Many of the ten big banks repaid their TARP preferred stock loans on Wednesday, June 17, 2009. This means that they can begin negotiations to repurchase the warrants that they issued to taxpayers. I estimate that the warrants issued by those banks are worth $4.2 billion. That is over one third of the value of the taxpayers' warrants. Yet, if history is any guide, we taxpayers could expect to get 50 cents or less on the dollar. My hope is that the U.S. Treasury will take a hard line in these negotiations and take steps to create markets in the TARP warrants. The best prices will come from taxpayers selling their warrants to third party investors.
One of the problems that plagues the valuation of the TARP warrants is the cancellation provisions. Apparently, to encourage banks to raise equity on their own, the purchase agreements for the TARP capital infusions say that half the taxpayers’ warrants will be cancelled if the bank issues equity in the amount of the U.S. Treasury's investment. Recent valuations by Credit Suisse and Bloomberg have assumed a zero chance of cancellation. These studies systematically overvalue the taxpayer's warrants for this reason. Moreover, they most overvalue the banks which are the most likely to cancel half the TARP warrants. Of the 10 big banks due to pay back TARP yesterday, Morgan Stanley (NYSE:MS) and Goldman Sachs (NYSE:GS) have such high share prices and have already issued so much equity that they have a 90 and 80 percent chance respectively of issuing enough equity to cancel half the taxpayer’s warrants by year end, according to my estimates. That means that both Bloomberg and Credit Suisse are overvaluing the taxpayers’ warrants in these companies by 45 to 40 percent, if you believe their estimates of volatility and dividend yield.
In my paper, “A Model for Estimating the Cancellation Probabilities of TARP Warrants”, I use the cost of equity capital, the stock price volatility, the current stock price, previous equity issues, and the future dividends to estimate the chances that the bank will find it in its economic interest by years’ end to issue enough equity to cancel half the warrants given it does not repurchase the TARP warrants by that time. Those probabilities and valuations of the TARP warrants of 68 of the largest banks, including the 10 banks due to repay TARP, are in Tables 2 and Tables 3 of my paper. I also list all the inputs of the option pricing model used in those tables. I estimate that these 68 banks have warrants worth approximately $11 billion.
Issuing equity leads to modest dilution of the share price, not because the number of shares is going up. Instead, it leads to modest falls in the share price because fees are paid to investment bankers (typically 5 percent of the issue amount) and the new stock is underpriced (on average about 2 to 3 percent) to encourage rapid purchase of the issue. Since the new number of shares is usually less than the number of shares outstanding this 8 percent transaction cost often leads to a one percent drop in the share price even if the number of shares issued is 10 percent or more of the shares outstanding. These issuing costs have to be weighed against the potential gains of cancelling half the taxpayer’s warrants. It is probably a wise move for rational managers to postpone the equity issuance decision to year end. Nevertheless, already I have identified seven banks have cancelled half the taxpayers warrants. I expect more of the 270 plus publically traded banks to follow in the coming months. I think there is nothing wrong with banks doing this. The government must honor the contracts it signs just like bank CEOs must honor the agreements that they sign.
Disclosure: This is not investment advice. Anyone using the valuation methods, data, or the valuation itself does so at his or her own risk. The author makes no warranties about the accuracy of the data or methods. The author only has long positions in broad-based index funds and does not hold individual securities issued by any company mentioned in this paper at the time of writing.