Corrupting Innovation

Transcription

1 Corrupting Innovation Ted (Taek-yul) Kim August 18, 2013 Abstract How do firms respond when operating within a corrupt environment? To answer this question, this paper analyzes the interaction between firm local area corruption, corporate lobbying, the cost of capital and firm value. Using a sample of almost 3,000 firms over a 13 year time period, we provide evidence that firms operating in corrupt political environments spend greater dollars on lobbying which helps secure safer cash flows. Consistent with a securing safer returns argument, these lobbying-intensive firms have less variable cash flows and stock returns as well as lower costs of equity and lower weighted average costs of capital. However, these benefits come at a cost: lobbying firms are associated with less spending on traditional investment such as research and development and capital expenditures. Overall, our evidence suggests that the misallocation of resources by firms operating in corrupt environments results in lower firm value, as measured by Tobin s q.

2 How do you become wealthy today? By getting government assistance. If you look at where MODERN wealth comes from, it almost always comes from political influence, which enables you to get benefits at the expense of the public at large. Now that IS a zero sum game. Milton Friedman I. Introduction Corruption, defined as abuse of public power for private benefit, is a global problem and has been fiercely debated in the literature. Early studies suggest that corruption is efficient, a means by which wheels can be greased. More recent literature suggests that corruption is more akin to sand in the wheels, and find that that corruption is inefficient. It has been associated with weak governments (Shleifer and Visnny, 1993) and slower macro-economic growth (Mauro, 1995; Aidt, 2009). Much of the literature about corruption is focused on industry or national level outcomes (Mauro, 1995; Shleifer and Vishny, 1993; Rose-Ackerman, 1999) and ignores the individual firm level response. In this paper, we explore how political corruption interferes with the resource allocation process at the firm level. In the absence of corruption, firms should allocate capital to value-maximizing projects. However, when faced with a corrupt political environment, managers are faced with two choices: 1) continue to compete as before while corrupt rival firms gain an advantage or 2) commit resources to lobbying in order to gain influence and extract rents. All else equal, firms operating in corrupt environments face the potential of suboptimal investment since political corruption may subvert resources away from some firms and towards those that are favored. Yu and Yu (2010) find that firms will lobby to enhance future revenues through favorable regulation and/or business conditions. In order to level the playing field in a corrupt political environment, firms have an even greater incentive to lobby to provide safe cash flows. While lobbying may prove beneficial, Parsley and Yang (2010) show that 2

3 excess returns of lobbying intensive firms diminish as time goes by, suggesting that lobbying has the largest effects only in the short term. Indeed, the greater spending on lobbying may simply substitute for those spent on innovation such as research and development costs ( R&D ) and capital expenditures. Using a sample of 2,913 firms over a 13 year time period with corruption convictions data from the Federal Justice department, we find that firms systematically spend more resources on lobbying when they operate in a corrupt political environment. We also document that these same firms are able to provide safer cash flows and exhibit operating and stock returns which are less variable and are perceived by investors to be less risky as evidenced by a lower cost of equity and lower weighted average cost of capital. However, the safer cash flows which we observe come at a cost. We find that when firms face corrupt environments, traditional investment is subverted to gain political influence. We provide evidence that when firms spend dollars on lobbying, they shift resources away from traditional investments which support innovation, and this ultimately harms firm value. That is, when firms operate in corrupt political environments, they spend more resources on lobbying, diverting precious resources away from capital spending and R&D. This misallocation of resources results in a loss of shareholder wealth. A firm that engages in lobbying typically spends thirty four million dollars less, on average, in investments than its non-lobbying counterpart. Also, when firms are located in highly corrupt areas, additional lobbying becomes a cost to the firm and the firm value measured by Tobin's Q is, on average, 5.8% lower for firms that lobby compared to their non-lobbying counterpart.. Our paper adds to the growing literature on corruption and provides a rationale of why firms lobby. Our paper is the first, to our knowledge that provides the rationale of this loss of 3

4 value: dollars spent on lobbying represent a misallocation of precious resources which is harmful to the firm. We begin by reviewing the current literature on corruption and lobbying and develop our hypotheses in Section II. In Section III, we describe our sample and data while Section IV reports our results. We conclude in Section V. II. Literature review and hypothesis development Corruption The literature has debated the impact of corruption on economic and financial outcomes. Some argue that corruption is efficient and can lead to economic growth and prosperity. Egger and Winner (2005) find that corruption stimulates foreign direct investment. Leff (1964) and Leys (1970) suggest that corruption can actually circumvent regulations, by introducing policies that would otherwise been unavailable. This line of argument is based on the notion of speed money or bribes, which may allow some to navigate through bureaucratic red tape more quickly. Egger and Winner (2005) find that corruption stimulates foreign direct investment, lending support to the greased wheel hypothesis of corruption. However, it should be noted that the greased wheel outcome is only available to those willing to expend the required resources to circumvent regulation and this consequently puts firms unwilling to make this expenditure at a relative disadvantage. Much of the recent literature suggests that corruption is inefficient and acts more like sand (not grease) in the wheels. In his seminal study, Mauro (1995) finds that corruption is negatively associated with investment/gdp and per capital GDP growth. Others (Knack and Keefer, 1995; Shleifer and Vishny, 1993) 4

5 corroborate these findings. Corruption is also associated with lower foreign investment (Globerman and Shapiro, 2003; Wei, 2000). Tanzi and Davoodi (1997) show that even when corruption is associated with higher public spending, the spending occurs for low-productivity projects at the expense of value enhancing ones. Galang (2011) stresses that corruption has a corrosive impact on global economy. Corruption not only makes it difficult for developing countries to attract foreign investment, but also adds uncertainty and risk to the effort of companies that enter corrupt environments (Galang, 2011; Shleifer and Vishny, 1993). The aforementioned studies tend to focus on the impact of corruption on macroeconomic measures of growth and development. While there are some studies on the impact of corruption on specific firms, they tend to be limited to certain countries (Africa: Bates, 1981; Pakistan: Khwaja and Mian, 2005, and Peru: DeSoto,1989). Lobbying Lobbying is the practice of seeking to influence a politician or public official regarding the passing or defeat of legislation or changing current laws to provide benefits to parties with special interests. Public firms regularly engage in lobbying. During the election cycle, expenditures on lobbying were $2.6 billion (Milyo, Primo, and Groseclose, 2000), and Hill, Kelly, Lockhart, and Van Ness (2011) find that 13% of Compustat firms engage in lobbying. The benefits of lobbying include lower effective tax rates (Richter, Samphantharak, and Timmons, 2008) and tax savings (Alexander, Scholz, and Mazza, 2009). Chen, Gao, and Huang (2011) report evidence that hedge funds with lobbying connections trade more heavily in politically-sensitive stocks and subsequently outperform funds with such connections. Yu and Yu (2010) find that fraudulent behavior is less likely to be discovered when firms lobby. 5

6 While lobbying can be beneficial, it may result in a misallocation of resources. Aggarwal, Meschke, and Wang (2008) find that political campaign donations are tantamount to a free cash flow problem. Firms that engage in political donations are associated with lower excess returns and lower firm value. The distortions upon corporate investment created by managers vying for political influence are evident in the conglomerate literature. Rajan, Servaes and Zingales (2000) show that mangers re-invest divisional cash flows in sub-optimal nonpoachable investments in an effort to both maintain control over those investments as well as secure the future cash flows provided by them. Thus, in an environment where claims upon future cash flows are not guaranteed, managers demonstrate a willingness to trade valuemaximizing investment for security in returns. Therefore, a firm may misallocate resources at the margin according to Rajan et al. (2000) to avoid even greater distortions. In this way, firms are willing to take a second-best outcome in order to avoid a third-best one. In a similar vein, firms residing in corrupt political environments may expend resources on lobbying in an effort to gain influence, but risk underinvestment on innovation. The above arguments lead to the following hypotheses: Hypothesis 1 Firms spend greater resources on lobbying when they operate in a corrupt environment Hypothesis 2 Firms in corrupt environments that spend large dollars on lobbying spend less on innovation Hypothesis 3 Firms in corrupt environments that spend large dollars on lobbying reduce the riskiness of their environment If firms lobby in corrupt political environments and reduce the riskiness of cash flows, their value may increase due to a lower risk-adjusted cost of capital. On the other hand, 6

7 misallocating dollars to lobbying and away from capital investment may outweigh the benefit of lower riskiness. Such a scenario would lead to lower firm value. Therefore, due to this ambiguity, we propose the following hypotheses related to firm value: Hypothesis 4a The reduction of risk associated with the cash flows leads to higher firm value. Hypothesis 4b The lack of investment in innovation (by firms operating in corrupt environments) leads to lower firm value III. Sample and data To test the hypotheses presented in Section II, we assemble a sample of firms with data on corruption and lobbying. Our measure for corruption is constructed from the number of federal convictions available from the Report to Congress on the Activities and Operations by the Public Integrity Section (PIN) of the U.S. Department of Justice (DOJ). The report provides annual statistics of convictions made by federal prosecutors against public officials for crimes on corruption across Federal Judicial Districts (FJD) of each of the fifty states. The focus of the crimes in the report are those which are usually inappropriate for the local U.S Attorney s office to handle which include, but are not limited to recusals by United States Attorneys Offices, sensitive cases, multi-district cases, referrals from federal agencies, and shared cases. The types of crime that the DOJ investigates include those such as conflict of interest, fraud, campaignfinance violations, and obstruction of justice. The individuals prosecuted may include any of federal, state and local officials as well as private citizens involved in public corruption offenses. On average, approximately 1,000 convictions are made by the DOJ concerning corruption every year across the United States. Using convictions as a measure of corruption has significant advantages to previously used survey evidence or international corruption proxies since U.S 7

8 convictions provide us with a concrete measure for corruption with no sampling errors of surveys and also makes state to state comparisons possible compared to international perception indices. Using the 2006 and 2009 reports we collect data on annual convictions by each FJD from 1997 to Out of the total of 94 FJDs in the data, we exclude the districts in U.S territories, which leave us with 89 FJDs across the 50 states and one for The District of Columbia ( D.C. ) Following previous studies (Glaeser and Saks, 2006; Butler, Fauver, and Mortal, 2008) we standardize the number of convictions using district population from the U.S Census Bureau to create conviction rates per 100,000 people (Corruption). The total number of convictions and average annual conviction rates by state and FJD are shown in Appendix Table A.1 and A.2, respectively. Our lobby data is collected from the Senate Office of Public Records (SOPR) via the Center for Responsive Politics (CRP). Following the Lobbying Disclosure Act of 1995, the SOPR records all lobbying activities by lobbying firms and payments made by corporations. The data from CRP covers the time period of 1998 to 2011 with 6,000 to 10,000 unique lobby expenditure observations per year. Appendix Table A.4 shows the annual statistics of lobbying in the final sample. In Appendix Table A.3, lobbying dollars shows large variation across industries. Also the amount of lobbying expended is quite small relative to a firm s asset or sales. The conviction rates and lobby amounts are merged with firm data from Compustat Fundamentals Annual database. First, the conviction rates are matched with the firms in Compustat by the location of the firms headquarter Federal Information Processing Standard (FIPS) county geographic code. Since the lobbying data from CRP includes all types of private firms and non-profit organizations in addition to public firms, merging with the main data set reduces the observations of lobbying significantly. Post merging, 27% of the final sample has a 8

9 reported lobby expenditure larger than zero. Our final sample consists of 22,993 firm-year observations with 2,913 unique Compustat firms over a 13 year period from 1997 to We describe our sample in Table I. The average firm in our sample has a conviction ratio (Corruption) of 0.37, measured as the ratio of convictions per 100,000 population, consistent with that found in prior studies (Glaeser and Saks, 2006; Butler, Fauver and Mortal, 2009). Twenty-seven percent of the firms engage in lobbying, the median firm has assets of 1.5 billion dollars, is thirty-six percent levered, and contributes eight percent of assets in the form of cash flows. The median firm demonstrates a Tobin s Q of 1.48 and a weighted average cost of capital of 9.79 percent. Our variable definitions also appear in Table I. [Insert Table I] IV. Results Our preliminary tests consist of splitting the sample at the median level of corruption and testing for differences in our main variables of interest. These findings, presented in Table II, give preliminary support for Hypothesis 1, that firms which operate in corrupt political environments are more likely to lobby. We also find evidence that firms in corrupt environments spend less on innovation, as measured by research and development expenses, consistent with Hypothesis 2. We also observe that firms in more corrupt political environments show lower cost of equity and lower cost of capital, consistent with Hypothesis 3, that these firms lower their risk through lobbying. We also exhibit evidence supportive of Hypothesis 4b, that firms in corrupt environments exhibit lower firm value, as evidenced by significantly lower Tobin s Q. Finally, we see that firms in corrupt political environments are larger, slightly older, 9

10 [Insert Table II] Our subsequent tests involve multivariate regressions. We begin by further testing Hypothesis 1 by estimating the following equation: Lobbying = β0 + β1 Corruption + β2 Herfindahl + β3 SP500 + β4 Cash Flow + β5 Age HQ_in_STCapitol + β7 Leverage + β8 Loss + β3 + β6 (1) We measure lobbying two ways. First we create a dummy variable, Lobbying Firm, which takes the value 1 if the firm engages in lobbying and 0 otherwise. Second we measure lobbying by taking the log of (lobbying dollars plus 1). 1 Our main variable of interest is Corruption, our measure of political corruption. Following Pitman (1976) and Hill et al. (2011), we also include a measure of market concentration as a control. The relationship of concentration to lobbying is unclear since the existing evidence is mixed. Pittman (1976) shows that firms in more concentrated industries make greater campaign contributions, which suggests they are used to curry political favor aimed at favorable changes in legislative policy. However, Zardkoohi (1988) suggests that firms in low concentration industries have an incentive to make contributions as they seek rents through governmental influence. We therefore, use the Herfindahl-Hirschman Index (Herfindahl) to measure industry concentration. Following Hill et al. (2011), we include size of the firm as they find that larger firms are more likely to lobby. Our proxy for size is inclusion in the Standard and Poor s 500 Index (SP500). 2 Hill et al. (2011) also show firms commit more resources to lobbying if the firm s headquarter resides in the state capital (HQ_in_STCapitol). Other control variables include a measure of the cash flow, leverage and previous period performance of the firm (Loss). We expect that firms with higher cash flows will be able to 1 We add the value of 1 to lobbying dollars because some firms have zero dollars of lobbying. 2 We use inclusion in the S&P 500 Index as our size variable instead of total assets because in our subsequent tests, we normalize our dependent variable by total assets (Equation 2). Therefore, in our subsequent regressions, we cannot include traditional measures of size such as assets because of its mechanical relation with our dependent variable (see equation 2). 10

11 spend more resources on lobbying and those with prior poor performance will spend less. We include firm leverage since previous literature indicates that highly levered firms are more likely to engage in lobbying activities (Kelly 1983, 1985). Finally, we control for year and industry fixed effects in all models. Our results are in Table III. We find, consistent with the univariate results, that firms which operate in corrupt political environments are more likely to lobby (Models (1)-(3)) and spend greater dollars on lobbying (Models (4)-(6)). We find that firms which are more concentrated, as measured by the Herfindahl-Hirschman Index, are less likely to lobby and spend less on lobbying. This finding suggests that firms in more competitive industries are more likely to benefit from lobbying. We also find that larger firms are more likely to lobby, consistent with Hill et al. (2011) and Agrawal and Knoeber (2010). These findings suggest economies of scale are present with lobbying dollars and that larger firms are more sensitive to political activities (Agrawal and Knoeber, 2010)). Consistent with Hill et al. (2011) we see that firms headquartered in the state capital are more likely to lobby and spend more dollars on lobbying. We find that firms with greater cash flows are both more likely to lobby and spend more dollars on lobbying, consistent with the corporate philanthropy findings of Brown, Helland, and Kiholm- Smith (2006). This result could simply derive from lower taxes that accrue to lobbying firms (Richter, Samphantharak, and Timmons, 2008; Alexander, Scholz, and Mazza, 2009). On the other hand, the finding could derive from the agency problems associated with free cash flow (Jensen, 1986). We also find that more levered firms are more likely to lobby and commit greater dollars to lobbying. Finally, firms that have performed poorly in the previous period are less likely to lobby in the following period, and for firms that do, the amount is likely to be much lower. 11

12 [Insert Table III] We now turn our analysis to whether firms divert resources away from investment in innovation when they spend large dollars on lobbying. We estimate the following equation: Innovation = β0 + β1 Corruption + β2 Lobbying + β3 Size + β4 Liquidity+ β5 Leverage + β6 Loss (2) Innovation is measured in two ways. First, following Hilary and Hui (2009), we measure investment (Investment) as investment rate in tangible capital (calculated as the ratio of Compustat item 128 to item 8 lagged by one year). Second, we measure intangible investment (R&D) as normalized research and development (the ratio of Compustat item 46 to item 6 lagged by one year). Our main variables of interest are our two measures of lobbying and conviction ratio. Following Hilary and Hui (2009), we control for other variables associated with investment: a measure of size (S&P500), liquidity (Cash flow), leverage and loss. 3. We first test equation (2) using OLS. As shown in Table IV, Columns (1), (2), (5), and (6), both the level of corruption and the amount of lobbying are associated with lower levels of innovation, as measured by capital expenditures and research and development. However, we recognize that we may have a potential endogeneity problem caused by self-selection. Firms which lobby may be associated with lower innovation, but this relation may be a function of the choice by the firm to lobby. Failure to account for self-selection amounts to an omitted variable bias, which results in biased and inconsistent parameter estimates. Therefore, to control for the endogenous choice of lobbying, we use a self-selection model to allow for these differences. Li and Prabhala (2007) outline the increased use of these models in finance. Ang and Nagel (2012) 3 As mentioned previously, we use S&P 500 for size rather than firm assets. Since our dependent variable is normalized by fixed assets, we do not want to include assets as an independent variable because of the potential mechanical relation. 12

13 use this procedure to control for the self-selection of inside versus outside CEO hires, and Campa and Kedia (2002) use this procedure to control for the endogenous choice to diversify. We therefore compute the inverse mills ratio (Lambda) from estimating equation (1), model (3) and use this as an instrument for lobbying (1/0) in equation (2). Critical to using this approach is our identification of an exclusion variable, which will identify our self-selection model. The exclusion variable must be (1) relevant, that is correlated with the firm s lobby decision (relevance), and (2) valid, not correlated with the firm s level of innovation. That is, the only role that our exclusion variable will play in influencing the outcome (level of innovation) is its impact on the firm s lobbying decision. We use HQ_in_StCapitol as our exclusion variable. 4 Our results for equation (2) using the Inverse mills ratio are in Table IV, columns (3) and (7). We find that firms which operate in corrupt political environments spend less on both tangible and intangible investment. Lobbying is also associated with lower levels of tangible investment. We also utilize an instrumental variables approach to control for endogenity. We estimate equation (1), model (6) and compute the fitted value of lobbying dollars. We use the fitted value as an instrument for lobbying in Equation (2). As shown in models (4) and (8) of Table IV, lobbying dollars are also negatively related to innovation. Size and leverage are associated with lower innovation as expected. [Insert Table IV] We have now established that firms in more corrupt political environments spend greater dollars on lobbying. We also document that these firms are associated with lower levels of investment. Therefore, these firms appear to be diverting resources away from investment and 4 To determine if the exclusion variable is valid, we use an empirical technique found in labor economics (Booker et al., 2011) and used in Ang and Nagel (2012). Specifically, when we include the exclusion variable HQ_in_StCapitol in our outcome equation (equation (2)) along with all other independent variables, it is insignificant. 13

14 towards lobbying. However, we know that lobbying could be beneficial to a firm because of the ability of a firm to possibly avoid regulation or possibly secure a government customer. These possibilities could reduce the variability of cash flow and returns, which are beneficial to the firm. We now explore whether this misallocation of resources impacts risk by estimating the following equation: Risk = β0 + β1 Corruption + β2 Lobbying + β3 Size + β4 Liquidity + β5 Age + β6 Leverage + β7 Loss (3) We measure risk a variety of ways. We calculate the standard deviation of returns (Stk Ret Vol), the standard deviation of ROA (ROA Vol), the cost of debt, the cost of equity, and the weighted average cost of capital. Our main variables of interest are measures of corruption and lobbying. 5 We control for liquidity, leverage and loss, three variables associated with our dependent variables in the literature (Hilary and Hui, 2009). Our results appear in Table V. We find that our variables of interest (Corruption and Lobbying) are associated with lower standard deviation of returns and ROA, lower cost of equity and lower WACC. These results give credence to the notion that firms lobby in this environment in an effort to provide safer cash flows to the firm. [Insert Table V] We now turn to direct tests of Hypothesis 3a and 3b, whether lobbying and corrupt environments improve or impede firm value. We test the following equation: Value = β0 + β1 Corruption + β2 Lobbying + β3 High Corruption* ln[lobbying ($)+1] + β4 Size + β5 Liquidity+ β6 Age+ β7 Leverage + β8 Loss (4) 5 We estimate equation (3) using lobbying dollars as well as the fitted value from the instrumental variables approach as described previously. 14

15 We measure value in two ways: Tobin s Q, calculated following Kaplan and Zingales (1997) and operating income to total assets (EBITDAdAT). We control for innovation (using our measure of investment, Investment), size, cash flow, leverage, and loss of the firm. Our results appear in Table VI. Consistent with Hypothesis 4b, we find that firms operating in corrupt environments are associated with lower firm value in Models (1) and (3). In general, greater lobbying dollars are associated with higher value. While lobbying dollars are associated with lower levels of investment (Table IV), we remind the reader that lobbying is also associated with safer cash flows (Table V). So while investment is lower, the safer cash flows and lower cost of capital provided by lobbying are instrumental in value improvement. We therefore want to know if the interaction of corruption and lobbying has an overall positive or negative impact on firm value. We create the interaction term, High Corruption* ln[lobbying ($)+1], which is the product of our lobbying measure and a dummy variable taking the value of 1 if firms operate in a district where corruption is higher than the median. We find that the interaction of corruption and lobbying has a negative impact on Tobin s Q. Therefore, the impact of lobbying on firm value cannot overcome the harm that a very corrupt environment causes. These results are supportive of Parsley and Yang (2010) which show that increases in excess returns due to lobbying are short term. Our results in Table IV show that lobbying is associated with lower levels of traditional investment. In Table VI, we demonstrate that lobbying is associated with greater firm value, yet lobbying in a corrupt environment harms firm value. Therefore, the misallocation of resources from traditional investment to lobbying when firms face corrupt environments is detrimental to the firm. [Insert Table VI] 15

16 Overall, our tests provide support for Hypotheses 1, 2 and 4b. that firms which operate in corrupt environments spend more dollars on lobbying and spend less on innovation. These actions ultimately are harmful to firm value. V. Conclusions Using a sample of almost 3,000 firms over a 13 year time period, we study how firms behave when operating in a corrupt environment. Utilizing federal conviction rates as a proxy for corruption, we are able to confirm the previous views of corruption of sanding the wheel at the firm level while providing evidence that firms operating in corrupt political environments spend greater dollars on lobbying. We find that lobbying leads to safer cash flows, lower costs of equity and lower weighted average costs of capital. However, these benefits come at a cost: the higher dollars spend on lobbying by these constrained firms (those operating in corrupt environments) are directly related to underinvestment on innovation. The misallocation of resources results in lower firm value, as measured by Tobin s q. We contribute to the existing body of literature by providing a comprehensive, firmspecific examination of corruption using data on U.S. firms. We link a firm s use of lobbying to corruption since firms need a means by which to gain influence in such an environment. Our study is the first, to our knowledge that examines the impact of firm behavior in corrupt environments and the subsequent impact on firm value. 16

17 References Aidt, T. S. (2009), Corruption, institutions, and economic development, Oxford Review of Economic Policy, 25, Alexander, R. M., Mazza, S. W. and Scholz, S., (2009) Measuring Rates of Return for Lobbying Expenditures: An Empirical Analysis under the American Jobs Creation Act, Working Paper Aggarwal, R. K., Meschke, F. and Wang, T. Y. (2011), Corporate Political Donations: Investment or Agency?, 2008 WFA Meeting Paper; AFA 2012 Chicago Meetings Paper. Agrawal, A. and Knoeber, C. (2001), Do Some Outside Directors Play a Political Role?, Journal of Law and Economics, 44, Bates, R.H. (1981), Markets and States in Tropical Africa: The Political Basis of Agricultural Policies, Berkeley: University of California Press. Brown, W. O., Helland, E. A. and Smith, J. K. (2006), Corporate Philanthropic Practices, Journal of Corporate Finance, 12, Butler, A. W., Fauver, L., and Mortal, S. (2009), Corruption, political connections, and municipal finance, Review of Financial Studies, 22, Chen, H., Parsley, D., and Yang, Y., (2010) Corporate Lobbying and Financial Performance, Working Paper, Vanderbilt University. De Soto, H. (1989), The Other Path: The Invisible Revolution in the Third World, New York: Harper. Egger, P. and Winner, H. (2005), Evidence on corruption as an incentive for foreign direct investment, European Journal of Political Economy, 21, Galang, R. M. N. (2011), Victim or Victimizer: Firm Responses to Government Corruption. Journal of Management Studies, doi: /j x 17

18 Galtung, F. (2006), Measuring the Immeasurable: Boundaries and Functions of (Macro) Corruption Indices, in Charles Sampford, Arthur Shacklock, Carmel Connors, and Fredrik Galtung, eds.: Measuring Corruption (Ashgate). Glaeser, E. and Saks, R. (2006), Corruption in America, Journal of Public Economics, 90, Globerman, S. and Shapiro, D. M. (2003), Governance Infrastructure and Us Foreign Direct Investment, Journal of International Business Studies, 34, Hilary, G. and Hui, K. W. (2008), Does Religion Matter in Corporate Decision Making in America?, Journal of Financial Economics, 93, Hill, M. D., Kelly, G. W., Lockhart, G. B. and Van Ness, R. A. (2011), Determinants and Effects of Corporate Lobbying, Working Paper. Huang, J. and Gao, M. (2011), Capitalizing on Capitol Hill: Informed Trading by Hedge Fund Managers, Fifth Singapore International Conference on Finance 2011; AFA 2012 Chicago Meetings Paper. Jensen, M. C. (1986), Agency Cost Of Free Cash Flow, Corporate Finance, and Takeovers, American Economic Review, 76, Kaplan, S. N. and Zingales, L. (1997), Do Investment Cash-Flow Sensitivities Provide Useful Measures of Financing Constraints?, Quarterly Journal of Economics, 112, Kaufmann, D. and Wei, S. (2000), Does 'Grease Money' Speed Up the Wheels of Commerce?, IMF Working Paper 00/64. Kelly, L. (1983). Corporate lobbying and changes in financing or operating activities in reaction to FAS No. 8. Journal of accounting and public policy, 1(2),

19 Kelly, L. (1985). Corporate management lobbying on FAS No. 8: some further evidence. Journal of Accounting Research, 23(2), Knack, S. and Keefer, P. (1995), Institutions and Economic Performance: cross-country tests using alternative institutional measures. Economics & Politics, 7, Leff, N. (1964), Economic development through bureaucratic corruption, American Behavioral Scientist, 8, Leys, C. (1970), What is the Problem About Corruption? In A.J. Heidenheimer, (ed.) Political Corruption: Readings in Comparative Analysis, 31-37, New York: Holt Reinehart. Mauro, P. (1995), Corruption and growth. Quarterly Journal of Economics, 110, Mian, A. R. and Khwaja, A. I. (2005), Do Lenders Favor Politically Connected Firms? Rent Provision in an Emerging Financial Market, Working Paper. Milyo, J., Primo, D., and Groseclose, T. (2000), Corporate PAC Campaign Contributions in Perspective, Business and Politics, 2, Pittman, R., (1976), The Effects of Industry Concentration and Regulation on Contributions in Three 1972 U.S. Senate Campaigns, Public Choice, 23, Pittman, R., (1977), Market Structure and Campaign Contributions, Public Choice, 31, Rajan, R. G., Servaes, H. and Zingales, L. (2000), The Cost of Diversity: The Diversification Discount and Inefficient Investment. Journal of Finance, 55, Richter, B., K., Samphantharak, K. and Timmons, J. F. (2008), Lobbying and Taxes, American Journal of Political Science, 53, Rose-Ackerman, S. (1975), The economics of corruption. Journal of Public Economics 4,

20 Rose-Ackerman, S. (1999), Corruption and Government, Causes, Consequences and Reform, Cambridge, U.K.: Cambridge University Press. Scharfstein, D., S. and Stein, J., C. (2000), The Dark Side of Internal Capital Markets: Divisional Rent-Seeking and Inefficient Investment, Journal of Finance, American Finance Association, 55, Shleifer, A., Vishny, R. (1993), Corruption, Quarterly Journal of Economics, 108, Smith, G. (1981), Investment and q in a Stock Valuation Model, Southern Economic Journal, 47, Tanzi, V. and Davoodi, H. R. (1997), Corruption, Public Investment, and Growth, IMF Working Paper 97/139. Tobin, J. and Brainard, W. C. (1968), Pitfalls in Financial Model Building, American Economic Review, 58, Tobin, J. and Brainard, W. C. (1977), Asset Markets and the Cost of Capital, Economic Progress, Private Values and Public Policy: Essays in Honor of William Fellner, North Holland, Wei, S. (1999), Corruption in Economic Development: Beneficial Grease, Minor Annoyance, or Major Obstacle?, World Bank Policy Research Working Paper No White, R. (2011), Greased Wheels or Just Greased Palms: Corruption and Firm Value in the U.S, Working Paper, University of Pittsburgh. Yu, F. and Yu, X. (2010), Corporate Lobbying and Fraud Detection. Journal of Financial and Quantitative Analysis, Forthcoming. Zaleski, P. (1992), Campaign Contributions from Corporate PACs, Atlantic Economic Journal, 20,

21 Zardkoohi, A. and Fraser, D. R. (1998), Geographical Deregulation and Competition in U.S. Banking Markets, The Financial Review, 33,

22 Table I. Descriptive Statistics This table presents summary statistics for our sample of firms for the period Variable definitions are provided in the last column of the table. N Mean Std Dev Median Description Corruption 21, Conviction rate (number of FJD convictions scaled by FJD population) Lobbying ($) 22, ,237 1,274,229 0 Lobby amount in dollars ($) Lobbying Firm 22, =Firm engages in lobbying, 0=Otherwise Investment 22, Investment rate in tangible capital ($mil) R&D 12, R&D to Total Assets ($mil) Stk Ret Vol 22, Standard Deviation of Returns ROA Vol 22, Standard Deviation of ROA Cost of Debt 20, Actual Cost of Debt Cost of Equity 20, Cost of Common Equity using the 36 Month Adjusted Beta WACC 18, Weighted Average Cost of Capital using Actual Cost of Debt and 36-Month Adj Beta Tobin s Q 19, Kaplan and Zingales (1997) Tobin's Q Return on Assets 22, EBITDA to Assets Firm Size 22, Log of Assets Cash Flow 16, Operating income before depreciation minus interest expense, income taxes, and common dividends divided by assets Leverage 22, Leverage ratio Loss 22, =ROA is negative, 0=ROA is positive Herfindahl 22, Industry concentration. Herfindahl Index Firm Age 22, Log of firm age in months (Years shown) SP500 22, Indicator of S&P500 index inclusion 22

23 Table II. Difference in firm characteristics: Firms in high corruption and low corruption areas This table presents the results of a univariate test of variables and firm characteristics depending on whether the firm is located in a high corruption or low corruption area, segregated by the median value of Corruption, the corruption measure. The description of variables can be found in Table I.,, and denote statistical significance at the 10%, 5%, and 1% levels, respectively. Variable N Mean Difference in means Low High Low-High Corruption N Corruption Corruption t Ln[Lobbying ($)+1] 11, , *** Lobbying Firm 11, , ** Investment 10, , *** R&D 11, , *** Stk Ret Vol 10, , *** ROA Vol 11, , *** Cost of Debt 10, , Cost of Equity 10, , *** WACC 9, , *** Tobin's Q 9, , *** Return on Assets 11, , Firm Size 11, , *** Cash Flow 8, , Leverage 11, , *** Loss 11, , *** Herfindahl 11, , *** Firm Age 11, , *** S&P500 11, , *** 23

24 Table III. Corrupt environment and firm lobbying This table shows regression results of corruption on lobbying. In models (1), (2) a Logit regression is estimated using the dependent variable Lobbying Firm. Models (3), (4) show OLS results on the dependent variable ln[lobbying ($)+1]. The description of variables can be found in Table I. Industry and year fixed effects are included in all regressions.,, and denote statistical significance at the 10%, 5%, and 1% levels, respectively. T-statistics are in the parentheses. Variables Lobbying Firm ln[lobbying($)+1] (1) (2) (3) (4) (5) (6) Corruption 0.21*** 0.22*** 0.21*** 0.57*** 0.54*** 0.53*** (6.11) (4.67) (4.52) (6.74) (6.17) (6.01) Herfindahl -2.31** -2.27** -4.24** -4.23** (-2.23) (-2.15) (-2.22) (-2.24) Cash Flow 1.18*** 0.89*** 0.98*** 0.69*** (5.40) (3.83) (4.93) (4.74) S&P *** 0.84*** 2.13*** 2.10*** (12.08) (11.81) (11.17) (11.01) Firm Age 0.74*** 0.72*** 1.78*** 1.75*** (22.75) (22.17) (26.87) (26.44) HQ in ST Capitol 0.37*** 0.37*** 0.89*** 0.90*** (5.68) (5.70) (5.52) (5.55) Leverage 0.19* 0.17*** (1.92) (2.80) Loss -0.32*** -0.86*** (-5.57) (-7.78) Constant *** -2.71*** -2.63*** 5.68*** (-11.82) (-6.39) (-6.12) (9.97) (-0.40) (-0.14) Year Fixed Effects Yes Yes Yes Yes Yes Yes Industry Fixed Effects Yes Yes Yes Yes Yes Yes Observations 21,431 15,387 15,361 21,488 15,433 15,407 Pseudo R Adjusted R

25 Table IV. Innovation and lobbying This table presents the results for models used to test the relationship between lobbying, corruption and innovation. The dependent variable for first four models is Investmets and the dependent variable for next four models is R&D. The OLS (1), (5) models are predicted with Lobbying Firm as the main variable. OLS (2), (6) models are predicted with ln[lobbying ($)+1]. The Heckman models are predicted using the inverse mills ratio from model (3) of Table III instead of Lobbying Firm. The IV models are predicted using the fitted value from model (6) of Table III as a proxy for ln[lobbying ($)+1]. The description of variables can be found in Table I. Industry and year fixed effects are included in all regressions.,, and denote statistical significance at the 10%, 5%, and 1% levels, respectively. T-statistics are in the parentheses. OLS (1) Investment R&D Heckman IV OLS OLS (3) (4) (5) (6) OLS (2) Heckman (7) Corruption -0.02*** -0.02*** -0.02*** *** *** *** *** (-2.70) (-2.69) (-2.88) (-0.88) (-6.54) (-6.53) (-6.66) (-3.38) Lobbying -0.03*** *** -0.01*** *** <-0.001*** *** (-9.39) (-10.14) (-2.63) (-0.60) (-3.38) (-3.45) (-1.44) (-8.44) IV (8) S&P *** -0.04*** -0.05*** ** 0.00*** (-3.10) (-3.09) (-3.53) (-0.39) (-1.52) (-1.52) (-2.04) (2.68) Cash Flow *** -0.09*** -0.09*** -0.09*** (-1.10) (-1.10) (-1.13) (-0.83) (-3.69) (-3.69) (-3.70) (-3.64) Leverage * -0.01* -0.01* (-1.55) (-1.54) (-1.61) (-0.81) (-1.72) (-1.72) (-1.73) (-1.63) Loss *** 0.03*** 0.03*** 0.02*** (0.45) (0.44) (0.50) (0.17) (7.73) (7.73) (7.74) (6.71) Constant -0.19*** -0.18*** 0.16*** *** 0.06*** 0.05*** 0.07*** (-14.10) (-13.86) (14.24) (1.35) (7.37) (7.42) (10.49) (12.91) Year Fixed Effects Industry Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Observations 15,105 15,105 15,104 15,104 15,408 15,408 15,407 15,407 Adjusted R

26 Table V. Lobbying, Corruption and firm risk This table presents the results for models used to explore the impact of misallocation of resources on firm risk. OLS models are predicted with ln[lobbying ($)+1]. The IV models are predicted using the fitted value from model (6) of Table III as a proxy for ln[lobbying ($)+1]. The dependent variables Stk Ret Vol, ROA Vol, Cost of Debt, Cost of Equity and WACC. The description of variables can be found in Table I. Industry and year fixed effects are included in all regressions.,, and denote statistical significance at the 10%, 5%, and 1% levels, respectively. T- statistics are in the parentheses. Stk Ret Vol ROA Vol Cost of Debt Cost of Equity WACC OLS IV OLS IV OLS IV OLS IV OLS IV Corruption -0.03** *** -0.61*** *** *** -0.09** (-2.08) (1.02) (-3.80) (-3.45) (-1.09) (1.26) (-3.12) (-0.94) (-5.24) (-1.98) Lobbying -0.02*** -0.15*** -0.10*** 0.52* -0.04** -0.84** -0.02*** -0.14* -0.02*** -0.15** (-12.01) (-4.16) (-8.36) (1.74) (-2.34) (-2.32) (-5.24) (-1.65) (-5.31) (-2.12) S&P *** ** *** 0.61*** ** (-5.16) (1.24) (0.76) (-1.63) (-2.35) (1.04) (4.42) (3.36) (1.35) (2.23) Cash Flow -2.63*** -2.26*** *** *** *** *** -5.75*** -5.39*** -4.43*** -3.99*** (-16.54) (-11.27) (-17.86) (-15.92) (-6.64) (-4.51) (-12.29) (-9.91) (-9.76) (-7.63) Firm Age -0.38*** -0.16*** -1.42*** -2.44*** -0.27* 1.05* -0.82*** -0.62*** -0.67*** -0.43*** (-27.76) (-2.60) (-12.34) (-4.80) (-1.66) (1.67) (-19.60) (-3.99) (-17.32) (-3.14) Leverage 0.08* 0.45*** * *** -8.75*** *** -3.41*** (1.75) (4.00) (-0.07) (-1.77) (-19.64) (-8.35) (-0.31) (1.04) (-28.52) (-13.79) Loss 1.10*** 0.97*** 5.31*** 5.92*** 1.59*** *** 1.41*** 1.12*** 1.01*** (33.74) (19.08) (20.22) (14.55) (4.87) (1.62) (17.53) (12.20) (14.28) (9.65) Constant 4.79*** 5.83*** 17.93*** 14.26*** 9.44*** 13.81*** 24.69*** 13.64*** 23.07*** 12.65*** (39.93) (38.96) (17.18) (14.68) (6.87) (10.16) (64.21) (29.48) (66.80) (29.22) Year F.E. Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Industry F.E. Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Observations 14,958 14,957 15,361 15,360 15,197 15,196 13,671 13,670 13,553 13,552 Adjusted R

27 Table VI. Corruption, lobbying and firm value This table presents the results for models used to explore the relationship between lobbying, corruption and firm value. OLS models are predicted with ln[lobbying ($)+1]. The IV models are predicted using the fitted value from model (6) of Table III as a proxy for ln[lobbying ($)+1]. The dependent variables are Tobin s Q and Return on Assets. The description of variables can be found in Table I. Industry and year fixed effects are included in all regressions.,, and denote statistical significance at the 10%, 5%, and 1% levels, respectively. T-statistics are in the parentheses. Tobin's Q Return on Assets OLS IV OLS IV (1) (2) (3) (4) (5) (6) (7) (8) Corruption -0.11*** -0.09*** -0.13*** -0.12*** *** ** ** (-5.07) (-4.48) (-4.36) (-3.73) (-2.87) (-2.10) (1.49) (2.20) Lobbying 0.03*** 0.04*** * *** *** -0.03*** -0.02*** (7.59) (7.86) (1.51) (1.73) (-3.66) (-2.77) (-2.96) (-2.92) High Corruption *Lobbying -0.01*** *** *** (-3.18) (-0.53) (-2.64) (-2.89) Size t *** -0.20*** -0.43** -0.45** 0.02*** 0.02*** 0.06*** 0.05*** (-7.24) (-7.23) (-2.24) (-2.52) (4.14) (4.13) (3.49) (3.47) Cash Flow -3.99** -4.00** -3.90** -3.89** 0.43*** 0.43*** 0.39*** 0.40*** (-2.01) (-2.01) (-1.96) (-1.96) (3.86) (3.86) (3.61) (3.65) Firm Age -0.27*** -0.27*** -0.33*** -0.33*** -0.01*** -0.01*** (-7.44) (-7.44) (-5.61) (-5.73) (-4.23) (-4.20) (-0.13) (-0.05) Leverage -0.21*** -0.21*** -0.21*** -0.21*** (-3.18) (-3.17) (-3.19) (-3.19) (0.98) (0.99) (1.09) (1.11) Loss -0.79*** -0.80*** -0.73*** -0.72*** -0.10*** -0.10*** -0.10*** -0.10*** (-4.08) (-4.08) (-3.49) (-3.52) (-9.77) (-9.77) (-9.72) (-9.84) Constant 5.30*** 5.32*** 5.90*** 5.98*** 0.16*** 0.17*** -0.13* -0.11* (15.22) (15.23) (7.37) (7.89) (3.86) (3.88) (-1.87) (-1.73) Year Fixed Effects Industry Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Observations 14,220 14,182 14, ,347 15,347 15,346 15,346 Adjusted R

28 Appendix A.1 Convictions by State State # of FJDs # of years Total Convictions Avg Convictions per year Avg annual Conviction per 100,000 pop. % of All Convictions Louisiana % North Dakota % Alaska % Mississippi % Virginia % South Dakota % Montana % Kentucky % Alabama % Delaware % Florida % New Jersey % Missouri % Illinois % Ohio % Maryland % Pennsylvania % Tennessee % West Virginia % Hawaii % New York % Oklahoma % Wyoming % Texas % Massachusetts % Arizona % Connecticut % Arkansas % Idaho % Georgia % Maine % New Mexico % Vermont % Indiana % Rhode Island % Michigan % 28

29 California % North Carolina % Wisconsin % Nevada % South Carolina % Utah % Colorado % Washington % Kansas % Minnesota % Iowa % New Hampshire % Nebraska % Oregon % District of Columbia % D.C., Guam & NMI, Puerto Rico, Virgin Islands were excluded 29

30 A.2 Conviction Rate (# of convictions per population) by federal Judicial District Federal Judicial District # of years Avg Conviction per 100,000 population Median Conviction per 100,000 population Louisiana Eastern Mississippi Northern Kentucky Eastern Missouri Eastern Florida Southern Virginia Eastern Tennessee Western North Dakota North Dakota Louisiana Middle Alaska Alaska South Dakota South Dakota Montana Montana Ohio Northern Pennsylvania Eastern Florida Northern New York Southern Alabama Northern Alabama Middle Illinois Southern West Virginia Southern Delaware Delaware Illinois Northern New Jersey New Jersey Pennsylvania Middle Mississippi Southern Oklahoma Eastern Alabama Southern Maryland Maryland Arkansas Eastern Virginia Western Texas Southern Indiana Northern Louisiana Western Georgia Southern Oklahoma Northern Hawaii Hawaii Wyoming Wyoming