مهار کردن جانبداری در مذاکرات حکم بازرسی برنامه کمک به دارایی های مشکل(TARP)
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|11017||2012||14 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Economics and Business, Volume 64, Issue 1, January–February 2012, Pages 63–76
This paper finds that banks that offered lower opening bids were rewarded with significantly lower warrant repurchase prices in transactions that raised $2.856 billion in 2009. These results were scaled by third-party consultants’ and the Congressional Oversight Panel's estimates of the warrants’ value. In contrast to the experimental psychology studies on anchoring bias in negotiations, these are real transactions involving large sums of money. This paper finds that larger banks paid significantly higher prices after controlling for other factors, and the U.S. Treasury obtained better prices over time. The results on anchoring bias are strong even after controlling for bank managers’ potential informational advantages over U.S. Treasury negotiators.
The Troubled Asset Relief Program (TARP) left taxpayers holding large stakes of preferred stock, subordinated debt, and warrants in hundreds of banks. This program allowed the U.S. Treasury Secretary to purchase up to $700 billion worth of capital and assets during the financial panic of 2008 and 2009. Blinder and Zandi (2010) argues that the combined financial sector rescue efforts by the Federal Reserve, U.S. Treasury by way of the TARP, and the Federal Deposit Insurance Corporation (FDIC) increased GDP by more than 2.5 percent in both 2009 and 2010. That study also projects that these financial sector rescues have reduced or will reduce unemployment by more than 2.5 percent in 2009, 2010, and 2011. Yet, Wilson and Wu (2011) documents that as soon as the TARP capital infusions were made many of the healthiest TARP recipients clamored to exit government ownership. The present study looks into the negotiation process of banks trying to end the last vestiges of public ownership of their firms. In the Emergency Economic Stabilization Act (EESA), which was signed into law on October 3, 2009, the U.S. Treasury was required to obtain warrants from banks benefiting from the government's assistance. Warrants are call options that give the owner the right but not the obligation to purchase newly issued shares at a preset price. According to Paulson (2010, p. 305) and Swagel (2009), this was one of the many deals that the U.S. Treasury made to win Congressional approval of the unpopular legislation that has often been referred to as the $700 billion bailout. The Capital Purchase Program (CPP) announced by Secretary of the U.S. Treasury Henry Paulson on Monday, October 13, 2008, proposed to inject up to $250 billion into U.S. banks. On its first day of the program, it passed out $125 billion in exchange for perpetual preferred stock and warrants that expire in ten years. The U.S. Treasury (2010) reports that the Capital Purchase Program closed for new investments on December 31, 2009, after passing out roughly $205 billion to 707 banks. Of those 707 banks, the author's analysis of transaction reports issued by the Office of Financial Stability of the U.S. Treasury found that 282 banks, insurance companies, and credit card issuers had issued warrants that allow those firms to negotiate with the U.S. Treasury to repurchase them at “fair market value.”1 Banks can only repurchase the warrants after they repaid the taxpayers’ investments and any remaining accrued dividends. By the end of 2009, fifty-two banks had repurchased the taxpayers’ preferred stock. Yet, through the end of 2009, only thirty-four negotiations had either concluded with banks repurchasing the warrants from the U.S. Treasury or the U.S. Treasury auctioning the warrants to third party investors. This paper studies the bargaining process of the twenty-eight warrant repurchases which had multiple bids through the end of 2009. It finds that the first bid significantly positively correlated with the subsequent prices paid for the warrants. Thus, banks that made lowball offers appear to have been rewarded with lower sales prices. This lends support to the anchoring bias theory, which says that prices in economic transactions can be affected by irrelevant early signals.2 The results are robust for controls that measure management's information advantage. Moreover, this study finds that the U.S. Treasury secured significantly better prices over time and from larger banks. These results provide some support for the hypothesis that the U.S. Treasury learned how to secure better deals over time and responded to political pressure to press for better deals after the early deals were criticized by Wilson (2009b) and then Congressional Oversight Panel (2009b). Nevertheless, the time trend results disappear when the deal prices are scaled by the Congressional Oversight Panel (2010)’s estimates. This study supports the anchoring literature in psychology. Tversky and Kahneman (1974) identify anchoring bias as one of common mistakes or heuristics to which people fall prey when making judgments under uncertainty. With anchoring bias, a participant's assessment of the value of an item is influenced by seemingly irrelevant pieces of information. For example, Chapman and Bornstein (1996) find that mock juries were more likely to see the merits of the plaintiff's complaint when the damages requested were higher. Ariely (2008, pp. 25–27) describes an experiment in which MBA students were asked to write down the last two digits of their social security numbers on a piece of paper. They then were asked to bid on items ranging from keyboards to bottles of wine. The students with the last two digits of their social security number between 80 and 99 bid the highest, and the students with the last two digits of 01 and 20 bid the lowest. There was often a wide gap between the two groups’ valuations. Anchoring bias has been demonstrated to exist in bargaining experiments. In particular, the first offer in a negotiation significantly affects the final price. This result was first uncovered by Liebert, Smith, and Hill (1968) who had students play bargaining games with each other. When the buyer made higher initial offers, the buyer paid significantly higher prices. This effect was most pronounced in the games where the buyer's reservation price was unknown to the seller. Galinsky and Mussweiler (2001) conducted three studies of MBA students who bargain as part of a class assignment for hypothetical settlements. That study affirms that anchoring bias leads to the negotiator's making the first offer to obtain better outcomes. Yet, Galinsky and Mussweiler (2001) find that anchoring bias disappears when the negotiators were instructed to focus on their opponents’ best alternative to a negotiated agreement (BATNA) or when the negotiators focus on their goals for the negotiation. The anchoring effect persists when the students were told to focus on their worst case scenario, their own BATNA. Northcraft and Neale (1987) found evidence of anchoring bias when experimental subjects, MBA students and real estate agents, toured houses with altered list prices. Thus, Northcraft and Neale (1987) confirm that anchoring bias is not just confined to the “sterile” laboratory setting. Yet, the participants as in Northcraft and Neale (1987) were not playing with their own money. They were not actually buying the houses. In contrast, this study involves real transactions. The U.S. Treasury raised about $2.856 billion from the twenty-eight warrant deals studied here! Yet, anchoring bias in the negotiated agreements of the TARP warrant is strongly supported by the regression analysis in this paper. In contrast to the present paper, which focuses on the bargaining process, studies of the Troubled Asset Relief Program (TARP) warrants have focused on valuation issues. Since Black and Scholes (1973) and the binomial model of Cox, Ross, and Rubinstein (1979), we have had a good methodology for valuing call options. Yet, there are unobservable parameters relating to the stock price's future standard deviation, volatility, and future dividends. This means that there is scope for disagreement as to the “fair market value” of options which are not traded. Wilson (2009a) argues that the long-dated warrants can be reliably valued using the adjustments to Black and Scholes (1973) for dividends by Merton (1973) and dilution by Galai and Schneller (1978). Congressional Oversight Panel (2009b) largely adopted Wilson, 2009a and Wilson, 2009b’s methodology and argued that the U.S. Treasury obtained only 66 percent of fair market value in the first eleven negotiated repurchases of the TARP warrants. Jarrow (2009) outlines the various option pricing methods used at the U.S. Treasury and discusses how transaction costs may affect the U.S. Treasury's and the banks’ estimates of fair market value. Wilson (2010) argues that, prior to the warrant auctions in December 2009, there were no call options or publically traded warrants that expired after 2014. Thus, using implied volatilities of traded warrants or call options of similar expirations to estimate volatility was not possible for the TARP warrants, which expire in 2018 and 2019, before the December 2009 auctions. Thus, prior to the auctions, the volatility of the stock price as estimated from traded options of a similar expiration was not available. That paper discusses using the secondary market trading of the publically traded TARP warrants to estimate the market values of other TARP warrants that have not been offered to investors. In Section 2, the data is explained. In Section 3, the results are discussed. In Section 4, we test for management's informational advantage. We find little evidence that banks repurchasing TARP warrants outperformed banks that rejected warrant deals in 2009. Moreover, the strength of the first bid in explaining deal prices is unaffected by measures of asymmetric information. Finally, in Section 5, the paper concludes.
نتیجه گیری انگلیسی
This study indicates that banks that started out offering lower prices in warrant negotiations with the U.S. Treasury paid lower prices. According to model 3 in Table 2, a bid that was 50 percent of a third-party consultant's opinion of fair market value, would lead to a 5.68 percent lower price than an opening bid of 60 percent of fair market value. (In other specifications, the anchoring bias effect was even stronger!) Unlike laboratory studies involving small sums of money, the twenty-eight negotiations studied led to nearly $2.9 billion changing hands. Thus, anchoring bias appears to be important for big negotiations as well as small ones. The U.S. Treasury's apparent susceptibility to anchoring bias could cost U.S. taxpayers hundreds of millions—if not billions—of dollars. Since banks make the first bid and all subsequent bids until acceptance in the TARP warrant negotiations, anchoring bias favors banks at the expense of taxpayers. These results of anchoring bias are strong even when controls for the presence of the asymmetric information are introduced. The data in the current study are, of course rely on correlation, not controlled experiments. The author did not randomly manipulate initial bids. Thus, no claims can be made about a causal relationship between initial bid and final purchase price. In order for that relationship to be considered an example of the anchoring bias, it would have to be the case that the initial bid is irrelevant information, unrelated to the true value of the warrant. Although banks are free to make any initial bid they like, it is possible that initial bids carry information about the true worth of the warrant. That is, initial bid and final purchase price may both be driven by other information about the warrant's value. I addressed this possibility by scaling bids and purchases prices by an objective evaluation of worth and by statistically controlling for asymmetric information held by bank managers but not the Treasury Department. Even after these statistical controls, the relationship between initial bid and final purchase price remained robust, suggestive of an anchoring effect. These results indicate that the U.S. Treasury could have obtained better prices for the TARP warrants if its negotiators were not swayed by the bids of the banks. The U.S. Treasury negotiators should follow the advice of Galinsky and Mussweiler (2001) about focusing not on the offer but on the alternatives to a negotiated settlement. These taxpayer representatives should ignore these irrelevant anchors and focus on the valuations of their advisors. Very little significance should be placed on the opening bid. The TARP warrants have value both as securities that can be sold to third party investors and as options that can be exercised at some point in the future. In short, taxpayers may have better options than any lowball initial offer by the warrant issuer may suggest. This study also found evidence that the prices as a percent of the third-party consultant's estimates fair market value trended upward over the period, albeit at a decreasing rate. This is consistent with the theory that the U.S. Treasury learned how to get better deals as time went on. This finding is also consistent with the hypothesis that a combination of the press coverage, the Congressional Oversight Panel (2009b)’s report, criticism from studies such as Wilson (2009b), and Congressional involvement pressured U.S. Treasury negotiators and bankers to make more lucrative deals for taxpayers. Ironically, any time trend effects cannot find any support in regressions scaled by the Congressional Oversight Panel (2010)’s estimates of the TARP warrants agreements. Thus, we should exercise some caution before we conclude that the U.S. Treasury has gotten better deals for taxpayers over time. Yet, there is evidence that larger banks paid a higher percent of the third party constants’ and Congressional Oversight Panel's (COP) estimates of fair market value. This also could be due to the greater political pressure on the larger banks.