The dictionary of the Royal Spanish Academy defines causality as “law by virtue of which effects are produced”. And it is evident that without knowledge of this type of law we cannot aspire to understand facts and phenomena of interest. In fact, scientific research consists of nothing more than the aspiration to understand how the world works from the data that history and experience leave us.
In the field of economics, identifying these types of laws (cause-effect relationships) is especially difficult for three reasons:
Many of the variables that describe a particular economic phenomenon are interrelated. Or put another way, cause-effect relationships are multidimensional and multidirectional because there are many possibilities by which economic phenomena originate.
The inability to perform randomized experiments in a generalized way as it is done in other sciences, for example, in medicine.
They do not exist economic laws immutable: cause-effect relationships that may exist in a given place and time do not have to be valid in different contexts.
The inability to establish and use cause-effect relationships is especially relevant when designing economic policy programs, which are nothing more than the implementation of a set of measures (instruments) to achieve certain ends (objectives).
The traditional way in which economists have tried to overcome this difficulty has been by constructing economic models. That is, a set of assumptions about the behavior of economic agents and their interactions that can be measured transmission mechanisms. That is, to what extent the economic variables respond to others.
Being aware that many models can be elaborated (and that not all of them capture relevant realities for the question of interest), they resorted to econometrics to develop a methodological instrument with which to discern which models are more in line with reality and, therefore, they can be more useful to inform economic policy decisions. As they say, although all inexpensive models are fake, some can be useful.
However, without the possibility of developing randomized experiments, the identification of cause-effect relationships is very limited and the empirical testing of economic models always requires strong and generally not very credible assumptions.
It was this dissatisfaction that, at the beginning of the nineties of the last century, led a group of economists to propose a different approach to cause-effect relationships in economics. Among them the most recognized were Alan Krueger, sadly passed away in 2019, and Joshua Angrist, David Card Y Guido Imbens, who have been awarded the 2021 Nobel Prize in Economics for their empirical contributions to labor economics (Card) and for their methodological contributions to the analysis of causal relationships (Angrist and Imbens).
The Card, Angrist and Imbens approach to the analysis of causal relationships, which is known by the label of the credibility revolution, has radically changed the way in which empirical research is done in economics and, also, in many other fields of the social sciences.
His method consists of exploring natural experiments. That is, events that, for purely random reasons, affect groups of individuals with similar characteristics differently and, therefore, comparison between these groups makes it possible to identify the effects of these events.
It is an approach without any kind of theoretical foundations or judgments a priori on the cause-effect relationships to be identified and, therefore, free from the limitations inherent to the empirical testing of economic models.
David Card pioneered the exploration of natural experiments, with studies on the effects of increases in the minimum wage, in collaboration with Alan Krueger, or on the economic effects of immigration. Angrist, who has also analyzed natural experiments, mainly in the field of educational economics, has developed together with Guido Imbens the methodological arsenal necessary to estimate cause-effect relationships from natural experiments.
Effects of the rise in the minimum wage?
Among all the works of the winners of the 2021 Nobel Prize in Economics, the most famous without any doubt is that of Card and Krueger on the effects of increases in the minimum wage.
The traditional view on the relevant cause-effect relationship in this case was very simple: if the minimum wage rises above the wage that balances the supply and demand of labor, employment decreases.
The transmission mechanism underlying this statement is also very simple: increases in the minimum wage mean increases in unit labor costs for companies (that is, what they have to pay workers in relation to their productivity) and, consequently, they will demand less work.
It happens, however, that increases in the minimum wage could also cause increases in productivity (and thus unit labor costs would not rise) or that the company had monopsony power (that is, it was not subject to competition from others). employers when hiring workers), in which case increases in the minimum wage may lead to increases in employment.
$ 4.25 vs. 5.05 ¢
With all these conditioning factors, the discussion about the effects on employment of increases in the minimum wage can only be resolved empirically. Card and Krueger realized that there was a natural experiment that could inform this discussion.
In April 1992 the minimum wage in New Jersey rose from $ 4.25 per hour to $ 5.05 per hour. They obtained data on employment in 410 fast food restaurants located in New Jersey and in eastern Pennsylvania, a state adjacent to New Jersey. Therefore, in this case the randomization is given by an administrative border that causes similar restaurants in similar locations to be affected differently by the increase in the minimum wage (those in New Jersey had to raise wages, those in Pennsylvania did not).
After comparing the growth of employment in the group of affected (treated) with that of the group of unaffected (control) they found that there were no significant differences, for which they concluded that the increase in the minimum wage had no negative effects on employment.
New ways of seeing
Since then, the use of natural experiments to identify cause-effect relationships has grown exponentially in both economics and other social sciences.
The advantages of the approach are great: a more credible identification of the relationships of interest and the potential to obtain empirical results that can make us think about transmission mechanisms that were not taken into account. The drawback is that, since the results refer to a specific geographical and historical context, their external validity is weak.
In any case, thanks to Card, Krueger, Angrist and Imbens we have another way of doing empirical analysis with which to advance our knowledge of reality.
We would like to say thanks to the writer of this post for this awesome web content
On cause-effect relationships and the 2021 Nobel Prize in Economics