by Michael Pettis, China Financial Markets
Tom Holland had an interesting piece in the South China Morning Post three weeks ago in which he discusses the low valuations of Chinese banks. About a decade ago, if I remember correctly, Chinese banks were trading between three and four times book. Those valuations have dropped considerably since then:
On Monday, the weighted average price-book value ratio for the 10 Chinese banks listed in Hong Kong fell to just 0.98. In other words, as an investor you would have been able to buy shares in Chinese banks for less than the cash you would have received – in theory – if the companies were wound up the following day and the residual value returned to shareholders. That’s a highly unusual state of affairs, especially in a rapidly developing economy like China’s where banks have traditionally been regarded as a geared play on future growth.
Holland goes on the argue that this has important implications about the quality of the banking balance sheets:
Some observers argue that the current abnormally low valuation of China’s banks is an aberration which represents a great buying opportunity. There is, however, another interpretation.
Imagine that the market has got the pricing of Chinese bank shares about right. That would imply the value at which they are carrying assets on their books is far too high. Or, to put it another way, the proportion of non-performing loans on the balance sheets of China’s commercial banks is a lot greater than they are admitting. A very rough back of an envelope calculation suggests how high the true level of bad loans might be.
If the long-term price-book ratio were an accurate representation, it would imply the real value of bank net assets should be some 40 per cent below their declared level.
While Holland is right to suggest that there are a lot of loans on the balance sheets of Chinese banks that should be, but aren’t, classified as NPLs, and that these NPLs should reduce the market value of bank shares, I think we need to think very differently about the informational content of share prices, especially for banks.
About six years ago, on October 26, 2006, ICBC completed its deal-of-the-year IPO, which, at $21.9 billion, was the largest IPO ever completed. In an article the next day, CFRI described the event like this:
Shares in Industrial & Commercial Bank of China (1398.HK), which is raising up to US$21.9 billion in the world’s largest IPO, rose as much as 18 percent in their Hong Kong debut on Friday after its stock sale generated huge investor demand.
…ICBC’s IPO values the lender at 2.23 times its forecast book value. By comparison, No. 2 mainland lender Bank of China trades at 2.35 times 2006 book, No. 3 China Construction Bank trades at 2.66, and No. 5 Bank of Communications trades at 3.04 times book.
Two weeks after the IPO was completed, while the market was still buzzing with excitement, the Far Eastern Economic Review published the following article of mine in its January/February 2007 issue:
Buying Into China’s Volatility
In November during a banking conference in Beijing, a senior manager from Bank of China noted humorously that his bank, whose recent IPO had been priced at just over three times book value and had since traded up to 3.4 times, had a significantly higher valuation than the conference’s host, HSBC, which was trading at roughly 2.1 times book value. Critics of China’s banking reform who continue to complain about bad loans, he suggested, might be missing the point. The market’s assessment was clear: Chinese banks are healthy enough for markets to assign very high values to their shares.
At first glance his suggestion seems reasonable. Share prices have informational content, after all, and market valuations must reflect real investor perceptions. Two weeks earlier the biggest public offering in history, the $22 billion IPO by the Industrial and Commercial Bank of China (ICBC), had been by almost any standard a success. Not only was the offering oversubscribed approximately 50 times, but immediately after its launch its share price surged 15%. Like other wildly successful recent Chinese bank IPOs, including those of the China Construction Bank and Bank of China, the ICBC transaction seemed to confirm that international stock markets love Chinese banks.
Markets process many kinds of information, however, and because shares in Chinese banks and shares in global banks represent different kinds of claims, it is a mistake to assume that their informational content is the same. On the contrary, whereas the share price performance of Citibank or HSBC may say a great deal about investors’ assessment of their loan and earnings quality, it says something very different in the case of Chinese banks.
This is because shares of companies with dramatically different levels of solvency trade on very different types of information, and while there have certainly been improvements in lending practices in recent years, Chinese banks have a long way to go before they are healthy and can be considered prudently managed. If nonperforming loans and other assets were valued correctly, these banks would be technically insolvent. In a November 2006 report, Fitch Ratings estimated the total amount of nonperforming loans in the system, and calculated, assuming a 30% recovery rate on nonperforming loans (though on average the recovery rate for Chinese banks has been closer to 20%), total unrealized losses in the banking system to be roughly $250 billion, which exceeds total capital and reserves by more than one-third.
This figure does not include estimates made for the rapid loan expansion of the past two years, with loans growing 10.2% in just the first half of 2006. Most analysts believe this breakneck growth will result in a surge in new nonperforming loans. Fitch’s figures also do not include another $300 billion in loan carve-outs by asset management companies, who paid for the loans with low-coupon bonds which, if marked to market, would involve a further write-down of 10-15%. After taking into account these adjustments, liabilities materially exceed the true value of assets for every major bank in China.
Optionality in Equity Prices
Because they are technically insolvent, the informational content of share prices for Chinese banks is different than for solvent global banks. High valuations normally indicate expectations of low volatility and smooth sailing ahead, but share prices can measure a number of things, and in fact under certain circumstances rising share prices may indicate more, not less, expected volatility. This is the case for Chinese banks, whose market valuations reflect a very strong component of optionality in the share price.
One useful way of understanding the valuation of equity shares is by comparing them to call options and decomposing their value. As is widely recognized, equity has a relationship to the operating assets of a company that is similar to the relationship between a call option and its underlying asset. Options are defined in part by their strike price. In the case of equity, the corresponding strike price is equal to the total liabilities of the company. This is because equity holders “own” whatever is left after all liabilities have been paid, just as owners of call options “own” the value of an asset above the strike price.
A call option has intrinsic value when the value of the underlying asset exceeds the strike price. Similarly, when a company is solvent-i.e. the market value of a company’s assets exceeds its liabilities-the company’s shares have intrinsic value equal to the difference between the two. If all assets and liabilities were recorded at their true market value, intrinsic value would be equal to a company’s book value. Since this is almost never the case, intrinsic value can be significantly more or less than book value. Only solvent companies have intrinsic value, and the greater the value of assets relative to liabilities, the more intrinsic value there is in the share price. For highly solvent companies, intrinsic value comprises by far the largest component of share price.
But intrinsic value does not fully explain share price. In the same way that the value of a call option must always exceed its intrinsic value, the value of a share is also always greater than its intrinsic value. The most important reason for this is that equity shareholders have limited exposure to a drop in asset value and unlimited exposure to an increase. This “extra” value over the intrinsic value, called time value, measures the expected volatility in the value of a company’s assets. Time value increases when expected volatility rises. The greater the reasonable range of possible outcomes for future asset values, the greater the time value. If there is a very high probability the value of the asset will soar, even if there is an equally high probability it will collapse, the limited exposure to a collapse combined with unlimited exposure to soaring values will ensure a very high time value.
A company’s share price, like the price of an option, is the sum of the intrinsic value and the time value. The ratio between the two varies as a function of solvency. Time value accounts for 100% of the share price of an insolvent company. However, for a solvent company, time value accounts for a steadily declining percentage of the share price as the value of assets rises relative to liabilities.
Solvent v. Insolvent
The sensitivity of the share price to changes in asset value increases as the value of a company’s assets rises relative to its liabilities. For an insolvent company this sensitivity is low, which is another way of saying that changes in the value of underlying assets have only a small impact on the share price of an insolvent company. For solvent companies, it is the opposite. Changes in equity value of these companies are highly sensitive to changes in asset value, and as market valuations of the company’s assets rise or fall, there will be a significant impact on the value of the company’s shares.
In the case of high intrinsic value companies, the main driver of changes in the share price is likely to be changes in investors’ assessment of asset values, which directly affects intrinsic value. Because companies are likely to have a wide mix of assets, on average the total value of these assets is not likely to vary very much since they may be largely uncorrelated. This is also true of well-managed banks, whose loans are highly diversified by industry and region. Generally for companies in the same industry, the greater the intrinsic value the less volatile share prices tend to be.
Sometimes shares have very little or no intrinsic value. The share price of new Internet companies, for example, consists almost entirely of time value since these companies have low intrinsic value. In spite of their lack of assets, however, Internet companies often receive high valuations. This may seem paradoxical at first, because we normally associate high valuations with low volatility, but Internet companies are valuable precisely because their future outlook is so uncertain-ranging from quick bankruptcy to massive future profitability.
Companies that are insolvent or nearly insolvent also have little to no intrinsic value. Their liabilities are not much less, and sometimes greater, than the value of their assets. Like Internet companies, the value of their shares consists only, or mostly, of time value. Nonetheless if the value of the underlying asset is very volatile, the shares of the insolvent company, like call options and Internet shares, can still be valuable. Their value consists not of net assets, but of the possibility that the company will generate very high revenues, or its assets rise sharply in value, at some future point.
This implies, conversely, that changes in the value of the underlying assets have only a minimal impact on the share price (until the value of assets has risen enough to approach or exceed liabilities). Share prices for insolvent companies are likely to rise or fall primarily because of changes in expected volatility, and because expected volatility can change often and dramatically, it is normal for their shares to rise and fall in price much more sharply than shares in general.
Chinese bank shares typify this kind of behavior. Their shares have no intrinsic value because their liabilities exceed the value of their assets. However even though investors may agree that Chinese banks are insolvent, they may still believe that the shares of these banks, like shares in Internet start-ups, are valuable. This is because in valuing these shares, the value of the underlying assets is much less important than the expected volatility of those assets.
The value of the Chinese banking franchise is closely tied to long-term economic growth in China. No matter how pessimistic he may be about the total amount of nonperforming loans, any investor who believes that China’s economic outlook is extremely volatile will place a high value on a call option giving him access to this volatility. Precisely because their shares have no intrinsic value, large Chinese banks are among the best assets with which to make pure bets on the volatility and growth of the underlying economy, and because China is undergoing radical reform this makes bank shares valuable even if the banks are in poor shape. As long as China’s economy lurches forward, and as long as expected GDP growth in the foreseeable future is very high, bank share prices will hold up or even strengthen. But when the economy shows danger signs, or when expected GDP growth is revised substantially downward, bank shares will suffer disproportionately.
This is not the first time markets have placed high values on the shares of insolvent banks. Large developing countries in the process of rapid reform and change are almost always likely to see extremely high values placed on their bank shares, even when (and this is often the case) the banks’ loan portfolios are doubtful. One recent example is the case of Mexico in 1990-92 when it privatized 18 state-owned commercial banks-the entire banking system-as part of a process of reform that promised to change the Mexican economic and political landscapes dramatically.
At the time Mexico was just emerging from nearly a decade of crisis, and as in China today, reform in Mexico created both great uncertainty and great optimism. After years of government mismanagement Mexico’s outlook seemed filled with great potential, although the reform experiment was also fraught with risk. With such high volatility around expected future GDP growth, investors placed an extremely high value on ways to access Mexican economic volatility, in the same way that investors would expect to pay a lot for an option on a very volatile asset.
They showed how much they were willing to pay when the banks were privatized. The prices at which the 18 banks were sold stunned even the government agencies responsible for the sales. Mexican banks, whose loan portfolios were doubtful at best and who were in nearly every case technically insolvent, sold for an average of more than three times book value-all the more remarkable given that global interest rates were much higher than they are today. Valuation criteria for the banks exceeded those of several of the largest banks in the world. After it was privatized Mexico’s largest bank, Banamex, had a price to book value ratio nearly twice that of Citibank, Mexico’s largest creditor.
Not surprisingly, the very high valuations placed on Mexican banks by Mexican and foreign investors were extremely unstable. As the first positive impacts of reform were felt, valuations rose even higher, but this was not to last. The Mexican crisis of 1994 saw bank share prices totter and then collapse, and for the next few years as the outlook for Mexico’s economy shifted, their share prices fluctuated violently.
By 1998, Banamex was trading at little more than half of book value as the impact of the Mexican and Asian crises slowed Mexico’s growth prospects. Not surprisingly, there was also a surge in nonperforming loans. By the late 1990s Mexican authorities were so concerned about the poor performance of the banks that they permitted and encouraged sales to better-capitalized foreigners, and today most major Mexican banks are foreign-owned. Citibank purchased Banamex in 2001.
There is an important lesson to be learned from the experiences of other developing countries undergoing reform. The success of the Chinese bank IPOs should not be seen as a referendum on the health of the banking system or on the process of banking reforms. As was the case in Mexico and in other developing countries going through dramatic reform, the success of the IPOs reflects primarily investor willingness to speculate on a very volatile set of outcomes. High expected volatility can lead to high option prices and high share prices, but just as already high share prices can soar at any good news about the economy, they will drop drastically at any bad news.
This has implications for bank regulators. For years economists have argued that China’s banking regulators need a more effective monitoring system. Although China’s regulatory bodies are filled with capable managers, the problems in the Chinese banking system are too deep to be easily managed and often involve political sensitivities that make it difficult to identify and resolve problems. It was thought that these issues might be partially resolved by the IPOs. One of the functions of a securities market, after all, is to provide clear signals about market perceptions of risk and value, and these market perceptions could be harnessed by Chinese regulators. By publicly listing the large Chinese banks in the international markets, regulators hoped to enlist the aid of thousands of sophisticated investors from around the world to assess and judge the operating performance of these banks.
It is not clear, however, that changes in bank share prices will in fact have much signaling value to regulators. Soaring prices do not indicate that investors are satisfied with the pace of reform or the resolution of nonperforming loans in the banking sector. They indicate that investors are betting on continued rapid GDP growth. If and when China enters into what investors think is a material slowdown in expected long-term growth, bank share prices will fall sharply even if bank loan portfolios are improving.
It will not be until the banks gain comfortable levels of solvency that the informational content of their share price behavior will resemble that of their developing-country peers. But for now, by avidly purchasing shares in Chinese bank IPOs, the market is not saying that it has evaluated the Chinese banking system and found it satisfactory. It is saying something very different-that Chinese bank shares represent a good way to speculate on Chinese economic volatility. Optimists, like me, may believe that in the long run this will turn out to be a good bet, but for now no one should take much comfort in recent price performance. The Chinese banking system is still a mess, and it will be years before we can decide otherwise. It will also be years before changes in share prices tell us much about the fundamental health of the banks.
The main point of the article is that owning shares in a bank is the functional equivalent of owning a call option on the bank’s future operational earnings, and if the share price contains little intrinsic value (i.e. the value of its assets does not exceed the value of its liabilities by a large margin, and may even be less than the value of liabilities), by definition most of its value consists of time value, and so is extremely sensitive to changes in expectations. This means at least two things.
- As expected volatility declines, the value of the shares should decline too, because time value is positive correlated with expected volatility. At first this may seem strange. Normally, when expected volatility declines, share prices rise, but this is because reductions in volatility increase the value of assets. In a company with significant intrinsic value, in other words, a decline in expected volatility can increase intrinsic value by more than it reduces time value. In a company with no intrinsic value, of course, the only impact of lower expected volatility is to reduce time value.
- As growth expectations decline, the value of the shares should decline very sharply. To put it in option terms, as expected growth declines, the forward value of assets declines with it, so that the “strike price” of the implicit option declines relative to the asset price, putting it more out of the money.
This is why back in 2006-07 I expected a rapid decline in the price of bank stocks relative to book value. As China began increasingly running into the limits of the current growth model, I assumed, growth would decline sharply, and over time the consensus would also drop, bringing down with it the volatility of growth expectations.
How far down can prices go? As the Mexican bank example showed (full disclosure: I was a senior member of the team at First Boston that advised the Mexican government in 1990-93 on the privatization of the banking system), even when prices have dropped from 3-4 times book down to book, they can still drop a lot more. Banamex, remember, was trading at half of book before it was purchased. If you believe that growth rates have further to drop, and that this drop has not been factored into the consensus, the option framework for evaluating stock prices suggest that we haven’t bottomed out yet in bank shares.
For those who are interested in thinking more about how the option framework can clarify thinking about bank share prices, after the piece above was published the Far Eastern Economic Review asked me to follow up with a piece evaluating the rumored purchase by the Bank of China of a US consumer finance company. Here is what I wrote:
Should Chinese Banks Acquire Banks Abroad?
In an article for the January/February 2007 issue of the Far Eastern Economic Review (“Buying into China’s Volatility”) I used an option framework to explain and predict the behavior of investors in China’s bank IPOs. Share prices of insolvent or nearly insolvent banks consist almost entirely of what option traders call time value, with little to no intrinsic value. The option framework predicts that in such a case investor perceptions of the quality of management or of levels of non-performing loans will have little impact on the share price performance of Chinese banks. Instead this will primarily reflect investor perceptions of changes in China’s underlying economic volatility.
Since the article was published, there have been reports suggesting that the Bank of China is considering acquiring a consumer finance company in the US. The reported acquisition size is fairly small – reported to be about $2 billion – so it is unlikely to have much impact on the value of the buyer. It is worth pointing out however that the option framework makes two powerful, and perhaps surprising, predictions: first, that the acquisition of a foreign financial institution is likely to have a significantly negative impact on the combined banks’ share prices; and second, that the largest shareholder, because of its multiple roles, will have a strong incentive nonetheless to encourage foreign acquisitions.
There are many good reasons why a Chinese bank may want to acquire a foreign bank. It may want to diversify its loan portfolio, to serve its Chinese customers abroad, to gain experience and technology, or simply to make a long-term bet in another market. At the time of the acquisition, however, the main effect on the value of the bank’s share price will be the impact of diversification – a Chinese bank with operations in the US will have a more diversified loan portfolio and earnings stream. Diversification reduces volatility, and so usually increases a company’s enterprise value. Since the intrinsic value in share prices consists of the difference between enterprise value and total liabilities, if enterprise value rises, the intrinsic value component of the share price must also rise. Normally we would expect that this would cause the combined market value of the merged companies also to rise.
But this is not always the case. A reduction in volatility may increase intrinsic value, but it always reduces time value (share prices always consist of more that just intrinsic value, and this excess is called time value). What actually happens to share prices depends on which effect is greater. When the share price of the acquirer has high intrinsic value – i.e. the company is highly solvent and the value of its assets comfortably exceeds the value of its liabilities – it will typically have low time value, and the positive impact on intrinsic value will exceed the negative impact on time value. This will cause the combined share price to rise.
However when the share price of the acquirer has low intrinsic value and high time value – i.e. it is in a highly volatile business and has few real assets, like an internet company, or its liabilities approach or exceed its assets, like an insolvent bank – the impact of an increase in intrinsic value can be much less than the reduction in time value. In that case an acquisition will cause the combined share price to drop.
This is true not just for insolvent banks but for any company whose share price consists mostly or wholly of time value. For example when AOL and Time Warner announced in January 2000 that they would merge to create a media super-company, the first excited response of the stock market was to run up the combined value of the two firms by over 10%. Within days, however, as investors began to understand the implications of the merger, market sentiment changed and the combined value of the two firms dropped to 5% below the pre-announcement levels – during a period in which the relevant market index rose nearly 10%. Mergers involving start-up internet companies have almost always resulted in declining market prices for exactly this reason.
For Chinese banks, like for internet startups, any large acquisition will almost certainly result in a lower combined share price once the implications to investors are digested. The negative impact on the Chinese bank’s time value will exceed the positive impact on its intrinsic value, or to put it another way, more than 100% of the increase in enterprise value will go to the bank’s creditors, who benefit from more stable and diversified earnings. Equity investors, unlike creditors, place a high value on Chinese banks precisely because China’s economic future is so uncertain.
Any reduction in the volatility around future expectations will reduce their value to equity investors. Chinese banks currently have much higher price-to-book ratios than highly solvent global banks, and this high ratio reflects the high component of optionality (time value) in their share price. By sharply reducing time value a large foreign acquisition will effectively drive the price-to-book ratio lower, and so reduce the combined market value of the two banks.
But this doesn’t this mean that a rational Chinese bank will never engage in a foreign acquisition. On the contrary, it makes sense for state- controlled Chinese banks to make foreign acquisitions because the main shareholder, the government of China, has a much more complex incentive structure than do other shareholders. As a shareholder, of course, the government would like to see rising share prices, and to that extent it should not encourage foreign acquisitions. However the government’s position is not so simple. In addition to its role as shareholder, the government has at least two other important roles. First, it guarantees the bank’s depositors, so it effectively absorbs any improvement or deterioration in the bank’s creditworthiness. Second, it regulates the banks as part of its overall responsibility for the health of the banking system.
It turns out that both roles involve optionality. Creditors and regulators are effectively short put options on the enterprise value of the company because their exposure to increases in enterprise value is limited, while their exposure to declines is unlimited. Because they have differing incentive structures, their objectives differ. This is just a variation on what is known as the agency problem in corporate finance, in which managers (whose incentives, incidentally, are very similar to those of creditors) have goals that often conflict with those of shareholders.
Because the government in its role as guarantor and regulator is effectively short a put option on the enterprise value of the bank, this creates a strong incentive to minimize volatility – lower volatility increases enterprise value, and so reduces the intrinsic value of the put option (the value of a put option always decreases as enterprise value rises), while lower volatility always reduces time value. Along with being long a call option as a shareholder, the government is short a put option as guarantor and regulator, and as such it unambiguously benefits from any reduction in volatility. In this case the option framework simply makes explicit what we intuitively know: unlike shareholders, creditors and regulators worry far more about downside risk than about upside profits.
What the framework adds to the analysis is that for nearly insolvent banks, the interests of creditors and regulators are diametrically opposed to those of shareholders. Since its interests as guarantor and regulator almost certainly exceed its interest as shareholder, the government has a strong incentive to encourage behavior which may hurt shareholders in general but will benefit the government and all other creditors. China’s state-controlled banks are likely, in other words, to behave in ways that benefit managers, regulators and creditors at the expense of shareholders because its largest shareholder has a very complex incentive structure.
China’s government is not the only government whose interest may conflict with that of bank shareholders – this always happens in the case of banks with questionable loan portfolios whose deposits are implicitly or explicitly guaranteed, as the S&L crisis in the US during the 1970s and 1980s demonstrates. But because of the government’s mixed role as guarantor, regulator, and principle shareholder, it is important that investors understand that although their long-term interests may be similar to that of the government – a rapidly growing economy which translates into an increasingly valuable banking franchise – in the short term incentives are aligned in very different ways.
This is an abbreviated version of the newsletter that went out three weeks ago. Academics, journalists, and government and NGO officials who want to subscribe to the newsletter should write to me at email@example.com, stating your affiliation, please. Investors who want to buy a subscription should write to me, also at that address.