New Economist

Nudging behaviour: When incentives do (and don't) work

Behavioural economics has blossomed in recent years, to the point that the Fall 2011 edition of the Journal of Economic Perspectives can, along with articles on neuronomics, genetics and retirement, include a fascinating 20 page review paper on incentives. That's not surprising; a lot of people have taken an interest in them lately, including President Obama and Britain's prime minister.

When and Why Incentives (Don’t) Work to Modify Behavior was written by Uri Gneezy from the University of California–San Diego, Stephan Meier from Columbia Business School, and Pedro Rey-Biel at the Universitat Autònoma de Barcelona. It discusses when extrinsic incentives are more or less likely to alter behaviors in the desired directions, focussing on incentives in education, incentives for 'pro-social' behaviour, and incentives and lifestyle habits. They argue that although "incentives to modify behavior can in some cases be cost effective", there is also a danger that direct extrinsic motivations might 'crowd out' intrinsic motivations. For example:

...an incentive for a child to learn to read might achieve that goal in the short term, but then be counterproductive as an incentive for students to enjoy reading and seek it out over their lifetimes.

The authors conclude:

Our message is that when economists discuss incentives, they should broaden their focus. A considerable and growing body of evidence suggests that the effects of incentives depend on how they are designed, the form in which they are given (especially monetary or nonmonetary), how they interact with intrinsic motivations and social motivations, and what happens after they are withdrawn. Incentives do matter, but in various and sometimes unexpected ways.

To paraphrase Sergeant Esterhaus: Choice architects, let's be be careful out there!

'The continuing muddles of monetary theory'

One of the more lively papers delivered at this week's 75th Anniversary Lionel Robbins Conference in London was by the LSE's Charles Goodhart: The Continuing Muddles of Monetary Theory: A Steadfast Refusal to Face Facts (PDF). Goodhart believes monetary economics has "a long way yet to go", and few economists "have bothered much to relate theory to reality." Here's the abstract to his short paper:

Lionel Robbins was much concerned about the methodology of economic science.
When he discussed the desirable relationship between theory and ‘reality’, two of the
three examples that he presented where the theoretical analysis was not sufficiently
based on a knowledge of historical fact were taken from monetary economics.
Indeed, monetary theory has remained prone to such shortcomings ever since.

Amongst the worst are:-
(1) IS/LM: the monetary authorities set the monetary base, and the interest rate is
determined in the market;
(2) The monetary base multiplier of bank deposits, and the role of reserve ratios;
(3) The current three equation neo-classical consensus, which not only assumes
perfect creditworthiness for all agents, but also an essentially non-monetary
system, e.g. no need for banks;
(4) The standard theory of the evolution of money.

Monetary economics can only get better, but it has a long way yet to go.

Interaction models - don't get caught!

In my naivety I had assumed most social scientists understood how to model interaction effects. Then I read this post by Omar on orgtheory.net, and realised maybe I was wrong:

So we are agreed interaction models are awesome. However, as your stats 101 teacher told you, you have to be careful about two things: (1) never omit the main effects. Thus you don’t test hypothesis 1 using any of these specifications:

Y=a+b1X+b2XZ+e

Y=a+b1Z+b2XZ+e

or Alanis forbid:

Y=a+b1XZ+e

And (2) when interpreting b1 and b2 in the fully specified model, remember that those effects are conditional on the value of the other variables. b1 is now the effect of X on Y when Z=0 and b2 is now the effect of Z on Y when X=0. If your variables don’t have a meaningful zero point (like a racial attitudes scale), center them at their mean so that you can say “b2 is the effect of being Southern on voting republican for those who have average levels of racial animus towards blacks.”

Seems simple. Everybody knows this. Why am I even explaining this to you? Well, as noted by Brambor, Clark and Golder (2006) in a recent article in Political Analysis, a survey of 156 articles published in the major Political Science journals shows that only 10% of researchers specified their interaction models correctly. A large chunk of them outright omitted main effects, which can lead to incorrect significance tests of the interaction term. In some of these articles the entire contribution was riding on the interaction term. So things are not so simple. Consider the horror:

In an award-winning article in the American Political Science Review, Boix (1999) examines the factors that determine electoral system choice in advanced democracies. He makes two main conclusions. First, ethnic or religious fragmentation encourages the adoption of proportional representation in small and medium-sized countries (621). He draws this conclusion based on a model that includes an interaction term between ethnoreligious fragmentation and country size. However, he does not include either of the constitutive terms. When these terms are included, there is no longer any evidence that ethno-religious fragmentation ever affects the adoption of proportional representation (italics added).

You should read the article to see other horror stories. The lesson: if your dissertation/paper is riding on an interaction effect, don’t be a fool. Estimate a fully specified model.

A salutory lesson. Here is a non-gated version of the Brambor, Clark and Golder paper, Understanding Interaction Models: Improving Empirical Analyses (PDF). NYU's Matt Golder has more useful material on interaction models on his web pages.

The economics of happiness: a progress report

Little by little, psychology's insights into happiness and well-being are being integrated into economic theory and debate. Happiness has become one of the hot topics in economics over the last decade, with both the size and depth of the literature increasing at a rapid rate.

Though US-based psychologists Daniel Kahneman and Ed Diener and economists Richard Easterlin and Alan Krueger can take much credit, so too can the British 'brat pack' consisting of Warwick's Andrew Oswald, Paris-based Andrew Clark, Dartmouth's Danny Blanchflower and (as an honourary elder statesman), the LSE's Richard Layard. The two Andrews were writing about this stuff years ago, when 'respectable' academic economists thought the topic barking.

What is particularly encouraging is that economists are starting seriously to tackle the major empirical and the theoretical challenges that this literature has uncovered. A new paper by Andrew Clark, Paul Fritjers and Michael A Shields, prepared for the Journal of Economic Literature, ups the ante. Relative Income, Happiness and Utility: An Explanation for the Easterlin Paradox and Other Puzzles (PDF) takes as its starting point Easterlin's seminal 1974 paper which posed a paradox - that happiness does not appear to increase with income, once basic needs are fulfilled. The paper covers a lot of ground, including adaptation, social comparisons, relative income, utility and much more besides. There is a short but insightful description of the key challenges for empirical work. And to wrap it up, the authors explore some implications for economic theory and policy design. As a sampler, here are the authors concluding remarks:

The interaction between economic theory and happiness is therefore the next milestone for the developing economics of happiness literature. However, it is clear that the empirical literature on happiness still faces several challenges, many of which are shared with other empirical literatures. Two of the key challenges are to deal with a general inability of survey data to precisely time changes in income with changes in happiness over long time periods, and the difficulty in mapping incomes into current and expected consumption. It is also the case that most datasets do not contain reliable (if any) ex ante information regarding the group (the reference point) to which individuals compare themselves. Similarly, no dataset can contain all the variables of importance, so that researchers will continue to face the issue of endogeneity with respect to income and other variables such as marriage, education, and the reference group. Finally, natural experiments producing exogenous variation in income are only rarely observed, making the issue of establishing the causal effect of income on happiness a major challenge.

Our final conclusion is that taking relative income seriously is an important step towards greater behavioural realism in Economics, such that our models and empirical analysis move closer to how real people feel and behave. Some may not like the insertion of additional arguments into individual utility, and remark that any behaviour can be rationalised by an appropriate manipulation of the utility function. While this is formally true, it does not apply wholesale to the issue of relative income. As we have tried to demonstrate, utility functions including relative income terms produce a wide variety of testable predictions regarding both well-being (measured by survey or neurologically) and observable behaviours: it is not true that “anything goes”. To our mind, this is precisely why we need to appeal to both direct measures of utility and observed behaviour in order to obtain a better idea of what the utility function looks like, and make policy recommendations in the best interest of society. Testing these predictions not only allies theory and empirical analysis in economics, it also spills across many disciplines in the social and natural sciences; it is arguably the most important and the most promising of the research avenues open to this thriving literature.

This is no light summer reading, but for those seeking to advance the debate, it's a big step forward.

Growth: the central questions

Daron Acemoglu's new textbook on economic growth concludes at Chapter 24 with a discussion of 'what we have learned', some answers to what he dubs 'the central questions', and a quick overview of 'some of the many remaining questions' faced by researchers. On page 1068 (sic) he writes:

The central questions are these:
(1) Why did the world economy not experience sustained growth before 1800?
(2) Why did economic takeoff start in 1800 and inWestern Europe?
(3) Why did some societies manage to benefit from the new technologies and organizational forms that emerged starting in 1800, while others steadfastly refused or failed to do so?

I will now offer a narrative that provides some tentative answers to these three questions.

And so he does. It's the best summary of the economic growth literature I have read - but alas, too long to include in this blog. To tempt readers, here is a small sample, from pages 1077f:

Why did some societies manage to benefit from the new technologies while others failed?

The economic takeoff started in Western Europe, but quickly spread to certain other parts of the world. The chief importer of economic institutions and economic growth was the United States. The United States already had participatory political institutions, founded by settler colonists, who had just defeated the British crown to gain their independence and set up a smallholder society. This was a society built by the people who would live in it, and they were particularly keen on creating checks and balances to prevent a strong political or economic elite. This environment turned out to be a perfect conduit for modern economic growth.

The lack of a strong political and economic elite meant that a broad cross-section of society could take part in economic activity, import technologies from Western Europe and then build their own technologies to quickly become the major industrial power in the world (Galenson, 1996, Engerman and Sokoloff, 1997, Keyssar, 2000, Acemoglu, Johnson and Robinson, 2002). In the context of this example, the importance of technology adoption from the world technology frontier is in line with the emphasis in Chapter 18, while the growth-promoting effects of a lack of elite creating entry barriers is consistent with the approach in Section 23.3 in Chapter 23.

Similar processes took place in other West European offshoots, for example, in Canada. Yet in other parts of the world, adoption of new technologies and the process of economic growth came as part of a movement towards defensive modernization. Japan started its economic and political modernization with the Meiji restoration (or perhaps even before) and a central element of this modernization effort was the importation of new technologies.

New book online: Daron Acemoglu's 'Introduction to Modern Economic Growth'

When Dani Rodrik recently asked his many readers which economist they most wanted to see start a blog, MIT's Daron Acemoglu came second only to Joseph Stiglitz. Rodrik asked him if he would start a blog:

Daron Acemoglu responds by saying blogging "seems like a lot of work." He adds "I am already overcommitted, so it's a bit scary to take something else on."

Busy? And how! Acemoglu has somehow found the time to write a new magnum opus. His Introduction to Modern Economic Growth is a whopping 1,169 pages. It covers the whol gamut, from the Solow growth model to the role of political regimes and institutions. Acemoglu explains the background to this textbook in his introduction:

This book grew out of the first graduate-level introduction to macroeconomics course I have taught at MIT. Parts of the book have also been taught as part of a second-year graduate macroeconomics.

In the preface, Acemoglu writes that the book is intended to serve two purposes:

First and foremost, this is a book about economic growth and long-run economic development. The process of economic growth and the sources of differences in economic performance across nations are some of the most interesting, important and challenging areas in modern social science. The primary purpose of this book is to introduce graduate students to these major questions and to the theoretical tools necessary for studying them. The book therefore strives to provide students with a strong background in dynamic economic analysis, since only such a background will enable a serious study of economic growth and economic development. It also tries to provide a clear discussion of the broad empirical patterns and historical processes underlying the current state of the world economy. This is motivated by my belief that to understand why some countries grow and some fail to do so, economists have to move beyond the mechanics of models and pose questions about the fundamental causes of economic growth.

In a somewhat different capacity, this book is also a graduate-level introduction to modern macroeconomics and dynamic economic analysis. It is sometimes commented that, unlike basic microeconomic theory, there is no core of current macroeconomic theory that is shared by all economists. This is not entirely true. While there is disagreement among macroeconomists about how to approach short-run macroeconomic phenomena and what the boundaries of macroeconomics should be, there is broad agreement about the workhorse models of dynamic macroeconomic analysis. These include the Solow growth model, the neoclassical growth model, the overlapping-generations model and models of technological change and technology adoption. Since these are all models of economic growth, a thorough treatment of modern economic growth can also provide (and perhaps should provide) an introduction to this core material of modern macroeconomics. Although there are several good graduate-level macroeconomic textbooks, they typically spend relatively little time on the basic core material and do not develop the links between modern macroeconomic analysis and economic dynamics on the one hand and general equilibrium theory on the other. In contrast, the current book does not cover any of the shortrun topics in macroeconomics, but provides a thorough and rigorous introduction to what I view to be the core of macroeconomics. Therefore, the second purpose of the book is to provide a first graduate-level course in modern macroeconomics.

A considerable achivement. Acemoglu asks readers to note that "this is a preliminary draft of the book manuscript. The draft certainly contains mistakes. Comments and suggestions for corrections are welcome."

Psychology and economics

Stefano DellaVigna from UC Berkeley has written a very useful review of Psychology and Economics: Evidence from the Field (PDF). Hat tip: Andrew Leigh

In the conclusions DellVigna specuates on future research topics:

..future research is likely to reduce the imbalance across fields in economics and across topics in psychology. While the research in behavioral finance and consumption savings is very active, relatively few studies, instead, have tackled mortgage markets, development, and political decisions, fields ripe for exploration. Future research is also likely to explore psychological phenomena that have been largely neglected. For example, emotions,automatic processing, and implicit discrimination are likely to matter for economic decisions such as divorce, judicial sentencing, and policing. Ten years from now, we will hopefully be able to assess quantitatively which psychological factors matter in which decisions.

I identify two specific areas for future research: the market interaction between standard and non-standard agents, as in Section 5, and public policy applications. The market interaction is likely to find several additional applications, for example to the interaction between politicians and voters. In addition, this area is likely to investigate the judgmental biases, such as overconfidence, of experienced agents such as managers and politicians.

For another recent review of the literature, see my earlier post, Who are the behavioral economists and what do they say?

New econometric textbook online

'Marshall Jevons' of the weblog The Bayesian Heresy alerts us to a new graduate-level econometrics textbook, available free online, by Professor Bruce E. Hansen from the University of Wisconsin. Here is a direct link to the PDF (1.2Mb).

Out with the new, in with the old...

Who needs new (or neo) Keynesian economics when you can have the old? UCLA's Roger Farmer is writing a book on Old-Keynesian Economics, and you can read the draft chapters online. The book - still a work in progress - junks the 'natural rate' and sticky price assumptions in favour of Keynes' "animal spirits". Here is his explanation:

This book is concerned with the question: Why do capitalist economies sometimes go very badly wrong? For several decades after the publication of Keynes' General Theory economists thought that they had an answer. But with the resurgence of classical ideas in the 1970's, the key premise of the General Theory, that market economies are not inherently self-stabilizing, has been called into question. Although there has been a recent resurgence of Keynesian ideas under the rubric of "new-Keynesian economics", the models studied by the new-Keynesians are hybrids that incorporate a classical core. New-Keynesian models allow for temporary deviations of unemployment from its "natural rate" as a consequence of sticky prices but they contain a stabilizing mechanism that causes a return to the natural rate over time.

In his 1966 book, Axel Leijonhufvud made the distinction between Keynesian economics and the economics of Keynes. By Keynesian economics, he meant the interpretations of Keynes that became incorporated into the IS/LM model and ultimately, into the new-Keynesian paradigm. Leijonhufvud pointed out that the assumption that the General Theory is about sticky prices is central to Keynesian economics but it is not a central argument of the text of the General Theory.

This book provides an alternative microfoundation to Keynesian economics that does not rely on sticky-prices. In successive chapters I construct a series of models that build on a single idea. Each of them is constructed around a conventional dynamic general equilibrium model in which real resources must be used to move unemployed workers into jobs using a "search technology". Although this technology is convex, I assume that the planning optimum cannot be decentralized as a competitive equilibrium because moral hazard prevents the creation of markets for the search inputs.

As an alternative, I introduce an equilibrium concept called demand constrained equilibrium, in which the level of economic activity is determined by investor confidence or "animal spirits". I refer to the resulting model as "old-Keynesian" to differentiate it from new-Keynesian economics that incorporates the natural rate hypothesis of Edmund Phelps and Milton Friedman. In contrast to new-Keynesian models, those described in this book display multiple stationary perfect foresight equilibria, and there is a different stationary unemployment rate, for each possible level of beliefs.

Who are the behavioral economists and what do they say?

That's the title of a new Tinbergen Institute paper by Floris Heukelom (Universiteit van Amsterdam). She identifies a divide within this economic school between those who see uncertainty as exogenous, and those who consider it endogenous:

The most important financial source for behavioral economics is the Russell Sage Foundation (RSF). The most prominent behavioral economists among the RSF’s twenty-six member Behavioral Economics Roundtable (BER) are Kahneman, Tversky, Thaler, Camerer, Loewenstein, Rabin, and Laibson. The theoretical core of behavioral economics made up of the work of these seven researchers is positioned in opposition to Adam Smith/Hayek type of economics, as exemplified by experimental economists Vernon Smith and Plott; and what is referred to as ‘mainstream’ or ‘traditional’ economics, meaning the neoclassical economics that roughly builds on Samuelson.

On the basis of an overview of the work of these seven behavioral economists, a theoretical division can be observed within behavioral economics. The first branch considers human decision-making to be a problem of exogenous uncertainty, which can be analyzed with decision theory. It employs traditional economics as a normative benchmark and favors a normative-descriptive (prescriptive) distinction for economics. The second branch considers human decision-making to be a problem of strategic interaction, in which the uncertainty is endogenous. Its main tool is game theory. It rejects traditional economics both positively and normatively.