My Blog

January 31, 2008

THU 31 JAN

More on why high share prices don’t mean Chinese banks are in good shape By Michael Pettis

In several earlier entries on this blog (for example see October 3: “Should Chinese banks acquire banks abroad?”), I have used the option framework to value Chinese banks and to try to correct what I believe to be a very widely-held but incorrect assumption – that the high prices investors are willing to pay for Chinese banks indicate that investors have judged the banking reforms in China to have been successful and the banks in good shape.  In fact, it is precisely because Chinese banks are still so uncertain and volatile that they are so expensive (and of course the fact that their home market may be experiencing a stock market bubble doesn’t hurt). 

 

As I have pointed out in those previous entries, shares in bankrupt banks are the closest things to call options on a country’s underlying economy, and in a rapidly reforming economy like China’s, these options should be extremely valuable.  In those entries I often mention the Mexican banking experience as a very illuminating example for China.  

 

Actually there are dozens of good examples of bad banks with high prices besides the Mexican one, but for personal reason I am most familiar with Mexico.  In the early 1990s I headed the Latin American bond and bank debt trading team at First Boston when it was hired in 1990, under the leadership of Pedro-Pablo Kuczynski, to advise the Mexican government on the privatization of its banking system via an auction process.  Although the “Chinese wall” between bankers and traders prevented me from being fully involved in the auctions, I was part of the team that advised the government on debt and market-related issues and very close to the whole process. 

 

Why am I bringing all this up again when I have already covered it so many times before?  Because I recently read a very interesting piece about the Mexican bank privatization, published in June 2004 by Stanford University’s Stephen Haber, (“Mexico’s Experiments with Bank Privatization and Liberalization, 1991-2002”, available at http://scid.stanford.edu/pdf/scid215.pdf). 

 

I strongly recommend that anyone interested in Chinese banks and banking reform read the piece.  One of my most regular complaints to my students at Peking University is that experts on the Chinese economy (and this is true of most other large countries too), think that everything that happens here is sui generis and cannot be illuminated or explained by events that have taken place in other times or other countries, whereas my many years of living and working in developing countries (and reading and writing about their financial histories) leaves me staggered by how little difference there often is.  Mexico is not China of course, but they do have in some cases an awful lot in common, and I am willing to bet that many Mexicans, looking at the Chinese banking system today, will find themselves remembering a lot of old and not-so-old stories.

 

For those who don’t know much about recent Mexican financial history, Mexico’s private banks were expropriated at the beginning of what was subsequently called the LDC Loan Crisis, a period of very difficult economic conditions for much of the developing world during which time loans to 32 countries, including much of Latin America, were either in default or in the process of forced restructuring.  At the time Mexican President Jose Lopez Portillo, seeking to halt what he called the ''looting'' of Mexico through the capital flight as Mexicans began withdrawing savings and converting them into dollars, nationalized the country's private banks on September 1, 1982, less than a month after his finance minister had announced in a phone call to the US Federal Reserve Bank that Mexico would be unable to meet its upcoming external debt maturities.

 

By the late 1980s the old import-substitution and nationalist models of economic development were terribly discredited (not always fairly) and the Mexican banks were bankrupt and mismanaged.  By the way I am often told – and what it is worse, told as if it were self-evident – that in China the banking system can never experience liquidity or credit crises because the banks are owned by the government.  This may well be true but for some reason it was never correctly explained to the Mexicans.  By the end of the 1980s their banking system was a mess, and a huge drag on the economy and on the fiscal balance.

 

By the time the administration of President Miguel de la Madrid (1982-88) began embarking towards the end of his term on the market reforms that were supposed to transform the Mexican economy, it was pretty clear that the banking system needed radical surgery.  His successor, President Carlos Salinas, determined to repair the country’s financial system and eager to raise revenues to cover the social costs of the major economic and financial reforms, began privatizing state-owned enterprises, including the banks (about half the government’s budget at the time went to subsidizing the losses of the state-owned sector).

 

The subsequent sale of the Mexican banks in 1991-92 was seen as a huge success by all the parties involved, not least by my friends at First Boston, with purchase prices coming in well above even the most optimistic private estimates.  The eighteen banks privatized were sold for an average price-to-book ratio of 3.04, even with all the uncertainty surrounding the Mexican reforms and the never-forgotten expropriation risk (twice before, in 1910 and 1982 the Mexican government had expropriated the private banks).  For comparison’s sake, bank acquisitions in the US during the 1990s occurred on average at a price-to-book ratio of 1.89, and in Europe the ratio was around 2.50.  Mexican banks sold at a huge premium to rich-country banks.

 

Even these high price-to-book ratios understate the prices at which the Mexican banks were valued because their book values were artificially boosted by the very loose accounting standards permitted Mexican banks, allowing them to carry at face value loans that in other countries would have been treated as non-performing.  To quote Stephen Haber:

 

One of the most lenient of Mexico’s bank accounting rules was that when a loan was past due, only the interest in arrears was counted as non-performing. The principal of such loans could be rolled over, and counted as a performing asset. Moreover, the past due interest could be rolled into the principal and the capitalized interest could be recorded as income. 

 

Reforming this rule (as well as others that inflated bank capital and assets) would have lowered the market value of the banks, because it would have increased the ratio of non-performing to total loans, lowered the banks’ reported rates of return, and decreased the book value of assets. How much lower the banks would have been valued is difficult to know.  It is known, however, that the government contracted outside consulting firms to provide it with a valuation of the banks.  It did not, however, make the results of those studies public.

 

The very high prices the Mexican banks received and the fact that Mexican banks were effectively bankrupt seems at first to fly in the face of rational investor behavior.  Why would investors pay a much higher price for risky, bankrupt Mexican banks than for healthier US or European banks operating in a much less risky environment?  

 

The answer, as I have discussed in the earlier postings, is neither because investors are crazy nor because the banks were actually much healthier than we supposed.  The reason investors paid so much was because bankrupt banks in a very volatile and uncertain economy have extremely high optionality.  

 

The fact that these banks were essentially bankrupt was important because it meant investors were not paying for net asset value (“intrinsic value”), they were paying only for time value.  The fact that Mexico was undergoing radical reforms was also important because it meant that there was a high probability of very high levels of success (many subsequent years of high, sustainable economic growth), but that this came with a high probability of high levels of failure (social unrest, followed by a reversion to former anti-growth policies).  Any investor would love to capture the upside without risking the downside.

 

It turns out that, whether or not they realized it explicitly, what investors wanted from Mexico was a call option on the success of its economic reforms.  Since bank profitability is highly correlated with economic growth, and since they didn’t have to risk buying good assets whose value might subsequently fall in case the reforms turned out badly (net asset value was low or even negative), shares in bankrupt banks gave them exactly that.  They were consequently, and not surprisingly, eager to buy those options and were willing to pay up.

 

I mention all this to argue – once again as I have perhaps too many times in this blog – that the high prices paid for Mexican banks had no more to do with investors’ beliefs that Mexican banks were healthy and well-managed than do the high prices of Chinese banks indicate that the “market” has judged banking reform and has judged it to be good.  As paradoxical as it may at first seem, investors are saying nothing – or perhaps more fairly they are saying very little – about the quality of Chinese banks when they value the banks at such high levels.

 

One of Haber’s points is that the way in which the banks were auctioned in Mexico had an important pricing effect.  The banks were auctioned in six batches, and each auction was so successful that prices paid in the subsequent auction were higher than in the previous one. Haber says:

 

All things being equal (size of bank, profitability, number of bidders) each additional round of bidding pushed up the bid-to-book ratio by .30. This ratio is stable across alternative specifications and is always significant at the one percent level.  In fact, bidding round is the only statistically significant variable that has a positive sign in the regressions. 

 

Surprisingly, neither the rate of return on assets, the rate of return on equity, nor the number of bidders is statistically significant.  Perhaps most surprisingly, the market power of a bank (measured as the log of bank assets) is statistically significant, but it has the wrong sign: market power is negatively correlated with the bid to book ratio.  This is not the outcome that one would expect from theory: one would usually expect that the market power of a bank would be capitalized in its auction price. 

 

Actually here is where I disagree with Haber.  It might at first seem rational that the higher the return on assets, the higher the return on equity, and the higher the market share, the more valuable a bank would be, but this is only true because these things should raise the intrinsic value of the bank share price.  For healthy banks, in other words, all these claims would be true, but not necessarily for bankrupt banks.  The intrinsic value for bankrupt banks is anyway low or zero.

 

The outcomes Haber finds surprising are not surprising if we consider that the value of these Mexican bank shares was largely the value of their optionality.  In that case neither the rate of return on assets nor the rate of return on equity is likely to be a terribly useful measure of the value of the bank since these are measures of intrinsic value, and what is really being valued has little to do with intrinsic value or the management of existing assets.  In fact if better managed banks had significantly higher real book values, so that these banks were “less” bankrupt and had high intrinsic value, it would not be at all surprising if their price-to-book ratios were actually lower than the worse banks. 

 

The fact that smaller banks were more valuable is also consistent with the optionality framework since smaller banks in Mexico tend to be regionally concentrated, and the regions of a country are always more volatile than the country itself.  Since for an option value increases with volatility, this gives more optional value to regional banks.  Of course if these banks had high intrinsic value, then Haber’s expectations would have been justified – more diversified banks are less risky and so more valuable, but that is because it is their intrinsic value that is more valuable.

 

Actually Haber does point out indirectly that some of the reason banks got such high valuations were option related:

 

Readers may wonder why bankers were willing to pay a substantial premium for the banks at auction.  The reason, as we shall discuss in detail below, is that the money that they were putting at risk was not their own.  Much of it was borrowed – some of it from the same banks that had just been purchased.   Moreover, the government’s commitment to guarantee all deposits (including interbank loans) meant that the group that ultimately bore most of the financial risk were not bank shareholders, but Mexico’s taxpayers.  

 

Buying on margin is of course like increasing the exercise price of a call option (reducing intrinsic value, if any, further).  And deposit guarantees automatically increase the value of risky behavior – ideal if you own a call option on the underlying assets.

 

Fascinating as the Mexican experience is in and of itself, we need to remember all this when we think of Chinese banks.  As with the Mexican banks in 1991 and 1992, what matters to the value of these banks is not likely to be measures that improve intrinsic value but rather measures that improve time value.  And bank share prices are going to be extremely volatile and highly sensitive to changes in expectations.

 

By the way the Mexican bank story does not end with their privatization – and it is worth remembering how volatile the share prices are of banks that have little intrinsic value.  After the banks were sold neither the underlying economy nor the ownership incentives encouraged prudent banking practices and by 1994, the banks were as bad as ever.  Haber does an excellent job of describing why.  As he says: “The combination of these two flaws – weak property rights and weak institutions to enforce prudent behavior – produced lending strategies that, at the very least, were reckless.  Even before the peso crisis of December 1994 (which is often blamed for the collapse of the banking system) many of Mexico’s banks were teetering on bankruptcy.” 

 

Haber also argues that there are two widely accepted sets of reasons that explain why the Mexican banks were managed so badly after their privatizations – one focusing on the lack of credit experience among Mexican bankers and the difficulty of the Mexican banking market, and the other focusing on the distorted incentive structures on otherwise very smart and capable businessmen.  He seems to prefer the latter explanation, and of course the option framework is very clearly on his side.  For those of us watching China some of this part of his article is a little chilling.

 

Needless to say within a short time Mexican banks were trading at price to book ratios way below their US and European counterparts.  That also should not be a surprise.