Labor Unemployment Benefits And Corporate Takeovers. Lixiong Guo Culverhouse College of Commerce, University of Alabama, United States

Labor Unemployment Benefits And Corporate Takeovers Lixiong Guo Culverhouse College of Commerce, University of Alabama, United States lguo@cba.ua.edu Jing Kong * Eli Broad College of Business, Michigan State University, United States kongjin1@msu.edu August 2017 Abstract Takeovers are often followed by layoffs of workers. However, little is known about how potential unemployment benefits to laid-off workers affect a firm s likelihood of being acquired and the synergy of the deal. We exploit changes in state unemployment insurance laws as a source of variation in labor unemployment benefits. We find that higher unemployment benefits increase the likelihood of firms being acquired, particularly for labor intensive firms, unionized firms, and firms with higher spending on corporate social responsibility. Moreover, we find that labor unemployment benefits also affect the synergy of the M&A deals, both the bidder s announcement returns and the combined bidder and target firm s announcement returns are higher when the target firm s state unemployment benefits are higher. Our findings suggest that labor unemployment benefits have a significant impact on takeover outcomes and state unemployment insurance laws have an unintended consequence on takeovers. JEL Classification: G32, G34 Keywords: Mergers and Acquisitions, Labor, Unemployment Benefits, Synergy, Layoff We are grateful to Ronald Masulis, Peter Swan, Oleg Chuprinin, Chang-Mo Kang, Hao Liang, Andrey Golubov and seminar participants at UNSW Sydney, conference participants at the 2017 AsianFA Conference, and the 2017 FIRN Conference for helpful discussions and valuable comments. 1

I. Introduction In labor contract negotiations, a firm s optimal risk-taking level depends on the tradeoff between benefits of risky projects and costs of human capital. 1 If contract agreements are breached (i.e. increase of firm risks), employees can observe it and ask for more compensations. Therefore, firms usually maintain their optimal risk levels until firms experience exogenous reasons that change employees risk aversion levels. Previous studies on labor unemployment benefits find that firms choose increase their risk-taking levels when workers costs of unemployment decrease (Agrawal and Matsa, 2013; Devos and Rahman, 2013; and Ellul, Wang, and Zhang, 2015). Compared with within-firm risk level changes, situations are more complicated in the context of takeover: labors face higher unemployment risk if firm is facing takeover; however, firms may not necessary pay more compensations because labor contract often need to be renegotiated. Following the study of labor unemployment benefits, our paper examines whether a firm s likelihood to be acquired is affected by employees costs of unemployment. Firms are likely to experience restructuring and labor contracts renegotiation after being acquired. Therefore, workers cost of unemployment is possible to have two opposite directions of impact. On the one hand, high cost of unemployment could place entrenchment effect on firms. This effect comes from several ways: first, implicit contracts may require potential target firms to resist takeover offer to protect employees from unemployment risks; second, workers with high cost of unemployment have more incentives to resist layoff, decreasing the expected synergy (Dessaint, Golubov and Volpin, 1 Workers would require certain contractual arrangements including a wage premium or higher benefits as compensation of the unemployment risk (Abowd and Ashenfelter, 1981; and Topel, 1984). 1

2017); third, the spill-over effect could increase human capital costs and decrease productivity (Brown and Matsa, 2012). Consequently, acquirers have less incentive to engage in takeovers that have less expected cost savings. By contrast, exogenous shocks that decrease workers cost of unemployment lead to both acquirers more likely to offer deal and target managers more likely to accept deal due to higher expected synergies. On the other hand, target firms that are bound by workers cost of unemployment may choose conservative policies at the risk of operating performance. When cost of unemployment decreases, firms can choose higher risk-taking level and prevent themselves to be acquired (Berger, Ofek, and Yermack, 1997). Thus, decreasing workers cost to unemployment decrease firm s likelihood of being acquired. Our study tests the above two hypotheses by examining whether decreasing cost of unemployment benefits can increase or decrease a firm s likelihood of being acquired. Following Agrawal and Matsa (2013), we use state unemployment insurance (UI) benefits as a proxy for workers cost of unemployment. This proxy largely solves the endogeneity issue since benefits amount is determined by state laws, and the changes result from different reasons. Apart from increases along with economic growth, political reasons also contribute to changes in UI benefits 2. Moreover, different states experience changes in different years. These features enable UI benefits to serve as an exogenous variable in our study. UI benefits are applicable to almost all workers who become involuntarily unemployed, and increases in maximum state UI benefits are strongly associated with greater state UI payouts (Agrawal and Matsa, 2013). Therefore, on average, the increase in 2 For instance, the UI benefits in California stayed unchanged at $230 from 1992 before it increased to $330 in 2002. Although economic experienced fluctuation during the prior ten years, the increase in 2002 coincides with year-end gubernatorial elections since it strengthened political support for the governor s reelection. (Kiplinger California Letter, 2001; and Oakland Tribune, 2002). 2

maximum UI benefits means workers in that state will receive higher unemployment benefits. First, we test if higher UI benefits increase the target firms likelihood of accepting takeover offers. At firm level, we find that higher UI benefits increase a firm s likelihood of being acquired. All else equal, a 10% increase in UI benefit increases the likelihood of a firm being acquired by 84 basis point, which means the average likelihood of becoming a target increased by around 25% 3. The inclusion of controls for firm fixed effects implies that the results reflect average, within-firm changes with increases in state UI benefits. The correlation is stronger when we exclude industries with a geographically dispersed workforce for which it is possible to measure eligible UI benefits with large error. Harford (2005) finds that industry merger wages can be driven by economic shocks, which could also lead to increases in UI benefits. We tackle this issue by replacing year fixed effects with industry year fixed effects. The correlation between UI benefits and the likelihood of being acquired remains significant, indicating the results are not driven by economic shocks. We rule out several alternative explanations through inclusion of control variables and placebo tests. Inclusion of state GDP growth and state unemployment rate as control variables rule out microeconomic factors that could both lead to UI benefits increase and merger waves. Our placebo tests rule out coincidence of the correlation and firm-operating factors. When we replace our main interested independent variable: UI benefits at the year before acquisition into UI benefits in the year after acquisition, and UI benefits in the bordering state, assuming bordering states experience similar economic environments. The 3 The mean likelihood of being acquired in our sample is 3.3%. 3

coefficient is insignificant in both hypothetical UI benefits. There are also insignificant correlations between UI benefits and firm s operating performance. We then explore the channel through which lower cost of unemployment lead to more likelihood of being acquired. If firm choose to maintain risk levels due to the compensation-unemployment risk equilibrium constraint, the effects should be stronger when requirement for maintaining unemployment risks is stronger (Berk, Stanton, and Zechner, 2010). Empirically, we find that the impact of UI benefits on the likelihood of being acquired is stronger in more labor-intensive firms and states have higher union coverage rates, suggesting the firm are subject to stronger binding effect when employees have more bargaining powers. However, the coefficient of UI benefits remains significant after adding union coverage rate in the regression, indicating our results are not solely driven by union s collective bargaining power, but through firm s discretionary choice. These results prove that firms with stronger employee powers have stronger constraints in wage-unemployment risk equilibrium. Apart from labor s requirement, we test if manager s likelihood to fulfill contract is correlated with their inherent care for employees. We use employee-related corporate social responsibility (CSR) index as a proxy for manager s extent to which they care about employees. We find that in a sample of firms with CSR index data available, the positive correlation between UI benefits and the likelihood of being acquired only exist in firms with high employee-related CSR index. This result suggests that the more manager care about employees, the more likely that firm will fulfill labor s requirements in implicit contract. 4

Then, we explore acquirer s incentive to takeover firms experience higher UI benefits. Acquirers have more incentive to make acquisitions with higher expected synergies. If workers with higher unemployment benefits concern less about unemployment, postacquisition layoff is easier and less costly. Therefore, acquirers can generate more synergies from cost-savings. Consistent with our hypothesis, we find that deals in which target firms have higher UI benefits experience more employment reduction within two years after the deal. We also find that higher UI benefits increase lead to both higher acquirer s announcement return. With average combined [-2, 2] CAR 0.037, a 10% increase in UI benefits lead to a 10% increase in combined announcement return. The announcement effect is more pronounced on acquirers. With average acquirer [-2, 2] CAR 0.008, a 10% increase in UI benefits lead to a 26% increase in acquirer announcement return. The increased announcement return suggests market s higher expected synergies for takeover in states with higher UI benefits. Results also prove that higher UI benefits decrease takeover costs: higher UI benefits lead to more post-acquisition worker reductions and lower restructuring costs. Our paper makes two major contributions to the literature. Firstly, this paper contributes to a growing literature in labor economics and finance. In exploring the employee protection effect of unemployment benefits, we link the effect of labor s costs of unemployment with takeover decisions. From acquirer s point of view, our paper is closely related to Dessaint, Golubov and Volpin (2016), who find that stronger employee protection decrease acquisition deal volumes and expected synergy. From target firm s point of view, our paper proposes a channel that stronger employee powers decrease firm risks: risk-sharing mechanism between firms and employees. Previous literature provides 5

evidence that labor s bargaining power can influence firm s financial policy choices and operation performances (e.g., Agrawal and Matsa, 2013, Simintzi, Vig, and Volpin, 2014; Chen, Kacperczyk, and Oritz-Molina, 2011; and Addessi and Busato, 2009). We provide novel empirical evidence that worker s cost of unemployment significantly impacts managers choice of accepting takeover offers. As human capital becomes an increasingly critical asset for firms (Zingales, 2000), firms choose financial policy partly as a means of mitigating labor s exposure to unemployment risk (Berk, Stanton, and Zechner,2010; Agrawal and Matsa, 2013; and Ellul, Wang, and Zhang, 2015). Our study shows that firms choose to be more entrenched when labor s cost of unemployment is high. Secondly, our paper contributes to the large literature on merger and acquisitions. Several recent papers identify the effect of both input human capital (Tate and Yang, 2016) and output product market on acquisition decisions (Ahern and Harford, 2014; Hoberg and Phillips, 2010). Our paper identifies the labor s cost of unemployment as determinants of merger decisions. Previous literature also reveals negative impact of labor protection on takeover outcomes (e.g. John, Knyazeva, and Knyazava, 2015) but did not disclose the driver of this correlation. Our paper identifies unemployment costs as an important driver that results in employee s negative impact. When workers face higher unemployment costs, both the probability of the firm being acquired and expected synergies are significantly lower. The rest of the paper is organized as follows. Section II describes the related literature and hypothesis development. Section III describes background of state unemployment benefits scheme and sample construction procedure. The empirical results are presented in Section IV to Section VI. Specifically, Section IV presents target s firms likelihood to 6

accept takeover offers when UI benefits changes, Section V presents identification tests, Section VI presents acquirers preference to UI benefits changes and Section VII shows robustness check. Section VIII concludes the paper. II. Related Literature and Hypothesis Development Workers concerns about unemployment affect firms policies on layoffs and wage setting (Topel, 1983, 1984; Li, 1986; and Hamermesh, and Wolfe 1990;). Concerns about unemployment also reduce workers labor supply, even when firms are far from bankruptcy (Brown and Matsa, 2012). Despite the magnitude of workers costs of unemployment, it is largely absent from theories in corporate finance. Berk, Stanton, and Zechner (2010) derive an optimal labor contract model in an economy with perfectly competitive capital and labor markets. In their model, employees become entrenched under labor contract and so face large human costs of bankruptcy. Therefore, employees pay for the insurance provided by the labor contract by accepting lower wages. In most cases, risk of unemployment is equal to firm bankruptcy risk, which increases along with firm s risk levels. Like bankruptcy, takeovers greatly increase worker s unemployment risks, because cost reductions through layoffs are both a major driver and a key source of takeover synergies (e.g., Houston, James, and Ryngaert, 2001). Due to far more restricted access to financial market, workers are more risk-averse than firms. Thus, workers request compensation levels to match their unemployment risk. The request for maintaining unemployment risk level could entrench firms when facing takeover. By contrast, lower cost of unemployment could increase a firm s likelihood of being acquired. This increased likelihood could result from different reasons. Firstly, from the 7

perspective of optimal contracting, we argue that there exists implicit contract between firms and employees. While employees accept lower wages, they also require lower risk of unemployment. This argument is supported by the evidence in Agrawal and Matsa (2013). They find the negative correlation between labor s cost of unemployment and financial leverage, and argue that firms choose conservative financial policies partly to mitigate workers exposure to unemployment risk. Consistent with this argument, Ellul, Wang, and Zhang (2015) provide further evidence that cost of unemployment also influence board s choice of for CEO incentive compensation. These studies imply that firm is bound by implicit contract arrangement with employees, and tradeoff between human capital cost savings and benefits of risky projects. Labor and finance literature documents many evidences that stronger labor bargaining power lead to lower firm risks. For instance, Chen, Kacperczyk, and Ortiz-Molina (2012) find that firms in unionized industries implement less risky investment policies. Bradley, Kim, and Tian (2016) find negative impact of unionization on firms innovation activities. Although these labor studies focus the conflict of interest between employee and shareholders, they suggest labor s implicit requirement for lower unemployment risks. When cost of unemployment decreases, target firm managers are released from the risk-level constraints in implicit labor contract, thus more likely to accept takeover offers. Main Hypothesis I: Firms experience increases in UI benefit are more likely to become target firms in acquisitions in the next year. Secondly, decreased cost of unemployment can facilitate post-acquisition layoff, and generate more synergies. Therefore, all else equal, acquirers prefer to find target firms with lower costs of unemployment. Previous findings show that firm with stronger labor rights 8

experience lower announcement return in M&A (John, Knyazeva, and Knyazava, 2015). This lower return could be market s expectation for reduced labor supply due to workers concerns about becoming unemployed (Brown and Matsa, 2012), higher wages required after being acquired (Conyon et al, 2004; and Ouimet and Zarutskie, 2016), and labor s resistance to layoff. Dessaint, Golubov and Volpin (2016) directly address the role of costreduction in deal synergies, and find that when pose-acquisition layoff is harder, both deal volume and deal expected synergies are lower. By contrast, if decrease in unemployment benefits could facilitate cost savings, it will also lead to higher synergies. Therefore, we hypothesize that high UI benefits also lead to higher announcement return. Main Hypothesis II: Expected deal synergies increase when target firms operate in a state with higher UI benefits. III. Sample Description Our sample starts with all U.S. public firms traded on NYSE, AMEX or NASDAQ from CRSP-Compustat merged dataset. This leads to an implicit requirement that all target firms and control firms are publicly traded within the United States. Acquisition transaction data comes from the Securities Data Corporation s (SDC) U.S. Merger and Acquisition database. Following Moeller, Schlingemann, and Stulz (2004) and Masulis, Wang and Xie (2007), our sample covers all transactions between 01/01/1990 and 31/12/2015 with the following criteria: 1. The transaction is completed and valued at least $1 million. 2. The acquirer owns less than 50% of target s share before announcement date, and owns 100% after transaction. 9

3. Deal Type: exclude tender offers, self-tenders, exchange offers, repurchases, spinoffs, leveraged buyouts (LBOs), and recapitalizations 4. 4. Neither acquirer nor target comes from financial (6000-6999) or utility (4900-4949) industries. We merge target firms that meet the above criteria into CRSP-Compustat merged dataset and winsorize all continuous variables at 1 st and 99 th percentile. In testing the likelihood of takeover, the dependent variable is acquisition indicator, which equals to one if the firm became the target of an acquisition in a year, and zero otherwise. There are 2,034 transactions that meet our criteria between 1990 and 2015. The Acquisition indicator has a mean value of 0.033, meaning that on average, 3.3% of our sample firms are acquired during the sample period. The main interested independent variable is the natural logarithm of state UI benefits amount prior to the acquisition deal. State unemployment insurance data comes from the US Department of Labor s significant Provisions of State UI Laws 5. There are two issues each year, on January and July respectively. Each issue provides detailed UI benefit schedules in each state, including a maximum weekly benefit amount and a maximum duration for eligible claimants. Agrawal and Matsa (2013) verify in their study that increases in state UI benefits are strongly associated with greater state UI payouts. 6 Thus, while the actual benefit and duration to receive the benefit are determined by a worker s employment history, we focus on the maximum benefit amount each claimant could receive (Maximum Weekly Rate Maximum Weeks). To merge UI 4 Our study focuses on target manager s choice when worker s cost of unemployment change. Therefore, we exclude tender offer to avoid potential agency problem. 5 Although the data is available from 1950, prior to 1990 the data is scanned document and hard to recognize. We choose to start from 1990 to avoid potential errors. 6 In test for the correlation between maximum UI benefits amounts and actual payout, R 2 equals to 92%. 10

benefits with Compustat annual data, we use the average amount of the two issues in each year. Our union coverage data are provided by Hirsch, Macpherson, and Vroman (2001) 7. We use state-year level of union coverage rate. High union coverage means that union coverage rate is above the median of each state in our sample period. The data for CSR ratings come from MSCI ESG Stats (formally KLD) Database, which contains yearly environmental, social, and governance ratings of large, publicly listed companies from 1991 to 2013. For the relevance of our study, we concentrate on a firm s CSR performance in employee category. ESG Stats compiles statistics on both strengths and concerns. Following the Bae, Kang, and Wang (2011) and Deng, Kang, and Low (2013), We construct a net CSR measure that adds all employee-related strengths and subtracts concerns. Our results are robust to using employee treatment index as proxy, which is constructed following Bae, Kang, and Wang (2011). Employee treatment index adds the indicator variables for union relations, cash profit-sharing, employee involvement, retire benefits strength and health and safety strength. For the test of market reaction to M&A announcements, we focus on the cumulative abnormal changes in share valuation around surprise announcements of acquisitions. This event study format filters out many time-invariant unobservable firm characteristics and provides a cleaner empirical setting for analyzing value implications of firm decisions. Cumulative abnormal return (CAR) on announcement is defined as the sum of daily abnormal returns from a market model estimated using CRSP value-weighted market return. Following Masulis, Wang and Xie (2007), we estimate the market model parameters 7 This database is updated annually. 11

over the 200-day period from event day 210 to event day 11. Then, we compute 3-day and 5-day cumulative abnormal returns (CARs) during the window encompassed by event days [-1, 1] and [ 2, +2], where event day 0 is the acquisition announcement date. Takeover premiums reflect the relative bargaining power between acquirer and target. It is equal to price per share paid by a bidder for a public target firm s shares relative to the target s preoffer announcement stock price 4 weeks prior to the announcement date. We control for a vector of firm characteristics that may affect the probability of being acquired. Motivated by previous literature (e.g., Song and Walkling, 2000), control variables include firm size, leverage, R&D expenditures, ROA, Tobin s Q, sales growth and firm tangibility. We also control for a vector of corporate governance characteristics, like independence board, CEO age, and CEO ownership 8. The corporate governance control variables come from Execucomp and Institutional Shareholder Services (ISS) database. To mitigate the concern that the both likelihood of acquisition and state UI benefits are driven by state-level economic factors, we include state unemployment rate and state GDP growth rate to control for contemporaneous local macroeconomic conditions. State unemployment data is annualized from Current Population Survey monthly data and state GDP growth rate come from U.S. Bureau of Economic Analysis. We also use firm fixed-effects and year fixed-effects to control for time-invariant differences in the likelihood of being acquired across firms and unobservable macroeconomic factors. An acquisition is defined as within industry acquisition if acquirer and target firms come from the same 3-digit SIC industry. 2-digit SIC industry is used to define industry fixed effects. Accounting data comes from Compustat. Since the variation in unemployment benefit is 8 Due to data limitation, we only include corporate governance control variables in a subsample regression in table 2. 12

at the state level, the estimated standard errors in all regressions are clustered at the state level (Agrawal and Matsa, 2013). This clustering method accounts for potential timevarying correlations in unobserved factors that affect all firms in the same state (Bertrand, Duflo, and Mullainathan, 2004). This method also incorporates within-firm error term correlations over time. Summary statistics are presented in Table 1 Panel A. Since we require target firms to be listed on major exchanges, median value of transaction is 217 million. UI benefits range from $3,484 to $31,410, with mean of $9,885, while the logarithm transformation of UI benefits ranges from 8,16 to 10.35. Our median sample firms have book value assets of $274 million, are moderately levered with a book leverage ratio of 16%. In terms of performance, sample firms have a median ROA of 12%, sales growth of 9%, and Tobin s Q of 1.58. For state characteristics, the median state GDP growth rate is 5% and average state unemployment rate is 6 %. In terms of union coverage, the median coverage rate of all states in our sample period is 15.6%. Consistent with previous literature (e.g., Li, 2013), acquirers on average experience negative cumulative abnormal returns around the announcement. Target realize large positive abnormal return on announcement. The average five-day announcement period abnormal return for acquirer firm is 0.8%, not significant different from zero. The average five-day announcement period abnormal return for target firms is about 8%. Combined CAR is the market-value averaged Target CAR and Acquirer CAR. The average five-day announcement period abnormal return for combined CAR is 2.8%. This seems to suggest that while target shareholders benefit more from takeovers than acquirers, mergers create values to shareholders as a whole. 13

IV. Unemployment Insurance and Target s Likelihood to Accept Takeover We begin showing the correlation between the likelihood of being acquired and UI benefits changes by two figures. Figure 1 plots the correlation between state-level acquisition ratio and UI benefits increases. Acquisition ratio is the number of public firms being acquired divided by total public firms in each state. The Y-Axis is the average acquisition ratio, and the X-Axis is the average UI benefits changes across all states in each year. The fitted line shows that in state level, ratio of firms being acquired increases when UI benefits increase. Figure 2 plots the co-movements between acquisition ratio and UI benefits changes in four states: Michigan, Illinois, Texas, and New Jersey. The comovement figures indicate that the likelihood of being acquired is highly correlated with UI benefits changes. Next, we show the firm-level likelihood of being acquired after controlling for a vector of factors. A. Firm-Level Tests In this section, we start our tests from a firm-level multivariate regression analyses. Specifically, we regress the acquisition indicator in the current year on the logarithm of state-level maximum UI benefits in the previous year while controlling for an array of state and firm variables. We implement our test using the linear probability regression 9. 9 With a large number of fixed effects, a non-linear model (such as a logit or probit model) is likely to produce biased estimates due to incidental parameter problem (Lancaster, 2000). Moreover, the marginal effects in a linear probability model are easier to compute and interpret relative to non-linear models. 14

Acquisition i,t = α + β 1 Benefit s,t 1 + β 2 Firm Characteristics i,t 1 + β 3 State Controls s,t 1 + Firm FE + Year FE + ε i,t 10 where i indexes firm, s indexes the state in which the target firm s operation is located, and t indexes the year. The Acquisition is an indicator variable that takes the value of one if the firm is acquired in year t, and zero otherwise. The Benefit s,t 1 denotes the logarithm of the maximum UI benefit in state s before the acquisition. We control for a series of state and firm characteristics that may affect the likelihood of takeover. The firm-level control variables include firm size, firm risks (measured by leverage), firm profitability (measured by return on assets, Tobin s Q and sales growth, firm growth opportunity (measured by R&D expenditure) and tangibility (measured by PPE scaled by book value of assets). When data is available, we also include corporate governance characteristics as additional control variables, such as independence board, CEO age, and CEO ownership as in Huang et al. (2014). To alleviate the concern that state UI benefits increase along with potential future economic changes that may also lead to more takeover activities, we also control for contemporaneous local economic factors such as state GDP growth rate and unemployment rate. We include firm fixed effect fixed effects to account for unobservable firm-specific and time-invariant attributes and year fixed effects to capture any time-trend of M&A. Table 2 presents regression analysis of the effect of UI benefit amount on the likelihood of a firm in that state being acquired. Column 1 and column 2 include firm fixed effects and year fixed effects, with standard errors clustered at state level. The correlation between UI benefits and acquisition indicator is positive and statistically significant at 1% level, 10 We tested different fixed-effects specifications 15

suggesting UI benefits can increase the likelihood of a firm being acquired. The economic effect of the UI benefit on the likelihood of being acquired is also significant. Specifically, a 10% increase in UI benefit increases the firm being targeted in acquisition by 0.84%, all else equal. The economic magnitude is also significant: the average likelihood of become a target is 3.3% means a 10% increase in UI benefits can increase the probability of a firm being acquired by more than 25%. State UI benefits cover workers in the state where they work, whereas the state information in the Compustat is the state where the firm is headquartered. This misspecification could attenuate our estimates if a large number of workers in a firm is located in a state different from a firm s headquarter. Column 2 repeats the estimation in column 1 but exclude dispersed industries 11. Consistent with expectation, after excluding those industries, the coefficient of UI benefits increases from 0.84 to 0.91 (comparing column 1 and column 2). In unreported summary statistics, total acquisition percentage changed little after eliminating those industries. Therefore, the regression results suggest that the likelihood of being acquired indeed increases after eliminating wide dispersed industries. With regards to control variables, firms with poorer value, smaller firms, and firms that are higher leveraged are more likely to be acquired. These results are broadly consistent with prior literature (e.g., Song and Walkling, 2000 and Bertrand, Schoar and Thesmar, 2007). The signs of state economic controls variables are also consistent with expectation. State with poorer GDP growth rates are associated with more firms being acquired in that state. The negative coefficient of state unemployment rate on likelihood of being acquired 11 Same as previous results, dispersed industries include, retail, wholesale, and transport industries 16

could also be explained by our hypothesis: when all else hold constant, employees in state with higher unemployment rate face more unemployment risk, thus have more incentive to resist the acquisition deal. Our main results are robust to sensitivity checks. Column 3 and 4 runs the same regression as column 1 and 2 but replace year fixed effects with industry by year fixed effects. In column 5 and column 6, we run a subsample regression and include corporate governance control variables. While the significance and magnitude of coefficients are impacted by the reduced sample size, the coefficient remains statistically significant at 10% level and increases when excluding wide dispersed industries. B. Placebo Tests Although state UI benefits are relative exogenous to firm-specific characteristics, it is still possible that omitted variables, such as state macroeconomic changes, are correlated with both higher UI benefits and the higher likelihood of being acquired. We address this concern through placebo tests. In the first placebo tests, we replace a firm s actual UI benefits with the bordering state s UI benefits amount, assuming a firm s bordering state experience similar microeconomic conditions 12. The results are presented in Table 3 column 1 and column 2. The correlation is insignificant both include and exclude dispersed industries. To address the reverse causality concern between UI benefits and the likelihood of being acquired, we regress the acquisition indicator on the amount UI benefits one year after the acquisition. Column 3 and column 4 shows that the only UI benefits increases in the prior year are associated with increases in likelihood of being 12 For any state that have more-than-one bordering states, we use the bordering state that has the closest population number. 17

targeted. There is no significant correlation between acquisition likelihood and UI benefits in the next year. Both placebo tests show that our results are not driven by omitted microeconomic conditions. UI benefits are financed by corporate unemployment insurance taxes. Thus, it is also possible that increases in state UI benefits increased the tax burden on firms operating in that state. The increased tax burden then reduces performance, which could lead to higher likelihood to be acquired. There are two reasons why we rule out this alternative mechanism. First, in all regressions, ROA is added as control variable. The results present a positive correlation between ROA and the likelihood of being acquired. Second, to test this alternative mechanism, we examine the correlation between UI benefits and ROA, proxy of firm performance. However, there is no evidence to suggest that increases in UI benefits lead to decreases in firm performance. The results are presented in Table 3 column 5 and column 4. V. Identification Tests: Employee Bargaining Power and Managers CSR If the mechanism through which the increase in UI benefits lead to increases in likelihood of being acquired is risk-sharing between risk-neutral firms and risk-averse employees, this effects should increase when employee numbers are large or they have higher bargaining power (Berk, Stanton, and Zechner, 2010). Therefore, we test whether the effect of UI benefits is stronger when employees have stronger bargaining power. The proxies for employee bargaining power are employee size and union coverage. Using the total size of employees is noisy to some extent because not all employees face the same level of unemployment risk. Employee layoffs 18

concentrated mostly among nonproduction workers (Li, 2013) and overlapped operations. Table 4 panel A presents the result of regress acquisition indicator on the interaction of workforce and UI benefits. Because a firm s employee size changes little unless it suffers significant change, in column 1 and 2, we use state fixed effects in the regression to explore an average within-state changes. To address industry the effect of industry waves, we use industry year fixed effects in column 3 and 4. While the UI benefits are still statistically significant at 5% level, the interaction term is statistically significant at 5% level and positively correlated with acquisition. The results suggest that the effects of UI benefits on the likelihood of being targeted are stronger in firms with more employees. Consistent with prediction, firms with more employees also have stronger requirements for job security. In unionized firms, labor contracts are negotiated through unions. To find out whether the influence of UI benefits on the likelihood of being acquired is implemented through union s bargaining power, we perform the subsequent identification tests. Table 4 Panel B shows the influence of union coverage use on the correlation between UI benefits and likelihood of being acquired. We regress the likelihood of being acquired on the indicator of high union coverage rate and its interaction with UI benefits. Consistent with current labor and finance literature, without the effect of UI benefits, high union coverage indicator has significant negative correlation with the likelihood of being acquired. The coefficient of UI benefits remains significant in all specifications, suggesting the results are not solely driven by union s collective bargaining powers, but come from firm s discretionary choice. Yet, the coefficient of the interaction term is positive and statistically significant. These 19

results suggest when labor has stronger collective bargaining power, firms are more sensitive to the changes on UI benefits. Apart from unions, stakeholder-focused managers would voluntarily act for employee s interest, and better fulfill the implicit contract with labors. From the perspective of stakeholders, managers have more incentive to decline takeover offers to protect workers job security when workers face high cost of unemployment. We use discretionary investment in employee-related CSR as a proxy for manager s care for employees. We test a subsample of firms with CSR reports available on MSCI ESG Stats Database, and divide this sample into high-csr firms and low-csr firms, with cutoff equals to the median CSR score for all firms in each year. The CSR score is strength concerns in employee-related categories. We then run a similar regression as the baseline test and the results are presented in Table 5. We find that the positive correlation between UI benefits and the likelihood of being acquired is only significant when firms have higher than median CSR score. For firms with lower-than-median CSR score, the correlation is insignificant. Our results suggest it is possible the implicit contracts that prevent worker from being unemployed are better fulfilled when managers care more about employees. In the robustness check, we use employee treatment index (Bae, Kang, and Wang, 2011) as the proxy for managers caring for employees. The coefficients are only significant in firms with high employee treatment index. VI. Empirical Tests: Unemployment Insurance and Acquirers Preference A. Cumulative Abnormal Return When searching for target firms, acquirer mainly consider two aspects: the likelihood of deal success and the expected deal synergy. When cost of unemployment is high, 20

workers concerns about becoming unemployed could lead to reduced labor supply (Brown and Matsa, 2012) and higher wages required after being acquired (Conyon et al, 2004; and Ouimet and Zarutskie, 2016). Therefore, human capital costs are likely to increase, and lead to lower expected synergies. By contrast, if higher UI benefits facilitate cost reduction through layoffs, more synergies can be generated. Therefore, we expect that market would react more positively higher UI benefits. To validate the argument that higher UI benefits reduce labor s sensitivity to layoffs, we also expect more post-acquisition layoffs following. Firstly, the results support the argument that high UI benefits lead to higher market s expected synergies. The results are presented in Table 6. We use state fixed effects, industry fixed effects and year fixed effects to control for the effect of unobservable within state characteristics and industrial waves on CAR. The standard errors are clustered at the state and industry level. Panel A presents the correlation between Combined [ 2, 2] CAR, a proxy for market s expectation to deal synergies, and UI benefits. Column 1 regress Combined [ 2, 2] CAR on UI benefits while column 2 to column 4 focus on the effects of UI benefits and labor intensity, high-layoff industry, and high union coverage rate interactions 13. After controlling for acquirer characteristics, target firm characteristics, and deal-specific characteristics, we find positive correlation between UI benefits and Combined [ 2, 2] CAR. According to the results in column 1, a 10% increase in UI benefits lead to 34 basis point increase in Combined [ 2, 2] CAR, which equals to around 12% increase relative to the average Combined [ 2, 2] CAR. Column 2 focuses on UI benefits Target Firm Workforce interactions. The positive coefficient shows that the positive correlation between Combined [ 2, 2] CAR and UI 13 Following Agrawal and Matsa (2013), we identify a high-layoff-rate industry indicator equal to one if an industry has above-the-median layoff rate. 21

benefits is stronger when target firms have more employees. This result strengthens our cost-saving argument, suggesting that more cost savings can generate from high labor intensive firms. Column 3 focuses on UI benefits High Layoff Rate Industries interactions. The insignificant coefficient show that target firms in high layoff industries bring no more cost savings. Column 4 focuses on UI benefits High Union Coverage Rate interactions. High union coverage rate means stronger employee bargaining power, which is normally hard to lay off employees. The positive correlation between the interaction term and Combined [ 2, 2] CAR shows that higher UI benefits can bring more synergies when target firms have stronger employee bargaining power. Following Agrawal and Matsa (2013), Column 5-8 repeats the regression in column 1-4 but excludes industries that are more likely to be geographically dispersed, that is, retail, wholesale, and transport industries 14. Because we use firm headquarter state as the proxy for firm operating locations, those dispersed industries are likely to have more measurement errors. In all specifications, the coefficient remains statistically significant and increases when we exclude potential measurement errors. To better show the effect of UI benefits on acquirers incentive to make deals. Panel B presents the correlation between UI benefits and acquirer s announcement return. The observation increases because we only include acquirer s firm characteristic and deal characteristic control variables. Consistent with our hypothesis, the correlation between UI benefits and acquirers CAR are statistically significant in all specifications. B. Post-Merger Effects 14 Industries with two-digit SIC code = 40-47; 50-59 22

Then, we test if higher UI benefits ultimately facilitate post-acquisition layoff. To observe deal-specific layoff, we keep acquirers that engage in only one M&A deal in a fiscal year and no deals in the following two years. This step leaves us 1,264 completed deals. For each deal, post-acquisition layoff equals to acquirer s employee size two years after deal completion, deduct the sum of acquirer and target firm s employee size at one year before deal announcement. Since employee number observation is limited for target firms, we only keep deals that exhibit layoffs. We regress post-acquisition layoff on target state UI benefits. Acquirer state UI benefits could be the same as target state UI if deal is within-state. To alleviate the concern that results are driven by acquirer state UI benefits, we add additional acquirer state UI benefits and acquirer state control variables. The results are presented in Table 7. Column 1 and column 2 show that when target firms operate in states with higher UI benefits, post-acquisition layoff size is larger. The results are statistically significant at 5% level. If higher UI benefits facilitation post-acquisition layoff, unit restructuring costs should be lower. Column 3 and column 4 show replace postacquisition employee change with restructuring costs. After controlling for employee size both before and after acquisition, the negative significant correlation between restructuring cost and UI benefits show that UI benefits could increase deal synergy through facilitating post-acquisition restructuring. VII. Robustness Tests In this section, we perform several additional tests to further support our argument. Firstly, if state UI benefits can increase a firm s likelihood of being acquired, in state level we should also observe the same effect. We show that in aggregated state level, 23

higher UI benefits lead to more firm being acquired in the next year. Higher UI benefits also lead to an increased fraction of within industry acquisition. Then, we test if results hold when we extend the acquisition sample to all initiated deals. When search for target firms, acquirers generally consider the likelihood of success and post-acquisition synergies. Therefore, if the increased UI benefits will increase a firm s likelihood of being acquired, it should also lead to increased likelihood of deal initiation. Next, we check whether the effects of UI benefits remain significant after match target firms with firms with similar characteristics. Finally, we test if the unemployment benefits of acquirer employees have impact the acquiring firms engage in M&A activities. Although target firm employees experience higher unemployment risks, acquirer employees also need to face increased unemployment risks. Therefore, acquirer may also bound with wage unemployment risks equilibrium and cannot making merger activities. We also perform other robustness checks to further validate our main results. The positive correlation holds if we replace linear probability model with profit model. We also rule out an alternative explanation that the results are driven by right-to-work law. Only three states passed the law within our sample period: Michigan, Indiana, and Oklahoma. In these three states, the results hold after including right-to-work law indicator as additional control variable in the regression. Two additional checks also validate our argue that firms are more likely to accept takeover when UI benefits decreass workers sensitivity to unemployment: the correlation is stronger if firms have limited local labor supplies, and the positive correlation only valid in firms with high employee treatment index. 24

A. State-Level Acquisition Ratio We validate our argument by testing the effect of UI benefits on the overall activity of the takeover market. If an increase in UI benefits can increase the likelihood of a firm operating in that state being acquired, in aggregate we should be able to observe a higher portion of firms in that state being acquired. Because each state has discretion in choosing changes in the amount of UI benefits change, every state can become treatment group in years with large change in UI benefits, and control group otherwise. For each state, we calculated the acquisition ratio, using the number of public target firms standardized by total public firms available in Compustat each year. For states without public target firms in a specific year, we replace the acquisition number with zero to balance panel data. Table 8 column 1 presents regression analysis of the effect of UI benefits amount on the acquisition likelihood. The coefficients of UI benefits are statistically significant at 5% level, suggesting in aggregate, an increase in state UI benefits can place positive effect on the likelihood of a firm being acquired. Economically, a one-standard-deviation increase in the logarithm of UI benefits leads to 37.5% increase in the likelihood of a firm being acquired. We also test the ratio of within industry acquisition in state level. If variation in cost of unemployment affects the likelihood of all takeovers, it should also influence within industry acquisition, which is subject to more unemployment risks to workers. We define an acquisition as within industry acquisition if acquirer and target firms come from the same 3-digit SIC industry 15. The number of within industry acquisition is standardized by total acquisition numbers in each state-year. Table 8 column 2 presents 15 Coefficient and significance level are quite similar if we use Fama-French 48 industry. 25

the correlation between UI benefits and within industry acquisition ratio. Consistently, there are significant positive correlation between UI benefits and within industry acquisition. These results support our hypothesis that decreased unemployment risks increase the likelihood if a firm being acquired. B. UI benefits and Deal Initiation We show that our results hold when we incorporate incomplete deals. In this case, we are testing the correlation between the UI benefits and the likelihood of being targeted. We use a sample with all deal transactions, including withdrew deals. The results are presented in Table 9. Column 1 only regress the likelihood of being targeted on the UI benefits while column 2 adds both state economic control variables, and column 3 and 4 add firm characteristic control variables. The coefficient of UI benefits is impacted little by including or excluding additional control variables. Similar with baseline results: there is significant positive correlation between UI benefits and the likelihood of being targeted. This finding suggests that acquirers will take the target s worker unemployment risks into consideration when choosing target firms. C. Matched Pseudo Targets We investigate all else equals, whether acquirers prefer to takeover firms with higher unemployment benefits. Similar to Rhodes-Kropf and Robinson (2008), we match each actual target firm by selecting pseudo target firms in the same 48 Fama-French industry, the same market capitalization and book-to-market ratio quintiles in the same year from CRSP-Compustat merged table. However, pseudo target firms may not come from the same state from actual target firms, thus their UI benefits are different. Unlike Rhodes- Kropf and Robinson (2008), we fix the acquirers because we are interested in the effect of 26

target firms UI benefits. We identify the actual target firm as treatment group with indicator dependent variable equals to 1 and the pseudo target firms as 0. On average, each actual target firm is paired with 8 pseudo target firms. Apart from firm-specific characteristic, we also include the geographical distance between acquirers and target firms (including actual target firms and pseudo target firms) as additional control variable. Table 10 presents our results for this analysis. Column 1 and 2 present the correlation with firm fixed effects and year fixed effects. According to the parameter in column 1, a 10% increase in UI benefits increase the mean likelihood of being acquired by 14%. Following Agrawal and Matsa (2013), column 2 repeats the estimation in column 1 but exclude industries that are more likely to be geographically dispersed, that is, retail, wholesale, and transport industries. In those industries, a firm s operation state may different from its headquarter state. This proxy error could attenuate our results. Consistent with expectation, after excluding those industries, the coefficient of UI benefits increased. Therefore, the regression results suggest that the likelihood of being acquired indeed increases after eliminating wide dispersed industries. Consistent with information asymmetry theory, longer distance between acquirer and target firms decrease the likelihood of takeover (Chung, Green, and Schmidt, 2016). Moreover, the positive coefficients of UI benefits Geographical Distance suggest that acquirers would give up certain geographical-related information advantages when a firm operate in a state with higher UI benefits. Column 3 and 4 replace firm fixed effects with state fixed effects. The interaction term is insignificant in these specifications, suggesting when UI benefits are same, acquirers won t bother to acquire further located 27

firms. Those evidences collectively indicate that acquirers do value UI benefits when choosing target firms. C. UI benefits for Acquirer Workers Although workers in acquirers generally face less unemployment risks than workers in target firms, they also face risks of reduce job benefits in post-acquisition restructuring. Therefore, unemployment cost for workers in acquirers would also impact the likelihood of engaging in takeover activities. In this robustness test, we regress acquirer worker s UI benefits on the indicator for acquisition 16. The results are presented in Table 11. Column 1 - Column 3 include all M&A deals while Column 4 - Column 6 present results for within industry deals. There is statistically significant correlation between UI benefit and likelihood of engaging in takeover activities in both cases. D. Other Robustness Checks Apart from linear probability model, we also test the effect of UI benefits on the likelihood of being acquired using probit model. Because probit model is subject to more bias under high-dimensional fixed effects. We include state fixed effects and year fixed effects in the regression. The results are presented in table 12. The coefficients of UI benefits both include and exclude dispersed industries are statistically significant at 1% level. According to the coefficient in column 1, a 10% increase in UI benefits lead to 50 basis points increase in the likelihood of being acquired, meaning a 15% increase in the likelihood of being acquired. We exploit the effect of state variation in the legal treatment of employee rights on the likelihood of being acquired. Labor unions have weaker bargaining power if a state pass 16 Indicator that acquirers successfully complete a takeover deal. 28

the right-to-work statue (Matsa, 2010). John, Knyazeva and Knyazeva (2015) find that acquirers that have weaker employee bargaining power experience higher announcement returns. Therefore, if it possible that lower union bargaining power increases a firm s likelihood of being acquired. Because most states with the right-to-work law passed this statue on or before 1990, only three states are relevant in this test. That is, Oklahoma (2001), Indiana (2012), and Michigan (2012). We test the correlation between UI benefits and the likelihood of being acquired in these three states with right-to-work law indicator (equals to 1 after a state passed this statue and 0 otherwise) as additional control variable. The results are presented in table 13. We find that the coefficient of UI benefits remains significant. Therefore, we exclude the alternative explanation that the positive correlation between UI benefits and a firm s likelihood of being acquired is driven by the right-towork law. We find that the effects of UI benefits are stronger when local labor supply is low. If managers in the targeted firm choose to decline takeover offers when workers have high costs of unemployment, and failed to do so may suffer increased human capital costs, this tendency will be weaken when local labor supply is abundant. Therefore, we use the population of the county where a firm s headquarter locates as the proxy for local labor supply. The results are presented in table 14. We find that UI benefits High Local Labor Supply interaction is negatively correlated with the likelihood of being acquired. We also use employee treatment index as the proxy of manager s caring about employees. Employee treatment index is constructed following Bae, Kang, and Wang (2011). Employee treatment index sums up a 0/1 rating in five employee relations category from KLD database: union relations, cash profit-sharing, employee involvement, 29

retirement benefits strength, and health and safety strength. The employee treatment index ranges between zero and five. Higher employee treatment index indicates better employee treatment. We identify a firm have high employment treatment index if a firm s total score is above the sample median. The results are presented in table 15. The first two columns include UI benefits High employee treatment index interaction. This interaction is positively correlated with a firm s likelihood of being acquired and statistically significant at 1% level (5% level after excluding dispersed industries), suggesting the effects of UI benefits and the likelihood of being acquired is only valid in firms that caring about employees. We then divide the sample into a low employee index subsample and a high employee index subsample. The results are consistent with the full sample regression: the positive correlation between UI benefits and the likelihood of being acquired is only valid in high employee index firms. VIII. Conclusion This paper investigates the effect of state unemployment benefits on the likelihood of being acquired and outcomes of takeover activities. From acquirer s perspective, higher unemployment benefits for workers in target firm could facilitate post-acquisition layoff and reduce restructuring costs, thus lead to higher deal synergies. From target firm s perspective, in labor contract negotiation, firm choose an optimal risk level that includes trade-off between human capital costs and benefits of risky projects. Therefore, firms have responsibility to maintain their choice of risk level in exchange for lower human capital costs. Changes in state law and state union coverage rates provide relative more 30

exogenous evidence on the role of unemployment risk on firm M&A activities. Empirically, we find that exogenous changes in state level labor unemployment benefits can affect the likelihood of firms in that state being acquired and stock market reaction to the deal announcement. Consistent with the argument of implicit contract in which labor requires wage unemployment risk equilibrium, we find that increases in state level labor unemployment benefits are associated with a significant increase in the likelihood of firms being acquired. The effect is stronger for firms in industries with high historical unemployment rates and for firms with larger employee size. Moreover, we find that labor unions and stakeholder-focused managers both contribute to the positive correlation, indicating the fulfilment of implicit contract also depends on labor bargaining power and manager s care about employees. The main effects continue to hold after a number of additional tests. The results are robust to inclusion of state economic controls and firm characteristics, and unobservable time-invariant state variation. In addition to the increased likelihood of being targeted, we also find that increased UI benefits also lead to higher announcement return, suggesting market s positive reactions to acquirer targets in which workers have less resistance. More broadly, the findings illustrate that companies are prevented from being acquired partly because of workers exposure to unemployment risk. Though lacking of decision-making rights, workers still present their influence on firm s strategic decisions through requirement in implicit labor contracts. The evidence emphasizes that labor market frictions are an important feature of the corporate environment. Reducing acquisition activities is one of many ways for firms to mitigate workers unemployment risk. Other ways, such as adopting lower leverages (Agrawal and Matsa, 2013), taking 31

less risky project (Hennessy and Whited, 2005) or reducing workers losses in distress by redesigning job tasks to require fewer firm-specific skills (Jaggia and Thakor, 1994) can reduce the workers exposure to unemployment risks. Labor s impact on firm s policy choices and firm valuations are documented by many labor and finance studies. This paper also provides evidence that the reason why unions could place impact on firms is possible through the implicit requirement in labor contract. However, due to unobservability of implicit labor contract, there lacks direct evidence that how labor could influence manager s decision makings. Exploring the channels through which labor place impact on corporate policies is an important area for future research. 32

References Acharya, V., Baghai, R., and Subramanian, K., 2014. Wrongful discharge laws and innovation. Review of Financial Studies 27, 301-346. Abowd, J., Ashenfelter, O., 1981. Anticipated unemployment, temporary layoffs, and compensating wage differentials. In: Rosen, S. (Ed.), Studies in Labor Markets. University of Chicago Press, Chicago, IL, pp. 141 170. Addessi, W., Busato, F., 2009. Fair wages, labor relations, and asset returns. Journal of Financial Stability 5, 410 430. Agrawal, A., Matsa, D., 2013. Labor unemployment risk and corporate financing decisions. Journal of Financial Economics 108(2), 449 470. Aktas, N., Bodt, E., Cousin, J., 2011. Do financial markets care about SRI? Evidence from mergers and acquisitions. Journal of Banking and Finance 35, 1753 1761. Atanassov, J., Kim, E. H., 2009. Labor and Corporate Governance: International Evidence from Restructuring Decisions. The Journal of Finance 64(1), 341-374 Bae, K., Kang, J., Wang, J., 2011. Employee treatment and firm leverage: a test of the stakeholder theory of capital structure. Journal of Financial Economics 100, 130 153 Berger, P., Ofek, E., Yermack, D., 1997. Managerial entrenchment and capital structure decisions, The Journal of Finance 52, 1411 1438. 33

Berk, J., Stanton, R., Zechner, J., 2010. Human capital, bankruptcy, and capital structure. The Journal of Finance 65, 891-926. Bertrand, M., Schoar, A., Thesmar, D., 2007. Banking deregulation and industry structure: Evidence from the French banking reforms of 1985. The Journal of Finance 62(2), 597 628. Bertrand, M., Duflo, E., Mullainathan, S., 2004. How much should we trust differences-indifferences estimates?. The Quarterly Journal of Economics 119(1), 249 275. Blaustein, S. J., 1993. Unemployment insurance in the United States: the first half century. With Wilbur J.Cohen and William Haber. W. E. Upjohn Institute for Employment Research, Kalamazoo, MI. Bradley, D., Kim, I., Tian, X., 2016. Do Unions Affect Innovation?. Management Science, Forthcoming Brown, J., Matsa, D., 2012. Boarding a sinking ship? An investigation of job applications to distressed firms. The Journal of Finance 71(3), 507-550. Chaplinsky, S., and Niehaus, G. R., 1994. The Role of ESOPs in Takeover Contests. The Journal of Finance 49, 1451-70. Chemmanur, T.J., Cheng, Y., and Zhang, T., 2013, Human capital, capital structure, and employee pay: An empirical analysis, Journal of Financial Economics 110, 478-502. Chen, H. J., Kacperczyk, M., Ortiz-Molina, H., 2011. Labor unions, operating flexibility, and the cost of equity. Journal of Financial and Quantitative Analysis 46 (1), 25 58. 34

Chen, H. J., Kacperczyk, M., Ortiz-Molina, H., 2012. Do non-financial stakeholders affect agency costs of debt? Evidence from unionized workers. Review of Finance 16 (2), 347 383. Chung, K., Green, T. and Schmidt, B., 2016, CEO Home Bias and Corporate Acquisitions Deng, X., Kang, J., Low, B., 2013. Corporate social responsibility and stakeholder value maximization: Evidence from mergers. Journal of Financial Economics 110(1), 87-109. Dessaint, O., Golubov, A., and Volpin P., 2016. Employment protection and takeovers. Journal of Financial Economics, Forthcoming Devos, E., Kadapakkam, P., and Krishnamurthy, S., 2009, How do mergers create value? A comparison of taxes, market power, and efficiency improvements as explanations for synergies, Review of Financial Studies 22, 1179-1211. Edmans, A., 2012. The Link Between Job Satisfaction and Firm Value, With Implications for Corporate Social Responsibility. AcadeOur of Management Perspectives 26(4), 1 19. Farber, H. S., 2005. What do we know about job loss in the United States? Evidence from the displaced workers survey, 1984-2004. Federal Reserve Bank of Chicago Economic Perspectives 29, 13 28. Franks, J., and Mayer, C., 1998, Bank control, takeovers, and corporate governance in Germany, Journal of Banking and Finance 22, 1385 1403. Freeman, R. 1984. Strategic Management: A Stakeholder Approach. Cambridge University Press Gibbons, R., Katz, L. F., 1991. Layoffs and lemons. Journal of Labor Economics 9, 351 380. 35

Gormley, T. A., Liu, H., Zhou, G., 2010. Limited participation and consumption-saving puzzles: a simple explanation and the role of insurance. Journal of Financial Economics 96, 331 344. Gruber, J., 1997. The consumption smoothing benefits of unemployment insurance. American Economic Review 87, 192 205. Harford, J., 2005. What drives merger waves?. Journal of Financial Economics: 77(3): 529-560 Heckman, J., 1976. The common structure of statistical models of truncation, sample selection and limited dependent variables and a simple estimator for such models. Annals of Economic and Social Measurement 5: 475 492. Heckman, JJ., 1979 Sample selection bias as a specification error. Econometrica 47:153 161. Hennessy, C. A., Whited, T. M., 2005. Debt dynamics. Journal of Finance 60, 1129 1165. Hirsch, B.T., Macpherson, D.A. and Vroman, W.G., 2001. Estimates of Union Density by State. Monthly Labor Review 124(7), 51-55. Houston, Joel F., James, C.M., and Ryngaert, M.D., 2001, Where do merger gains come from? Bank mergers from the perspective of insiders and outsiders, Journal of Financial Economics 60, 285-331. Jaggia, P. B., Thakor, A. V., 1994. Firm-specific human capital and optimal capital structure. International Economic Review 35, 283 308. Jensen, M., Meckling, W., 1976. Theory of the firm: managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3 (4), 305 360 36

John, K., Knyazeva, A., Knyazeva, D., 2015. Employee rights and acquisitions. Journal of Financial Economics, 118(1), 49 69. Katz, L. F., Meyer, B. D., 1990. The impact of the potential duration of unemployment benefits on the duration of unemployment. Journal of Public Economics 41, 45 72. Kim, E. H., Maug, E. G., Schneider, Christoph., 2016.Labor Representation in Governance as an Insurance Mechanism. European Corporate Governance Institute (ECGI) - Finance Working Paper No. 411/2014. Kiplinger California Letter, 2001. Prospects for several California bills. Kiplinger Washington Editors, Inc., 37(16). Klasa, S., Maxwell, W. F., Ortiz-Molina, H., 2009. The strategic use of corporate cash holdings in collective bargaining with labor unions. Journal of Financial Economics 92: 421-42. Lancaster, T., 2000. The incidental parameters problem since 1948, Journal of Econometrics 95, 391-414. Larsson, R., Finkelstein., S., 1999. Integrating Strategic, Organizational, and Human Resource Perspectives on Mergers and Acquisition: A Case Survey of Synergy Realization. Organizational Science,10(1): 1-26. Lee, D. S., Mas, A., 2012. Long-Run Impacts of Unions on Firms: New Evidence from Financial Markets, 1961-1999. The Quarterly Journal of Economics 127(1), 333 378. 37

Li, X., 2013. Productivity, restructuring and gains from takeovers. Journal of Financial Economics 109:250 271 Maksimovic, V., Phillips, G., Prabhala, N. R., 2011, Post-merger restructuring and the boundaries of the firm, Journal of Financial Economics 102, 317-343. Malmendier, U., Tate, G., 2008. Who makes acquisitions? CEO overconfidence and the market s reaction. Journal of Financial Economics 89:20 43. Masulis, R. W., Wang, C., Xie, F., 2007. Corporate Governance and Acquirer Returns Corporate Governance and Acquirer Returns. Journal of Finance 62(4), 1851 1889. Masulis, R. W., Wang, C., 2016. Employee-Manager Alliances and Shareholder Returns from Acquisitions. Unpublished Working Paper. Masulis, R.W., Simsir, S. A., 2015. Deal Initiation in Mergers and Acquisitions. FIRN Research Paper. Matsa, D., 2010. Capital structure as a strategic variable: evidence from collective bargaining. Journal of Finance 65 (3), 1197 1232. McWilliams, A., Siegel, D., 2001. Corporate Social Responsibility: A Theory of the Firm Perspective. The Academy of Management Review 26(1), 117-127. Moeller, S. B., Schlingemann, F. P., Stulz, R. M., 2004. Firm size and the gains from acquisitions. Journal of Financial Economics 73(2), 201 228. 38

Oakland Tribune, 2002. Davis rallies labor, signs bills, raises cash in Oakland. J. Richman, September 22. Pagano, M., Volpin, P. F., 2005. Managers, workers, and corporate control. The Journal of Finance 60(2), 841 868. Pagano, M., Volpin, P. F., 2008. Labor and Finance. Working Paper, London Business School, (June). Rhodes-Kropf, M., Robinson, D. T., 2008. The market for mergers and the boundaries of the firm. The Journal of Finance 63, 1169 1211. Song, M. H., Walkling, R. A., 2000. Abnormal returns to rivals of acquisition targets: A test of the acquisition probability hypothesis. Journal of Financial Economics 55(2), 143 171. Tian, X., Wang, W., 2015. Hard marriage with heavy burdens: Labor unions as takeover deterrents. Working paper. Tirole, J., 2001, Corporate governance, Econometrica 69, 1 35. Titman, S., 1984, The effect of capital structure on a firm's liquidation decision, Journal of Financial Economics 13, 137-151. Topel, R. H., 1983. On layoffs and unemployment insurance. American Economic Review, 73(4), 541 559. Topel, R. H.,1984. Equilibrium earnings, turnover, and unemployment: new evidence. Journal of Labor Economics, 2(4), 500 522. 39

Wooldridge, J., 2005. Introductory economics: a modern approach. South Western College Publishing 40

Figures and Tables Figure 1: UI Benefits Increases and Acquisition Ratio This figure plots the positive correlation between UI benefits increase and acquisition ratio across all states in each year. Acquisition ratio is the number of public firms being acquired in each year in each state, divided by a state s total public firms available in CRSP. UI benefits increases are the annual UI benefits changes ratio in each state. The Y-axis is the average acquisition ratio and the X-axis is the average UI benefits changes of all states in each year. Each dot on the graph represents a given year s UI benefits increases and the Acquisition Ratio. 41

Figure 2 Time-Trend UI Benefits Increases and Acquisition Ratio This figure plots the co-movement of UI benefits changes and acquisition ratio in four states. Acquisition ratio is the number of public firms being acquired in each year in each state, divided by a state s total public firms available in CRSP. UI benefits increases are the annual UI benefits changes ratio in each state. 42