Find the original here: The American
I find this discussion of rational and nonrational market behavior – animal spirits – very interesting. His article also illuminates how he ran the Fed over almost two decades. He concludes by pointing out that as the Standard Deviation of stock prices increases, so bubbles become more prominent. I have seen some chatter of late around the idea that bubbles are “good” or “normal” for the economy, and Greenspan provides some perspective: some bubbles are better than others.
September 15, 2008 was not just the beginning of the financial crisis, which may very well have been the greatest financial crisis ever (the Great Depression of the 1930s was a far more devastating economic problem). In the hours and days following the Lehman Brothers default that day, there was a virtual breakdown of the global financial system.
For the first time ever, I believe, the short-term and overnight credit markets in the United States and around the world shut down. Not only did prices collapse, but the markets ceased to function. Bids virtually disappeared. Money market mutual funds, within hours of the default, failed or began to fail. The Federal Reserve had to intervene there, as it did in the commercial paper market and even in the repo market when investors became wary of all counterparties. Trade credit virtually disappeared, which essentially disabled global trade. Within days, ships were backing up out of Singapore because buyers could not make their payments.
The result was that, for the first time since 1907, overnight markets shut down. In 1907, it was the call money market, the dominant overnight market at the time. Bid rates, unmatched by offers, rose to 125 percent, but markets were restored by the next day. There were no similar problems during the Great Depression or any other time in history of which I am aware.
In the hours and days following the Lehman Brothers default, there was a virtual breakdown of the global financial system.
Bubbles, the root of the 2008 crisis and earlier ones, for reasons I will discuss later, have become more prevalent. All bubbles deflate, by definition. But most deflate without debilitating economic consequences. On the fateful day of October 19, 1987, for example, the Dow Jones Industrial average plunged 23 percent, the all-time daily record. I defy anyone to find economic disruption in the GDP figures following that date. I believe there was none. Similarly, in the dot-com boom, capital losses, as in 1987, were huge. But the owners of the toxic assets were mainly pension funds, households, and mutual funds; all suffered very large capital losses, but few ever got to the point where they defaulted on debt. And it turns out that the only two times in recent history in which we experienced real contagion in the financial markets were in late 2008, when, as AEI’s Peter Wallison has documented, there was a very destructive subprime mortgage crisis, the toxic assets of which, of course, were highly leveraged. And in 1929 we experienced a leveraged broker loan crisis. The result was contagious defaults.
Contagious defaults by financial institutions, I believe, are both a necessary and sufficient condition for significant financial and economic contraction. The nonfinancial parts of economies are rarely highly leveraged; their capital is generally 45 to 50 percent of assets. The nonfinancial sector was not where the problems lay in 2008. Indeed, just prior to the onset of crisis, nonfinancial corporate balance sheets were in good shape. The economic breakdown was initiated by the crashing of the financial system.
I emphasize in The Map and the Territory that it was not subprime mortgages alone that caused the crisis. Subprimes were indeed the toxic asset, but if they had been held by mutual funds or in 401(k)s, we would not have seen the serial contagion we did. It is not the toxic security that is critical, but the degree of leverage of the holders of the asset. A heavily debt-burdened economy is in danger of serial defaults. In 2008, tangible capital on the part of many investment banks was around 3 percent of assets. That level of capital can disappear in hours, and it did. And the system imploded.
It is not the toxic security that is critical, but the degree of leverage of the holders of the asset.
The question before policymakers is how to avoid such breakdowns in the future. As far as I can see, the only way to address such issues is to recognize that euphoria-driven bubbles are an inherent consequence of human nature over which we have little or no control. Successful financial policy, in my experience, ironically spawns the emergence of bubbles. There was never anything resembling financial euphoria, or the bubbles it creates, in the old Soviet Union, nor is there in today’s North Korea. At the Federal Reserve during my tenure, we often joked that our greatest fear was that policy might be toosuccessful. Achieving an underlying stable rate of growth and low inflation appears to have been a necessary and sufficient condition for the emergence of a bubble. We would conclude with mock seriousness that optimum monetary policy for bubble prevention was to create destabilizing inflation.
Can bubbles be prevented from rising once markets are in the grip of euphoria? At the Fed, we tried to defuse the nascent dot-com bubble of 1994. We failed. We raised the federal funds rate by 300 basis points and stopped a budding financial boom, as we called bubbles back then — stopped it dead. For the first time, we believed we had achieved a “soft landing”: a tightening of monetary policy that defuses a bubble but that is not strong enough to precipitate a recession. Or so we thought. For as we were patting ourselves on the back, the markets apparently assumed that, because the 300 basis point rise did not break the back of the economy, the economy must be far stronger than investors contemplated. The equilibrium level of the Dow Jones Industrial Average had apparently been significantly elevated, and the market took off shortly after we stopped tightening. The presumption that monetary policy can incrementally defuse a bubble is true only when an econometric model is constructed with such a mandate. It became clear in the latter part of the 1990s that the very act of defusing a bubble could alter its subsequent trajectory.
So how can we avoid the type of crisis that occurred in 2008? Serial defaults require the existence of debt, by definition. If we require that financial institutions hold a significant amount of contingent convertible debt (co-co bonds) that will automatically become equity under pre-determined crisis conditions, there may be a major loss of capital as a bubble deflates, but no serial defaults.
All bubbles deflate, by definition. But most deflate without debilitating economic consequence.
To be sure, co-co bonds, on issuance, may require a 200 or 300 basis points premium over straight debentures.1 But they measure the true cost of such risk if the alternative is a taxpayer bailout. In The Map and the Territory, I discuss why a significant increase in regulatory risk-adjusted capital ratios, whether through pure equity or co-co bonds, need not cause the rate of return on equity to decline. Over the course of 100 years following the Civil War, there was a significant decline in the equity-to-assets ratio of banks in the United States, but also a proportionate decline in the ratio of net income to assets. Thus, annual net income per dollar of equity remained in the range of 5 to 10 percent during that century. Markets adjusted. If we were to raise capital requirements now, it is not too farfetched to believe that the net income-to-equity ratios would not fall importantly as the equity-to-assets ratio rose.
In my book, I derive many equations that assess the impact of animal spirits on economic and financial activity. I sought particularly to determine how animal spirits leave their imprint on the extensive published body of market transactions. How, for example, can we separate, statistically, rational decisions from biased intuitions that impact the economy and especially finance? Defining the dividing line between what can be described as rational and nonrational has been, and will continue to be, the subject of large areas of disagreement.
Reason-driven individuals will not always interpret the reality they perceive accurately, but in most cases, they eventually do. Mistaken actions thus tend to be randomly clustered around rational goals as people learn from their experiences. Individuals driven by animal spirits, especially fear, euphoria, and herding, however, are driven by their biased perceptions of the probabilities of economic outcomes. A whole school of (behavioral) economics has arisen to identify these human propensities driven by animal spirits. Their mistakes are not randomly distributed about reality as it is, but around their distorted perceptions of reality.
The only two times in recent history that we experienced real contagion in the financial markets were in late 2008 and in 1929.
How do we judge the impact of economic events engendered by judgments based on real-world probabilities relative to those warped by fear, euphoria, and herding?
To better understand this complex issue, I have chosen a small but important segment of an economy — stock prices — that reflects the broad interaction between rationality and animal spirits. Stock prices are illustrative, if not necessarily representative, of the marketplace as a whole.
A key determinant of stock prices is productivity growth, conventionally measured by output per hour. Output per hour is necessarily a product of reason. As I note in the book, the insights leading to the invention of the steam engine and telegraph, and the discovery of atomic energy, for example, could not have arisen from irrational musings. To be sure, our long-term forecasts of technology (and stock prices), for another example, can go astray, but the process we employ to make such judgments is predominantly rational. Annual percentage changes in output per hour since 1889 appear to be distributed randomly around the long-term annual growth rate of 2.2 percent, a pattern associated with rational (but sometimes mistaken) reasoning. There is no evident fattening of tails of probability distributions or asymmetry between tails that would imply the bias of fear over euphoria. Quarterly data since 1952 tell the same story.
To probe further, I first remove the long-term growth rate of stock prices, largely the result of output per hour growth whose cause cannot be other than rational decision-making. I would expect that actual stock price movements, stripped of such long-term rational growth components, would reflect a combination of short-term rational decision-making interlaced with a large dollop of animal spirits behavior. I illustrate this by trend-adjusting the daily Standard and Poor’s 500 stock price change for each day the markets were open during the years 1951 to 2013 by subtracting a fixed average daily uptrend of 0.028 percent.2 The resulting series of trend-adjusted daily stock price changes since 1951 traces a bell curve where the x-axis is the size of the trend-adjusted daily percentage stock price change (in incremental buckets ranging from “greater than 5 percent decrease” to “greater than 5 percent increase”), and the y-axis is the number of market days in the six-decade sample that fall in each specified bucket. I then calculated a normal distribution whose peak value (smoothed) and mean (zero) were set to match those of the distribution of the sample of stock price changes. The results are plotted in Exhibit 1.
In short, human decision-making as represented by our near 16,000-day sample has characteristics similar to a reason-associated normal distribution. There are, however, some important differences. For example, trend-adjusted daily price changes of more than 1.2 percent, both positive and negative, display a persistent tendency to be far more common than observations of that magnitude in a “normal distribution” created by unbiased coin-tossing. I attribute this fattening of the distribution’s tails to the herding instinct (that is, following the crowd), a response of spirit-driven market participants who demonstrably tend to exaggerate stock price movements, especially large movements, in both directions.
Defining the dividing line between what can be described as rational and nonrational has been, and will continue to be, the subject of large areas of disagreement.
But the data also portray the asymmetric bias of fear being more powerful than euphoria, a conclusion also reached with nonfinancial data sets, as I note in The Map and the Territory. The asymmetry is particularly evident in the fact that daily losses of 5 percent or more significantly outnumber daily gains of 5 percent or more (22 versus 16) over a 63-year period of daily price changes.
But what I find most intriguing about these data is that they convey a pronounced increase, since the mid-1960s, in the proportion of stock price decisions implicitly driven by animal spirits, as suggested by an ever widening standard deviation of the 63 separate annual sample distributions covering the years 1951 to 2013. A five-year centered moving average of those 63 annual observations is shown in Exhibit 2.3 The sample data exhibit a pronounced shift of daily price change to ever larger percentage increases — in other words, they exhibit increasingly fattened tails. The greater the standard deviation, presumably the larger the proportion of economic outcomes determined by animal spirits, because the more decision-making that is detached from reality, the more likely it is to be infected by spirits — euphoria-driven bubbles followed by their fear-driven collapse.
As the standard deviation of stock price changes increases from decade to decade, bubbles seemingly have become more common. If true, the implication is that the impact of animal spirits as a driver of economic outcomes has been patently rising relative to the prevalence of rational decisions. And that could conceivably explain why we experienced the housing bubble right on top of the dot-com bubble.
Alan Greenspan served as chairman of the Federal Reserve Board for more than 18 years. He currently heads Greenspan Associates and is the author of The Map and the Territory: Risk, Human Nature, and the Future of Forecasting.