THE TOULOUSE SCHOOL OF ECONOMICS MAGAZINE

Living economics

Specia� issue Development economics

Mariacristina Daniel Garrett Esther Dufl o & Jerôme Bolte De Nardi on on overbooking Abhijit Banerjee on the future the origins of in the airline on fi ghting of optimisation inequality industry poverty algorithms Editor�' message #12 Content�

he new term 2016-17 has brought New� & Event� bustling academic life back to our 4 Appointment & Prizes Manufacture campus, and some ex- T 5 Strengthening French venture capital citing developments. We have launched our new international masters programme taught 6 Newcomers 2016-17 in English, welcoming 135 students for the 7 Save the date first year. We have also welcomed four new faculty members: Johannes Hörner, one of the most gifted economic theorists of his genera- tion, who joins us as full professor from Yale University, and three new assistant profes- � inker� sors. Isis Durrmeyer, Ana Gazmuri and Tim 8 What shapes inequality? Lee, whose profiles you can discover on page Specia� issue Mariacristina De Nardi 5. These arrivals bring our tenure-track junior DEVELOPMENT 10 Why do airlines overbook? faculty base to 17, a record for us. ECONOMICS Daniel Garrett 12 An outstanding counter-example A remarkable confirmation of the high calibre of our tenure-track faculty came Bruno Ziliotto in September with the award of two prestigious ERC starting grants to Takuro Yamashita and Daniel Garrett, two of our assistant professors in theoretical 14 The experts economics and industrial organisation. Congratulations to them both! TSE development team

18 The field Actor� In this issue you’ll find a special dossier featuring our research activities in “Development economics has been one of the most Esther Duflo & Abhijit Banerjee 26 Big data & the military development economics. Understanding economic development has been Jérôme Bolte exciting and innovative areas at the heart of our field ever since Adam Smith wrote The Wealth of Nations, 22 The study of economic research in recent drawing inspiration from his long stay in Toulouse where he witnessed first- Emmanuelle Auriol, years, at the forefront of the Josepa (Pepita) Miquel-Florensa hand the world’s largest infrastructure project at the time, the Canal de Midi. use of new empirical tools” Development economics has been one of the most exciting and innovative 24 The future areas of economic research in recent years, at the forefront of the use of new TSE PhD students Campu� empirical tools such as randomized control trials, discussed in this issue. Going 28 Masters in public policy 25 The figures back to TSE’s forefather Jean-Jacques Laffont who had a keen interest in the and development field, our thriving development group has made great contributions, as the ar- Jean-Paul Azam ticles featuring the work of Emmanuelle Auriol, Matteo Bobba, Pepita Miquel 30 TSE Inaugural lecture 2016-17 Florensa and Stéphane Straub show. This summer, the group brought some Alexandre de Cornière of the world’s most eminent researchers in the field together in Toulouse for 31 TSE Polls a state-of-the-art workshop, including Abhijit Banerjee and Esther Duflo, and we’re delighted to offer you interviews with them in this special issue.

Wishing you an enjoyable read, and hope to see you soon, Trimestrial magazine of Toulouse School of Economics 21, allée de Brienne - 31 015 Toulouse Cedex 6 - FRANCE - Tél. : +33 (0)5 67 73 27 68 Ulrich Hege, TSE Director Publication Director: Ulrich Hege - Managing Editor: Joël Echevarria Editor in chief: Jennifer Stephenson - Production Manager: Jean-Baptiste Grossetti Jean Tirole, TSE Chairman With the help of: Claire Navarro - James Nash Graphic design and layout: Yapak Pictures: ©Studio Tchiz, ©Peter Tenzer, ©US Air Force, ©Fotolia, ©Shutterstock

1000 magazines printed on offset paper from renewable forests. ISSN number incoming. tse-fr.eu 3 News & Events

Appointments & Prizes Strengthening French Jean-François Michel Moreaux venture capital work honoured by the EAERE Bonnefon Marie Ekeland (Paris School of Economics), Augustin Landier Michel Moreaux (TSE - UTC) received work published in Science (TSE-UTC) and Jean Tirole (TSE) have submitted the following the 2016 Lifetime Achievement Award recommendations to the Economic Analysis Council with the Are you ready to drive a car that's pro- in Environmental Economics from aim of reinforcing French venture capital: grammed to kill you? That is the question the European association EAERE that Jean-François Bonnefon (TSE-IAST (Association of Environmental and 1 Clarify and streamline the objectives and investment -CNRS-CRM) and his co-authors Iyad Resource Economists). Rahwan (MIT) and Azim Shariff (Uni- policies of Bpifrance and the “Investissements d'Avenir” versity of Oregon) answer in their study, The prize rewards his enormous scientific programmes. unveiling a moral paradox related to using productivity over more than 30 years, which driverless cars. helped to develop current knowledge in envi- 2 Rethink the operation of the French Public Investment ronmental economics. Bank (Bpifrance) as an industrial policy aiming at the emergence of an autonomous venture capital industry.

3 Assign Bpifrance a governing body that ensures its tse-fr.eu/people/michel-moreaux independence and long-term responsibility, in line with the best international practices. Articulate its strategy with other key actors to optimize public intervention and Emmanuelle Auriol joins its assessment. Marc Ivaldi the Institut Universitaire 4 Develop and encourage links between the scientific (TSE-EHESS) named Vice-President community and entrepreneurial industry in France. of the International Transportation de France Economics Association (ITEA) By order of the Minister of National 5 Clarify the fiscal situation for foreign investors in French Education, Higher Education and Research venture capital, whether they are individuals or In June, Marc Ivaldi became the Vice- on 11 April 2016, Emmanuelle Auriol was institutions, and simplify their access to these funds. tse-fr.eu/people/ President of this association dedicated to  promote the development and application appointed as a senior member of the jean-francois-bonnefon Ensure the tax structure for entrepreneurs encourages of transport economics. The annual ITEA Institut Universitaire de France as of 1 6 reinvestment of capital gains in the ecosystem. conference was hosted by Toulouse School October 2016, for a five year period. of Economics in 2014. Teaching researchers appointed at the Institut 7 Assess the efficiency of current public policy (both fiscal Universitaire de France continue to work at their and industrial) on venture capital to get the most out of Emmanuelle Auriol "home" institution and are discharged of two- the budget for creating an autonomous entrepreneurial thirds of their teaching duties. ecosystem. winner of the Grand Prix du livre tse-fr.eu/people/marc-ivaldi éco 2016 Emmanuelle Auriol, TSE-UTC professor, won the first Grand Prix BFM BUSINESS Maddalena Ferranna 15 other TSE professors are also IUF members: du livre éco (BFM BUSINESS Grand Prize for Economics Books) for her book "Pour wins prize • Jean-Paul Azam • Augustin Landier en finir avec les mafias - sexe, drogue et Doctoral candidate Maddalena Ferranna won the prestigious • Bruno Biais • Michel Le Breton • Helmuth Crémer • Thierry Magnac clandestins : et si on légalisait ?” (Put- 2015 Ernst Meyer Prize from the Geneva Association • Pierre Dubois • Michel Moreaux ting an end to mafias - sex, drugs and for her thesis on decision-making. This prize is awarded • Patrick Fève • Franck Portier people smuggling: what if we legalised every year and recognises particularly promising doctoral • Jean-Pierre Florens • Patrick Rey them?) published by Armand Colin in May. theses in risk and insurance. Christian Gollier (TSE-UTC) was her PhD supervisor. • Christian Gollier • Jean-Charles Rochet • Christian Hellwig tse-fr.eu/people/emanuelle-auriol

4 tse-fr.eu tse-fr.eu 5 News & Events

Newcomers 2016-17 Save the date A warm welcome to our new TSE faculty members: TSE is also proud to welcome five long-term visitors this year:

Professor Johannes Hörner Assistant professors Yale University 17 17-18 Visiting researchers Postdoctoral research fellows Specialised in economic theory, NOVEMBER NOVEMBER Johannes Hörner has been 2016 2016 a professor at Yale for the last nine years.

Ana Gazmuri Wharton School Conference on Sustainable Specialised in Industrial Alessandro Bonatti Jean Tirole's book Agro-food Sectors Organisation, Ana Gazmuri MIT Sloan School of Management works mostly on education and Public Policies inequalities.

Isis Durrmeyer Tim Lee University of Mannheim University of Mannheim Specialised in Empirical Industrial Organisation, Tim Lee’s work in 6 most of Isis Durrmeyer’s research deals with macroeconomics focuses Mikhail Drugov 16-17 the automotive industry and French regulation. on inequalities. DECEMBER DECEMBER New Economic School, Moscow 2016 2016

Health Economics European Simone Maciel Antoine Hochart Pinar Keskin Workshop Conference of Cuiabano École Polytechnique Wellesley College, USA the Econometrics Zheng Chen CADE, Brazil Antoine Hochart’s research Community Université Paris Sud, Orsay Specialised in Industrial is mainly dedicated to game Zheng Chen’s main domain Organisation, Simone theory, as well as work of research is applied Maciel Cuiabano is Stefan Lamp on nonlinear analysis and mathematics. He recently interested in cartels and Yale University optimisation and optimal Joining the environmental completed his PhD on the monopolistic situations. control. “Minimization for Mechanical economics group, Stefan 12-13 Systems”. Lamp is more precisely Marta Troya Martinez JANUARY working on energy issues New Economic School, Moscow 2016 related to solar electricity.

Giulia Pavan University of Rome Giulia Pavan specialises in Industrial Organisation 10th Conference on the and environmental economics and works on Economics of Intellectual issues related to both fields, such as green cars Dan Yang or emission trading systems. Property, Software Dennis Rickert Southwest University, China University of Düsseldorf and the Internet Joining the food economics research Lei Xu group, Dennis Rickert works on McGill University consumption issues. Lei Xu is working on the digitalisation of Events by invitation only: tse-fr.eu/events our societies and its economic impact.

6 tse-fr.eu tse-fr.eu 7 Thinkers

SED CONFERENCE 2016 What shapes inequality?

his summer, TSE hosted the 27th annual meeting for the Society of One of your papers praises Thomas Economic Dynamics, welcoming more than 600 participants. One of Piketty for revitalizing the debate on the brightest and most industrious minds at the event was that of inequality. How are other economists T trying to build on his research? Mariacristina De Nardi, invited to give one of three fascinating plenary talks. Piketty had an incredible impact. He made Professor at University College London as well as a senior economist and inequality front and centre of a stimula- research adviser at the Federal Reserve Bank of Chicago, she broke from her ting political and economic debate. He busy schedule to talk to TSE Mag about inequality, the secrets of academic and his team also worked really hard to success and the impact of Brexit. provide better data and facts, helping How does it feel to be in Toulouse? What drives your work? to put together the World Wealth and This is a great moment for me. The SED Ultimately, my goal as an economist is Income Database that is now available has an incredibly high quality of parti- to try to make people’s lives better. But to everyone. We have a lot more data cipants and to be invited to speak in to propose better policies, you need to now, but we need more solid models Toulouse has special meaning to me. understand why people behave the way to interpret it – for instance, to com- Although I grew up in Italy, my father is they do. For instance, I started studying pare the importance of intergenerational French and I always felt strongly about Social Security reforms in the US with bequests and entrepreneurship in deter- my French identity. My kids have been [SED founder] Tom Sargent. Our models mining inequality. in a French school abroad since they predict that the savings of retired people were three. go down very fast when people get very What role does family have old. But the data shows that highly edu- in shaping inequality? cated people don’t run down their assets We have a lot of exciting empirical evi- dence that the family is an important until they are well past 90. You have com- the market they face, they can use the being productive: their passion is their determinant of key economic outcomes. pletely counterfactual saving behaviour labour supply of both partners – if one work. You can have ideas but it’s a ton of There is work suggesting that by age and that poses a big problem. gets unemployed, the other can take a work to go from an idea to a convincing 16-23 at least half of inequality in life- second job or work more and can share finished product. My biggest advan- “Brexit was mystifying to me. time income is already realized. People Your plenary talk focused on savings savings and housing. In addition, couples, tage is that I never give up. Satisfactions It’s a big mistake that will turn are shifting the focus to younger and and wealth inequality. To what extent and especially older ones, are also exposed happen very rarely and every day you’re out to be very costly” younger ages to investigate the crea- can governments make a difference? to the health risks of both partners. On faced with frustrations. You have to keep tion of human capital, which relates to A lot of what governments do is deci- the other hand, the healthier one can going and take it one step at a time, but earning capacity and health. ding how to take resources help the one who is sick. you also need to listen to feedback from For instance, it seems that from people who are rich “Piketty had an others. You talk with people, you present mathematical and IQ capa- to give them to others. A incredible impact, Does your research influence your to people, and almost every single one be very costly. It’s likely to get worse if bilities are shaped at age large part of government relationships with your own family? of them has something that will make article 50 is triggered, because right now but we need more three. The family has a cru- budgets is thus made up An important lesson from economics is you a better economist. At the same the markets are hoping that this will not solid models to cial role in education and of transfers. So we care that people are complicated but respond time, you cannot be pushed in all direc- happen. We need to find a way to work interpret the data” shaping attitudes. a lot about inequality. In to incentives. I was trying to teach my tions. You need to develop a good sense with the people who voted for Brexit. If

addition, the rise of Trump children about the importance of giving to of what people are saying and why they we do not, it is going to be hard – the How do relationships and marriage and the Brexit vote make you think that others, but it wasn’t working. So I decided are saying it in deciding to what extent pound and the markets will drop further. act as a source of both risk and inequality has repercussions that go to show I had some ‘skin in the game’. I that should influence what you do and What happened was mystifying to me. I insurance? beyond aggregate output and consump- said: “I will put up a matching contribu- how you do it. didn’t have the sense that people were One thing that hasn’t been studied a lot tion – they have much broader reach. tion every time you decide that a toy can ignoring what the vast majority of eco- is how husbands and wives are exposed Governments need to worry about all be given to someone who needs it more.” As a Franco-Italian who has spent nomists were saying in the UK. Clearly to each other’s wage risk. Depending on of these aspects. several years in London, what was we have failed to communicate. But I Do you have any advice for TSE your reaction to Brexit? know my colleagues at UCL will make students? Every economist I know, and especially sure UCL remains a top-tier economics Mariacristina De Nardi department and stays at the very heart Professor at University College London The views expressed in this article are those of Professor De Nardi alone and do not Successful people work incredibly hard. those most involved in Europe, thinks reflect those of the Federal Reserve Bank of Chicago, IFS, TSE, UCL, or NBER. Look at Nobel prize winners who keep this is a big mistake that will turn out to of economics research in Europe.

8 tse-fr.eu tse-fr.eu 9 � inker�

CONSUMER VALUE Why do airlines overbook?

aniel Garrett joined TSE in 2012 as assistant professor (UTC) after com- they will be willing to pay for a speci- How do companies compensate What are your next research projects? particular, passengers may want to pur- pleting a PhD at Northwestern. A specialist in microeconomic theory fic flight. This makes selling tickets in passengers for overbooking? One project related to the above is chase tickets early because they have and industrial organisation, much of his research deals with dynamic advance a good strategy for the air- We know that several companies ask with Toomas Hinnosaar. We are trying to plan their activities and make other D line, and indeed the airline may want passengers how much they would be to understand better why people buy purchases (such as hotel bookings) mechanism design. He tells us about his work and two recently accepted willing to take to give up their seat, tickets in advance, given that they in advance. Perhaps surprisingly, this articles on consumer value. to overbook the flight by selling more tickets than there are seats. Tickets pur- effectively using an auction to determine buy complementary services such as consideration has not been studied in One of your recently accepted articles What is the reason behind chased in advance are then the buyback price of the overbooked hotels to go with their flight tickets. In the existing literature. deals with overbooking in the airline this behaviour? “options to fly”; on the travel tickets. This auctioning seems to be industry. Can you tell us more? Airlines typically explain “Data indicates date, some passengers will the optimal way to compensate bum- We know that airlines sometimes sell that a certain random num- that about 1% of choose to be seated while ped passengers. more tickets than they have seats on ber of passengers can be passengers fi nd others will give up their seat their flights. Should all the passengers expected not to show up for compensation paid by You also individually wrote an article show up at the airport, they then try to for a flight and overbooking themselves in this the airline. To get passen- on the pricing of durable goods. Can negotiate with passengers to get some allows companies to antici- situation on any gers to purchase tickets in you tell us more? of them to take the next flight. Data pate these defections. This fl ight in the US” The idea was to build a new theory advance, the airline raises indicates that about 1% of passengers is also the view expressed in explaining price fluctuations or sales the price of a seat as the find themselves in this situation on any the extensive body of lite- based on consumer values, which is the flight approaches, a pattern we see in flight in the US. rature in operations research. Our article price consumers are willing to pay at a many airline markets. tries to develop an alternative, comple- given time for a product. In the bench- mentary theory. mark model of the literature, the seller’s What are the limits of your model? optimal price does not fluctuate: i.e., it is Our theory is developed under the What is this new theory? constant over time. I show that, when assumption that there is a monopolis- Along with Jeff Ely, who was one of my consumer values change randomly over supervisors at Northwestern, and Too- tic company with no competition. While time, optimal prices fluctuate. There are mas Hinnosaar (Collegio Carlo Alberto, most air routes are subject to some com- high prices at some dates and tempo- Turin), a former classmate, we tried to petition, we believe such competition is rary “sales” at others. model the ideal strategy for airlines to often imperfect (e.g., simply because each maximize profit. The twist we introduce carrier provides a somewhat different How do changing consumer values to existing theories is that consumers service), and so the key ideas we deve- affect the existing theory? don’t know in advance exactly how much lop might be expected to apply. Once you accept that people don’t necessarily know how much they will be Number of passengers denied boarding by the largest U.S air carriers willing to pay for something in advance, from 2000 to 2013, by type (in 1,000s of passengers) it becomes much more evident that you

2000 467 57 Voluntary need to hold sales on your products. 2001 539 59 Involuntary 2002 578 48 Why did you focus on durable 2003 681 65 goods? 2004 651 67 The interesting thing about durable 2005 620 64 goods is that there is a dimension of 2006 621 64

Year choice for customers about when to 2007 619 55 buy, because you usually need only 2008 552 45 one durable good. The fact that cus- Consumer value 2009 702 45 tomers only need one of many durable 2010 727 42 Consumer value is how much someone is ready to pay for a particular good. If 2011 803 34 goods makes the details of the theory a consumer is willing to buy a good, it implies that the customer places a higher 2012 861 39 easier to work out, but the idea seems value on the good than the market price. The difference between the value to the

2013 1,062 57 of Transportation US Department Source: to apply in most settings where cus- consumer and the market price is called "consumer surplus". Daniel Garrett 0 200 400 600 800 1000 1250 tomers have a choice about how to time TSE-UTC assistant professor Number of passengers in thousands their purchases.

tse-fr.eu 11 � inker�

SHAPLEY LECTURE

An outstanding counter-example Specia� issue

runo Ziliotto was attending a conference on game theory in Singapore when he heard the news: he had been chosen to present In October, Ziliotto moved to Paris Dau- the prestigious Shapley Lecture at the Fifth World Congress of the phine University, initially as a postdoctoral B laureate then as a CNRS researcher, Game Theory Society, held in Maastricht in late July. The Shapley Lecture is Development presented every four years by a researcher aged 40 or under whose work where he continues to study conti- nuous-time repeated games and hopes represents an exceptional breakthrough in game theory. At age 27, Ziliotto to find further promising results. is by far the youngest researcher to have presented it. decisions. "These models and these economics Bruno Ziliotto conjectures were presented in 1986, Former TSE PhD the discovery of a counter-example will Game theory student help us to better understand how they Game theory appeared in 1945 and work and the limits to their application," was heavily developed from 1950 explained Ziliotto. This discovery helps to onwards by John Nash, winner of narrow down the conditions under which the Nobel Prize in Economics, and Merten's conjectures can be verified. others. Their work has been used in many fields: Ziliotto points out that he came across this counter-example by following sug- Geopolitics gestions given to him by Jérôme Renault e.g. modelling nuclear (TSE-UTC), his PhD supervisor. "Jérôme deterrence suggested that I focus my work on a Economics specific repeated game model, where e.g. matching students players receive no information about their with universities After earning a masters in probability and environment throughout the game. He Political Sciences macro-economics at the École Normale thought I might be able to find something, e.g. modelling spending Supérieure in Lyon, Bruno undertook a and he was right." options for election challenging PhD at Toulouse School of campaigns Economics. He started the thesis, entitled Renault is still very impressed by the Biology “Long term strategies in repeated two- young mathematician: "Bruno is a great e.g. understanding and player games”, in 2012 and defended in young mathematician and an excellent predicting the results of the September 2015. It looked at a mathe- problem-solver. I obviously guided him evolution of species matical study of game models repeated in his research and helped him to finish Game theory is also sometimes used a large amount of times, involving two a few proofs, but he was exceptionally in sociology, history and philosophy. players who compete against each other creative and discerning, and less than (zero-sum game). It was for this work that a year after the start of his thesis, he he was chosen to present the renowned was able to present a wholly remarkable Shapley Lecture. counter-example. In my opinion his PhD at TSE was an extraordinary success, The experts The fi eld The study The future “These models and these Ziliotto's work generated a lot of atten- and I would like to add that it would be conjectures were presented 14 Meet the TSE 18 Esther Dufl o 22 Should aid 24 TSE's PhD students tion, in particular because it included a difficult to find a game theory thesis development team & Abhijit Banerjee donors be taxed? counter-example to two conjectures with more profound results.” in 1986, the discovery of a 16 The future of research on fi ghting poverty formulated by Jean-François Mertens, a counter-example will help Belgian economist and pioneer in repeated While the discovery of this counter- us to better understand how is increasingly 21 The devil is The fi gures borderless in the detail games. These conjectures concerned a example does not have a direct practical they work and the limits to 25 Facts & statistics very general repeated game model, where application, it makes it possible to recons- their application” 17 Ideas to change the players have limited knowledge of ider part of the repeated game theory the world their environment and their opponent's which is used in many fields (see below).

12 tse-fr.eu Developmen� economic�: the � pert�

RESEARCH AT TSE Meet the TSE development team Stéphane Straub evelopment economics was a driving passion of Mohamed Saleh I am a professor of economics at TSE, and head of its development research group. My TSE’s charismatic founder Emmanuelle Auriol D I am an assistant professor (junior chair) work revolves around the impact of physi- Jean-Jacques Laffont. Today, TSE I’ve been a professor at TSE since Matteo Bobba at TSE and the Institute for Advanced cal and institutional infrastructure issues regularly plays host to the world’s 1998. I am also a fellow of the EEA, CEPR, Josepa (Pepita) Study in Toulouse. My research interests I’m an applied microeconomist and my on the process of economic development. best researchers in this fi eld, EUDN, CESifo and a member of the IUF. are focused on economic history, develop- research interests lie at the intersection Miquel-Florensa Having spent 10 years in Paraguay, some and boasts a rich crop of its own My research interests include industrial ment economics, and empirical political between development and labour econo- of which I recall in my book Frontières, I also creative thinkers who are proud to organisation, regulation, labour economy, I am an assistant professor at TSE and economy. In particular, I am interested mics. I received my PhD in economics from retain a particular interest in that beauti- keep this fl ame alive. Meet some collective decision-making and develop- a member of the Institute for Advanced in understanding the historical origins of the Paris School of Economics in 2011. In ful country at the heart of Latin America’s ment economics. I rely both on theory and Study in Toulouse. In addition to journal the observed socioeconomic differences of its members. the same year, I joined the Inter-American southern cone. I am also secretary of the empirical studies to derive policy recom- publications on water services in Tanza- between religious groups in the Middle Development Bank (IDB) in Washington Jean-Jacques Laffont association, a chari- mendations on industrial organisation nia and signalling with limited resources, East and in the effects of state indus- DC and worked as a research economist table organisation created after his death issues such as privatization, regula- I have written working papers that inves- trialization and mass education on these before becoming an assistant professor by his wife Colette to provide scholarships tion or markets design. Since in practice tigate transfers in rural Mexico, and the differences. I approach these questions (junior chair) at TSE in 2015. to students from developing countries policy implementation matters as much challenges faced by Costa Rican coffee using new datasets constructed from both studying at the University of Toulouse. as policy design, I also study incentive farmers and Colombian fishermen. My primary and secondary data sources. OTHER RESEARCH INTERESTS: in public organisations and government current work extends this research to OTHER RESEARCH INTERESTS: structure. • Labor economics cover trustworthiness in Haiti and cor- OTHER RESEARCH INTERESTS: • Applied microeconomics • Industrial organisation ruption in Paraguay. • Economic history • Contract theory OTHER RESEARCH INTERESTS: • Political economy • Applied microeconomics • Regulation OTHER RESEARCH INTERESTS: • Middle East, especially Egypt • Latin America • Theory of organisations • Industrial organisation • Industrial organisation • Behavioural economics

OTHER MEMBERS OF THE TSE DEVELOPMENT GROUP Philippe Alby, Stefan Ambec, Jean-Paul Azam (featured in the Campus pages), Milo Bianchi, Liliane Bonnal, Rachid Boumahdi, Marie- Françoise Calmette, Jean-Luc Demonsant, Andrezy Dudzinski, Farid Gasmi, Jean-Louis Guy, Marc Ivaldi, Christine Maurel, Georges Molins-Ysal, Isabelle Pechoux, Vincent Rebeyrol, Michel Simioni

14 tse-fr.eu Developmen� economic�: the � pert�

POLICY SOLUTIONS The future of research is increasingly borderless Ideas to change the world

In some sense, development economics from outside of the household, in terms SE’s development economists Mohamed Saleh used to be at the centre of all of econo- of environmental factors, relative prices, are committed to producing mics. The classical economists of the 17th, and local policies. world-class research that can Does building schools 18th and early 19th centuries were all T improve the lives of others. Here we reduce inequality? development economists, as they were A fourth development is the alternative present a snapshot of their sugges- usually writing about a developing country strategies adopted to evaluate the effects “Our findings from Jordan indicate that improving the tions to help policymakers address the (in many cases, Britain) going through a of policy initiatives, including the selection supply of public basic schools significantly increases intractable problems that affl ict those process of industrial transformation. Yet, of participants in the policy programmes intergenerational mobility, especially among daugh- on the lower rungs of the global ladder. development took off as a separate field from among those who are eligible to ters. At secondary level, improving the availability within economics only in the middle of the participate. The different methods for pro- of schools increases boys’ educational achie- 20th century with the pioneering work of gramme evaluation are founded in classical Josepa Miquel-Florensa vements, but girls may be more constrained Hirschman, Myrdal, Lewis, Prebisch and statistics and econometrics, but ultimately by conservative parental attitudes.” others. These scholars departed from rely on different identification strategies How can we protect neoclassical economics in its standard for estimating key parameters, such as small farmers? Walrasian form with constant returns to the response or outcome changes attri- Emmanuelle Auriol scale, pure competition, perfect informa- butable to programme treatment. Beyond “Our analysis of the Costa Rican coffee sec- Can we eliminate tion and so on, by attempting to identify structural or reduced-form econometric tor suggests market volatility can encourage specific rigidities, market failures, shor- approaches to evaluation, the profession excessive integration of export-oriented human smuggling? agricultural chains, to the detriment of tages and surpluses, and other structural has increasingly relied on the design and “Our analysis shows that legalisation of migra- farmers’ welfare. By vertically integra- characteristics of countries in Africa, Asia implementation of social experiments, also tion through the sale of visas, combined with ting, firms can shift profits downstream and Latin America that affect economic called randomized control trials (RCTs), repressive measures, is an efficient way to dis- by paying lower prices, undermining small adjustments and the choice of develop- whereby the policy treatments are admi- mantle smuggling networks. In past decades, producers. In Costa Rica, guaranteeing ment policy. nistered randomly at the individual, family, border controls have been constantly reinforced payments at the end of harvest reduces or community level (you will see more on in spite of very small effects. Reinforcing systematic price-risk exposure while contract enfor- The field has much evolved since then. The this topic in the upcoming pages). controls of undocumented workers at the workplace cement curbs excessive entry. Alternative microeconomic orientation is increasingly may be much more effective.” evident. Economic behaviour is traced from The development group at TSE well reflects organisational forms, such as networks, the individual to the family, and in some the rich diversity in both topics and metho- alliances and consortia, may help promote cases to local schools or communities and dological approaches that permeates the secure market participation and achieve Stéphane Straub labour markets. Because most people in field today. In recent years, we have used higher exporters’ margins.” poor countries continue to work in agricul- both theoretical and empirical tools to pro- Jean-Paul Azam Does grade repetition ture, the focus is often on the family farm duce research on topics such as education, Can economic incentives tame jihad? help students? and includes household production, as well aid, regulation and infrastructure, corruption, Matteo Bobba “Our research suggests economic incentives are often stronger “Looking at Brazilian primary schools, we find that students as the emphasis assigned to human capi- social policies, and conflict, to mention only What widens the reach than ethno-religious ties. The recent history of Sudan and Chad progress more rapidly, and are less likely to fall behind in their tal as a constraint on development. a few. We have presented and discussed our shows that the so-called ‘ancient hatred’ between Muslims and studies, when exposed to episodic social promotion than work in seminars and conferences around of social policies? Christians gives way to pragmatism when economic calculations when they face the threatEmmanuelle of repetition Auriol every (USA) year. A signi- A second emerging feature of the field is an the world and have been one of the focal “When potential beneficiaries share their dictate a cooperative strategy. For over 15 years, Darfur waged ficant fraction of studentsJean-Paul who are Azam pushed (Somalia, ahead India, Pakistan, – despite Bengladesh) interest in explaining how institutions deve- points in Europe for development scholars knowledge and attitudes about a policy war on rebels in southern Sudan. But the northern coalition fell low achievement – appear to make up their learning short- lop and operate in low-income countries, and practitioners from international orga- intervention, this may influence their parti- Jean-Paul Azam, Emmanuelle Auriol apart and peace was quickly agreed when the Chinese built an fall before the subsequent(Senegal, retention Soudan , Tchad, year.” Nigeria, Cote-d’ivoire) and how they potentially enhance or hin- nisations. We are extremely proud of the cipation and the effectiveness of the policy. oil pipeline across the country.” Emmanuelle Auriol (USA) Matteo Bobba (Peru) der the process of economic development. academic work done by our research group Studying conditional cash transfers in rural Jean-Paul Azam (Somalia, India, Pakistan, Bengladesh) Matteo Bobba, Emmanuelle Auriol (Mexico) and we hope this dossier will give the rea- Mexico, we find spillover effects have impli- A third change is the vast improvement in der a sense of our scientific contribution 19 countries studied by the TSE Development EconomicsJean-Paul Azam, research Emmanuelle group: Auriol Josepa Miquel-Florensa (Colombia, Costa Rica, Tanzania) cations for the design, implementation and (Senegal, Soudan , Tchad, Nigeria, Cote-d’ivoire) data, primarily in the form of household in the field, and in economics in general. evaluation of social policies. In particular, the Emmanuelle Auriol (USA) Matteo Bobba (Peru) Mohamed Saleh (Egypt, Jordan) surveys and censuses combined with spatial concentration of the target popula- Jean-Paul Azam (Somalia, India, Pakistan, Bengladesh) Stéphane Straub (Brazil) local community modules which facili- Matteo Bobba, Stéphane Straub tion may offer large gains, as local networks Matteo Bobba, Emmanuelle Auriol (Mexico) Jean-Paul Azam, Emmanuelle Auriol tate the identification of causal effects TSE researchers Josepa Miquel-Florensa (Colombia, Costa Rica, Tanzania) Stéphane Straub, Emmanuelle Auriol (Paraguay) can act as social multipliers.” (Senegal, Soudan , Tchad, Nigeria, Cote-d’ivoire)

Matteo Bobba (Peru) Mohamed Saleh (Egypt, Jordan) 16 tse-fr.eu tse-fr.eu 17 Matteo Bobba, Emmanuelle Auriol (Mexico) Stéphane Straub (Brazil)

Josepa Miquel-Florensa (Colombia, Costa Rica, Tanzania) Stéphane Straub, Emmanuelle Auriol (Paraguay)

Mohamed Saleh (Egypt, Jordan)

Stéphane Straub (Brazil)

Stéphane Straub, Emmanuelle Auriol (Paraguay) Developmen� economic�: the fi el� RUSSIAN FEDERATION 0.06%

MEXICO Measuring 0.06% COLOMBIA 12.81% CHINA aid 14.11%

Despite the billions spent on aid pro- LIBERIA INDIA

Data unavailable Economics © 2011 Poor jects, little has been understood about 79.98% 26.46% 0%>13% their impact. “We don’t know if they 14%>30% BOLIVIA TANZANIA fail or succeed because we don’t mea- 17.84% 73.16% sure,” shrugs Duflo. “It’s hard to measure 31%>65% the counterfactual, so you don’t know More than 65% what would have happened if you didn’t Poor people living under $1 a day around the world implement the project.” Banerjee and Duflo’s response has been to appro- priate the scientific rigour used to test new medicine, subjecting aid projects for intelligence, self-control and the abi- to randomized control trials (RCTs, see Poor FIGHTING POVERTY lity to delay gratification. They do much page 21). In 2003 they established the better after harvest, when they have Poverty Action Lab (J-PAL), with the aim decisions more money.” of reducing poverty by ensuring that Understanding the impact of poverty on policy is informed by scientific evidence. opportunities and behaviour is a central Thanks in no small part to the work of The empiricists J-PAL is now a network of 136 affiliated concern. “The poor are not just like eve- J-PAL, this micro perspective has revo- professors, with an online database of rybody else but with less money,” says lutionized attitudes about what works. 770 randomized evaluations. Duflo. “Say two people have the same “I used to think a lot of problems were idea, but one of them will not be able structural,” says Banerjee, “but I realised strike back The study of poverty offers the pair a to borrow to execute this idea because we were just getting some details wrong. unique opportunity to pursue simul- they are poor. That’s a very foundatio- This makes me optimistic because I don’t taneous objectives: doing nal development problem. have to change the entire framing of the good research and doing “The poor are not Why? Because being poor economic system, but it also makes our good. “The questions are just like everybody cuts you off from opportuni- intellectual challenge very different. In important,” says Duflo. ties. And these problems are ew economists have had as much impact on the Esther Duflo else but with less Indonesia, we gave out cards which said, “That’s why I became an much bigger outside richer way we think about poverty as Esther Dufl o money. Being poor ‘This is the amount of rice subsidy you Abhijit Banerjee economist. As it turns out, countries, where informa- cuts you off from are entitled to’ and the amount of words and Abhijit Banerjee. Tipped as a future Nobel it’s also a delightful field to tion flows and access to F you print – three lines versus four lines prizewinner, Dufl o has already won the John Bates Clark be in, serious without being opportunities” credit are better.” – makes a huge difference. The card Medal, for the best economist in America under 40, dogmatic. Development nearly doubled the delivery of rice. So I and a MacArthur “genius” fellowship. The Independent economists work with very different In their bestselling book Poor Economics used to want to address the big ques- called her “the new face of French intellectualism” and It’s a Franco-Indian partnership that marries contras- tools but we are still able to talk to each (2011), Duflo and Banerjee show how tions. Now, I don’t know what the big Foreign Policy included her in its Top 100 global thin- ting styles and abilities, a combination well suited other. The research is defined by ques- poverty can both hinder and improve things are because small things appear kers. Not bad for a 43-year-old who hated economics to the intellectual, logistical and physical challen- tions, not methods.” decision-making. “Imagine an item in to have enormous consequences, and ges of their punishing global schedule. “Abhijit a shop costs $10, but at the other end as an undergraduate. vice versa. It’s a more confusing world." has a much better understanding of the conceptual, theo- Banerjee agrees: “Why there is so much of town it costs $8,” Duflo suggests. The career trajectory of Banerjee – her mentor, partner retical side,” says Duflo. “My background is the empirical poverty in the world? What can we do to “People are much more likely to go to and MIT colleague – is no less impressive. Once an stuff and I’m more focused on the day-to-day implemen- fix it? Where is it going? Anything rele- the other shop when the price goes from Female unpromising student living next to a Calcutta slum, tation of things. We work together to bridge the span.” vant to these questions is development $10 to $8 than from $100 to $98. This his dedication ultimately earned him a Harvard PhD. Banerjee agrees: “In terms of skillsets we’re completely economics. It’s good that a development is of course a mistake because it’s $2 leaders Together with Dufl o, their revolutionary work has won different. If you’re doing lots of experiments, it’s about course at MIT is very different from one regardless. People often get fooled by Duflo counters this “confusing world” by them the ear of world leaders, aid agencies and billio- keeping your eye on all the balls in the air and Esther is at Harvard. But the fundamental insights this kind of stuff, but not the poor. But refusing to be swept along by received naire philanthropists. amazingly good at that. I can write down models, think are often very similar. In Toulouse, theory the poor might get other decisions wrong wisdom. “I don’t have very strong ideas through the implications, organise and interpret findings, is done in a way that’s very relevant to because they have less information or or priors about things. Ex ante, I’m able In Toulouse in June, they told TSE Mag how cutting-edge design the intervention... But for keeping the train run- all questions in the world – and that’s are so focused on having no money. In to persuade myself that this could economic methods can change lives for the better. ning on time, I’m disastrous!” a wonderful thing.” another experiment, farmers take tests happen, or that could happen.”

18 tse-fr.eu tse-fr.eu 19 Developmen� economic�: the fi el�

RCTS ROUND TABLE

... The devil is in the detail And she’s particularly careful not to make because you are increasing the pool of from 50% to 20%. These are massive inferences that go beyond her findings. talent. Voters realised that these women, improvements.” andomized controlled trials This cautious pragmatism is evident despite all their disadvantages, were Progress in persuading politicians has (RCTs), in which a treatment in Duflo’s assessment of the exclusive actually just as good as men at running been hit-and-miss, says Duflo, but both R group is compared with a focus on women by some microcredit the show. And this evidence is much bet- researchers have generally been sur- control group, have become a gold and other aid projects. “I’m all in favour ter at convincing people about female prised at the willingness of policymakers standard in the assessment of of female empowerment. I don’t see any leadership than any discourse.” to bring them on board. “When I started, social policies. Here, we present reason why half the population should my first instinct was that governments excerpts from a stimulating debate have less rights and opportunities. But Low-hanging would say, ‘We don’t want your research on the subject at a high-profi le TSE I’m sceptical of the argument that there because it’s not politically viable’,” says workshop in June. is a business case for this. Lending exclu- fruit Banerjee. “We discovered that, in many sively to women can improve female cases, they’ll say, ‘Great. Can you come Orazio Attanasio (University College Lon- bargaining power and investment in Banerjee is similarly sceptical about the and implement this?’ So the problem is don): The main role economists can have kids, maybe even household efficiency. prospects for one “big idea” to save the often lack of capacity. We’ve put together in the evaluation of policy interventions But I don’t think there are large effects world, but he believes there is plenty of a pot of money for embedding people comes from the fact that we are trained leading to a virtuous circle of happiness. low-hanging fruit. “We have much bet- inside governments to help them try out to model human behaviour, with all its When you redistribute to a girl, you’re ter technologies now. There’s no reason and scale up successful programmes. But problems and difficulties. RCTs give us not giving to a boy; and it would be an for service delivery to be as abysmal as there’s so much demand for the money a powerful tool to isolate and remove – we can decide not to paint the walls Will this be naturally solved over time? illusion to think the pie will be so much it is. If a teacher doesn’t show up, they that we’re running out fast.” some of the biases of simple observa- and learn if the painting has any effect. Or do we have to create some inde- greater that men will be better off, even are cheating the poorest out of things tional data. From that we can estimate pendent, low-ranked journal to host with a smaller share.” they are owed. We have very good ways very simply the impact of a programme replication studies? of monitoring that people do their job. in a given context. But, most importantly, Price of precision Evidence from experiments in West In India, most people thought that the we need a model that explains what we Duflo: It’s a very powerful tool and get- Duflo: If there is a problem, it’s matching Bengal is more clear cut, showing that food grain distribution system was totally ”In Toulouse, theory is done observe and helps us to understand how ting cheaper. It’s because it’s useful to the money and the people. Coordinated quotas for female politicians are effec- broken, but World Bank data shows the in a way that’s relevant to all policies work. RCTs can make the identi- policymakers that we’re being allowed activities are costly and low return tive at countering voter prejudice. “This proportion of food reaching people has questions in the world – and fication of such models easier. To do so, in. When I started, I thought policy- in terms of ink in a journal; but they is an example where the research really gone up sharply. The money that was when running an RCT we need to collect makers would ask us what to do. But that’s wonderful” are high return in terms of influence. changed my view. I am absolutely in vanishing from the government’s giant much more data. We can also construct typically, policymakers are interested Uncoordinated work is also useful, but favour of quotas for women in politics, labour programme has gone down many more measurement tools, which in how to do what they’re planning to we need a place where we can see it you need to validate and pilot. do anyway. all. The American Economic Associa- tion has successfully started a new Esther Duflo (MIT): RCTs have grown enor- Banerjee: It has become very difficult registry. If it sticks, we’ll have a good mously in scope and number. They’re also to do a useful, low-budget policy eva- sense of the universe of studies that Cho� in� � charit� crowding in other things, partly because luation. Ten years ago, you could do an are out there. they’re a benchmark. If you do a shoddy RCT with $20,000, but now expecta- How can the public maximise the impact of their donations to development job, someone is going to do an RCT and tions and standards have gone through projects? “Effective altruism is a great movement that’s helped young people Pascaline Dupas (Stanford): Incentives show that. We’ve all realised that the the roof. It’s completely perverse and it become aware of many issues about poverty,” says Duflo. “If you want to to replicate are very low. But it’s also devil is in the detail. discourages people from trying. Lots of give money, be very careful of organisations that rate charities on the ratio very important to ensure we can do questions can be answered for $20,000 of their overheads over what’s been given to the people directly. It’s a stu- meta analysis later on. We’ve been Abhijit Banerjee (MIT): The great bene- and the world would be a better place. pid measure because it’s the impact that is important.” fit is we can test exactly what we are compiling evidence from 16 rando- modelling. I know that test scores can mized experiments and we had to The following websites have useful www.thelifeyoucansave.org be improved by hiring and training Replication crisis write to the authors asking for very recommendations and use ratings teachers, changing textbooks, the cur- Frederico Finan (University of Califor- basic facts. In public health, they have based on impact: www.poverty-action.org riculum, painting walls, buying school nia-Berkeley): Multi-country studies a template for how to write a paper. buses and providing meals. But which might solve external validity issues, but We need more information to extrapo- www.givewell.org piece of this is important? Experiments there’s a lack of researchers who want late about whether the experiment will allow us to get inside that black box to take on the replication conundrum. work in another context.

20 tse-fr.eu tse-fr.eu 21 Developmen� economic�: the stud�

HEALING AID Should aid donors be taxed?

he aid business is a maze,” say TSE researchers Emmanuelle Auriol reports to donors and 1,000 delega- “The novelty of our approach is to look “ and Josepa Miquel-Florensa. This maze is overcrowded with donors, tions every year. Each of the donors at the fragmentation problem from including Western nations, emerging countries such as China, Brazil have different accountability and pro- the recipient’s point of view,” say the T curement rules, and each project must TSE researchers. “Our paper explores and Arab states, the World Bank and other multilateral organisations, as well as thousands of NGOs, foundations and private entities. This proliferation of be matched with the recipient country’s the actions the recipient might take to existing portfolio.” As a result, adminis- discipline the donors. In our setting, donors and the persistence of bilateral, micro-managed aid projects places a trative costs absorbed a massive 4.7% the recipient does not reject unilateral heavy administrative burden on often ill-equipped recipient countries, wasting of bilateral aid in 2005 (only 9.6% went projects if they are useless, but instead valuable resources. to humanitarian and food aid and just taxes them to redirect the aid where 2.2% went to NGOs). it is most needed. The tax instrument therefore fills a double purpose. First Efforts to fix the ‘donor coordination’ pro- it helps the recipient country to collect blem have had limited success. In fact, funds to strengthen its administrative since the Paris Declaration on Aid Effec- capacity. Second, it helps to align the tiveness in 2005, fragmentation of aid interests of the donors with the inte- has increased. This makes it increasingly rest of the recipient. It is a stick used to difficult for recipient government repre- keep the donors in line and to modu- sentatives to coordinate and manage aid late their aid flows.” flows. Confusion on available resources and procedures to reach them is common. From a policy perspective, this mecha- nism is appealing as it is simple and To improve aid effectiveness, Auriol and might help to finance administrative Miquel-Florensa propose a simple, origi- costs. Auriol and Miquel-Florensa’s nal tool. In a recent working paper titled model also proves the existence of an complementary principles, this group allowing recipients to arbitrarily tax uni- So why don’t donor countries pool their “From the point of view of the reci- ‘Taxing Fragmented Aid to Improve Aid upper bound on the donors’ aggregate aims to reduce transactions costs for lateral aid would facilitate corruption. efforts? Benevolent donors may want pients, aid fragmentation generates Efficiency’, they argue that if a donor budget above which the recipient stops recipients and to increase aid effec- to circumvent the corruption and inef- huge transaction costs,” explain Auriol country insists on pushing its own deve- using the aid money. This comes from tiveness, even at the cost of reduced “A better solution would be to nego- ficiency of recipients. Self-interested and Miquel-Florensa, who specialize in lopment goals, it should pay a fee for it. the agent’s management costs and the visibility for the donors. tiate corrective taxation rates at the donors may devise projects and rules industrial organisation and development Recipient countries, their model sug- decreasing returns of the projects’ pro- international level,” the TSE researchers to promote their own exports or maxi- economics. “For example, the Tanzanian gests, should be allowed to impose a tax ductivity. Waste of resources comes both The paper shows that the optimal tax propose. “Coordination at an earlier stage mize media coverage. Either way, most government has to prepare over 2,000 on uncoordinated, unilateral aid. from agent’s and principals’ choices. rates should be different for different of programme design, allocating tasks to prefer bilateral aid because they want So increasing the aid budget does not donors, depending on the donors’ bias donors according to their comparative control. In 2009, for example, bilateral necessarily lead to higher development and on the productivity of their unilateral advantage and budgets, and allowing aid was almost three times greater than production. projects. But Auriol and Miquel-Florensa recipients of aid to tax unilateral aid multilateral aid. admit it is unlikely that donors will agree and micro-aid schemes, could greatly The empirical evidence presented to this in practice. Another concern is that improve the effectiveness of aid.” suggests that even when donors are altruistic, their heterogeneity and lack of coordination causes inefficiencies. This result militates against “cosme- tic” micro-aid schemes. Instead, say Find out more:

“For recipients, aid fragmentation the researchers, donors should follow ‘Taxing Fragmented Aid to Improve Aid Efficiency’ generates huge transaction costs. the example of the Nordic Plus group, by Emmanuelle Auriol & Josepa Miquel-Florensa, see: Tanzania has to prepare over 2,000 which includes Denmark, Finland, www.tse-fr.eu/sites/default/files/TSE/documents reports to donors every year” Norway, Sweden, Ireland, the Nether- /doc/wp/2015/wp_tse_600.pdf lands and Britain. Through joint financing arrangements, procurement policy, and

22 tse-fr.eu tse-fr.eu 23 24 tse-fr.eu TOMORROW'S ECONOMISTS Developmen� economic�economic�:the� ture 4 3 2 1 her work next year. gious beliefs.She isexpected topresent to Ghanaconduct afieldstudyonreli- of institutions.Shealsorecently travelled issues andquestionsrelated tothequality cal techniquestoinvestigate governance Her work usesboth theoryandempiri- then came back to TSE to pursue her PhD. tant atRBBEconomics inBrussels.She Lassebie worked as an economics consul- tics at Toulouse School of Economics, Julie After studyingeconomics and mathema- Economics” Development “Essays in Julie Lassebie should present herwork next year. in Toulouse for her last year of thesis and develop afieldexperience. Sheiscurrently spent three months inNamibiaorder to dual's choice inautocracies”. Shehasalso Thesis “Government strategiesandindivi- London andCityUniversity working onher months in London at the University College and Development Masters. She spent 6 mathematics) degree andaPublicPolicy Toulouse SchoolofEconomics MASS (Applied Born inToulouse, AnaïsDahmaniholdsa in autocracies” individual's choice strategies and “Government Anaïs Dahmani TSE's PhDstudents PHD SUPERVISOR: EMMANUELLE AURIOL PHD SUPERVISOR: EMMANUELLE AURIOL 2 1 of Latvia before joining the Stockholm Petersburg State University and University and studied theoretical physics at the St. Vitalijs Jascisensgrew upinRiga,Latvia Procurement” “Essays onPublic Vitalijs Jascisens complex situations. conflict studiestobetter understand work combines history, economics and of Economics for his PhD. His research mics before arriving to Toulouse School Barcelona Graduate SchoolofEcono- State University. He laterjoinedthe to enjoy economics attheMoscow Born in Voronezh, Russia, Oleg started and development” infrastructure, conflict “Three essays on Oleg Polivin rity andtaxation systems. the designof procurement, social secu- rically analyse various policyfailures in large administrative datasets toempi- ters degree. InhisPhDthesis,heuses to Toulouse where hereceived a Mas- School of Economics. He then moved PHD SUPERVISOR: JEAN-PAUL AZAM PHD SUPERVISOR: STÉPHANE STRAUB

on theTSE website Find outmore 4 3 62 62 peopleownthesameashalfworld, but remains highinlow-income countries Undernourishment hasdeclinedglobally 39 IN 1960 Life expectancy atbirth,total(years) 43 IN 1995 reveals OxfamDavos report in low-income countries UNDERNOURISHMENT (* %of population, approximately) * (* Approximately) % IN 1990 +10 * * 27 IN 2015 59 +20 IN 2010 3.7 BILLIONS % * * *

Source: data.worldbank.org/ Source: data.worldbank.org/ Source: https://www.oxfam.org Developmen� economic�:thefigure� World’s populationliving belowtheinternational 37 population willlive inurbanareas by 2030 75 An estimated60percent oftheworld's Access toelectricityishigherthanever IN 1990 IN 1990 60 poverty lineof$1.90 aday declines % % (% of population) % A DAY A $1.90 2030 85 13 IN 2012 IN 2012 % %

Source: data.worldbank.org/ Source: data.worldbank.org/ Source: data.worldbank.org/ Actor�

US AIR FORCE - JÉRÔME BOLTE Big data & the military US AIR FORCE RESEARCH LABORATORY The laboratory and its predecessors have overseen more than 80 years of critical research efforts for the Air Force and érôme Bolte (TSE – UTC) is a mathematician specialized in optimisation Department of Defense. Its technological breakthroughs can be algorithms. His work, and more precisely an article co-written with Hedy found in all of today’s modern aircraft and weapons systems. Attouch (University of Montpellier) and Benar Svaiter (Instituto Nacio- J It has also contributed to signifi cant advances in modern nal de Matemática Pura e Aplicada, Rio de Janeiro), has caught the attention communications, electronics, manufacturing, of researchers at the US Air Force, which has decided to fund his research. and medical research and products. In 2011, Bolte and his co-authors If the US Air Force is interested in received a surprising email. “The Bolte’s work, it’s largely because of sender’s email address was quite the challenges they face in dealing intriguing, listing all sorts of ranks with the gigantic datasets gathered and serial numbers,” the researcher by modern technology. $4.4 10000 recalls. The message explained that Jérôme Bolte BILLION BUDGET RESEARCHERS the US Air Force was very interested Applications vary greatly and range TSE – UTC researcher (2014) in their work and was willing to fund from digital imagery to artificial intel- several trips to the US to exchange ligence issues, deep learning, to the ideas and solutions regarding optimi- simulation of quantic calculus. “With sation algorithms. traditional data analysis Bolte has been pleasantly surprised by "With traditional techniques it would take the mindset of the US Air Force. “After meeting with an operational research “The US Air Force owns data analysis centuries to class and use research departments the data we have today. director, I must say that they seem in many fields, inclu- techniques it would The storage capabilities very enthusiastic for each and every ding mathematics and take centuries to are colossal and we need discipline. They are very open-minded calculus. Their theoreti- class and use the new ways of exploiting and deeply convinced that solutions cal topics are very close data we have today" big data.” can come from anywhere and that to ours regarding opti- nothing is useless. This humility and misation,” Bolte explains. “They are Because of its 6.1 status, Bolte’s curiosity is fascinating and refreshing.” constantly looking for new research partnership with the US Air Force and our article caught their attention.” allows him complete freedom in his His contract with the US Air Force will 3 TYPES 6.3 research. “I can publish my work in be up for renewal in 2017. Meanwhile, 6.1 OF RESEARCH ADVANCED The three researchers now fly regu- the journals of my choice. The US Air as the accumulation of data conti- BASIC TECHNOLOGY RESEARCH larly to a research centre at the former Force isn’t interested in the exclusivity nues to rise, optimisation algorithms PROGRAMMES DEVELOPMENT Griffiss Air Force Base, outside New of my research but is willing to deve- will become increasingly important York, to exchange on their latest pro- lop new theoretical approaches and to organisations in every sector of Basic research in which new scientific Advanced technology development society. jects. Those discussions led the US Air techniques in optimisation.” phenomena are sought in an effort to in which multiple technologies are Force to agree an extended, three-year discover and advance fundamental integrated and demonstrated to enable research partnership with the TSE- knowledge in fields important to the development of a new military UTC researcher which began in 2014. “Although it may be surprising for researchers in France, national defence. Such research is capability to satisfy a military need. the US Air Force bureau in London AFOSR/EOARD does generally broad in nature can be “The US Air Force has three funding fund projects all over Europe and I believe the area of considered inherently “high risk”. plans named 6.1, 6.2 and 6.3. The 6.1 6.2 research of Jérôme Bolte, Optimisation with a focus APPLIED plan, which I received, gathers theo- on Lojasiewicz inequalities, will be important for many RESEARCH retical research; 6.2 is for projects future applications in many areas, including machine with possible applications; while 6.3 learning and big data at large.” funds research projects directly lin- Applied research in which technology is developed based on a newly ked to military application.” Bruce Suter, Rome Laboratory, NY discovered scientific phenomena or by the application of scientific phenomena in a totally different manner than currently applied.

26 tse-fr.eu tse-fr.eu 27 Campu�

TEACHING AT TSE Masters in public policy and development he TSE Public Policy and Development masters programme aims to teach students how to analyse social and economic issues related to Overseas Development T developing countries. Here's a quick presentation of the programme as well as testimonies from alumni Laura Berméo and Anaïs Dahmani. Institute Under the direction of Jean-Paul Azam public policies in developing countries The Overseas Development Institute (TSE-UTC), the teachers encourage stu- more favourable to economic develop- (ODI) is the UK’s leading independent dents to address the core question of ment. Many economists and public policy EGYPT thinktank on international development development economics: why are there consultants work in these institutions. and humanitarian issues. so many inequalities between countries Through this masters, TSE aims to help “The ODI Fellowship Scheme offers postgraduate in terms of wealth? The answers can its students become actors in this field, NAMIBIA be found both at the level of states, by able to use theoretical knowledge and economists and statisticians the opportunity Jean-Paul Azam Anaïs Dahmani of working as a civil servant in a developing studying economic policies; and among empirical data to help millions of people Programme Director Alumni - PhD firms, by observation of their strategies. in developing countries. country for two years. We are looking for people who have a good grasp of applied economics Ever since the Second World War, the and statistics, who can communicate technical international community has become After publishing mainly on the macroeconomics of Africa, Jean-Paul Azam has Anais Dahmani concepts to non-specialists, who can work in a more focused on helping developing focused since the mid-1990s on explaining violent conflict and its prevention, with My research interests relate to both development economics and politi- foreign bureaucracy and live in a very different countries out of poverty. International application to foreign aid, civil war, and transnational terrorism. His recent article cal economy, with an emphasis on government strategies and citizens’ culture, and who can thrive in challenging envi- organisations such as the World Bank “Killing for the Sake of Infamy: The Herostratos Syndrome and What to Do about it” choices in autocracies. I look at how political and economic forces affect ronments generally. Over the past few years the or the UN Development Programme will be published by the Peace Economics, Peace Science and Public Policy journal. political institutions, when there is a conflict. Fellowship Scheme has recruited several gra- were created in an attempt to make duates from TSE who have proven well-suited As an example, my work “Social media, heterogeneous networks and to the demands of the scheme. Five graduates government strategy: The Egyptian Arab Spring” focuses on the impact from TSE are currently of information and communication technologies (ICTs, such as mobile enjoying successful phones or the internet) on popular uprisings, depending on existing fellowships in some Laura Bermeo social networks and government propaganda. of the toughest coun- Where do you work and which tries we operate in.” position do you hold? This work is highly correlated with my masters thesis. In 2012 my advi- Since completing my masters degree, I’ve HONDURAS ser Emmanuelle Auriol suggested many themes, and I chose “ICTs and ODI Head of the been working in Honduras for almost a governance”. My research then evolved towards political economy, as Fellowship Scheme - year as a field coordinator of an impact I drew inspiration from the Arab Spring events. Robin Sherbourne evaluation (IE) for a World Bank research group. Honduras has one of the world’s It has often been said that social media were used as revolutionary highest homicide rates. In this context, tools by rebels aspiring to regime change – ICTs allow easier and faster the IE studies the protective role of a Laura Bermeo communication among citizens, improving both the spread of informa- Teaching inspired labour market intervention that aims Alumni - Policy tion and coordination of groups. Despite widespread optimism that ICTs to move at-risk youth away from the would help spread democracy around the world, most autocracies have by research market of crime. now realised the power of such tools and use them to channel pro- Most of the teachers - Emmanuelle Auriol, team and the project implementation strong emphasis on programme eva- paganda and repress rebels. There is an important trade-off between Jean-Paul Azam, Matteo Bobba, Sylvain How do you apply the skills you unit. Additionally, I provide input to the luation. I had three different courses on users’ freedom and government surveillance. Chabé-Ferret, Pepita Miquel-Florenza, Mohamed learned during your TSE masters design of the intervention and help adapt this subject as well as other complemen- Saleh, Paul Seabright, Stéphane Straub - degree to your new role? the study to the local, Honduran context. tary courses which provided a good mix As my research is related to both political economy and develop- are leading lights in their field of research, My job responsibilities relate to the imple- I am also involved in data analysis and of theory and its “real life” application. ment, I decided to get a glimpse of more applied development with experience of consulting for international mentation of the project intervention administrative tasks. I also consider the TSE masters pro- work and accepted a three-month job as a research assistant for organisations such as the World Bank or the and the collection of data used for IE. gramme as very demanding, with high IPA Namibia. Being involved in a local project has helped me a lot French Development Agency. I am mainly responsible for bridging the The M2 PPD degree has been extre- standards that enabled me to develop in understanding development and political issues, such as power requests and needs of the investigation mely helpful for my work because of its many valuable skills. and bargaining games.

28 tse-fr.eu tse-fr.eu 29 Campus TSE POLLS

TSE INAUGURAL LECTURE 2016-17 Kicking off the academic year IS ECONOMIC GROWTH lexandre de Cornière (TSE-UTC) gave new and returning TSE students NECESSARY TO ELIMINATE the traditional Inaugural Lecture for the academic year 2016-2017, A presenting his latest work on “The Economics of Digital Media Plat- forms”. Here, we present some of the highlights of his talk. ABSOLUTE POVERTY? “In many ways, the internet has deeply An ever-increasing number of online impacted our daily lives, forever changing media are now fighting for consumers’ GIVE YOUR OPINION ON DEBATE.TSE-FR.EU/POLL our ways of communicating (for exa- attention as the internet abolished the mple, email, Skype...), shopping (Amazon, entry cost for the industry. There is no Ebay…), moving (Uber), planning holi- longer any need to invest heavily in a days (online tour operators, AirBnB...) distribution network or printing. and socializing (Facebook, Twitter, Meetic...). I’d like to Another massive change “On the internet address another area that brought about by the inter- is being completely rede- you’re more likely net is the introduction of signed by the internet: the to read articles and media gateways or plat- media industry. opinions that are forms such as Google News similar to yours” or, increasingly, Facebook. Historically, the media These gateways select industry has been financed articles based on complex through a mix of advertising and consu- algorithms which try to present you with mer fees (subscription and per-issue) as articles suited to your profile. This “cus- Alexandre de Cornière well as public subsidies. For example, in tomization” of the news is facilitated by TSE-UTC researcher France, around 70% of the printed press’ the enormous amount of data those revenue have come from advertising, and companies have about their customers. 30% from sales. But many media out- The internet is forever changing the lets relying on sales were also relying on One side-effect of this development is media landscape and it is very hard advertising, creating what economists that you’re more likely to read articles to know whether this is for better or call “multi-sided-platforms” which can and opinions that are similar to yours. worse. These profound changes are generate more revenue from advertisers Researchers call this the “echo chamber still unfolding before us as the media if they have more consumers. effect”, in which people are presented industry struggles to develop new, sus- the same opinions repeatedly and rarely tainable business models for the 21st Academic literature on multi-sided-plat- come across alternative views. century. forms suggest that we can divide media into two categories: 1. Relying more on advertisements and thus proposing cheap content that is RECENT POLL RESULTS FROM TSE DEBATE: attractive to many viewers 2. Relying on sales and therefore deve- lop high-quality, premium content that of47% time spent is very attractive to a few viewers on media is online (UK, 2015) 66% 52% 71% 68% By making most news coverage avai- think the fall of oil prices had anticipated don’t see “uberisation” think intelligent software lable for free, the internet has eased the is a good news for the the COP21 as an economic and robots won't create media towards the first category. While global economy agreement threat mass unemployment about 30% of people were paying for news in the pre-internet era, this fell to 9% in English-speaking countries in 2015. debate.tse-fr.eu 30 tse-fr.eu