Judgmental Forecasting and Fundamental Analysis

Judgmental Forecasting

and Fundamental Analysis

© Martin A. Armstrong

Each year I personally meet with more than 1,000 corporations and I lecture before another 5,000 to 6,000 at various business gatherings. With the exception of our long standing clients, better than 90% regard technical models for forecasting as not quite on par with good old fundamental analysis. Many believe that fundamental analysis is far more sound due to its roots being based upon some sort of fact in regard to supply and demand.

While the vast majority of corporate, business, investment and speculative decisions are based upon fundamental news, this same majority remains quite skeptical about fundamental forecasts. Nevertheless, forecasts based upon technical models offer no fundamental explanation by themselves, which means they cannot respond to the simple question of “why?”. When I was visiting a brokerage house I saw one of those joke signs posted above the screen on the bond desk. It read: “The same people who believe in economists believe in fortune-tellers.” Although this is perhaps merely one of those funny signs posted in all offices, it also takes a stab at a perceived truth. I have often stated that the profession of being an economist is rather unique. As far as I am aware, it is the only profession where you can become world renown and win the Nobel Prize without ever being correct.

Numerous articles are appearing around the world tell the saga of how economics is looking for another John Maynard Keynes. All the macroeconomic theories have crumbled before our very eyes. Not one Nobel Prize winner’s theories have succeeded in even identifying our present state of economic affairs. At the end of last year, I cannot begin to count the number of stories in the media, which began with the conviction that a recession was going to hit right away starting day one in 1989.

We even find major financial newspapers in several countries who market themselves as the source for business information which will be the key to your success. Yet these same financial papers have yet to forecast a single future event with any degree of accuracy. Instead, following the ’87 stock market crash they talked up the ’29 fears. When oil tumbled into last October the ran stories about how it would fall to $5. Perhaps if one uses the media as a perfect forecaster of what will not happen, success can at last be achieved.

Fundamental analysis has one major draw back – human interpretation. It is not that the facts about supply or demand are necessarily incorrect, the true culprit lies in the human judgmental forecasting. As a human being, there isn’t a soul alive or dead who has ever been able to forecast the future consistently based upon personal gut feelings, interpretations or with ESP. Nevertheless, judgmental forecasting still remains the number one method of planning for the future.

Ben Franklin, while in his ’70s, was asked by his colleagues for his opinion on a matter of state. He replied that with all his wisdom and experience he had arrived at the conclusion that he should not trust his own judgment. For with each passing year, he came to realize how little he knew the year before.

There have been several important studies undertaken to determine the performance of various types of forecasting. One of the best known papers on the subject was published in 1981 by R.M. Hogarth and S. Makridakis (The Value of Decision Making in a Complex Environment). Hogarth and Makridakis arrived at some very interesting conclusions after studying a variety of forecasts rendered during the 1970s. They reported their findings along these lines:

1) Forecasts beyond the short-term (three months) can be very inaccurate due to changes in long-established trends, systematic bias or errors and that historical data may provided conflicting scenarios with future trends.

2) Forecasts based upon different methods, systems, assumptions or models usually vary considerably. The problem which this introduces then comes back to the evaluation of which forecast shall be used. This can often present insurmountable problems in decision making or take longer that the process of forecasting itself.

3) Objective approaches to forecasting have performed as well or better than judgmental forecasts which was also the conclusions of Camerer 1981 and Dawes 1976.

4) Short-term forecasts tend to be more reliable due to the considerable inertia which prevails during the course of a given trend.

One example that Hogarth and Makridakis provided was a study of oil forecasts. In 1972 the generally accepted forecasts called for a continued trend in oil prices with no substantial rise. Oil was trading at the time under $2 per barrel. Following the OPEC situation, sentiment began to turn very negative, meaning that oil prices were perceived to be on the advance between 1974 and 1979 with no end in sight. Forecasts during this period for $100 a barrel oil by 1990 were in the majority. During 1979-1982, the oil glut perception emerged and forecasts were revised to $20-15 a barrel by 1990. Of course we all know how bearish the general crowd became on oil just last October. One so-called “oil industry specialist” forecast that oil would drop to $5 and stay under $10 until the end of the century.

In 1976, V.A. Mabert published his paper “Statistical Versus Sales-Force Executive Opinion Short Range Forecasts: A Time Series Analysis Case Study.” Mabert took sales forecasts which had been based upon the opinions of corporate and sales force personnel. He then studied the accuracy of these fundamentally based forecasts and compared them with three different quantitative methods. The comparisons were conducted on a mean absolute deviation and on a mean absolute percentage error basis. He found that over the case study period of 5 years, the judgmental forecasts proved to be the most inaccurate while the Box-Jenkins method provided the best forecast of the four methods involved.

There have been numerous studies on this subject and the single conclusion that one finds is that judgmental forecasting is inherently the worst. The human emotion tends to cloud the perspective. Whenever a market or economic trend is falling rapidly, fear wins over reason. Hence the fundamentalist cannot see any reason why the current trend will change. This emotional basis provides the error in judgment.

Regardless of how many case studies reveal that judgmental forecasting produces the least accurate result, judgmental analysis is the accepted norm. Unless we realize that our mistakes are caused by errors in our own judgment – there will never be any hope of avoiding mistakes in the future.