Economic Science and The Holy Grail of Forecasting
Alan Greenspan has recently alighted from his daily formaldehyde bath (h/t to Vidoq for the imagery) to promote his new book, the main topic of which is "forecasting" of all things. The obsession of the economics profession with forecasting, which has become extremely pronounced ever since Milton Friedman pronounced it to be the holy grail, is inappropriate to economic science – especially considering the form it nowadays takes.
The approach to economic forecasting by today's mainstream economists, even the approach to ascertaining economic laws, is based on econometrics (this is to say, statistical data) and mathematical modeling. The original motto of the Econometric Society was "Science is Prediction", the implication being that any endeavor in the sciences not focused on prediction shouldn't be considered proper science. Obviously, something went wrong along the way. A 1990s study by the Economist magazine gauged the economic predictions of a group of academic economists and a group of housewives over the span of a decade. The housewives won the contest hands down.
We have recently come across a few articles (one of which was published on Acting Man in fact), in which it was argued that "economics is not a science". Let us first look at the example published in these pages, namely Bill Bonner's recent article “What a Wheeze”. Bill Bonner discusses Harvard economics professor, Raj Chetty to make his point. Chetty defended economics against its critics in the NYT, arguing that economics is indeed a science, which he then attempted to prove by citing the following example:
“Consider the politically charged question of whether extending unemployment benefits increases unemployment rates by reducing workers’ incentives to return to work. Nearly a dozen economic studies have analyzed this question by comparing unemployment rates in states that have extended unemployment benefits with those in states that do not. These studies approximate medical experiments in which some groups receive a treatment – in this case, extended unemployment benefits – while “control” groups don’t. These studies have uniformly found that a 10-week extension in unemployment benefits raises the average amount of time people spend out of work by at most one week. This simple, unassailable finding implies that policy makers can extend unemployment benefits to provide assistance to those out of work without substantially increasing unemployment rates.”
It is no wonder that after reading such absurdities, people come to sweeping conclusions such as "economics is not a science". Chetty's is a rather odd defense of economics as a science. The example he gives is simply meaningless in this context. Chetty is not describing a controlled experiment, he is describing the gathering of specific slices of historical data. A myriad of other factors that might have influenced the results were left out. Not only is this approach methodologically unsound, its conclusion is wrong to boot (as it flies into the face of economic logic). Bill Bonner's critique of this paragraph is entirely on the mark.
Two more articles presenting arguments along the lines discussed by Bill Bonner and coming to the same conclusions were recently published at the Daily Bell (one was a critique of Eugene Fama, the other an article on the mythical "QE Taper"). We want to point out that once again, we entirely agree with the critiques as such, which endeavor to establish why economics is different from the natural sciences (a point that apparently eludes Mr. Chetty). However, we cannot agree with the conclusion that therefore, “the only trouble is that economics is not a science”, as the Daily Bell puts it, or with Bill Bonner's remark that “only Harvard professors could possibly believe that there is anything the least bit scientific about economics.”
Contrary to such assertions, economics is indeed a science. What should have been said is "economics differs from the natural sciences" and "economic theory is distinct from the study of history". That however does not mean that is is not a science – it merely means that as a social science, it must employ a methodology that is different from the empiricism that is employed in the natural sciences.
As Ludwig von Mises has pointed out, economics is a branch of praxeology which is an aprioristic science (originally, Mises referred to the overarching science of human action as sociology, however he later adopted the term "praxeology" as many sociologists had adopted an anti-economic bias in the first half of the 20th century). Its method is not the erection of hypotheses that are then tested empirically in controlled experiments. Its method is deductive reasoning based on the action axiom. Economic laws do not require empirical testing to ascertain their validity – they are discovered by means of ratiocination. As Murray Rothbard explained in the foreword to Mises "Theory and History":
“Mises saw that students of human action are at once in better and in worse, and certainly in different, shape from students of natural science. The physical scientist looks at homogenous bits of events, and gropes his way toward finding and testing explanatory or causal theories for those empirical events. But in human history, we, as human beings ourselves, are in a position to know the cause of events already; namely, the primordial fact that human beings have goals and purposes and act to attain them. And this fact is known not tentatively and hesitantly, but absolutely and apodictically.”
As an illustration of this idea, think e.g. about the law of marginal utility. Do we need to empirically test the fact that an additional unit of a homogeneous good will be used for purposes that are lower on the value scale of the individual obtaining it than the preceding unit? Surely the very idea of such "testing" must strike one as absurd. Let us consider gallons of water as an example. One may use the first gallon one obtains to slake one's thirst. The second gallon may be used for washing. The third may be used for watering the garden, the fourth for filling an aquarium, and so on. We do not need to go out and "test" the fact that people will probably first slake their thirst before they employ water in less valued uses such as filling up an aquarium (always keeping in mind that the value scales of individuals do of course differ; it is the principle, the underlying economic law that does not require 'testing").
What then, does this mean for economic forecasting? It certainly does not mean that forecasting is "impossible". The implication is rather that forecasting can only ever be qualitative rather than quantitative in nature and that forecasts are constrained by praxeological laws as Hans-Hermann Hoppe has formulated it (in his monograph "Economics and the Austrian Method"). Correct praxeological reasoning is a sine qua non if one wants to arrive at useful economic forecasts, but its utility is limited – since as Hoppe also points out, future states of knowledge upon which people will act are inherently unknowable. Economic forecasting is an interdisciplinary endeavor in the sense that it requires not only a firm grasp of economic laws, i.e. the formal body of economic theory, but also involves "understanding" – an approach similar to that the historian must apply to the study of history. The economic forecaster must be both an economist and a thymologist in other words. As Rothbard wrote inthis context (in "Praxeology as the Method of the Social Sciences"):
“For the praxeologist, forecasting is a task very similar to the work of the historian. The latter attempts to “predict” the events of the past by explaining their antecedent causes; similarly, the forecaster attempts to predict the events of the future on the basis of present and past events already known. He uses all his nomothetic knowledge, economic, political, military, psychological, and technological; but at best his work is an art rather than an exact science.”
In a natural science like physics, predictions can indeed be made with great accuracy, since the objects of its inquiries are inanimate: tell a physicist at what angle and acceleration a rock is thrown, and he will be able to precisely predict where and when it will land. In the social sciences, the object of study are human beings and their purposeful, goal-oriented conduct. Since they have volition, we cannot possibly make precise predictions in this field. Friedman and other positivists were and are wrong: it is not possible to empirically "test" hypotheses in economics in order to assign a predictive value to them (which is what Mr. Chetty attempts to do in the example cited above).
Let us summarize: economics is indeed a science. Praxeology, of which economics is a branch, is “the complete formal analysis of human action in all its aspects” (Rothbard, "Man, Economy and State"). Economic forecasting is not a 'systematically teachable ability" (Hoppe), it is more art than science, but forecasts are more likely to be accurate over time if they are based on correct praxeological reasoning.
Alan Greenspan and the Stock Market
Attempting to defend his dismal forecasting record while at the same time hawking his new book about forecasting, Dr. Greenspan remarked at one point in a Bloomberg interview that he has "nothing to apologize for", even if he did get a number of major forecasts wrong, since after all, he is no "superhuman" and not omniscient. Unfortunately the old "nobody saw it coming" saw (which he reformulated as "I"m no better than anyone else at forecasting") doesn't really cut it, since a number of people did correctly forecast all the things Greenspan missed. Virtually the entire Austrian school foresaw the crashes of the 2000ds, since its proponents focused on the credit expansion the Greenspan (and later Bernanke) Fed aided and abetted. Greenspan's excuses sound hollow; the reality is that he simply tried to please his audiences, telling them what they wanted to hear.
Ludwig von Mises would probably scoff at this as irrelevant, but allow us to point out that members of the Austrian school have beaten mainstream economists at the forecasting game since forever. Mises severely criticized Irving Fisher's "price indexes" and the "stabilization policy" based on them, and forecast in the late 1920s that a major crash and depression would ensue as a result (Hayek also made such a forecast in 1928). Fisher by contrast argued that stocks had reached a "permanent plateau" and urged people to buy after the initial crash because he thought that the stabilization policy whose champion he was, implemented by the wise men at the Fed, had done away with business cycles.
Greenspan and Bernanke incidentally fell prey to the exact same conceit – to wit, Bernanke's in hindsight rather embarrassing "Great Moderation" speech of 2004, or Greenspan's "New Era" speech of early 2000.
However, we want to focus here on Alan Greenspan's forecasting record with regards to stocks specifically, mainly because he has once again commented on the stock market in the course of his book peddling tour.
From Bloomberg we learn:
“Former Federal Reserve Chairman Alan Greenspan said the stock market has room to rise from record levels. “In a sense, we are actually at relatively low stock prices,” Greenspan, who guided the central bank for more than 18 years, said in an interview with Sara Eisen on Bloomberg Television today. “So-called equity premiums are still at a very high level, and that means that the momentum of the market is still ultimately up.”
Greenspan said the stock market is “just barely above 2007” and the average annual increase in stock prices “throughout the postwar period” is 7 percent, which leaves room for a rise. “Price-earnings ratios are not hugely up,” he said. The market has “gone up a huge amount, but it’s not bubbly,” according to Greenspan."
Uh-oh. For a great overview as to why this valuation argument is complete hokum, we point you to this recent article at Zerohedge that presents the analysis of Damien Cleusix. As Cleusix points out, back in 1996, when Alan Greenspan worried about "irrational exuberance" in the stock market, the valuation ratios Cleusix watches were a full 40% below current levels.
As a matter of fact, Alan Greenspan's record as a stock market forecaster is especially dismal. Here is what he said in the New York Times on January 7 1973, exactly four calendar days prior to the second most important stock market peak of the entire 20th century:
“[I]t is very rare that you can be as unqualifiedly bullish as you can be now.”
As stock market forecasts in the 20th century go, only Irving Fisher's forecast of late 1929 was worse, and that was mainly because the 1930s bear market was accompanied by genuine deflation, which weighed greatly on nominal stock prices. By contrast, the inflationary backdrop of the 1970s ensured that nominal stock prices held up somewhat better.
Alan Greenspan: "unqualifiedly bullish" just four calendar days before the beginning of the second worst bear market of the 20th century.
It is true: Alan Greenspan is not exactly the Superman among forecasters. He is Bizarro from the Phantom Zone.
His stock market forecasting record didn't get any better as time passed. We already mentioned his "irrational exuberance" speech of late 1996 (the S&P 500 proceeded to advance by more than 100% from that point without hesitation; the Nasdaq Composite even rose by a stunning 400% in the four years following the "exuberance" speech). Then, as the market finally approached its manic top in early 2000, Greenspan decided it was a "new era" after all. From his January 13, 2000 speech before the Economic Club of New York:
“Four or five years into this expansion, in the middle of the 1990s, it was unclear whether, going forward, this cycle would differ significantly from the many others that have characterized post-World War II America. More recently, however, it has become increasingly difficult to deny that something profoundly different from the typical postwar business cycle has emerged. Not only is the expansion reaching record length, but it is doing so with far stronger-than-expected economic growth. Most remarkably, inflation has remained subdued in the face of labor markets tighter than any we have experienced in a generation. Analysts are struggling to create a credible conceptual framework to fit a pattern of interrelationships that has defied conventional wisdom based on our economy's history of the past half century."
It is true that there were large gains in productivity in the 1990s due to large capital investment in new technologies. The advances in computerization had a notable effect. Moreover, the fall of communism resulted at least for a time in a "peace dividend" in the form of declining government expenditures on defense. Moreover, it opened new markets and an expansion in trade ensued, with the concomitant gains delivered by the law of comparative advantage.
But it was a "new era" in the same sense that the roaring 20's were a "new era". It is astonishing that Greenspan couldn't see the parallel: precisely because large productivity gains led to lower prices of consumer goods, the credit expansion egged on by the Fed's accommodative policy became extremely large. After all, the only way to keep price indexes 'stable" when they would normally decline is by inflating all out! Even while price indexes seemed to be well behaved, the underlying economic distortions accumulated, as capital was malinvested while at the same time overconsumption increasingly depleted the pool of real savings. To his credit, Greenspan did have a few reservations, as can be seen from this remark later in the same speech:
“Productivity-driven supply growth has, by raising long-term profit expectations, engendered a huge gain in equity prices. Through the so-called "wealth effect," these gains have tended to foster increases in aggregate demand beyond the increases in supply. It is this imbalance between growth of supply and growth of demand that contains the potential seeds of rising inflationary and financial pressures that could undermine the current expansion."
In reality, the expansion had already become unsustainable - the major credit and asset boom Greenspan was in no small part responsible for was about to burst. All it took were a few small hikes in the Federal Funds rate and the house of cards came crashing down. Below you can see a historical chart with annotations, showing moves in the Fed Funds rate and various comments Greenspan made along the way.
The bubble blowers' chart.
We don't remember Dr. Greenspan hitting the talk show circuit in early 2009 to tell everyone that stocks were cheap. His timing, to say the least, is rather suspect. And once again, he fails to see the underlying driving force of the Bernanke echo bubble. Here it is:
Broad U.S. money supply TMS-2, with quarterly annualized and year-on-year change rates. Increase since 2000: 230%. Increase since 2008: nearly 82%.
We would submit that it is highly likely that Greenspan's remarks on the stock market will once again prove to be a contrary indicator. Maybe not immediately, but the clock is ticking.
Charts by: Chartstore, BigCharts, Michael Pollaro