The discipline of economics is in better shape today than it was in the 1970s, 80s and 90s. Here are five reasons why:
1. Now, economists test their theories. In the 1960s, the majority of published economics papers were entirely theoretical. Even in the 1980s, the typical top-10 journal published mostly theoretical work (reference here). Today, top journals like American Economics Review, the Journal of Political Economy and the Quarterly Journal of Economics are overwhelmingly dominated by papers that put theory to the test.
There is more data available than ever before, econometric software has improved immeasurably, and the cost of data, software and computing power has plummeted. Economists are doing more and better empirical research, so applied work is crowding out the second-rate theorizing once common in academic journals.
This is a good thing, for three reasons.
First, until a theory has been tested, there is no way of knowing whether or not it describes reality.
Second, theory scales. The same theory can be used to explain people's labour supply decisions in Manitoba and Mongolia. Empirical work, however, is particular: because Manitobans and Mongolians live in different environments and face different constraints, there is a value to carrying out individual studies of each countries' labour markets.
Third, policy makers often need to know "how much", for example, how much will labour supply change in response to a change in the after-tax wage rate. The increased volume of empirical research is a good thing because multiple studies using a variety of data sets and time periods may be needed to get a reliable estimate of the size of any given economic parameter.
I find deep and insightful theory far more satisfying than most empirical research. Yet because a marginal empirical piece is generally of greater value than a marginal theoretical piece, I celebrate the rise of empiricism.
2. Now, economists are better at establishing causality. Back in 1980s, McCloskey could, with some justification, take aim at economists who ignored the economic or policy significance of their research, and focussed solely on statistical significance (here and here).
Mere statistical significance doesn't cut it any more: good papers identify causal relationships. Economics deals too much with real people in messy real world situations to ever attain the precision of the natural sciences. But with greater use of natural experiments, creative use of truly exogenous phenomena such as climate or geography, as well as more behavioural and experimental research, economic methods are closer to "science" than they once were. The acceptance of these different types of research methods points to another way in which economics is better than it used to be:
3. Now, economics is (somewhat) more open to a range of ideologies and methodologies
On this blog, Nick Rowe has argued that everyone should be forced to take Econ 1000. I am not sure I agree. People who do not take Econ 1000 are not exposed to Mankiw's ten principles. They do not have the rational choice paradigm and the virtues of trade and markets drilled into them at a young and impressionable age.
Today, some people who call themselves economists have little exposure to the discipline until they start their PhDs. Grad schools, for the most part, spend more time instructing students on econometric and other technique than teaching people "how to think like an economist". Indeed, one could argue that increasingly papers are judged more by their technique and their data than by their adherence to the mainstream paradigm as summarized by Mankiw.
But people whose primary interests are in data and modelling, and who judge a paper by its treatment of endogeneity, are less likely to be bound by a strict neo-classical adherence to rational choice theory. They tend to be more open to a variety of methodological approaches, as long as there is room for some neat econometric analysis or experimentation: neuroeconomics, explorations of identity, norms, culture and values, application of economic methods any place data can be found.
4. Now, economics is engaging
People have always wanted to know what is likely to happen to interest rates, or why people are out of work, but once upon a time it was hard to engage people in microeconomic debates. That's no longer true. Economic research can be satisfying the way that a good detective story is: a mystery is created, and solved. Moreover, people can see themselves in research on the economics of everyday life, and think about how their own behaviour, or that of those around them, responds to incentives, or is shaped by systemic biases.
Not only is the research more engaging, but economists are working harder at engaging people with it - in books like YouNameIt-Onomics, and blogs like this one. More and more media outlets - the New York Times, the Atlantic, Macleans, and so on - feature regular contributions from serious economists. No longer are economists placed in the position of talking for 20 minutes on the phone to a reporter and then cringing at the results. We can communicate directly with the typical reader - who quite often has a better understanding of basic economic principles than the typical journalist.
5. Now, economic research is (in some ways) more democratic.
Thirty years ago, the only way to know what other researchers in one's area were working on was to physically travel to an academic conference. Those who could not afford to travel (or who did not get invited to the right kind of conferences) did not see papers until they appeared in hard copy journals, by which time the literature had often moved on. Even hard copy journals were not easy to access - one actually had to walk to the university library, and if the university did not subscribe to a specialized journal in one's area, the only option was to get it through inter-library loan, or buy it oneself.
Twenty years ago, simply submitting a paper to a journal took a not inconsiderable amount of time and effort - going to the library to look up the mailing address of the journal, printing out three hard copies of one's paper, writing and printing out a submission letter on university letterhead, organizing international postage. Now it takes half an hour, at most, at the computer.
Back then, accessing data internet connection has tens of thousands of datasets just a click away, the development of R means high quality statistical software is, literally, free, and thousands of economics articles are freely available on-line.
All of this has narrowed the gap between elite and non-elite institutions: a 2011 paper by Winkler et al found a noticeable lessening of the difference in publication rates at elite and non-elite institutions between 1991 and 2007, providing evidence that a combination of technological change and other factors, such as increased emphasis on research output at the non-elite institutions, has, in fact, made the profession more democratic than it once was.
Still, every silver lining has a cloud. Each one of these positive trends has a not-so-positive flip side.
1*. Economists test their theories. But only some test results get published.
Papers are more likely to get published if they find startlingly large, new or different results, if they find statistically significant effects, or if they find effects that accord with economic theory. This phenomenon is called publication bias, and has been documented, for example, here here or here. Of course, these studies of publication bias are themselves subject to publication bias, suggesting publication bias may not be as much of a problem as one might think.
Still, the discipline is not characterized by the kind of formulate-hypothesis-then-test-it scientific method taught in introductory econometrics textbooks, and never will be.
2*. Causality isn't everything. Correlations are interesting too.
Maternal employment has been linked to childhood obesity. But does maternal employment cause obesity, or is there some underlying, unobservable characteristic that is associated with both employment and chubby children? I don't know, but that doesn't bother me in the slightest.
Obsessing about causality can blind us to the fascinating, complex, if ultimately unknowable relationships around us.
3*. Economics is not that open to methodological diversity.
Well established economists like George Akerlof and Robert Frank can get away with pushing the boundaries of economics, as can those at more heterodox institutions, such as Nancy Folbre.
The young and untenured are best advised to write papers that stick to the standard conventions.
4*. Public engagement makes the senior administration happy, but no one ever got tenure by blogging.
The hard currency of academic success is still publication in refereed journals.
5*. Old barriers have been replaced by new ones.
As the cost of producing economic research has fallen, and the number of active academic researchers, particularly in Europe and Asia, has risen, the volume of research produced is too much for anyone to cope with.
An author's name, institutional affiliations, location, education and professional reputation, provide a signal of his or her ability. Referees would be irrational to ignore that information when assessing an individual's research. Winkler et al's study of the gap between publication rates in elite and non-elite institutions concluded, in the end, that "the story is more of constancy than of change, even in the face of changing technology and rising research expectations."
So the profession is far from perfect. But it is better than it used to be.
Happy Canada Day!