existing research has focused on price-based cost shifting as the only mechanism for “sharing

of the pain” and therefore may miss other important aspects of nonprofit behavior. In addition,

most previous work has not focused on whether firms may differin the amount of rents available

for redistribution, which may cause them to respond to shocks using different mechanisms. In this

article, we examine a range of potential channels for cost shifting in the hospital sector and note

that if a hospital cost shifts in some manner, we can distinguish true nonprofits from for-profits

in disguise. We also examine heterogeneity across hospitals in their reaction to a financial shock

based on measures of market power,which should be positively correlated with differences in the

rents that are available for redistribution.

The distinctive behavior of nonprofit organizations can have important policy implications,

as illustrated by the recent debate about the Affordable Care Act (ACA). The ACA extends

health insurance to millions of previously uninsured Americans, and it is expected that hospitals

will realize a substantial positive shock from the reduction in uncompensated care (Garthwaite,

Gross, and Notowidigdo, 2015). Many analysts and policy makers believe that, in a sort of

“reverse cost shift,” nonprofit hospitals will reduce their prices for privately insured patients,

thereby benefitting taxpayers whohave subsidized the expansion of insurance coverage under the

ACA. President Barack Obama offered this argument in defense of the ACA, as did the United

States Supreme Court when it upheld its constitutionality. In the majority opinion, Justice John

Roberts offered what is, essentially, a cost-shifting argument, stating that hospitals recoup the

losses from uncompensated care by, “pass(ing) on the cost to insurers through higher rates, and

insurers, in turn, pass on the cost to policy holders in the form of higher premiums.” It follows

that if hospitals pare their losses from uncompensated care, they will reduce their rates, and that

insurers will pass on the savings to policy holders. Though this argument rests on nonprofit cost

shifting, as we discuss below, there is at best mixed evidence to support it in the existing literature.

In this article, we examine the performance of nonprofit firms by exploiting the 2008 stock

market collapse (henceforth, we will refer to this as the “2008 collapse” or, when unambiguous,

the “collapse”) as an exogenous negativefinancial shock to nonprofit hospitals—the largest single

class of nonprofit organizations. Wefirst document that the 2008 collapse caused an economically

1Weisbrod(1988) summarizes the mixed evidence on nonprofit behavior.

2If firms funded investments through internal capital markets (Stein, 1997); the negative shock could curtail

investments. Borenstein and Farrell (2000) posit that profitable firms maysuffer from “X-inefficiency” and will attempt

to remove slack if profits fall.

C

The RAND Corporation 2017.

DRANOVE, GARTHWAITE AND ODY / 487

meaningful decline in hospital assets.3We also show that there was great variation in the asset

declines that was unrelated to local economic factors, and not driven by a change in the real estate

holdings of the hospital. Thus, the shock that we study is unlikely to be correlated with changing

local demand conditions for medical services and serves as an exogenous event that we can use

to assess how nonprofits respond to financial shocks.

At a broad level, wedocument that the assets of the average hospital remain lower for several

years, suggesting they cannot recoup losses from financial shocks as easily as the theory of cost

shifting would suggest. We then examine the specific channels through which hospitals may have

been able to recover some of their losses. Given that assets remain depressed for several years, it

should not be surprising that we find that the average nonprofit hospital suffering a largefinancial

loss did not raise prices, cut operating costs, or decrease their offering of charity care. We find

price changes for only a small subset of nonprofit hospitals with arguably the greatest market

power and rents available for redistribution.4We do find that the average nonprofit decreased its

offerings of relatively unprofitable services, some of which may provide substantial community

benefits. Overall, our results suggest that the average nonprofit adjusts its provision of service-

related community benefits but does not otherwise behave differentlythan for-profits, particularly

with respect to pricing. However, hospitals with more market power, and presumably more rents

to redistribute, adjust prices without changing service offerings.

The next section provides a theoretical framework for understanding how nonprofits may

respond to financial shocks. Section 3 describes our data. Section 4 shows the effect of the 2008

stock market collapse on hospital finances while Section 5 shows the effect of the collapse on

hospital behavior. Section 6 concludes the paper.

2. Differences in how firms respond to financial shocks

Healthcare providersmake up the largest category of nonprofit organizations. In 2010, nearly

40% of all nonprofit revenue was earned by hospitals and nonhospital primary care facilities

(Blackwood, Roeger, and Pettijohn, 2012). As a result, economists attempting to understand the

objective functions of nonprofits firms have often concentrated their research on hospitals and

have advanced several conceptually distinct theories of their behavior (Cutler, 2012). Weisbrod

(1988) posits that some nonprofits are actually “for-profits in disguise,” behaving in all ways

like for-profits but benefitting from preferential tax treatment. Pauly and Redisch (1973) observe

that physicians have de facto control of hospital production and hypothesize that the hospital is

therefore a “physicians’ cooperative,” managed to maximize physician profits.

A number of economists offer positive theories of nonprofits. Weisbrod (1988) argues that

nonprofits represent an alternative to government subsidies and government provision of public

goods. Thus, nonprofits might provide subsidized services to low-income individuals and offer

unprofitable services that are deemed to be socially valuable. Importantly, under this theory, we

would expect that the response of hospitals to a financial shock would not involve changes in

prices. Arrow (1963) suggests that nonprofit hospitals are an institutional response to consumer

uncertainty about quality. Hansmann (1980) elaborates on this idea by pointing out that nonprofit

hospitals are barred from compensating managers on the basis of financial returns. This limits

managerial incentives to shirk on hard-to-measure aspects of quality. Patients are presumably

aware of this and favor nonprofits, allowing these firms to outcompete for-profit rivals. A few

studies document differences on hard-to-measure dimensions of quality. For example, Svarstad

and Bond (1984) find that nonprofit nursing homes use fewer sedatives than for-profit facilities,

3Throughout this article, we use “assets” to describe the net assets of a hospital. This includes financial assets

(including endowment), physical assets (i.e., property,plant, equipment), and debts. We confirm that the financial shock

that we study is driven bychanges to a hospital’s financial assets.

4Market power maynot directly translate into profits, as these nonprofits may dissipate their rents on other utility-

enhancing activities.

C

The RAND Corporation 2017.