Financial Economics: Science or Alchemy?

Non-Certeris Paribus Markets

   Introductory economics textbooks usually begin with a certeris paribus assumptions. That is, holding everything else equal, if a increases then b decreases. “To beginners, it can seem odd to claim that economics is a science. After all, economists do not work with test tubes or telescopes. The essence of science, however, is the scientific method – the dispassionate development and testing of theories about how the world works. This method of inquiry is as applicable to studying a nation’s economy as it is to studying the earth’s gravity or a species’ evolution.” (Markiw’s Principles of Economics)

He continues,

“Although economists use theory and observation like other scientists, they face an obstacle that makes their task especially challenging: In economics, conducting experiments is often difficult and sometimes impossible. Economists, like astronomers and evolutionary biologists, usually have to make do with whatever data the world happens to give them.” […] “To find a substitute for laboratory experiments, economists pay close attention to the natural experiments offered by history. Economists make assumptions for the same reason: Assumptions can simplify the complex world and make it easier to understand.”

However, the scientific method cannot be applied in macroeconomics, because theories are not verifiable as the “everything else equal” assumption does not hold.

Again, the scientific method assumes that if everything else is kept without change then we can see the effect of changing variable x on variable y. Thus, the structure of the scientific method is:

If A, and everything else without change, then B.

However, if one of the two conditions of the logical statement is always false, no meaningful conclusion can be reached. The statement above cannot be verified, because we can never affirm with observations that if A is true then B is true, as we will never be able to hold everything else without change.

This is not a small matter. We can and do have academic results that contradict each other. A paper claims that if A increases then B increases, and another one finding that if A increases B decreases. The environment in which the two experiments took place were different and were changing, perhaps they were different countries, or different time periods; perhaps a country was at war, there was a global crisis or different economic policies, and thus we can reach opposing conclusions using the scientific method in a non-scientific environment.

In physics we cannot have the apple falling from the tree and another scientist concluding that it should fly up from the tree. There might be disconfirming evidence that requires the development of a new theory or a refinement of the current theory, but there is no two or more widely accepted theories that reach different conclusions over the same event with supporting data.

Thus, it is better to stop pretending that economics is a science and think how we can better understand the economics of the world. One alternative is to realize that economics is not a science, and its models are “parables” with varying degrees of usefulness which in turn depend on their application.

In financial markets, the conclusions reached by economics and finance are far off the mark in terms of investing. If one were to listen to efficient market theories one would reach the conclusion that there it is futile to try to beat the market, and yet some people consistently beat it. Another way to see the world is to use differentiated analysis to untangle cause and effect. For instance, we observe that banking profitability increases in Brazil from 2003 to 2007. On the other hand, concentration of the industry increased, so an financier-scientist gathers data, constructs measures of profitability and concentration, runs some regressions and some tests and reached the conclusion that indeed increased concentration and be associated with increased profitability. However, more knowledge about the world will reveal that indeed during 2003 and 2007 profitability increased in most countries. Now, the common sense analysis will tell us that if most banking systems profitability increased, it is probably due to some global factor during those years and not to banking concentration in Brazil. It could be the case that banking concentration increased in all countries during those years, which although plausible does not seem to be the most reasonable explanation as each country has its own competition regulator with different criteria to allow mergers and acquisitions and different willingness of the banks themselves to consolidate. So before spending months analyzing this hypothesis, it would be better to investigate the system overall going back decades of history to try to disentangle the most probable factor that might affect many banking systems at the same time. Perhaps, a decrease in interest rates in the US. This explanation seems more reasonable.

The next step would be to analyze banking profitability in other instances in which the Fed decreased interest rates. We will observe that that during the early 80s to mid-80s interest rates were decreasing and Reagan’s tax cuts and star wars spending led to increasing deficits, at the same time bank profitability increases and shares increased. However, not at the global level, as in the 80s episode the dollar was rising and thus negatively affected commodity prices and emerging-market banks. So as you can see, the analysis is not so simple. However, with a lot of knowledge of different episodes, variables, and countries one can begin to reach the most likely to be correct conclusions and not the oftentimes spurious conclusions of so-called scientific conclusion of tests.

Thus somebody like George Soros, without the technical tools of the academic scientist can beat the market. Soros and other investors like him do not assume that everything else is equal, but analyze the whole situation and reach a more profitable and accurate conclusion for the specific situation at hand.

In the meantime, economists are busy trying to reach generalizations, knowledge useful for the long-run, for the average to try to obtain tendencies, which are much less accurate and profitable for analyzing a specific situation.

Predicting Human Behavior and Knowledge

Moreover, trying to analyze complex problems requires, not only information, but knowledge. That is the way in which that information is processed to reach conclusions. Knowledge is virtually impossible to model with equations and math. The creation of financial and business knowledge is long process that takes years of thinking, reading, and using differentiating analysis.

For instance, economists do not study the fundamentals of different sectors of the economy. Balance sheets, income statements, and how specific businesses make money. Thus they cannot know if a company will generate alfa. Economist-scientists simply do not have the tools to make those predictions. Those predictions are based on having the correct knowledge, not on econometric tests or simulations.

So there are two main reasons economists do not get the most profitable and accurate answers. At the global macro level, they do not take into account the situation in its unique context, and they do not apply differentiating analysis to generate knowledge that can serve to predict the evolution of a specific problem in a specific context.

Second, economists trying to have far-reaching tools that can analyze from the labor market to international finance, to game theory, simply do not have the specific knowledge required to understand specific businesses in specific sectors.

On the other hand, we have Warren Buffett how has studied specific businesses all his life and is able to make a few accurate predictions among thousands of businesses. This is enough to generate outstanding returns. Moreover, as he demonstrated in his piece from 1984, there is a group of investors with common style of investing that can produce superior returns which is difficult to attribute to chance. It is easier to attribute to some knowledge they share in common, namely, having a similar style of processing information acquired from the same professor.

Simply, economics tries to do too much and ends up transmitting inaccurate knowledge and inaccurate reach of its tools to those who studied it, resulting on lost social welfare economic-scientific thinking has led to costly financial crisis, among other things.

The scientific method consists of,

1.   Observation

2.   Hypothesis

3.   Testing

4.   Conclusion

Economics cannot do scientific testing as the environment is not kept equal. Thus, the scientific method cannot be applied and reach true and accurate conclusions. Economists can pretend to scientifically test by using sophisticated statistical and econometrics tests on a complex mathematical model, but that is just shifting the attention from the real problem. There is no real scientific environment to be able to test the hypothesis.

Economists in academia can life until retirement pretending to do science as there is colleagues confirm the conclusions that they are performing scientific experiments. However, if an economist were to reach investing conclusions and put the money in their scientific theories and lost it, I think they would rethink the seriousness of the science.

It is easy to say that economics is wrong, but it is more difficult to say with solid arguments in the language of economics why economics is wrong.