«Johan Sulaeman National University of Singapore Sheridan Titman University of Texas at Austin April 10, 2015 Abstract We ﬁnd that a corrupt local ...»
Struggling firms in corrupt cities∗
Christopher A. Parsons
University of California at San Diego
National University of Singapore
University of Texas at Austin
April 10, 2015
We ﬁnd that a corrupt local environment ampliﬁes the eﬀects of ﬁnancial distress.
Following regional spikes in ﬁnancial misconduct, credit becomes both more expensive
and harder to obtain for nearby borrowers – even those not implicated themselves. This is particularly harmful for cash-constrained ﬁrms, which cut investment more sharply and lay oﬀ more workers during industry downturns. Moreover, we ﬁnd that local waves of ﬁnancial misconduct are a risk factor for bankruptcy.
Keywords: ﬁnancial misconduct, corporate failure, bankruptcy, loan spread, security issuance, trust We especially thank Jonathan Karpoﬀ, Allison Koester, Scott Lee, and Gerald Martin for making their ∗ data on ﬁnancial misconduct available, and Justin Murﬁn for sharing his data on covenant strictness. All errors are ours.
Send correspondence to Christopher Parsons, Rady School of Management, University of California,
San Diego, Otterson Hall, Room 3S137 9500 Gilman Drive #0553 La Jolla, CA 92093-0553; telephone:
858-534-8782; fax: 858-534-0745. E-mail: firstname.lastname@example.org.
1 Introduction Dallas and Minneapolis are fairly comparable cities in the central United States, each with a fast growing population, vibrant business center, and reputation for cultivating business- friendly climates. Over 100 public ﬁrms currently call the Minneapolis region home, placing it 5th in ratio to population among large U.S. cities, with Dallas-Fort Worth (7th ) close behind.
These cities, however, tend to be very diﬀerent along one very important dimension. From 1970-2010, ﬁrms in the Dallas metropolitan area were over twice as likely to be prosecuted for ﬁnancial misconduct as those headquartered in and around Minneapolis (2.21% versus 0.93%), a disparity peaking in the 1998-2002 time period, during which Dallas produced more cases of ﬁnancial misconduct (14) than were produced in Minneapolis (10) over the entire four decades.
In this paper, we ask whether proximity to a rash of ﬁnancial misconduct – like Dallas in the late 1990s – creates a unique set of challenges for resident companies. Similar to foreign investors pulling out of countries following civil unrest or nationalization (Schneider and Frey (1985)), we hypothesize that city-level waves of ﬁnancial misconduct may temporarily taint an entire region, reducing ﬁnanciers’ willingness to provide capital to local ﬁrms. Among those most reliant on external ﬁnance, such frictions may extend to inﬂuence real business decisions such as investment and employment, and in extreme cases, even survival.
That banks (and other providers of ﬁnance) would be wary of ﬁnancial misconduct is not a new idea. For one, when a ﬁrm’s ﬁnancial statements are not viewed as credible, it is diﬃcult to estimate its ability to repay, or should it default, the value of its assets upon liquidation. Either will increase the cost of borrowing.1 A second consideration is that a history of ﬁnancial misconduct reﬂects poorly on the ‘trustworthiness’ of the ﬁrm’s executives.2 A manager that does not respect explicit rules – think about misrepresenting 1 See Francis, LaFond, Olsson, and Schipper (2004) and Hribar and Jenkins (2004) for evidence that concerns regarding earnings quality and restatements, respectively, increase the ﬁrm’s cost of capital.
2 The theoretical contracting literature distinguishes between the concepts of trust and reputation; see Cabral (2012). Trust is typically invoked in moral hazard problems (involving hidden actions), whereas the concept of reputation arises in contexts involving adverse selection (involving hidden types). In a pure case earnings or trading on inside information – is unlikely to respect implicit agreements, which are often featured in debt ﬁnancing.3 At best, a lack of trust imposes additional costs (e.g., writing more complete contracts or enforcing them), and at worst, may prevent ﬁnancing altogether.
What is new, and the subject of this paper, is whether a ﬁrm’s ﬁnancial misconduct imposes negative local spillovers on nearby ﬁrms not (perhaps yet) implicated themselves.
In a recent paper, Parsons, Sulaeman, and Titman (2014) ﬁnd that ﬁnancial misconduct occurs in distinct, regionally-concentrated waves, involving ﬁrms in a wide variety of sectors.
Regardless of the mechanism responsible for this geographic clustering – common environmental factors in the region, or peer eﬀects between neighboring managers – these results suggest that the determinants of ﬁnancial misconduct have a regional component, making it possible to predict a ﬁrm’s likelihood of committing ﬁnancial misconduct by observing the behavior of its regional peers. Moreover, as we show in this paper, evidence of ﬁnancial fraud in a region has a material eﬀect on the credit worthiness of its resident ﬁrms.
In our analysis, we will pay special attention to ﬁrms with already poor prospects. The reason is that although a corrupt environment may encourage misbehavior even by healthy companies (e.g., manipulating stock prices to maximize the value of option compensation), misrepresentation by struggling ﬁrms clearly imposes higher risk to creditors. Accordingly, following a rash of ﬁnancial misconduct, we expect the weakest ﬁrms in the region to bear the brunt of lenders’ skepticism.4 of the former, “trusting” the agent is equivalent to understanding her payoﬀs given diﬀerent actions, and thereby being able to predict her actions, e.g., trigger strategies in a repeated prisoner’s dilemma. When both adverse selection and moral hazard are at work, agents can choose actions to build their reputations. This paper makes no distinction between pure trust (moral hazard) games and those where trust and reputation are mixed. What matters is that there exists regional and temporal variation in their determinants, e.g., social punishment being an especially eﬀective deterrent in some cities compared to others.
3 See for example, Eaton and Gersovitz (1981), Gale and Hellwig (1985), Sharpe (1990), and Bharath, Dahiya, Sanders, and Srinivasan (2011) for discussions of implicit contracts in lending relationships.
4 It may also be the case that ﬁnancially distressed ﬁrms have higher incentives to commit ﬁnancial misconduct, and that being surrounded by similarly misbehaving peers lowers the cost of doing so (e.g., via reduced social stigma or sharing of fraud-related information or techniques). Although the literature is ambiguous regarding the relationship between a ﬁrm’s ﬁnancial condition and its likelihood of engaging in ﬁnancial misconduct (see Dechow, Ge, and Schrand (2010) for a review of this literature), a negative relation would further reinforce the hypothesis proposed above.
2 The source for our ﬁnancial misconduct data is the hand-collected sample by Karpoﬀ, Koester, Lee, and Martin (2013), hereafter KKLM. We describe the sample, and discuss the types of events identiﬁed as ﬁnancial misconduct in Section 2.
We begin our analysis by exploring the link between regional rates of ﬁnancial misconduct and the costs of syndicated bank loans in Section 3. When a region experiences a wave of ﬁnancial misconduct, bank credit becomes temporarily more expensive for nearby ﬁrms. For regions in the top quartile of ﬁnancial misconduct, resident ﬁrms pay, on average, about 14 basis points higher in interest costs. Consistent with our hypothesis that ﬁrms with poor prospects are most sensitive to spikes in local corruption, the eﬀect is over twice as strong for borrowers with either low or no credit ratings (17 basis points, t = 10.01) versus those more creditworthy (8 basis points, t=2.08). Note also that because our regressions include city ﬁxed eﬀects (in addition to standard ﬁrm-speciﬁc controls), these estimates are identiﬁed by time-series variation in ﬁnancial misconduct (i.e., local waves) within each area.
A second indication of lenders’ collective wariness is that loan covenants become stricter following a rash of local ﬁnancial misconduct. As with the cost of credit, the covenant eﬀects are stronger for ﬁrms with either low or no credit ratings, for which information and monitoring concerns are likely the most severe.
We also observe eﬀects in the quantity of credit supplied, with ﬁrms in cities that have higher fraud rates borrowing less frequently and/or lower amounts. Though the results diﬀer across speciﬁcations, ﬁrms above the median rate of ﬁnancial misconduct are about 10-15% less likely to issue signiﬁcant amounts of debt (deﬁned as greater than 5% of total assets).
These reductions are not oﬀset by additional equity issuance; consequently, regional waves of ﬁnancial misconduct are associated with less overall inﬂow of external capital.
In summary, we ﬁnd the following: (1) ﬁrms are less likely to obtain external ﬁnancing following regional spikes in ﬁnancial misconduct; (2) for those that are able to obtain ﬁnancing, ﬁrms in cities that have higher fraud rates are less likely to issue debt; and (3) for those that are able to obtain loans, they face higher rates and/or stricter covenants. It is worth reiterating that when predicting these credit market eﬀects, none of the ﬁrms we consider were prosecuted for ﬁnancial misconduct themselves; rather, these are due to regional spillovers from ﬁrms (revealed to have) engaged in ﬁnancial misconduct to those that have not.
The next part of the paper explores whether these elevated credit market frictions have real eﬀects, inﬂuencing either the ﬁrm’s investment (Section 4.1) or employment (4.2) policies. As before, we focus our attention on ﬁrms with poor prospects, but now measure these using industry-wide declines in performance.5 Also, in an attempt to focus on ﬁrms most reliant on external capital markets, for which any type of credit market friction should have the largest real eﬀects, we identify “ﬁnancially constrained” ﬁrms as those in the top decile of the ‘Kaplan-Zingales’ index, as described by Lamont, Polk, and Sa´-Requejo (2001). For a these ﬁrms, we ﬁnd dramatically elevated sensitivities to industry downturns. In response to a decline of at least 10% in industry-level investment, constrained ﬁrms headquartered in cities experiencing the top quartile of ﬁnancial misconduct cut investment by almost twice as much compared to constrained ﬁrms in less corrupt locations.
We observe similar patterns when analyzing employment changes in response to negative industry shocks. Again focusing on ﬁrms in the top KZ-index decile, we ﬁnd that constrained ﬁrms in areas ranking in the top quartile for ﬁnancial misconduct reduce employment by about 50% more in downturns. Interestingly, we observe consistent eﬀects using a diﬀerent data set on city-level employment, collected by the Bureau of Labor Statistics. Following regional spikes in ﬁnancial misconduct, we observe slower employment growth at the city level, even accounting for prior trends in the dependent variable. In addition to providing robustness to our earlier analysis, these latter results explicitly relate spikes in regional misconduct to local labor markets, a feat not possible using ﬁrm-level employment data from COMPUSTAT.
In our ﬁnal analysis (Section 5), we consider the ultimate sequelae of local waves of 5 Traditional models of investment include ﬁrm-level variables like Tobin’s q and cash ﬂows. This is less desirable in our setting, as the ﬁrm’s own stock price might already capitalize the negative eﬀects of being headquartered in a region with high rates of corporate fraud.
4 ﬁnancial misconduct – bankruptcies in the region. A cursory look at the data indicates a strong cross-city correlation between rates of ﬁnancial misconduct and rates of corporate failure, as shown in Figure 1. The vertical-axis graphs the average bankruptcy rate for ﬁrms headquartered in each of the 20 largest cities in the U.S. from 1970-2010, and the horizontal axis plots the average rates of ﬁnancial misconduct for each city over the same horizon.
The positive cross-city correlation (0.65) is clearly apparent, indicating that cities with high average rates of ﬁnancial misconduct are strongly associated with high average failure rates.
Similar patterns emerge in the time-series. That is, it is not simply that cities such as Dallas have higher than average bankruptcy rates, but rather that ﬁrms headquartered in Dallas are particularly likely to fail following spikes in ﬁnancial misconduct, such as 1998Moving from the 25th to 75th percentile in the regional rate of ﬁnancial misconduct, or an increase of about 1.5 percentage points, increases by about 10% (0.15-0.20 percentage points) the likelihood that a resident ﬁrm will declare bankruptcy over the following year.
Thus, in addition to ﬁrm-speciﬁc variables typically employed in failure models, these results indicate that regional information on ﬁnancial misconduct provides additional explanatory power.
While previous research has studied the costs, both direct (e.g., SEC penalties) and indirect (e.g., increased cost of capital), of committing ﬁnancial misconduct (Karpoﬀ, Lee, and Martin (2008)), the eﬀect of any spillovers to nearby ﬁrms is less understood. In this respect, the most similar paper to ours is Giannetti and Wang (2014), which ﬁnds that following accounting scandals, local investors – such as those in Houston after the Enron debacle – lose faith in the stock market, reducing their exposure to equity in general. Complementing their results for retail investors, our work suggests that even highly sophisticated ﬁnancial institutions can be put oﬀ by ﬁnancial misconduct; moreover, as we show, ﬁrm policies and even survival are aﬀected by these perceptions.