Forecasting Crashes and Recessions

"This expansion will run forever." Not two years, or three, or ten. Forever. That was how an MIT Professor of Economics summarized his vision of the U.S. economic expansion in the July 30th, 1998 issue of the Wall Street Journal.

Nowadays his assertion appears extreme. It did not then. This exuberance was rationalized by the obvious fact this was a New Economy with no room for recessions. Dornbusch himself said the American economy would "not see a recession for years to come"? Mmmm, why? Because, Dornbusch again, "We don't want one, we don't need one, and, as we have the tools to keep the current expansion going, we won't have one." "We", apparently, are the macroeconomists.

More Dornbusch: "Only natural causes, and not the Fed, can bring the economy to a standstill. Fortunately, we have the monetary and fiscal resources to keep that from happening, as well as a policy team that won't hesitate to use them for continued expansion."

In the latter part of the 1990s, euphoria was rampant and not only in the States. "It is hard to imagine any article with worse timing than, say, 'Asia's Bright Future,' by Harvard Professors Steven Radelet & Jeffrey Sachs, writing in the November/December 1997 issue of Foreign Affairs". So J. Orlin Grabbe told us. Their article was published at the precise moment East Asian financial markets and economies were deepening their collapses. As Grabbe put it: "Of course Asia probably does have a bright future, much as Europe could have been said to have had a bright future during the Black Death years of the 14th Century."[1]

It is one thing to say crises are undesirable, but another to say macroeconomists are, firstly, so skilled at forecasting they can predict trend breaks to the downside; and secondly, they have the tools and the policy teams to avoid economic and financial crises. If this were so, why do they utterly fail over and over again at forecasting economic downturns? Why do they have to adjust their projections over and over in times of trend change? Why has Japan stagnated for 11 years and had three recessions during that time?

The dismal record of forecasting crashes and recessions we economists have is not new. The crash of 1929 and the Great Depression came as an unexpected avalanche to economists, particularly those in the hall of fame.

Fourteen days before Wall Street crashed on Black Tuesday, October 29, 1929, Irving Fisher, America's most famous economist, Professor of Economics at Yale University, said: "In a few months I expect to see the stock market much higher than today".

Days after the crash, the Harvard Economic Society informed its subscribers: "A severe depression such as 1920-21 is outside the range of probability. We are not facing a protracted liquidation." After continuously issuing erroneously optimistic forecasts, the Society closed its doors in 1932. The two most renowned economic forecasting institutes in America at the time failed to understand that a crash and depression were forthcoming, and continued their optimism even as the Great Depression swept over America.

Irving Fisher lost 140 million U.S. dollars[2] in the stock market crash. Fisher was a man of many talents, a great economist, excellent theoretician, one of the founders of econometrics and pioneer in index number analysis. He was also the inventor of the c-kardex file system which he sold to Remington Rand for millions, which he subsequently lost in the crash.

John Maynard Keynes, the most famous British economist and the father of macroeconomics, who made fortunes in the financial markets for himself and Cambridge University, lost one million English pounds[3] in the crash.

With two exceptions, no academic economists forecasted the crash of 1929 and the following depression.

Seven decades have gone by. Surely we must know more today? In 1988, sixty years after the crash and the depression, Kathryn Domínguez, Ray Fair and Matthew Shapiro concluded in the American Economic Review, the leading journal of the American Economic Association, that employing sophisticated econometric techniques of the late nineteen-eighties and even using data unavailable in 1929, the Great Depression could not have been forecasted.

In October 2000, sixty economists gathered at the Minneapolis Fed to present papers and discuss the Great Depression of the 1930s. The cream of the macroeconomic profession was present: Nobel prize-winner Robert Lucas, Ed Prescott, Tom Sargent, Ben Bernanke, Finn Kydland, Nancy Stokey, Kevin Murphy and many others. The gist of the conclusions may be found in the headline that the Minneapolis Fed's review The Region used for the conference's article: "Something Unanticipated Happened". "In his summary remarks at the close of the conference, Robert Lucas made a pitch for the continuing investigation of macro fundamentals. . . . 'We should continue to seek common factors,' he said, and offered monetary instability as one area for further exploration. Big deflations are related to depressions, he said, and everywhere in the 1930s there was deflation."[4]

The concluding paragraph of the article states: "In the end, if the Great Depression is, indeed, a story, it has all the trappings of a mystery that is loaded with suspects and difficult to solve, even when we know the ending; the kind we read again and again, and each time come up with another explanation. At least for now."[5]

Economists, especially since 1936, and as Bob Lucas' quote reveals, look at macroeconomic fundamentals. Yet history teaches us no financial collapse has ever happened when things look bad. On the contrary, macroeconomic flows look good before crashes. Before every collapse, economists find the economy in excellent shape. In a major boom, the economy is a "New Economy". As President Hoover tells us in hisMemoirs about the period preceding the Great Depression: "With increasing optimism, they gave birth to a silly idea called the New Economic Era. This notion spread all over the country. We were assured we were in a new era where the old laws of economics no longer applied."[6] Too familiar, perhaps.

In these new eras, everything looks rosy, stock markets go up and up, and macroeconomic flows (output, employment, etc.) appear to be improving. Macroeconomic fundamentals, however, tell us about the past, and the good times are invariably extrapolated linearly into the future.

Friedrich von Hayek, 1974 Nobel Laureate, was the only academic economist who wrote prior to the Great Depression that a crisis and downturn in America were imminent. Interest rates in the world would not fall, he wrote, until the American boom collapsed. And "the boom will collapse within the next few months". This prediction, printed in the Austrian Institute of Economic Research Report, February, 1929, generated interest in Austrian economics and Hayek was offered a Professorship at the London School of Economics in the early 1930s.

Ludwig von Mises, also an Austrian, anticipated a worldwide depression in the 1930s, as reported by Fritz Machlup, Mises' assistant at the time. Mises' wife Margit wrote, in her husband's biography, that in the summer of 1929 he had rejected a high position in Kredit Anstalt, one of the largest banks in Europe at the time. His explanation was "a great crash is coming and I do not want my name in any way connected with it". Less than two years later, Kredit Anstalt was bankrupt.

Did rational economists adopt the economics of Hayek and Mises? Alas no, they adopted the economics of Keynes.[7]

But Hayek and Mises were exceptions. Not only did economists fail to forecast the Great Depression of the 1930s, but they have also failed to forecast economic contractions in general. The present contraction (in 2001) is only the latest example. "Economists have a dismal record in predicting recession" is the subtitle of an article in the November 29th, 2001, issue of The Economist.

Why is this so? What is it that economists do not know? Or what truths did they once know—ones since forgotten or neglected?

First, the facts.

Three centuries of financial crises and economic contractions yield four sets of empirical observations.

1. Money and Debt, Financial Markets and Business Cycles.

This first set is consistent with postulates of Business Cycles Theories well-known to economists but neglected nowadays. Sizable and sustained increases in money and private sector debt accompany financial and economic booms. Important contractions in money and private sector debt accompany financial and economic contractions. This observation is consistent with Monetary (Milton Friedman) and Austrian (Ludwig von Mises and Friedrich von Hayek) theories of the business cycle. Bob Lucas' remarks indicate a renewed interest in these theories may be forthcoming.

Debt accumulation speeds up during booms. Changes in the level of private debt correlate with changes in stock markets indices and economic activity, especially at higher degrees.[8] Debt increases heavily and rapidly in times of financial and economic booms, and it decreases importantly in times of major financial and economic contractions. This observation on debt accumulation and deflation was highlighted by Irving Fisher in his 1933 Econometrica article, "Debt-Deflation Theory of Depressions", but it has received little attention since. In a lifetime of work neglected by mainstream economics, Hyman Minsky emphasized the financial instability created by mounting indebtedness building up through time.

In the high degree bull market which has just ended with the 20th Century, debt accumulation in the private sector proceeded at a very fast pace and reached very high levels, especially in the latter years of the boom. However, little if any attention was paid by most economists to this phenomena.

The next two sets of observations, not considered by economists, are of course the key to understanding economic contractions of all degrees. Their not considering them isthe explanation for economists' dismal forecasting failures.

2. Stock markets indices are patterned.

Stock markets trend and reverse in recognizable patterns.[9] Structures are clear and definite in form [not in time or amplitude]. Patterns of smaller degree link together to form similar patterns at a larger degree. These insights date to the careful inductive analysis published in the 1930s by Ralph N. Elliott[10], and to the important and breathtaking work Robert R. Prechter, Jr. has pursued over the last three decades, applying and extending these principles to a wide variety of phenomena.[11]

Markets, in other words, are hierarchical. These insights have been rediscovered by physicists studying financial markets. Markets' movements, like earthquakes, have different degrees.[12][13] More importantly, markets are fractals (robust fractals or quasi-fractals).[14] Markets proceed relentlessly according to form. Elliott and Prechter are in good company: Pythagoras stood for inquiry into pattern rather than inquiry into substance.[15]

"The mysterious changes in market psychology"[16] proceed according to pattern. Stock markets are not random walks, as is still taught in many top graduate business schools. Rather, their changes exhibit fractal behavior.[17] Stock markets as complex systems show discrete levels or scales in a global hierarchy.

Viewing markets as dynamic, complex systems, Sornette and Johnson conclude markets proceed unabatedly toward a crash, "the market anticipating the crash in a subtle self-organized and cooperative fashion, releasing precursory fingertips observable in stock market prices."[18]

There is, therefore, an element of predictability in markets not despite, but because of, their complexity.[19]

3. Financial market changes precede changes in economic variables.

High degree bull markets in stock indices precede economic booms of high degree.Bull markets of lower degree precede economic expansions of lower degree.

High degree bear markets in stocks precede economic contractions of high degree.Bear markets of lower degree precede economic contractions of lower degree.

Changes in stock market indices thus precede, not follow, changes in economic fundamentals or news about them. Changes in stock market indices are a leading indicator of changes in economic activity. This was, of course, recognized by Wesley Mitchell and the National Bureau of Economic Research eight decades ago. But failure in identifying markets' different degrees has made Mitchell's insight less useful. Witness Nobel Prizewinner Paul Samuelson' s famous (and wrong) quip[20], a result of completely missing the hierarchical nature of stock markets.

Finally,

4. Booms are followed by Contractions.

High degree bull markets in stocks are followed by high degree bear markets in stocks. Correspondingly, high degree booms in economic activity are followed by high degree contractions in economic activity.

Milton Friedman's "Plucking Model", where contractions are related to succeeding expansions and unrelated to previous expansions, is not consistent with this observation. The Austrian Theory of the Business Cycle (Mises and Hayek), where the excesses of the prior booms are the sources of the following bust, is consistent with this observation.

After the facts, a story.

"Economic reasoning will be of no value in cases of uncertainty". These are Robert Lucas' words.[21] Mises, Knight and Hayek have taught us, however, uncertainty is normal and pervasive in society.[22] Markets deal with uncertainty, coordinating information and knowledge. In a world of uncertainty expectations are a critical variable in most decisions, particularly for investment decisions, financing (increasing debt levels) and other financial and economic decisions.

Nobel Laureate Robert Solow has recently said, "it is acutely uncomfortable to have so much in macroeconomics depend on how one deals with a concept like expectations, for which there is (inevitably?) so little empirical understanding and so much room for invention".[23] To have a better grasp of a relevant concept of expectations and confidence is therefore crucial. As Bob Prechter says, however, although expectations may imply rationality they are usually the product of rationalization.[24]

An alternative assumption to the macroeconomists' "representative agent" with rational expectations can be fruitfully substituted to obtain a more coherent explanation of observed patterns. In a world of uncertainty, changes in expectations (in confidence, rather, in moods) are reflected more rapidly in changes in stock market prices. The latter can thus be used as a proxy for changes in confidence. Changes in the trend of stock market prices become useful as a barometer of optimism and pessimism. As changes in optimistic or pessimistic expectations lead most changes in financial and economic decisions, it is no surprise that changes in stock market prices are useful as leading indicators of changes in the state of the economy.

Changes in stock market indices do not respond to, nor are they caused by, exogenous changes in economic fundamentals or news about them; they are not random walks. Changes in stock market prices reflect changes in confidence, moves from optimism to pessimism and from pessimism to optimism. They precede (rather than follow) changes in economic fundamentals.

The preceding sentences provide an explanation for why negative changes in stock market indices — of a very high degree — anticipate economic depressions. Negative changes of lesser degree, in turn, anticipate economic recessions and milder downturns. Stock markets are thus led by waves of optimism and pessimism.[25]

Optimism during the boom leads to higher indebtedness. Increasing private sector debt is an important manifestation of optimism and euphoria in the latter part of the boom. Excessive debt leads to financial fragility in banks, business enterprises, and individual households. As the boom ends and pessimism replaces optimism, lenders recall loans, banks contract credit, bankruptcies are stepped up and a major economic contraction ensues. Fragility turns into insolvency.[26]

The final leg of a stock market boom — as seen in the NASDAQ in the years prior to 2000, or in the Dow prior to late 1929 — is correlated with a high degree of indebtedness, and is correlated in turn with the severity of the subsequent economic contraction.

If this is correct, an ex-post explanation of a major economic contraction, for example, would start with a dramatic and sustained fall in stock price indices, a proxy for a major breakdown in confidence, a change from extreme optimism to pessimism. The high levels of debt, accumulated during the boom at a very fast pace in the later years, would be responsible for generalized financial fragility all over the economy. This is what we have now.

The change to pessimism, announced by the trend change in the stock markets, will trigger loan recalling, bankruptcies, unemployment and generalized economic contraction. This is what is beginning now. Only when debt reaches very low levels, as a consequence of bankruptcies or of inflation, is the economy ready for recovery. This will be some years into the future.

Economic Science?

Now, is this economics?

I use here the two tests posed by Nobel Laureate James Buchanan in his "Economics and its Scientific Neighbors" to answer that question[27]:

1. Does this theory provide the economist with an additional set of tools? By understanding the nature and the hierarchy of stock market changes, patterns of stock market price changes can be predicted in form and sometimes in time. Also, the elucidation of the degree of change in stock market prices allows a prediction to be made on the degree of the subsequent economic contraction. Thus, not only can economic contractions be anticipated, but also their degree. More clearly, a trend break in stock markets can be predicted in form, and that in turn precedes a trend break in economic activity. Timing on the other hand can only sometimes be pinned down with precision.

The linear extrapolation so commonly used by macroeconomists ("the crudest form of technical analysis")[28] is substituted by a non-linear framework — the Wave Principle — which allows prediction of trend breaks of different degrees.

I believe it provides the economist with an additional set of tools.

2. Does this extend the application of the central principles of the discipline?

There is clearly no contradiction with the statement: "More of any good will be chosen, the lower its price relative to other goods", a central tenet of economics.

Under uncertainty, however, dealing with future prices involves not actual but expected prices, and expected prices are highly dependent on whether confidence is high or low. As stock markets are patterned, their changes are — to a first degree — exogenous to economic variables.[29] We can state this as an assumption: To a first degree, changes in optimism and pessimism, measured by changes in stock market indices, are exogenous — independent of changes in economic variables (or "macro fundamentals", as Bob Lucas called them).

As stock markets are patterned, so are the true causal forces.[30] They are not random. Stock market patterns are predictable in form and sometimes predictable in time. The economy goes into an economic contraction not because of random shocks, as stated by real-business-cycle theory, but because extreme optimism, euphoria, is replaced by growing pessimism.

Is this extending the application of the central principles of economics? I am not sure.

Does this theory have predictive implications? Most certainly. This indicates it may become a science. But, again, is it economic science?

Most likely not.

Footnotes

[1] J. Orlin Grabbe, "And Now, the Financial Apocalypse", orlingrabbe.com/Apocalyp.htm.

[2] Calculated by the author on the basis of nominal figures provided by Irving Fisher´s biographer son, Irving Norton Fisher.

[3] Figures provided by Keynes´ biographer Professor Skidelski.

[4] "Something Unanticipated Happened", The Region, Minneapolis Fed, December 2000.

[5] Ibid.

[6] As I have lost the English text, and translated back into English from Spanish, some words may be different in the original.

[7] They did not however pay attention to key insights contained in Chapter 12 of Keynes' The General Theory.

[8] Financial booms and contractions involve a hierarchy with different degrees of change in financial markets, as earthquakes do.

[9] As well as financial markets in general.

[10] Ralph N. Elliott, The Wave Principle, 1938. Reprinted in Robert R. Prechter, Jr., editor, R. N. Elliott's Masterworks, Elliott Wave International. Also, Robert R. Prechter, Jr., editor, R. N. Elliott's Market Letters, Elliott Wave International.

[11] Robert R. Prechter, Jr., Wave Principle of Human Social Behavior, 1999; At the Crest of the Tidal Wave, 1995; Popular Culture and the Stock Market, 1992; Prechter's Perspective; and with R. Frost, Elliott Wave Principle, 1979.

[12] A. Arneodo et al, "Fibonacci Sequences in Difussion-Limited Aggregation", in J.M. García-Ruiz et al, editors, Growth Patterns in Physical Sciences and Biology, Plenum Press, 1993.

[13] Discrete scale invariance, as developed by D. Sornette, "Generic Mechanisms for Hierarchies", InterJournal Complex Systems 127, October 15, 1997; "Discrete Scale Invariance and Complex Dimensions", Physics Reports 297, 1999.

[14] Arneodo et al., put it this way: "[T]here is room for "quasi-fractals" between the well- ordered fractal hierarchy of snowflakes and the disordered structure of chaotic or random aggregates". Prechter uses the term robust fractal. They differ from fractals as defined by Mandelbrot, in that there is no self-similarity.

[15] G. Bateson , "Form, Substance and Difference" in Steps Toward an Ecology of Mind, Chandler, 1972, p. 449. Also R. G. Collingwood, The Idea of Nature, Oxford, 1945.

[16] A theme running through accounts of the 1920s, according to R. Schiller, Irrational Exuberance, Princeton University Press, 2000, p. 115.

[17] Non self-similar fractal behavior, i.e., not simple but complex fractal behavior.

[18] D. Sornette and A. Johansen, "Large Financial Crashes", Physics A, 245, 3-4, 1997.

[19] Robert R. Prechter, Jr., Wave Principle of Human Social Behavior, 1999.

[20] "The market has anticipated 12 of the last 9 recessions"

[21] Robert Lucas , "Understanding Business Cycles", in K. Brunner and A. H. Meltzer, Eds., Stabilization of the Domestic and International Economy, North Holland, 1977, p.15.

[22] Frank Knight, a leading Professor of Economics at the University of Chicago, teacher of several Nobel Laureates.

[23] R. Solow, "Toward a Macroeconomics of the Medium Run", J. Economic Perspectives, Winter 2000.

[24] Robert R. Prechter, Jr., private communication with the author.

[25] Early in the century, some economists were well aware of the importance of these waves of optimism and pessimism. A.C. Pigou, Industrial Fluctuations, London, Cass, 1927; The Economics of Welfare, London, Cass, 1920. And J.M. Keynes, The General Theory of Employment, Interest and Money, London: Macmillan, 1936, chapter 12.

[26] An expanded treatment of this process appears in H. Cortés Douglas, "Forewarnings", processed, Catholic University of Chile, January, 2001.

[27] James Buchanan, What Should Economists Do?, Liberty Press, 1979. Buchanan is clear that so-called macroeconomics does not pass the test. It is neither economics nor science. I agree.

[28] Robert R. Prechter, Jr., Wave Principle of Human Social Behavior, 1999

[29] "To a first degree" allows for subsequent feedback.

[30] There may be several alternative explanations of why stock market patterns are exogenous. Most likely the explanation may have to do with interactions among individuals in a context of uncertainty. Bob Prechter has a powerful hypothesis, as presented in his Wave Principle of Human Social Behavior, 1999. Among economists, Robert J. Shiller, Professor of Economics at Yale, is the leading representative of the view that "solid psychological research does show that there are patterns of human behavior that suggest anchors for the market that would not be expected if markets worked entirely rationally", Irrational Exuberance, Princeton University Press, 2000

Hernán Cortés Douglas is Professor of Economics at the Catholic University of Chile. He thanks Bob Prechter for valuable comments on an earlier version. His email address is [email protected].

Chile, 24 January 2002