The 2021 Nobel Prize in Economics — All About “Natural Experiments”
Stockholm has had a pretty busy week. This year’s Nobel Prize winners were announced recently and three of them, in particular, stood out for their empirical contribution to labour market economics.
David Card, Joshua Angrist and Guido Imbens, three economists based out of the US were awarded the Nobel Prize in Economics for 2021 due to their groundbreaking work with the use of “natural experiments”. Their research has been instrumental in offering new and improved insights about the labour market, particularly, its effects on minimum wages, immigration, education etc.
Let’s see what natural experiments are all about.
Au Naturel
The definition as given in The New Palgrave Dictionary of Economics is this:
Natural experiments or quasi-natural experiments in Economics are serendipitous situations in which persons are assigned randomly to a treatment (or multiple treatments) and a control group, and outcomes are analysed for the purposes of putting a hypothesis to a severe test.
What this means is that it is a method to study real-world situations and their effects on individuals or groups, similar to the way one studies the participants in a clinical trial.
Let’s try and simplify this. In experimental sciences and studies like, say, clinical trials, scientists can decide who receives a treatment or application and who doesn’t. They can accordingly check and measure for various metrics (like sensitivity, responsiveness) in different participants.
These are called randomised experiments. In fact, the three recipients of the 2019 Nobel Prize in Economics were recognised for their work in these randomised experiments that lead to alleviating poverty.
But in social sciences, scientists don’t have control over which individuals shall receive a treatment or application, or essentially be chosen to become part of a control group.
Economists, therefore, often draw their conclusions by studying policy changes or historical events and trying to determine their causes and effects with contemporary development. Although they can use natural experiments to study these effects, these are difficult to conduct for two reasons:
- The question of who or what will serve as the experiment.
- The absence of clear demarcation between the experiment and the control groups.
Unlike a laboratory setting, in a society, people choose their own groups, circles and communities and even move across those groups. There is no way to determine these group dynamics without artificially influencing their lives which would effectively compromise the study and leave the results moot (remove the “natural” from natural experiments, precisely).
But David Card found an innovative and pioneering way to do so by comparing fast food workers in Pennsylvania and New Jersey. He showed that one can indeed “control” who receives a benefit and who doesn’t in a social set-up through natural experiments.
Fast Food and Fidel Castro
There was a long-standing assumption among economists in the 20th century about there being a tradeoff between higher wages and jobs. It was believed that if the minimum wages went up then some workers would get higher wages at the expense of others who would be laid off.
Card, along with his Princeton colleague Alan Krueger (now deceased), debunked this theory in the 1990s. They studied the actual effect of higher wages on fast food workers in two American states and found no substantial drop in employment. The New Jersey restaurants raised wages (from $4.25 to $5.05 per hour) whereas the ones in neighbouring Pennsylvania didn’t.
In another study, Card examined the impact of a decision made by Castro in Cuba in 1980 that allowed all Cubans who wished to leave the country the freedom to do so. As a result, Cubans began thronging the shores of Miami in droves, causing mass migrations. But even in this instance, Card found no negative effect on wage or labour demand for Miami residents with low levels of education.
So, essentially, the broad questions regarding cause and effect of social factors (like education and immigation) and policy/welfare measures (like minimum wage limits) weren’t impossible to measure after all, owing to Card’s research. He showed how natural experiments could be used to study and measure these indicators.
Angrist and Imbens
If Card proved the means of conducting natural experiments then the Angrist-Imbens duo proved the method to do so and fetch conclusive results.
The methodology they developed is called LATE which stands for Local Average Treatment Effect. Let’s consider education as an indicator for individual development. By applying the LATE model, Angrist and Imbens showed that an extra year of schooling leads to an increase in income by approximately 9%.
Although this doesn’t indicate personal incomes of individuals per se, the approximation is an estimated size of the impact that education is shown to have on people’s lives.
This approach has since become extremely relevant to measuring the cause and effect relationship between policies and consequences. Especially during the ongoing pandemic when natural experiments like the above have allowed researchers to study the effect of mask mandates, social distancing policies and supplemental unemployment benefits (stimulus checks, food subsidies etc.).
Revolutionising empirical research in this manner has enabled scientists and economists to answer important socio-economic questions which earlier remained unanswered due to lack of comparison tools. How would a nation’s immigraton policy impact its labour market? How are national income levels affected if public schooling benefits are extended for another year? How often and moderately should the minimum wage limits be revised to avoid mass retrenchment?
Questions like these not only impact policymaking but also demonstrate the causality of indicators like education, training and human capital on earnings and future incomes. They also debunk popular notions in developed economies against policies like immigration (sometimes bordering on xenophobia) which stir up anxiety amongst the natives who are worried about losing their jobs to immigrants.
The Nobel laureates this year have shown that the key to determining the impact of such outcomes is to treat these chance events or policy changes (like wage raises and migration) in a way that resembles clinical trials in medicines. With increased digitisation and dissemination of archival records, the data generated from natural experiments has also become increasingly easier to interpret and employ in studies in the 21st century.
Thanks to these noble causes, there are quite a few policy lessons to be learnt from these natural experiments.
(Originally published October 13th 2021 in transfin.in)
