Mastering Metrics

I went through a copy of Mastering Metrics by Joshua Angrist and Steve Pischke. Their previous book Mostly Harmless Econometrics is widely used as a companion to advanced undergraduate and introductory graduate courses in Microeconometrics. Both books are very useful. Mastering Metrics has six chapters on: 1. Randomised Trials; 2. Regression; 3. Instrumental Variables; 4. Regression Discontinuity Designs; 5. Differences in Differences and 6. The Wages of Schooling.

The RCT chapter is a clean and concise overview of the main economic parameters that can be estimated with an RCT and provides a good and well worked-through example. The chapters on regression and IV don't cover too much more ground than in their previous work but have some very interesting extra examples that students will find useful. Similarly, RDD and Difference-in-Difference designs are covered in Mostly Harmless Econometrics but are extended with more examples in this book. The final chapter goes through in detail the case of estimating returns to education and goes through the problem of omitted variable bias in depth. Each of the chapters will certainly be very useful companion chapters to lectures on the topics.

The style of both MHE and MM (whether you appreciate the jokes or not) is rightly popular among students, combining informal text with worked examples and appendices providing the mathematical foundations. Many people's first introduction to topics like IV or RDD is in the context of highly formal and stylised mathematical models and the approach here has probably led to many students figuring out for the first time that the key concepts are interesting and applicable. Armed with these insights it is easier to step back into the fight with the notation.

It is clear they have the potential to continue this into a series of short and accessible books.  I think a clear demand from my students would be to extend the RCT chapter into a book that addressed not just the basic parameters but the wider set of parameters that can obtained from RCTs as well as how to think about common problems with RCT designs in a more formal way.

Raj Chetty's slides on BE & Public Policy

Raj Chetty of Harvard has a new, very accessible 62-slide presentation online called Behavioral Economics and Public Policy A Pragmatic Perspective. His main takeaways are that BE makes three contributions to public policy (i) New policy tools (e.g. defaults, framing), (ii) Better predictions of the effects of existing policies (e.g. taxes), (iii) New welfare implications.

Study on Self-Control - Stirling Participants

We are searching for students and staff from Stirling University to participate in a study on self-control. Participants told me that it is an interesting study and you'll get Amazon.co.uk Gift Certificates in return. The study will take place in the end of January. More details below. 

Please feel free to register for the study here and spread the word.

BEHAVIOURAL SCIENCE STUDY ON SELF-CONTROL

*NEW SLOTS*

Complete a questionnaire on self-control (and your behaviours, emotions, motivations, and preferences). In Stirling, already 96 participated (you cannot participate again).

When:   Thursday 29th January, 11.00 - 13.00 (Room 2A15)

              Thursday 29th January, 14.00 - 16.00 (Room 2A15)

              Friday 30th January, 11.00 - 13.00 (Room 2A15)

              Friday 30th January, 14.00 - 16.00 (Room 1A11)

In return, we’ll send you Amazon.co.uk Gift Certificates worth between £8 and £16 (average so far: £14.40 for  70 minutes)

Please follow this link or register on the Stirling Portal.

The study is organized by Professor Liam Delaney and Dr Leonhard Lades (l.k.lades@stir.ac.uk) from the Stirling Behavioural Science Centre.

Behavioural Economics and Soap Operas

From Brendan Greeley at Businessweek [h/t Marginal Revolution]. I added references to the papers being discussed.

"Before she won an Academy Award in 2014 for her role in 12 Years a Slave, Lupita Nyong’o starred in two seasons of the TV drama Shuga. Set first in Nairobi and then in Lagos, Shuga features young, attractive people who sleep with each other. It’s wildly popular and shown on broadcast channels that reach 500 million people, mostly in Africa. ...Now in its fourth season, the show recently added a new member to its production team: Eliana La Ferrara, a professor at the University of Bocconi in Italy who specializes in a mix of behavioral and development economics. La Ferrara wasn’t hired for her writing talent. MTV and its donors want to apply a more rigorous approach to make sure Shuga’s message actually creates change where it airs.

Governments have long known that TV can have an impact on poverty by changing behavior. In the 1970s, after the launch in Peru of Simplemente María, a telenovela about an aspirational maid, the country’s government noticed a rise in demand for literacy classes. More recently, economists have tried to measure these effects with greater precision. In 2009, Emily Oster, an economist at Brown University, found that the arrival of cable television in rural India had decreased the acceptability of domestic violence against women and led to a drop in fertility rates. At about the same time, La Ferrara began working on a paper that uncovered a similar effect in Brazil, where birthrates had declined in households within the signal range of Rede Globo, a Brazilian broadcaster specializing in telenovelas.

The phenomenon of people changing their behavior as they identify with characters became known as the telenovela effect. It’s seen not just in the developing world. In 2014 the U.S. National Bureau of Economic Research released a paper suggesting the MTV reality show 16 and Pregnant had led to a 5.7 percent reduction in births among teenage mothers during the 18 months after it premiered. In the U.S., TV has been on the leading edge of evolving social trends at least since The Mary Tyler Moore Show and Maude.

The news isn’t universally good: A paper by Benjamin Olken of MIT showed that social ties in Indonesia weakened with the arrival of TV. The challenge, La Ferrara says, is to take a passive effect and turn it into an active policy. “Now that we know what happened, we can leverage the good side.”

With Shuga, La Ferrara applies the rigorous standards of economic research to the development of the show. She pulls existing data on attitudes to get a baseline before programming starts and suggests themes for the writers to consider. When the show’s third season was ready to air in 2014, La Ferrara screened it for community groups. This spring, those groups will respond to a survey measuring what behaviors actually changed. Such data-driven decisions are increasingly important for donors, in particular the Gates Foundation.

Last season, while studying a Nigerian survey on attitudes about HIV, La Ferrara found that only 47 percent of women and 61 percent of men had heard of antiretroviral drugs, which can prevent the onset of AIDS. A plot line on the show then featured a woman who discovered the drugs in her lover’s dresser and asked her friends what they were. TV producers can’t always share an economist’s rigor. La Ferrara suggested that MTV produce several plot lines on domestic violence to see which was most effective, but the network decided not to shoot multiple versions of the same scene.

In 2013, Bilal Zia, an economist with the World Bank, published the results of a randomized field trial around Scandal!, a soap opera on South Africa’s E.TV. He worked with the show’s writers on a plot line about personal finance. At first, “we were worlds apart,” he says. “They wanted something with a lot of spice, and we wanted a lot of messaging. They said if you put on a show with a lot of information, no one’s going to watch.” Eventually the writers agreed to do a plot about a woman who buys furniture on an expensive installment plan, gambles, and then calls a government debt hotline for help. Zia found in field studies that people who watched were less likely to buy on installment or gamble.

The World Bank is beginning to think about behaviors in the developing world, in addition to its traditional focus on infrastructure and the workforce. The bank’s 2015 World Development Report includes a section on the work of La Ferrara, Oster, and Zia [see p32]. Karla Hoff, the report’s author, says the bank is intrigued by the telenovela effect, but more research is needed before it becomes a regular part of funding decisions."

References

Berg & Zia (2013), Harnessing Emotional Connections to Improve Financial Decisions, Policy Research Working Paper

Jensen & Oster (2009), The Power of TV: Cable Television and Women's Status in India, QJE.
--> There's a writeup of this paper on the Chicago Booth website.

Kearney & Levine (2014), Media Influences on Social Outcomes: The Impact of MTV's 16 and Pregnant on Teen Childbearing, NBER Working Paper.

La Ferrara et al. (2012), Soap Operas and Fertility: Evidence from Brazil, American Economic Journal: Applied Economics

Olken (2007), Monitoring Corruption: Evidence from a Field Experiment in Indonesia, Journal of Public Economics

World Bank (2015), World Development Report 2015: Mind, Society and Behaviour, World Bank Group.

Heads-up limit hold ‘em poker is solved

Researchers at the University of Alberta have just announced that “heads-up limit hold ‘em”, the simplest form of poker commonly played for cash stakes, has been “solved” (science paper, summary on nature.com). This means that the researchers have found a complete strategy, which describes what plays to make in every possible situation of the game, and that is essentially unbeatable. This strategy will not lose against any other strategy, a defensive strategy that therefore makes it almost certain to win over the long run. This is a landmark in artificial intelligence, as another game is dominated by computers and not humans at the highest level (think chess). But what does this mean for the psychology of how humans play poker, and does it mean that downloading the perfect strategy will make you rich?

Game theory is the study of optimal decisions involving two or more competing interests (players). Games such as noughts-and-crosses, or indeed poker, were the initial inspiration for game theory, but now the theory of games covers any situation where the strategies and payoffs for two or more players can be formally modeled. The current research marks a landmark in at least one way: this is the first time that a significant game of “incomplete information” has been solved. Previous games to be solved (e.g. checkers) are games of “complete information”, where both players know the exact state of the game. In poker, each player’s private cards mean that neither player knows the exact game state (until the hand is over), and this leads to the phenomenon of bluffing where players pretend to be strong when they’re actually weak. Games of incomplete information are more realistic as models of real-life games, where the assumption of complete information can be far-fetched.

Computers solve games via brute-force, trialing different strategies until they find one that cannot be beaten. The complexity of most games is what makes this difficult. To make progress with complex games, researchers generally need to combine a massive amount of computing power with some clever programming tricks to design strong strategies. These were the main factors also underlying this breakthrough: a new solution algorithm and over two months of calculations over 4,000 CPUs were enough to do the job. This was enough computing power to surpass previous heads-up limit hold ‘em agents (which have been involved in annual competitions for some time).

This means that the latest breakthrough is excellent work but is not a real game-changer. In terms of playing style, the latest agent plays pretty much the same strategy to a neutral observer as previous agents. The only difference is now that the new agent is so close to the optimal game theoretic strategy, that researchers have called this form of poker “solved” in a bid to move research to more complex games. There are many more complex forms of poker that have not been solved, and will become a larger object of focus after this result. Merely adding one extra player to the game (the term “heads-up” refers to when only two players are involved in a game), or changing the betting rules slightly (in “limit” poker a player can only bet a set amount at each situation, in “no-limit” players can choose their bet size) is sufficient to create a much more complex game. These more complex forms of poker are unlikely to be solved any time soon.

As a game played for cash stakes, heads-up limit hold ‘em poker has been declining in popularity for some time. One reason is the relative simplicity of this form of poker; in fact poker agents have been (illegally) playing this game online for some time against unsuspecting punters across various poker networks. You can even play this form of poker against computers on modern betting terminals in Las Vegas. These games are free of charge, but the owners are so confident their strategy is better than any human that all-comers are accepted.  So this breakthrough is unlikely to make you a fortune if you get hands on a copy: the low-hanging fruit have been harvested some time ago in this form of poker.

Of course, as humans we have no way of remembering the terabyte-or-so of information (in heavily compressed form) required to describe this unbeatable strategy. So how can humans possibly hope to compete? Well, humans will never be as good as the computer’s strategy in a straight battle, but the approximations and heuristics that skilled humans use have proven remarkably effective. Many professional players use two key heuristics on the first round of heads-up limit hold ‘em: raise-or-fold as the first player (never calling), and always call as the first player if the second player re-raises. It turns out that the computer agrees with these simple heuristics well over 99% of the time. A boundedly-rational human plays almost exactly the same way as a computer with complete knowledge! (Without requiring the million hours of CPU time.) These heuristics have been an open secret among top players for more than five years.

Humans have further advantages over computers. Tweak a few parameters of the game and most expert humans can easily adjust, but an optimising agent has to completely start from scratch. Secondly, humans can often win a lot more money against imperfect opponents than this agent (using what are called exploitive strategies), an important criterion in the real-world where winning money and not prestige are the main goals of a professional gambler. To sum up, solving heads-up limit hold ‘em is a great achievement in artificial intelligence, but one that just goes to underscore how amazing human intelligence can be.

(In a past life I was a professional heads-up limit hold ‘em player and I have written two books on how humans can approximate optimal strategies in poker, specifically focusing on heads-up limit hold ‘em poker as a model for more complex forms of poker.)

Readings on the Irish Economy post-2007: Suggestions welcome

Below is an incomplete list of articles from various sources that I think are important for understanding what is happening in the Irish economy at present.  Suggestions welcome. The podcasts and slides from January 2012's session  February 2013 session  and January 2014 session contain a wealth of information about the Irish economy.

Cormac O'Grada's brilliant non-technical summary of the five major economic crises faced by the Irish state since its formation gives the historical context. A good reference book for the Irish economy is the 11th edition of Economy of Ireland edited by John O'Hagan and Carol Newman. Standard of articles high in general but Jonathan Haughton's history of the Irish economy is particularly recommended. Patrick Honohan's short history of the Irish pound, written just as Ireland entered the single currency, is well-written and informative. In the spirit of "twas ever thus", Moore McDowell's piece on the devaluation of 1492 is an excellent read.

Morgan Kelly's UCD working paper on the Irish credit bubble is a very clear discussion of the collapse. His famous paper predicting the collapse of the Irish housing market is available here For a sense of the details of how Anglo Irish bank's loan book expanded so much during the 2003-2007 period and the subsequent unravelling, see two books by Tom Lyons/Brian Carey and Simon Carswell.  Michael Lewis's Vanity Fair article is one of the most detailed accounts written by someone in the international press.

The currently available documents relating to advice provided to the Department of Finance in the run-up to the bank guarantee are available here 

Patrick Honohan's ESR paper on the the Irish banking crisis, written before he took over as Governor of the Central Bank, is the definitive account of what happened and I believe will be cited in 100 years time when people seek to understand what went on.

One of the main policy decisions taken by the FF/Green government was the creation of a 'bad bank' called NAMA, National Assets Management Agency, to take property development loans out of the banking system. The blog Namawinelake contains extraordinarily detailed accounts of several aspects of the assets being managed by NAMA. Will add more readings on NAMA at a later stage.

Karl Whelan's 2010 article "Policy Lessons from Ireland's Latest Depression" gives the policy context of the period leading up to the economic collapse (thanks to Dan Davis for suggestion).

UCD Politics academic Niamh Hardiman has written a number of pieces on the political and institutional context of the Irish economic crisis. Her JSSISI paper on this topic is here

Philip Lane's article the Irish Crisis (working paper here) (other version here) examines origins of Irish crisis and conclusions for other countries in monetary unions (h/t Rob Farhat). His ESR paper from 2009 on a new fiscal strategy for Ireland is linked here

Morgan Kelly's May 2011 Irish Times article argued that the bailout deal would bankrupt Ireland and the main hope for rescuing the economy was to break out of it. "Cutting Government borrowing to zero immediately is not painless but it is the only way of disentangling ourselves from the loan sharks who are intent on making an example of us. In contrast, the new Government’s current policy of lying on the ground with a begging bowl and hoping that someone takes pity on us does not make for a particularly strong negotiating position."

Lorenzo Bini Smaghi's famous (at least in Ireland) FT article makes the case that Irish taxpayers should shoulder the bank debt as a fair consequence of having their own national bank regulator. His article in Foreign Affairs July 2011 gives a sense of his general view of what caused the euro crisis and might be a good hint as to what has been driving ECB policy.

Constantin Gurdgiev has been a strong critic of government policy throughout the period. An example of his critique is this December 2010 article about the inevitability of an Irish default under the bailout conditions "Instead of resolving the core problem of catastrophic losses within our banking sector and the related problem of the fiscal insolvency of our Exchequer, the ECB/EU/IMF loan created an internationally binding agreement that officially transferred the debts of our private banks onto the shoulders of Irish taxpayers. By doing so, the EU, with the complicity of the Irish Government, has delivered a full-blown contagion across the entire Irish economy." (see also  Debt Restructuring: Orderly, Selective and Unavoidable )

Gurdgiev, Lucey, Mac an Bhaird and Kelly (2011);  "The Irish Economy: Three Strikes and You’re Out?""We examine the three interlinked Irish crises: the competitiveness, fiscal and banking crises, showing how all three combined to lay a lethal trap for Ireland. Starting from a point of economic balance, a series of poor government decisions led to the country once dubbed the Celtic tiger become the second eurozone state after Greece to seek a bailout, with the EFSF/IMF intervening in late 2010."

Kevin O'Rourke's irisheconomy posts also give a lot of details and food for thought on the problems with austerity and the poor evidence on expansionary fiscal contractions. 

Colm McCarthy's recent article on the potential of the bailout process in Ireland recently is also important reading. The key point from many commentators is that, while the Irish are sticking to the programme, the nature of the programme has piled so much banking debt onto the sovereign that it has made the programme self-defeating.

Stephen Kinsella's recent article in Cambridge Journal of Economics on austerity in Ireland critiques the policy. His recent article in Foreign Affairs a lay account of austerity in Ireland.

Pretty much all of Karl Whelan's irisheconomy blogposts and on his own blog are important for understanding what has been happening over the last three years.

Frank Barry's December 2011 Bloomberg article makes the case that the euro project contains substantial design flaws that are damaging the prospects of recovery in Ireland.

The Target 2 debate has been a big part of the recent debate about Ireland in europe. See the article by Tornell and Westermann arguing that Germany's banks are propping up the periphery banks and Karl Whelan's refutation of this.

Karl Whelan's briefing note to the Irish Parliament on promissory notes is one of the most lucid and cogent accounts of how to solve Irish economic crisis.

June 2012 briefing from the INET council on the euro discusses potential rescue plans.

The "Manifesto for Economic Sense" outlines another model of rescuing the European economy.

The most recent IMF statements on the Irish economy are linked here by Philip Lane

Philip Lane's summer 2012 JEP paper on the the European Sovereign debt crisis is a must-read

Donal Donovan and Antoin E. Murphy "The Fall of the Celtic Tiger: Ireland and the Euro Debt Crisis"

Philip Lane's statement to the Irish Banking inquiry and list of his relevant papers are here. 

Research Assistant position in Behavioural Science Centre

Post Details

Full time, fixed term contract for 1 year 

Grade 6 (spine point 22) £24,775

Closing date: midnight on 1 February 2015

The Post

A highly motivated researcher with a strong background in psychology and/or economics is required to work on an ESRC funded project on self-control and health. This exciting ESRC Future Leaders project focuses on how childhood self-control contributes to patterns of health behaviour, health problems, social mobility, and health inequalities over the lifespan. The project lead is Dr. Michael Daly along with Professor Liam Delaney as internal collaborator and several international collaborators on this topic including Professor Roy Baumeister. 

This post will be based in the Behavioural Science Centre in the Stirling Management School. The Behavioural Science Centre is an interdisciplinary research centre which brings together approaches from psychology and economics to address key questions in society including how to better understand and foster health and well-being. Your primary responsibilities will involve assisting in data analysis and in the preparation of literature reviews in collaboration with the PI and internal and international collaborators. You will also be involved in the preparation of data access applications, preparing and distributing online surveys, and the preparation of presentations and dissemination of findings.

Applicants should have completed an honours degree or master’s degree in a relevant area. Expertise in analysis of large scale survey data or other complex datasets and evidence of strong academic writing skills would be an advantage.

Description of Duties

Job Purpose

To set up and run analyses on several existing data sources following pre-specified data protocols, with particular responsibility for data cleaning, coding, and management. In addition, be responsible for preparing literature reviews and distributing surveys to validate questionnaire items for this interdisciplinary project. This post represents an excellent opportunity for a talented and enthusiastic researcher to become embedded in a productive team and contribute to publications focused on how self-control may foster social mobility, health and well-being.

Description of Duties

Assist in data analysis of large scale population survey datasets.

Conduct literature reviews focused on key areas of the project.

Assist in development of a validation study to be conducted with teachers and parents.

Assist in development of external presentations of the research findings.

Assist in organising seminars and workshops relating to the project.

Assist in development of on-going funding applications to continue the project.

Assist in drafting and submission of journal articles relating to the findings.

Assist in reporting on the outcomes of the project.

Actively engage in the intellectual life of the Behavioural Science Centre including initiating journal clubs, making suggestions on the work of other researchers, participating in seminars and making suggestions for social media outputs.

Essential Criteria

An Honours degree, or equivalent professional qualification.

Strong intrinsic interest in research in a relevant area (e.g. Behavioural Science, Behavioural Medicine, Behavioural Economics, Health Economics)

High level proficiency in STATA and/or SPSS.

Ability and willingness to contribute to the intellectual life of the Behavioural Science Centre including participating in seminars, journal clubs, group discussions and related activity.

Excellent written and oral communication skills.

Ability to work individually and autonomously as well as potential to work as part of a team.

Ability to manage time and work to deadlines.

Commitment to academic research and professional development.

Desirable Criteria

Postgraduate training in Psychology or Economics with a strong emphasis on statistics.

Specific knowledge of techniques for multilevel and panel data analysis.

Existing experience directly in the area of statistical analysis of determinants of health and well-being.

Evidence of active engagement with the area of behavioural science including student publications, internship experience and social media activity.

Proven ability in preparing high quality original research papers for publication.

Experience of preparing research funding applications.

Apply for this position here.