Empirical = scientific ... ?

I am puzzled by this ope-ed by Raj Chetty in the New York Times today. So, apparently, economics is a science because economists run experiments. Well, actually, most economists do not run experiments, including most if not all Nobel (memorial) prizes in economic science...

And the last sentence is the most puzzling: "it is simplistic and irresponsible to use disagreements among economists on a handful of difficult questions as an excuse to ignore the field’s many topics of consensus and its ability to inform policy decisions on the basis of evidence instead of ideology." My problem comes from the fact this year's Nobel prize addresses a "difficult question" for which the three awardees give pretty different answers, with at least one of them being very ideological (if not in his research papers, at the very least in his public positions since the advent of the financial crisis in 2008...).

Yes, Economics Is a Science

By RAJ CHETTY

Published: October 20, 2013 105 Comments

• FACEBOOK
• TWITTER
• GOOGLE+
• SAVE
• E-MAIL
• SHARE
• PRINT
• REPRINTS
• 

CAMBRIDGE, Mass. — THERE’S an old lament about my profession: if you ask three economists a question, you’ll get three different answers.

Connect With Us on Twitter

For Op-Ed, follow@nytopinion and to hear from the editorial page editor, Andrew Rosenthal, follow@andyrNYT.

Readers’ Comments

Share your thoughts.

• Post a Comment »
• Read All Comments (5) »

This saying came to mind last week, when the Nobel Memorial Prize in Economic Science was awarded to three economists, two of whom,Robert J. Shiller of Yale and Eugene F. Fama of the University of Chicago, might be seen as having conflicting views about the workings of financial markets. At first blush, Mr. Shiller’s thinking about the role of “irrational exuberance” in stock markets and housing markets appears to contradict Mr. Fama’s work showing that such markets efficiently incorporate news into prices.

What kind of science, people wondered, bestows its most distinguished honor on scholars with opposing ideas? “They should make these politically balanced awards in physics, chemistry and medicine, too,” the Duke sociologist Kieran Healy wrote sardonically on Twitter.

But the headline-grabbing differences between the findings of these Nobel laureates are less significant than the profound agreement in their scientific approach to economic questions, which is characterized by formulating and testing precise hypotheses. I’m troubled by the sense among skeptics that disagreements about the answers to certain questions suggest that economics is a confused discipline, a fake science whose findings cannot be a useful basis for making policy decisions.

That view is unfair and uninformed. It makes demands on economics that are not made of other empirical disciplines, like medicine, and it ignores an emerging body of work, building on the scientific approach of last week’s winners, that is transforming economics into a field firmly grounded in fact.

It is true that the answers to many “big picture” macroeconomic questions — like the causes of recessions or the determinants of growth — remain elusive. But in this respect, the challenges faced by economists are no different from those encountered in medicine and public health. Health researchers have worked for more than a century to understand the “big picture” questions of how diet and lifestyle affect health and aging, yet they still do not have a full scientific understanding of these connections. Some studies tell us to consume more coffee, wine and chocolate; others recommend the opposite. But few people would argue that medicine should not be approached as a science or that doctors should not make decisions based on the best available evidence.

As is the case with epidemiologists, the fundamental challenge faced by economists — and a root cause of many disagreements in the field — is our limited ability to run experiments. If we could randomize policy decisions and then observe what happens to the economy and people’s lives, we would be able to get a precise understanding of how the economy works and how to improve policy. But the practical and ethical costs of such experiments preclude this sort of approach. (Surely we don’t want to create more financial crises just to understand how they work.)

Nonetheless, economists have recently begun to overcome these challenges by developing tools that approximate scientific experiments to obtain compelling answers to specific policy questions. In previous decades the most prominent economists were typically theorists like Paul Krugman and Janet L. Yellen, whose models continue to guide economic thinking. Today, the most prominent economists are often empiricists like David Card of the University of California, Berkeley, and Esther Duflo of the Massachusetts Institute of Technology, who focus on testing old theories and formulating new ones that fit the evidence.

This kind of empirical work in economics might be compared to the “micro” advances in medicine (like research on therapies for heart disease) that have contributed enormously to increasing longevity and quality of life, even as the “macro” questions of the determinants of health remain contested.

Consider the politically charged question of whether extending unemployment benefits increases unemployment rates by reducing workers’ incentives to return to work. Nearly a dozen economic studies have analyzed this question by comparing unemployment rates in states that have extended unemployment benefits with those in states that do not. These studies approximate medical experiments in which some groups receive a treatment — in this case, extended unemployment benefits — while “control” groups don’t.

These studies have uniformly found that a 10-week extension in unemployment benefits raises the average amount of time people spend out of work by at most one week. This simple, unassailable finding implies that policy makers can extend unemployment benefits to provide assistance to those out of work without substantially increasing unemployment rates.

Other economic studies have taken advantage of the constraints inherent in a particular policy to obtain scientific evidence. An excellent recent example concerned health insurance in Oregon. In 2008, the state of Oregon decided to expand its state health insurance program to cover additional low-income individuals, but it had funding to cover only a small fraction of the eligible families. In collaboration with economics researchers, the state designed a lottery procedure by which individuals who received the insurance could be compared with those who did not, creating in effect a first-rate randomized experiment.

The study found that getting insurance coverage increased the use of health care, reduced financial strain and improved well-being — results that now provide invaluable guidance in understanding what we should expect from the Affordable Care Act.

Even when such experiments are unfeasible, there are ways to use “big data” to help answer policy questions. In a study that I conducted with two colleagues, we analyzed the impacts of high-quality elementary school teachers on their students’ outcomes as adults. You might think that it would be nearly impossible to isolate the causal effect of a third-grade teacher while accounting for all the other factors that affect a child’s life outcomes. Yet we were able to develop methods to identify the causal effect of teachers by comparing students in consecutive cohorts within a school. Suppose, for example, that an excellent teacher taught third grade in a given school in 1995 but then went on maternity leave in 1996. Since the teacher’s maternity leave is essentially a random event, by comparing the outcomes of students who happened to reach third grade in 1995 versus 1996, we are able to isolate the causal effect of teacher quality on students’ outcomes.

Using a data set with anonymous records on 2.5 million students, we found that high-quality teachers significantly improved their students’ performance on standardized tests and, more important, increased their earnings and college attendance rates, and reduced their risk of teenage pregnancy. These findings — which have since been replicated in other school districts — provide policy makers with guidance on how to measure and improve teacher quality.

These examples are not anomalous. And as the availability of data increases, economics will continue to become a more empirical, scientific field. In the meantime, it is simplistic and irresponsible to use disagreements among economists on a handful of difficult questions as an excuse to ignore the field’s many topics of consensus and its ability to inform policy decisions on the basis of evidence instead of ideology.

Raj Chetty is a professor of economics at Harvard.

Income inequality: country comparison and evolution

France has been doing very well since the mid-1980, much better than average, in terms of the evolution of income inequality ... even though the absolute level of income inequality remains quite high, especially for a country that taxes so much.

In-kind benefits, incentives, and labor supply

My former grad school colleague François Maniquet (from CORE) provides an excellent short summary of what we know about (the effects on labor supply of) in-kind benefits, based on a survey by my co-author Firouz Gahvari.

In-kind benefits, incentives, and labor supply by fmaniquet

In-kind benefits, incentives, and labor supply

Policies dedicated to directly increasing the material well-being of poor people can broadly be ranked in three categories. The first category is the direct money transfers to poor people. The second category is the subsidy to prices of goods that are consumed by poor people. The third category is the in-kind transfers, that is, the free provision of goods to poor people. The goods that are typically provided to poor people are health care, food, housing and childcare.[1]

The relative merits of the three categories of policies have since long been debated. There is no final consensus on that question, as it is illustrated by the fact that European states resort typically more often to money transfers whereas in-kind transfers seem to be more popular in the US.

Three main justifications of in-kind transfers are usually discussed. The first justification works as follows. There is a risk that poor people “misuse” the money they receive. Assuming that the service providers know better than the poor people themselves what is good for them, it is more effective to provide the necessary goods directly rather than transferring money. This is the paternalistic justification. It is a challenge to the moral autonomy of poor people. This justification is less and less used, and, to our opinion, it is fortunate.

The second justification is based on the so-called self-selection issue. It is related to the objective of poverty policies that only those who really need welfare benefits receive them. While simple money transfers are believed to give incentives to non-poor people to claim to be poor in order to receive the money, there are theoretical arguments and there is some evidence that it is less the case with in-kind transfers. Those who do not really need them are less willing to try to obtain them, as these transfers are of low value to them.

The third justification is a political feasibility argument. Even if policymakers and scientists are less and less paternalistic, voters still believe that poor people are likely to make “wrong” use of their money. As a consequence, in-kind transfers are easier to justify in front of the voters than money transfers.

As a matter of fact, the long debate on the optimal shape of welfare programs has led to the existence of massive in-kind transfer programs. In a fascinating review, Janet Currie and Firouz Gahvari discuss a very long list of theoretical and empirical works that assess the justifications, merits and consequences of in-kind programs (see J. Currie and F. Gahvari, “Transfers in cash and in-kind: Theory meets the data,” NBER Working Paper No 13557). Their review focuses on evidence from the US.

One important dimension of the empirical assessment of in-kind transfers is their influence on labor supply. Evaluating the influence of a welfare program on labor supply is important because it is often claimed that a welfare program should not be implemented if it gives the incentive to the poor people to decrease their labor supply, and therefore, to undermine their ability to get themselves out of poverty.

How could welfare programs in general, and in-kind transfers in particular, affect the poor people’s labor supply? First, welfare programs may increase the well-being of the families, and make working less necessary. Second, these programs may decrease the uncertainty associated with material well-being, and free the adults from the fear of falling into deep deprivation and from very short term subsistence behavior, so that it is easier for them to look for regular jobs. Third, the goods that are provided in-kind may be complement to labor, for instance because they decrease the cost of looking for jobs or the cost of having a job.

There are many technical difficulties to estimate the effect of in-kind transfers on labor supply, and, naturally, the literature has concentrated on identifying ways to circumvent them. Here we review some of the main lessons that the evidence allows us to draw.

Medicaid, the free provision of health care to poor families, is a major in-kind transfer program. This is a means tested welfare program, which means that there is a threshold of financial resources above which families are no longer eligible for the program. There is no clear evidence that it affects the labor supply. A major fear associated to that program was that parents just below the threshold would be reluctant to accept jobs and be excluded from the program. This does not seem to happen, among other things because employment often goes together with private health insurance.

Food stamps are another in-kind transfer program. In this case, it is very difficult to estimate the effect of the program, as food stamps typically go together with other (cash) programs. All in all, in studies in which the effect is significant, it is negative: the incentives to stay out of the labor market are stronger than the assistance to look for jobs.

Providing public housing is another major in-kind program. There is surprisingly not much research on this theme, neither in the US nor in Europe, in spite of the general prevalence of housing programs. The studies did not succeed in identifying any significant effect. This is partly due to the fact that providing public housing goes together with asking people to live in a specific area, and living in this area may by itself help or make it more difficult to find a job.

The last major in-kind program that Currie and Gahvari review is the childcare program. Many studies have estimated the effect of making childcare more easily available on labor supply of, especially, young mothers. One dimension of free public childcare is determined by the cutoff age for eligibility for Kindergarten. Policies affecting that cutoff age have given the opportunity to researchers to estimate that free childcare for at least one half of the day increases labor supply of young mothers by a small but statistically significant amount.

More research needs to be done. Nevertheless, a general conclusion can already be drawn. It concerns the claim often made by policymakers or observers that welfare programs are bad for poor people. The argument is that these programs lower the incentives that poor people have to look for jobs and to work. The evidence just described teaches us that such a claim is largely exaggerated.

---

[1] Education is another good that is freely provided by the State. We don’t include it in the list here, because public schooling is provided to all citizens and it is not targeted towards poor people. Also, the arguments that justify free and public provision of education are based on the special nature of that good (the many externalities associated to it), which make it different from the goods we are interested in.

Europe’s future, such as it is, may belong to France (dixit Krugman)

Amid all this gloom and doom about France, I can't resist a second posting on the "Coming French Imperium" ;-)

More on the Coming French Imperium

I wrote a little while back about the little-discussed French demographic advantage within Europe; its relatively high fertility rate means that it should eventually overtake Germany and become the largest continental European economy. It turns out that the French Economic Observatory has been on this case. Their projections are fairly startling:

And to the extent that we believe advanced countries will suffer from the burden of large elderly populations, France will suffer less:

Germany should enjoy its hegemony while it lasts; Europe’s future, such as it is, may belong to France.

Let's (not) have a beer

Contrary to what my co-author David Bardey says, it is not because I am currently in Colombia that I find the following article by The Economist interesting...

Drugs that cause most harm

Scoring drugs

A new study suggests alcohol is more harmful than heroin or crack

MOST people would agree that some drugs are worse than others: heroin is probably considered to be more dangerous than marijuana, for instance. Because governments formulate criminal and social policies based upon classifications of harm, a new studypublished by the Lancet on November 1st makes interesting reading. Researchers led by Professor David Nutt, a former chief drugs adviser to the British government, asked drug-harm experts to rank 20 drugs (legal and illegal) on 16 measures of harm to the user and to wider society, such as damage to health, drug dependency, economic costs and crime. Alcohol is the most harmful drug in Britain, scoring 72 out of a possible 100, far more damaging than heroin (55) or crack cocaine (54). It is the most harmful to others by a wide margin, and is ranked fourth behind heroin, crack, and methamphetamine (crystal meth) for harm to the individual. The authors point out that the model's weightings, though based on judgment, were analysed and found to be stable as large changes would be needed to change the overall rankings.

"Drug harms in the UK: a multi-criteria decision analysis", by David Nutt, Leslie King and Lawrence Phillips, on behalf of the Independent Scientific Committee on Drugs. TheLancet.

Gerrymandering: have your cake and eat it too?

The Economist looks at gerrymandering in the US (the practice whereby those who hold the majority at a given time redesign the voting districts with a view to the results of the subsequent elections) and stresses the trade off between maximizing the chance that your candidate is elected in a given district, and the number of districts that your party can expect to win. They show (or rather claim...) that you can reach both goals by designing the districts in such a way that your opponent wins overwhelmingly in a minority of districts, while your party wins with a sizable majority in a majority of districts. This practice then makes Congress a very polarized institution, since most members do not face a real challenge from the opposition (but rather from more extreme members of their own party, as the Tea Party demonstrates).

Gerrymandering

How can Republicans be both safer and more numerous?

HERE'S a truth of American politics that at first seems a bit paradoxical. In the National Journal, Ron Brownstein, David Wasserman and Ben Terris write that Republicans in the House don't have to worry about the backlash against the shutdown because compared to a few years ago, gerrymandering has guaranteed that they're now in much safer, more deeply Republican districts. At the same time, as we wrote just after the 2012 elections, gerrymandering helped increase the number of seats won by Republicans such that they retained a solid 33-seat majority in the House despite losing the overall popular congressional vote by 1.4m votes. One might wonder: how can both these things be true?

After all, if you're trying to gerrymander electoral districts, you face a choice. You can either cluster all of your voters into fewer districts, which will guarantee that you have very safe districts that the other party cannot win. Or you can spread your voters out more thinly across many districts while ensuring that you gain a majority wherever they are present, which will probably win you more elections but puts each of your members at greater risk. But it's pretty hard to do both at once—concentrate your voters so that each of your members faces little risk of losing, but also maximise your voters' impact so that you win the greatest possible number of elections with the voters you have.

Yet that is what these reports are saying Republicans achieved with the 2010 redistricting, which they largely controlled, since they held most state legislatures. TheRepublican State Leadership Committee itself boasted that clever GOP redistricting efforts were behind the party's retention of the House last year despite losing the popular vote. Meanwhile, Mr Brownstein et al say they also ensured that GOP congressmen, on several distinct metrics, are in far safer districts than they were during the last government shutdown in 1995. The number of GOP congressmen representing districts won by the Democratic presidential candidate (Bill Clinton in 1992, Barack Obama in 2012) has dropped from 79 to 17. Just 71 Republican congressmen represent districts where Mitt Romney got less than 55% of the vote last year, compared to 141 who were in the equivalent situation in 1995.

Finally, the authoritative Cook Political Report produces a Partisan Voting Index for each district, showing how strongly it trends towards its preferred party. In 1995, according to the National Journal article, the Cook report's average Partisan Voting Index score for Republican-held congressional districts was 6.6. Today, the average is 11.1. Further:

Beyond those averages, the PVI data also show that the share of House Republicans in overwhelmingly safe districts has soared, while the portion in even marginally competitive seats has plummeted. In 1995, 12 House Republicans represented ruby-red districts whose index score leaned toward the GOP by at least 20 points; now 24 represent such districts. In 1995, 25 House Republicans represented districts with a Republican-leaning index score of at least 15; now 61 represent such districts.

Conversely, back then, more than two-fifths of the Republican caucus (105 members in all) represented at least somewhat competitive seats with a Republican-leaning index score of 5 points or less. Today only about one-fifth of Republicans (53 in all) represent districts so closely balanced.

Republicans have managed to both make their seats safer, and ensure there are more of them, despite the fact that they lost the overall popular congressional vote. How did they do that?

By finding the golden mean. The ideal strategy for elections is to make sure your districts have just enough of a partisan tilt to ensure you'll almost certainly win them, but not so much that you win them overwhelmingly and waste your votes. Meanwhile, you want to cram the opposition's voters into districts which they win by overwhelming margins and thus waste their votes. Republicans can make sure their seats are both safer and more numerous by achieving lots of districts where they're likely to win by a safe but not extravagant margin, say 15-30%. If they pursue this strategy, they should wind up with relatively fewer seats that tilt overwhelmingly Republican. Meanwhile for Democrats, whose votes have been "cracked" or "packed" such that they lose more districts, the districts that they do hold would be more likely to be overwhelmingly Democratic than is the case for Republicans.

And this is what the Republicans' redistricting appear to have achieved. Of members of congress who won their districts with a margin of 60% or more in 2012, 18 were Republicans, while 29 were Democrats. In the crucial safe-but-not-overwhelming zone, with victory margins between 15% and 30%, Republicans won 92 seats while Democrats won 42. The average margin of victory for Republicans was 28.6%; for Democrats, it was 35.7%.

So there you go. This is one big reason why Republicans in the House are likely to react to widespread anger over the shutdown by becoming more, rather than less, confrontational. The vote distributions in their districts would have to swing by 10% or more for any sizable number of them to lose their seats, and that's not very likely to happen.

NPR on privacy in the digital era

NPR has aired a segment on "Your digital trail: does the fourth amendment protect us?" This amendment of the Bill Rights, ratified in 1791, is about "The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures". The piece explains how, basically, the interplay of information technologies and of interpretation by the courts have gutted this amendment:

"But since the 1960s and 1970s, the Supreme Court and other courts have issued a series of rulings declaring that the government does not need a search warrant to obtain your personal documents if you have already shared them with somebody else. For instance, since you allow your bank and credit card company to know what you buy, and since you let your phone company know whom you call, you can't claim that information is private."

So, when you send an email, you have shared it with the Internet provider, so that it is not private anymore. You can apply the same interpretation to your location: if you use your mobile phone, you share your location with the phone company, so that it is not a private information anymore.


This is important because, once you have shared information with someone,

"instead of a search warrant, the police might just need a subpoena — which is "trivially easy to issue," says Bankston of the Center for Democracy and Technology. Law enforcement doesn't need a judge's approval to obtain subpoenas — prosecutors can sign them on their own, as can authorized employees at federal and state agencies. And law enforcement agents don't need evidence that there's likely a crime. They need only to be able to show that the records they want are relevant to an investigation."

For instance, even divorce lawyers in many states can issue subpoenas themselves and obtain that kind of information.


The bottom line to me is that lawyers have made the US a kind of scary country...

Dani Rodrik on the state of economics

Here are a few excerpts I have especially liked from the interview that Dani Rodrik has given to the World economic Association on the state of economics.


"PhD programs now train applied mathematicians and statisticians rather than real economists. To become a true economist, you need to do all sorts of reading – from history, sociology, and political science among other disciplines – that you are never required to do as a graduate student. The best economists today find a way of filling this gap in their education."

"There are powerful forces having to do with the sociology of the profession and the socialization process that tend to push economists to think alike. Most economists start graduate school not having spent much time thinking about social problems or having studied much else besides math and economics. The incentive and hierarchy systems tend to reward those with the technical skills rather than interesting questions or research agendas. An in-group versus out-group mentality develops rather early on that pits economists against other social scientists. All economists tend to imbue a set of values that tends to glorify the market and demonize public action."

"The two most exciting developments in Economics in the last two decades are the behavioural and experimental revolutions. The first of these has made a significant dent in the rationality postulate of neoclassical economics, while the latter has taken the profession in a profoundly empirical and policy-oriented direction. These are significant changes in how one does mainstream economics, and the fact that they have happened suggests there is room for methodological changes. Not plurality, perhaps, but some degree of evolution in methods. I am not necessarily a great fan of either of these methodological innovations, but they show the profession is able to adapt and change. Note also that both sets of new methods came from outside Economics -- psychology and medicine, respectively. Young economists made these methods their own, and changed the discipline from within."

Get a life!

I have found the following piece from The Economist interesting on several accounts. First, it reminds us that, notwithstanding what we repeatedly read in (french) newspapers, French workers (I guess it is rather France residents!) work on average more than many of their neighbors, and especially from the East (Germany) or North (Denmark, Netherlands, Norway) that are always presented as so hard working.

Second, the relationship between productivity and number of hours worked is fascinating. Of course, the causality can go both ways, but I like the quote by Adams Smith: [T]he man who works so moderately as to be able to work constantly, not only preserves his health the longest, but in the course of the year, executes the greatest quantity of works.

OK, Time to get back to work ;-)

Working hours

Get a life

Sep 24th 2013, 12:19 by C.W. and A.J.K.D. | LONDON

BERTRAND RUSSELL, the English philosopher, was not a fan of work. In his 1932 essay, “In Praise of Idleness”, he reckoned that if society were better managed the average person would only need to work four hours a day. Such a small working day would “entitle a man to the necessities and elementary comforts of life.” The rest of the day could be devoted to the pursuit of science, painting and writing.

Russell thought that technological advancement could free people from toil. John Maynard Keynes mooted a similar idea in a 1930 essay, "Economic possibilities for our grandchildren", in which he reckoned people might need work no more than 15 hours per week by 2030. But over eighty years after these speculations people seem to be working harder than ever. The Financial Times reports today that Workaholics Anonymous groups are taking off. Over the summer Bank of America faced intense criticism after a Stakhanovite intern died.

But data from the OECD, a club of rich countries, tell a more positive story. For the countries for which data are available the vast majority of people work fewer hours than they did in 1990: 

And it seems that more productive—and, consequently, better-paid—workers put in less time in at the office. The graph below shows the relationship between productivity (GDP per hour worked) and annual working hours:

The Greeks are some of the most hardworking in the OECD, putting in over 2,000 hours a year on average. Germans, on the other hand, are comparative slackers, working about 1,400 hours each year. But German productivity is about 70% higher.

One important question concerns whether appetite for work actually diminishes as people earn more. There are countervailing effects. On the one hand, a higher wage raises the opportunity cost of leisure time and should lead people to work more. On the other hand, a higher income should lead a worker to consume more of the stuff he or she enjoys, which presumably includes leisure.

Some research shows that higher pay does not, on net, lead workers to do more. Rather, they may work less. A famous study by Colin Camerer and colleagues, which looked at taxi drivers, reached a controversial conclusion. The authors suggested that taxi drivers had a daily income "target", and that:

When wages are high, drivers will reach their target more quickly and quit early; on low-wage days they will drive longer hours to reach the target.

 Alternatively, the graph above might suggest that people who work fewer hours are more productive. This idea is not new. Adam Smith reckoned that

[T]he man who works so moderately as to be able to work constantly, not only preserves his health the longest, but in the course of the year, executes the greatest quantity of works.

There are aberrations, of course. Americans are relatively productive and work relatively long hours. And within the American labour force hours worked among the rich have risen while those of the poor have fallen. But aA paper released yesterday by the New Zealand Productivity Commission showed that even if you work more hours, you do not necessarily work better. The paper made envious comparisons between Kiwis and Australians—the latter group has more efficient workers.

So maybe we should be more self-critical about how much we work. Working less may make us more productive. And, as Russell argued, working less will guarantee “happiness and joy of life, instead of frayed nerves, weariness, and dyspepsia".