By Ron Robins. First published April 21, 2011, in his weekly economics and finance column at alrroya.com
“…both risk models and econometric models… are still too simple to capture the full array of governing variables that drive global economic reality,” wrote Alan Greenspan, former chairman of the US Federal Reserve in the Financial Times on March 16, 2008. And if anyone should know about the quality and predictive validity of such models, it would be Mr. Greenspan. Time and again it has been shown that reliance on the predictions from such models is foolhardy.
It was the reliance on, and failure of their predictions, that caused enormous global financial and economic carnage in 2008 and 2009. Yet today dependence on these models seems greater than ever. I suggest our overt focus and use of them is often a wasted effort.
A truth that many modellers and their followers seem to have difficulty accepting is that the past—which most modellers use to prognosticate the future—has frequently been shown to be a poor basis upon which to determine future outcomes. Modellers can continue to refine their models in great detail, and then some unusual event occurs with a one in a million chance of happening—such as the US sub-prime mortgage fiasco—and their models fail. Sadly, the variables which may encompass a one in a million event are numerous. Among them are sudden changes of investor attitudes, weather patterns, geological events, and political and social upheavals.
If we look around today from the sudden movements in sovereign bond markets to the extraordinary weather recently in Australia, to the horrific Japanese earthquake, tsunami and nuclear reactor troubles, to the political upheavals in North Africa and the Middle East—all are kinds of exogenous events that can trash the predictions of the most exacting risk or econometric model.
Furthermore, a ‘perfect’ econometric model would only be possible, metaphorically speaking, if the modeller had ‘the mind of the creator.’ Only then perhaps, could all be known and predicted. Sadly—and I do not mean any disrespect to the modellers—I do not believe that many (if any) of them have that level of intelligence and consciousness at this time. So those constituencies that trust in these models are doomed to suffer continuing disappointments.
Another problem with these models is how to model for human behaviour, as it is both rational and irrational at different and unpredictable times. Therefore, before such modelling can ever hope to fully succeed, it must completely understand human consciousness: who we are, and how and why we act. And the modellers are a long, long way from such an understanding. Incidentally, there is a branch of economics, ‘behavioural economics,’ that is moving in that direction. I wish them good luck with that!
Economists today, unlike those of earlier eras, seem to believe that the only way they can be perceived as legitimate is to be scientifically oriented. Hence their passion for increasingly complex models and their statistician-like orientation.
The type of economic modelling that incorporates mathematics and statistical relationships to economic data, is termed econometrics. Google econometrics and you will probably find over 5,000,000 links. They are largely links to innumerable academics, research institutions, studies, papers and journals. With so much effort put into this field, any independent observer could conclude that econometrics must be a highly successful and seemingly scientific endeavour. It reminds me of the enormous quest for artificial intelligence (AI) to recreate the abilities of the human mind in computers. At least AI is somewhat plausible as it advances the field of computing and robotics which have many, many practical applications that we all know about.
But unlike AI research, economic and econometric models—with their significant variances and failures—have much less to offer society at this time. Mark Thoma, Professor of Economics at the University of Oregon offers these pertinent remarks in his blog, Economist’s View, on February 8. “Much of the uncertainty in economics derives from our inability to do laboratory experiments, and that includes uncertainty about which model best describes the macroeconomy. When the present crisis is finally over, those who advocated fiscal policy, those who advocated monetary policy, and those who advocated no policy at all will all say ‘I told you so’ based upon their reading of the evidence… the answers you get are only as good as the model used to get them, and considerable uncertainty remains over which macroeconomic model is best.”
In the 19th century’s Europe and North America, there were no econometric models (not in the way we know of them today), yet those continents experienced unprecedented economic growth. And the concept of gross domestic product (GDP)—which is usually a top concern in econometric modelling—was not created and used until World War II.
We know that econometric models are unreliable in providing information on how economies behave as well as their projections of future economic activity. Similarly, modelling for financial risk has been shown to be more than problematic and history shows reliance on risk models brings eventual failure and grief.
Therefore, given the facts, we need to be much, much less anxious about trying to create perfect risk and econometric models—and not rely on these models, generally. After all, it was mostly intuition and drive, not decisions based on risk and econometric models that led our greatest inventors, financiers, entrepreneurs and leaders to great success, thereby creating our modern economies.
Copyright alrroya.com
Enlightened Economics 