Merton Miller and Modern Finance

Financial Management (Financial Management Association), Winter, 2000 by Rene M. Stulz

over the last few years. For instance, capital flows to East Asian countries experienced a swing of more than $100 billion in one year. When such events take place, trades by some group of investors can end up destabilizing markets unless Merton Miller's arbitrageurs can stand up to the herd and take advantage of it. Paradoxically, however, while more arbitrage capital is needed to permit other investors the luxury to exit from the markets at a moment's notice, developments over the last few years show that instead of having more arbitrage capital, we may actua lly have less and perhaps not enough of it.

Among recent events suggesting that there are limitations to the arbitrage mechanism, the fall of Long-Term Capital Management (LTCM) stands out. LTCM was mostly engaged in transactions that would be close to the Modigliani-Miller arbitrage transactions if markets were perfect. Further, the practices of LTCM show that the growth in financial engineering made it possible to create arbitrage transactions where, before, it would not have been possible to do so. LTCM would hold a long position in securities that were underpriced and hedge these positions, thereby capturing the mispricing of the securities. A typical example of such a strategy would be to go long in an agency bond and go short in a similar Treasury bond. [9] Bonds issued by government agencies should have yields fairly close to bonds issued by the Treasury. At times, however, the yield of agency bonds is much higher than the yield of comparable Treasury bonds. Assume for a moment that the federal guarantee to the agencies is strong enough that the re is no default risk on the agency bonds. In such a situation, a long position in an agency bond trading at par and a short position in a comparable Treasury bond held to maturity earns a positive cash flow for sure in perfect markets since the coupon payments of the agency bond exceed the coupon payments of the bonds held short. There would be full use of the proceeds and no transaction costs in perfect markets.

In the real world, there are three difficulties with such an arbitrage transaction that are relevant here. First, there might be some default risk on the agency bond. With a credit derivative, one would be able to construct a default-free synthetic agency bond. Hence, financial engineering removes that obstacle to creating an arbitrage. Second, there would be transaction costs. Transaction costs simply would require that the yield differential between agency and Treasury bonds at which an arbitrage trade becomes profitable is higher than otherwise. Without full use of the proceeds on the short-sale, the investor would have to use some capital to implement the trade which would increase its cost. Finally, and most importantly, this trade would be an arbitrage only if it was sure that it could be held to the bonds' maturity. Remember that at maturity the long position and the short position have exactly the same value since both bonds are at par. However, if a month after the trade is put in place, the yield on agency bonds increases sharply relative to the yield on Treasury bonds, the value of the position becomes negative in that the bond held long falls in value relative to the bond held short. If the position has to be liquidated when its value falls, then the arbitrage is no longer an arbitrage. In a perfect market world, such a situation would not arise or, if it did arise, it would last for only a very short period of time. This is because an increase in yields on agency bonds relative to Treasury bonds would lead to arbitrage trades which would bring the yields closer together. The certainty that arbitrage trades would take place later if arbitrage opportunities arose would therefore make arbitrage trades feasible now. Absent this certainty, an arbitrageur would not know whether he would be able to maintain his position over time since seemingly irrational changes might force a trade to be undone at a loss.