A professor at my university once asked the master class during the middle of an "Economic Development" lecture, "Does trade liberalization increase or decrease wage inequality in a developing country?" As a proactive person and a good, participating student, I raised my hand to answer the question with the following reasoning: It depends on which sector of the country is being affected by trade liberalization. If the lesser share of the employment force, or the higher-wage workers, are affected by the trade liberalization, then it can increase inequality. If the bigger share of the employment force, or lower-wage workers, are affected by trade liberalization, then inequality might decrease.
Regardless of whether my answer was wrong or right, my professor, looking puzzled, asked the following: "Okay, yes, but which economic model do you base your assumptions on?"
To which I just replied, “Well, I used my previous knowledge, common sense, and the assumptions we made so far." He replied:
"That's good, but here we are thinking in models now, and the Stomper-Samuelson theorem suggests that inequality will increase as a result of trade liberalization.” ”
The Limits of Models
It is not hard to argue that not having enough variables leads to incomplete assumptions, yet having too many variables can be misleading and overfitting.
Economists, just as social scholars in general, tend to think in models in order to help and understand how society works, but from an economic perspective. In cases, they do help to understand trends, and they also explain to a certain extent the behavior of economic actors. But in models, some assumptions and factors are taken into consideration and some are not. Nonetheless, they believe they can conclude how economic actors think. Different scholars take various numbers of variables. Some like to keep it simple and try to explain the measured economic benefits/losses with only a few variables, while others like to complicate it and bring in as many variables as humanly comprehensible, which can be in connection with the measured economic benefits/losses. It is not hard to argue that not having enough variables leads to incomplete assumptions, yet having too many variables can be misleading and overfitting. However, we tend to learn and teach these economic models, and we are disregarding reality and common sense in our thinking and acceptance of absurd models under the thinking that:
"This is the economic theory. This is the model. This is the set of data we have. These are the assumptions."
But is it accurate to limit our thinking to these factors and not think outside the box?
Many economic theories, like the most famously known Solow-Swan model of endogenous growth or Keynesian economic models of employment, interest, and money make underlying assumptions and presumptions like:
- rational decision-makers in the economy
- agents have symmetrical information flow
- markets clear without friction
- excluding externalities
One might say yes, they are generally estimating the benefits/losses of economies with relative accuracy. However, with any scholarly research, if we do not take into account the necessary assumptions of the empirical reality, the results might be correct but will result in the phenomenon of GIGO effect—referred to by economists as "Garbage In - Garbage Out"—as a general critique of Computable General Equilibrium (CGE).
The Case of NAFTA
One tangible example where the limits of economic models allegedly collided with reality but were nevertheless accepted as "evidence" in public policy debates involved models to simulate the effects of NAFTA, the North American Free Trade Agreement.
Many times, models are accused of leaving out externalities because of economists’ pro-free market bias, which is not necessarily incorrect to advocate for but results in half-true assumptions and half-complete models. NAFTA is a good case study. In recent years, we have seen a rising skepticism toward the deal and a generalized dissatisfaction because of firm and output losses, which was topped by Donald Trump being elected president of the United States. Even today, economists warn of the serious implications of Trump leaving NAFTA or the fact that he actually left TPP, but these economists were reluctant to warn against the serious implications of having these deals in the first place and only saw the benefits of their model-based calculations.
Even economist Joseph Stiglitz argued in his Nobel Prize-winning lecture that the information symmetry we base our general models on is outdated, and the neoclassical paradigm of the past isn't viable in today's economy and can lead to faulty policy implications and recommendations that arise from their unrealistic assumptions.
Focus on the Tangible
"There are no limitations to the mind except those that we acknowledge."
Of course, we discuss in brief the shortcomings of models in understanding economies, but we tend to focus less on this and more on the tangible and comprehensible implications of a model; therefore, incorrect conclusions are made.
Well, Napoleon Hill (not to be confused with the French Napoleon) might have been right in his self-help book Think and Grow Rich, where he discussed individuals who seek to motivate themselves but acknowledges that deliberate shortcomings will still result in incorrect answers. So, instead of trying to understand the economy through models based on a set of assumptions to simplify it to our brain, to get a full picture, we need to understand the economy and discuss certain issues while also involving and comprehending the shortcomings of models. Of course, we discuss in brief the shortcomings of models in understanding economies, but we tend to focus less on this and more on the tangible and comprehensible implications of a model; therefore, incorrect conclusions are made.
Yes, there are no limitations to the mind, but there are also no limits to how many limitations we can impose on the mind that misguide us in the process of better understanding of economics.