Select text to annotate, Click play in YouTube to begin
00:00:12
INAUGURAL LECTURE POVERTY AND PUBLIC POLICY NOVEMBER 24, 2022 WITH THE SUPPORT OF THE COLLÈGE DE FRANCE FOUNDATION It's rare to hear so much applause before we've even begun. Welcome to the Defender of Rights, to the Governor of the Bank of France, Madam Minister, the members of the President of the Republic's cabinet
00:00:58
and ministerial cabinets... the Chief of Staff for the mayor of Paris' cabinet... the various presidents, directors, and heads of institutions... the professors, and to all of you who are also present. My dear colleagues and friends, Ladies and Gentlemen... it is my great pleasure to welcome all of you to Esther Duflo's very first lecture. She is taking over as chair at the Collège de France
00:01:27
with her talk entitled "Poverty and Public Policy." Actually, this is a... I don't want to say "repeat" event, but her second inaugural lecture at the Collège de France. In 2008 and 2009, she also held the annual chair with "Knowledge Against Poverty." Many of you present will certainly remember that moment. The fight against poverty is a constant source of concern
00:01:53
throughout Esther Duflo's research, as you will hear for yourself in a moment. We can all agree poverty has been a major challenge for humanity since the origins of civilization. Throughout antiquity many attempts were made to explain poverty, such as in ancient Chinese, Mesopotamian, Egyptian, and many other texts. There is often a focus on the need to... "save" the poor
00:02:24
or to "banish" poverty. For example, in a text attributed to Confucius, we can read: "Poverty is a disgrace under a good government, "but wealth is also a disgrace under a bad government." In the Book of Deuteronomy, in the Hebrew Bible, we can read a sermon that has been attributed to Moses, in which he speaks about wanting to eliminate poverty, while also stating: "The poor will always exist in our nation,
00:02:53
"which is why I am giving you the following Commandment: "'Thou shalt surely open thy hand unto thy brother, "to thy needy, and to thy poor.'" Though it is difficult, and even impossible, to provide a concise definition of what poverty is, the last few decades were marked by a desire to see it reduced. Ultimately, the number of people living in extreme poverty has been halved, mortality rates for infants and mothers
00:03:21
have also been reduced considerably. Almost all of the world's children now go to school, and diseases like malaria and AIDS are now being treated better. These advances were not made by a few countries or through the exceptional generosity of wealthy nations, but through more reasonable and efficient policies in many parts of the world.
00:03:47
However, this positive trend is now endangered. The reduction of poverty has slowed in the past few years. The COVID crises, and the subsequent economic crisis... pushed many individuals into a very precarious situation, which they had never expected. And the effects of global warming, which will affect the poorest nations very disproportionately, now threaten to erase much of the hard-won progress
00:04:15
in the fight against poverty. As such, the Collège de France must contribute to the fight through the foundations of its mission: research and the sharing of knowledge from that research. We are very happy that Esther Duflo has joined the Collège de France to give us a better understanding of poverty, its causes, and what we can be done to reduce it. Esther Duflo holds degrees in History and Economy from the École Normale Supérieure of Paris.
00:04:44
She then obtained her doctorate from MIT in 1999, where she then held various teaching and research positions. Since 2005, she has been a professor on the fight against poverty and the economy of development in the Economics Department at MIT. Her research is based on investigations and action. She co-founded J-PAL, an action laboratory against poverty.
00:05:09
Esther Duflo and her colleagues conducted a research investigation, inter alia in India, in the state of Tamil Nadu, where for six months they tested various methods to end the isolation of elderly people. Her study revealed the necessity of providing a sufficient pension for the elderly, while also offering psychotherapeutic services to people suffering from isolation. Esther Duflo attempts to understand the economic life cycle of the poor,
00:05:37
with the intent to design and evaluate social policies and programs. She has worked in the fields of health, education, financial inclusion, the environment, and governance. As an economist, Esther Duflo wants to answer some concrete questions by working in the field to try and offer solutions that are based on observations and above all on testing. Her research has earned her a significant number of distinctions,
00:06:03
which would be impossible to fully list right now. However, I will still mention a few, of those: in 2019, the Sveriges Riksbank Prize in Economic Sciences, given in honor of Alfred Nobel, which she won with Abhijit Banerjee and Michael Kremer, the Golden Plate Award in 2022, the Princess of Asturias Social Sciences award, the Angela and Shu Kai Chan Social Science award,
00:06:30
the Infosys Prize, the David N. Kershaw prize, the John Bates Clark Medal... I could keep going for 30 minutes, so I'll stop there. Since 2021, Esther Duflo has been a Dame Commander... Or is it "Commandress?" No, I don't think so. Of the Legion of Honor. She has published many books, including the two volumes of Fighting Poverty, published by Seuil in 2010, and Rethinking Poverty, with Abhijit Banerjee, in 2012,
00:06:59
also published by Seuil. In 2010, the original English version of their book Good Economics for Hard Times won the Deutscher Wirtschaftsbuchpreis for the best book of the year on economics. She has also just published a series of illustrated books for young readers, in which she provides a simple illustration of the economic issues and challenges that we are currently facing. This year Esther Duflo will be offering a course called
00:07:26
"Fighting poverty: "from science to public policies." So the course will begin tonight, with this first inaugural lecture, "Experience, Science, and the Fight Against Poverty "almost 15 years later." We are very excited to hear from you, Esther Duflo, so I will now hand it over to you. Hello, Mr. Administrator, dear colleagues and friends. Hello, everyone.
00:08:05
Almost 15 years ago now, you gave me the... exceptional honor... Just a young girl of 36 back then, of inviting me into these halls to give my first inaugural lecture. I took that moment, that chance... to place all of my hopes on a more humane and pragmatic sort of economic sciences,
00:08:33
along with more humane and pragmatic policies. My hope was to surpass... all the obstacles, with the despair, for example, on the one hand that one might feel when confronting the problem of poverty throughout the world, and, on the other, the great ease there is with accepting cookie-cutter solutions
00:08:58
that can supposedly solve all our problems at once. My hope was to provide a hand to each person, as needed, to solve the concrete problems that... are associated with extreme poverty. At that time, I had mentioned... that one cornerstone of the project could be made up of random experiments.
00:09:24
But all those experiments were obviously in their infancy then. My projects had just begun and J-PAL, which you heard about before, only had about 100 active projects. So the idea that I had for today, 15 years later, was to look back and perhaps draw, not a conclusion, but more of an idea of what... The things that have happened since.
00:09:49
Have we made any progress on the issues around poverty, and, in particular, in regards to solutions... for poverty? And of course, so as not to attribute everything to experiments, I want to look at the... roles that this method actually played both in defining economic problems and in regards to... economic policies.
00:10:15
Let's start with the good news. As we heard earlier, the... It might be harder to notice if you live in France or a Western country, but poverty has dropped around the world in a very... significant and systematic way. Between 2010 and 2019, it was halved with regard to those in extreme poverty, meaning the number of people who live on less than two dollars a day per person.
00:10:43
That figure dropped from 15% to 8%. And actually, contrary to belief, this reduction in poverty is not only because of the rapid growth being seen in China and India, with their huge populations and many people who have risen out of poverty. Actually, if you look at this graphic, you will see that, in terms of percentage points, the greatest... reductions in poverty
00:11:09
occurred in countries on every continent and in countries you might not ever expect as being victors in the fight against poverty, such as Tanzania, where... we saw some very fast progress... as well as the Democratic Republic of the Congo... Pakistan, the Republic of Kyrgyzstan, for example. The countries are distributed over... almost every continent,
00:11:35
and not just in Europe, where people are much wealthier. So, this reduction in... extreme poverty is not only a victory that is because of China's rapid growth. You might be thinking, "Sure, it's all nice and good, "but that triumph is really a bit... limited. "People are only making "a tiny bit over two dollars a day, after making a tiny bit less, "since two dollars means so little to us,
00:12:03
"that just a few pennies... from our joint prosperity "would be enough to get... many people over that threshold." Actually, even though that threshold is derived from fixed prices, meaning that it also takes rising prices into account... Can someone who lives on just two dollars a day lead a decent life? So, is that rapid increase in poverty--
00:12:31
Pardon, that rapid decrease in... extreme poverty not simply the result of a line that has been set... that is unacceptably low? The best way to answer that question is to... not only focus on economic... poverty, in terms of consumption or income, but also on the quality of life for the poorest. And obviously... a very large part of quality of life
00:13:00
is being alive. That progress... As you heard in the introduction, there has also been some rapid progress here, with the mortality rate of infants and mothers dropping by half since 1990. Once again, the phenomenon of reduced mortality rates has not only occurred in countries where growth has been the fastest, but also in countries that have had no or very little economic growth,
00:13:28
such as Malawi, for example... which, despite that, still saw some of the quickest progress. And overall, in terms of percentage points, progress has been the fastest in Sub-Saharan Africa and in Southern Asia. So, where does this progress stem from? Well, in part... I believe, from the advance... of pragmatism... as much in the governments of developing countries
00:13:59
as well as within the international community, which have managed to shift away from the consensus of Washington, focused on the importance of economic growth, the reduction... of deficits... and so and so on, on macroeconomic indicators, which are based on a certain fiscal discipline, towards a wider definition of a country's economic and political success... overall,
00:14:27
which incorporates the eight Millennium Goals, along with the Sustainable Development Goals, which include education, health, education, respect for women, and so on. These Millennium Goals, as well as the Sustainable Development Goals, helped to focus the attention... of political decision-makers on objectives that are much more concrete,
00:14:55
specific, and... reachable... than was possible... before. The result is that there has been much less Manichaeism, like before, when there was this sort of opposition with boundless growth or foreign aid being the answer to all problems, when the act of overthrowing capitalism was presented as the only way of resolving all problems...
00:15:21
Basically, all sorts of cookie-cutter solutions. There are of course exceptions... such as Venezuela, for example, which acted on its ideology, without any great success from a development perspective. This has had two advantages. Firstly, such as the generation... of my parents said, back in 1968, because no one "falls in love" with growth rates. So economic growth, in and of itself
00:15:48
doesn't have any... doesn't provide any benefits. Only in the sense that it leads, or can lead, to an overall rise... in general well-being, whereas the reduction in mortality rates for infants, however, is precisely the type of goal we can "fall in love with." And then the other advantage has been that questions that are phrased well
00:16:15
require much more precise answers. Already back in 2009, I spoke of how statistical analyses of various nations' performance did not help us to discern which factors and which set of policies a country could implement as a sure-fire way to ensure economic growth. Various countries, such as China and Korea,
00:16:43
have experienced... entirely different trajectories, which are hardly applicable to any other country. However, we do now have more and more solid knowledge about the effective ways to vaccinate children, prevent maternal death, ensure children go to school, and perhaps even learn something... while there. And so that focus on... concrete objectives
00:17:13
allowed for progress to reach them, in areas where it was never easy to explain to African countries on how to make progress in their growth, for example. Our experiments played a role in creating this... toolbox, which now exists and is available to countries and... governments that would like to implement and use it.
00:17:42
Just a quick reminder as to what is meant by an experiment based... on geographical illustrations. This is the district of Udaipur, in Rajasthan, the site of one of our experiments... several years ago. In order to conduct a random experiment, villages are also chosen at random. The villages tagged in white saw no change to the status quo. This... experiment here
00:18:08
dealt with vaccinating children. The villages tagged in blue saw an improvement in their vaccine infrastructure. The villages tagged in red are also places where parents were incited in some way to take their children... to get vaccinated. And so, since the... choice was completely random, the points are sort of spread out all over the map,
00:18:35
but more importantly there's no room to think there might be systematic differences between the villages before and during the experiment, except for those due to the various interventions implemented in each village. That's the basic principle of a randomized experiment. In my lecture from 2009, I had mentioned Roosevelt as a way of stating the strong demand for experiments,
00:19:02
that we needed experiments, everyone wanted them, but to be perfectly honest... it still wasn't all that clear that that need was as widespread as it is. J-PAL already existed at that time, without about 100 projects or so, but, as you can see, out of those 100 or so projects, only eight of them, to our knowledge, made enough of an impact
00:19:27
to be adopted in the form of policies. Overall, they affected... 23 million people... around the world. That's a lot, but... it's not so much for the entire globe. But it was a start. At that time, there was also a certain skepticism... which was evoked by some people, like Lant Pritchett, Bill Easterly,
00:19:55
around the fact that that demand for experiments existed, and that anyone would be interested by the results. The idea was that political restraints are too strong, that some programs are very important to some political decision makers, and that they do not care if they work or not. They can only be held responsible if it doesn't work out, especially since most things don't work out as we would like.
00:20:20
So is that true? Have the last 15 years shown us that this is the case? There was a very interesting experiment in which the decision-makers were actually the participants. It was conducted in Brazil by a team of researchers that had attended the national mayor's conference, with 1,500 mayors present.
00:20:47
In Brazil, mayors are... They control regions that are a lot bigger than what we might imagine. So all 1,500 mayors in Brazil came together, and they took part in two experiments. The first was more of a thought experiment. They were asked if they would be willing to give up a portion of their remuneration
00:21:12
for coming to the conference in exchange for experimental data on young childhood. There were several variations on the types of results available in exchange for... their contribution. First answer: The mayors were very excited about receiving the information, and even displayed a sense of sophistication, since they preferred random experiments
00:21:36
to the various... data. They preferred large samples over much smaller ones, and they preferred places that were close to Brazil. So they displayed a certain sophistication in their request... for information. That's the first thing, and a very good result. In another experiment, during a special session, the mayors were told...
00:22:01
Their teams gave them presentations on... experiments to help them raise more money from their constituents by sending reminder letters. Almost two years later they met again with the mayors who had taken part in the session and others who had not, and those who had participated had made use of the results very efficiently, so they appeared... to be ready to...
00:22:29
to learn... To be interested in them and follow up on them. I generally try to avoid speaking about my own personal experience, but that study... really echoed... one that I worked on... during the COVID crisis. At the start of the crisis, I got a call from Mr. Shegun Bakari,
00:22:58
an adviser to the president of Togo, whom I didn't know, and he told me, "Listen... "I need your advice. We want to start a support program... "for the people living in areas... that we have to quarantine. "And my idea would be to distribute food packages. "What do you think?" I told him that I didn't think it was a good idea,
00:23:27
because we now have data from about a dozen studies telling us that, for the same amount of money, you could actually give money directly to families, rather than giving them food packages. It's much easier, there are far fewer logistical issues. Logistical issues during COVID could become very problematic. And the overall effect on nutrition turns out to be the same. He listened very carefully and then called back a few days later:
00:23:54
"OK, we're going to do what you said. "So... how do we do it?" The second thing was, "Do you have..." As we spoke, I learned that they had already set up an electronic phone payment system. I suggested that we use that technique, and they pieced together their program in three weeks. They convinced telephone companies all to sign up for the platform,
00:24:22
and they launched their program. Then... That was the first thing he told me, actually. As I spoke with him, I found out the data existed, so I was able to give him a clear answer to his question... and he ended up using it. On top of that there was already a lot of data indicating that money was a good idea and that electronic wallets worked great.
00:24:48
But there was one question... I couldn't answer for him, or at least not give him a clear-cut answer, which was identifying the best way for them to target the populations that needed it the most. Keep in mind that Togo did not have any sort of registry for... poor people. I put him in contact with a team from Berkley and one from IPA,
00:25:15
who helped him to develop a targeting system, first through satellite images, to identify the poorest regions, and then by using telephone data, telephone metadata, to find out if mobile phones are used a lot, and so on, as a way to predict the local poverty level. And then what they did while implementing the program
00:25:41
was that they compared their system with a more traditional one, which was based on occupation from a previous census, to find out if their system was better, worse, and so on. So from both perspectives the experiment... reassured me, I would say, with regard to these people I didn't know at all. It gives us the possibility of listening... to experimental results
00:26:07
and then setting up a plan even during a tense moment, such as the COVID crisis. THE INFLUENCE OF J-PAL SINCE 2009 In sum, we can definitely say... We can measure the influence of experiments simply by looking at... the number of people affected in an indirect way by the experimental results.
00:26:32
And what we do at J-PAL is that we try to follow what happens to experiments that have grown in scope, in order to see how many people are affected... not directly by us, but indirectly, by the government. And today we're at-- It was 23 million in 2008, and now we've reached 540 million as of today. For... Another calculation was done by Michael Kremer,
00:27:02
who calculated... the social returns of the Development Innovation Ventures Fund, a fund created by the American Agency for International Development. The portfolio of the 40 first evaluations that they conducted cost them 19 million dollars. After that they calculated social return in extended evaluations,
00:27:28
and they... discovered a profit of... more than 86 million dollars, simply by focusing on a few evaluations. So, if you take those 86 million, earned from just four projects, and then look at the entire cost of the portfolio, we see that the level of social returns is at 77%, surpassing the overall performance of markets by a lot.
00:27:56
To give you an example of how this could be, I'd like to go back to an example that I had cited during my inaugural lecture in 2009. I'd mentioned the work done by Pascaline Dupas, who shone the light on the debate around the matter of: Should people have to pay for the mosquito nets that protect them from malaria or should they receive them for free? On the one hand, if you hand them out for free, they'll be distributed more widely,
00:28:25
but, on the other, there was a fear that people would incorrectly use the object they'd received for free, since they had no concept of its value. In order to find out she varied the cost of mosquito nets between zero and a subsidy of 50% of the overall price. Her results were extremely clear. First off, when people had to buy a mosquito net, fewer people would do so,
00:28:52
but if people actually do have a mosquito net at home they use them just as properly if they'd received them for free as if they'd paid for one. Hence, the very clear effects of studies like this, which was a stronger push for freely accessible mosquito nets. I wanted to talk. For those of you who missed the first part, what I wanted to discuss with you here today
00:29:18
was what happened after that study. Dupas' articles were ready and debated, copied and so on, and they did end up making a very convincing argument. As such, widespread distribution of free nets became a shared goal in the methods used to fight against malaria. The Roll Back Malaria partnership, in particular, emerged in April 2008, with 500 public, private, and international partners.
00:29:43
It launched a campaign for the massive distribution of nets, with the goal being universal mosquito net coverage. Universal coverage wasn't exactly obtained, but, in 2019, directly before the COVID crisis, WHO estimated that 68% of all African households had at least one treated mosquito net compared to 5% in 2000. So there was this... massive distribution of nets. An article was then published in Nature,
00:30:10
comparing very small, defined regions, in terms of mosquito net coverage, and estimated that 450 million cases of mosquito ne-- of... malaria had been avoided thanks to the distribution of those mosquito nets. So... Bill Easterly, who had advocated making people pay for nets, finally admitted his defeat on this one point only,
00:30:38
in a display of magnanimity: He illustrated... his tweet with this graph showing the drop in malaria cases throughout the regions. Here you have a child sleeping beneath a mosquito net. That being said this example is a bit misleading, because it's too simple, too perfect, too direct...
00:31:18
It goes directly from the results to convincing many people, with a few replications in between, but in fact not so many. So this massive impact from just a few studies. That's really the ideal. You conduct a well-controlled experience, you analyze the results, you write up a nice summary, you send it all over the place, and a large-scale adoption follows. That was the plan that I had in mind back in... 2009,
00:31:50
which I thought would take the world by storm. But that plan was... was also criticized, for example, with comments about how a well-controlled experiment... cannot be replicated exactly the same way. Everything changes on a larger scale. The men and women who conduct subsequent replications might not display the same energy as the initial people who...
00:32:19
really felt invested in the project. There might be issues of balance: If a few people go to school, that works, but if everyone goes to school, in the end, competency values begin to fall apart. A whole slew of questions we could... Of critiques you might have if this plan were indeed the model. But, ultimately, this is not at all how influence is passed on,
00:32:45
or how... This... A simplistic model like this isn't what happens in reality. So those 540 million lives that were impacted weren't generally affected by following such a simple schema. I'm going to talk about three ways... we can use... to go from the research stage
00:33:11
to large-scale implementation. The first I call the "pointillist painter-economist," followed by the matter of... experimentation on a larger scale, and then, the "plumber-economist." Let's start with what I call the "pointillist economist." For those of you who might not know what pointillist painting is, this is a painting by Georges Seurat. He painted it using dots. The painting was made dot by dot, each of them... just a single dot,
00:33:42
but when you see all of the dots you start to see the image better, or the subject, so, seeing the Île de la Jatte here and all the people having picnics. I think of randomized experiments in the same way: Each of them provides us with a new dot. When they are in the same area we get a good understanding of an area
00:34:09
or of a problem. For example: Why do children have difficulty learning in school? A whole series of experiments can be carried out on education, and over time we can... We're going to get all these dots, different dots, spread out, mixed in with other colors, and... in doing so, we can get a better understanding of a set of problems and the various ways in which they are expressed.
00:34:36
An example of how to multiply dots to understand a problem better in doing so in a small area of that painting would be to look at microcredit. Initially praised in 2000, microcredit quickly went from being a miracle solution to a danger for the poor. Within a time period of perhaps just two years, it went from Muhammad Yunus receiving the Nobel Prize
00:35:00
to documentaries that now accuse microloans of pushing people towards suicide... without there actually being any data to clarify one position or the other. So, what is the actual situation? Well... We can't really know after just one experiment. The contexts are extremely different, as are the programs themselves,
00:35:28
so the initial evaluations that were made... had very little impact, primarily because the results were quite mixed and also because it was too easy to say, "Well, the results are mixed, because of..." One of the very first experiments I took part in was in India. In the state of Andhra Pradesh, the cradle of microcredit, so it seemed natural. There were many institutions,
00:35:53
so it was saturated. RESULTS ON SEVEN EXPERIMENTS But the experiments began to multiply over time, and after a while there were seven of them... in different regions, conducted by different teams following different economic models. Those seven experiments provided us with some interesting results. The first was that, on average... in every single country,
00:36:20
microcredit is disappointing. It does not lead to any sort of increase in income, nor to increased consumption within villages. So on average, for the average person, microcredit is by no means a way to rise from poverty. However, another... lesson is that, for people who had already shown some entrepreneurship, in that they already had a business beforehand,
00:36:45
we can see some stronger results. These results are all shown on the same graph, since they were published at the same time in just one magazine, and were examined by Rachael Meager, a young statistician and economist, who began to wonder: Can these studies show us, if just one point on the graph... one dot in the whole painting can tell us a lot, if anything, about the dot next to it?
00:37:14
Does a study... Do these studies want to give us different messages, or are all of these studies actually teaching us the same lesson? In this case, regarding microcredit, we can observe very strong uniformity among the results. So this would be a solid and very robust conclusion that could certainly be applied to the next study: Microcredit is a disappointing option for most people.
00:37:42
So, following these seven experiments and the joint analysis, the microcredit movement was sort of forced, in a way, to look at these lessons, and the debate then went from "Does it work or not? Is it good or bad?" to "How can microcredit be turned into a... more useful tool?" So either by finding an average to focus on, like those who had displayed some entrepreneurship,
00:38:07
if that's what we want, or by offering customers other financial services they might need more than a loan, such as a savings account, insurance, and so on. So after a while, once you have enough information to clearly define at least one section of the overall picture, you can start to spread out more, maybe use a lighter green to present... Even for the microcredit movement,
00:38:34
could go from "all or nothing" to a more pragmatic approach on how to improve on products. Secondly, economists... The first series of experiments that I conducted, as well as those I presented in 2009, tended to be relatively smaller experiments. "Relatively," because they were very big in comparison to clinical trials for medications,
00:39:01
but they only concerned a few hundred schools and villages. But when the experiment's scope was widened many things did not go the same way they had in the original experiment. Firstly, there's no... When you go from a pilot experiment to a policy-level experiment, there isn't complete adhesion to the original program. There could be equilibrium effects, as I mentioned earlier, or... the political economy could bring down the entire program
00:39:29
on a wider scale or actually cause it to take on a dimension it didn't have before. We can't answer these questions with abstract thinking. If our question is, "How would this program, "which appears promising on a smaller scale, "perform if it were expanded to the scale of France?", then we need to widen the experiment to include all of France. That all requires a scope-level evaluation,
00:39:55
which can sometimes lead to different results. I don't have much time, so I won't list examples... of these techniques in detail, but one experiment that I conducted with... Bruno Crépon and others, and the French unemployment agency revealed, for example, that the benefits provided to the unemployed in... extreme situations...
00:40:22
To young people who have been unemployed for a long time appear to be beneficial to each of those young people, when compared to young people who did not receive any aid, but, if you compare everything, and widen the scope of the experiment, do we then rob Pierre to pay Paul? Or is everyone making progress? Unfortunately, what this experiment showed us is that it's actually the second... the first outcome,
00:40:49
meaning that the gains observed when comparing... a young person receiving aid and another not receiving aid are mostly due to... to the effects of passing a job from one person to another... especially during a crisis. That's just one example, and there are several others. Ben Olken and Rema Hanna work very closely with the Indonesian government
00:41:14
to evaluate many of their projects on a wider scope before they even happen. We've worked with various state governments... in India, and so on. This type of large-scale evaluation that we're seeing now within the context of the proper functioning of administration has taught us to take all of the details more seriously, because it has shown us that, many times, the differences in success
00:41:44
were not necessarily due to the enthusiasm of teams, really, but to the matter of: Was everything properly put into place for the program to have an outcome? In general, economists, and not only those who conduct experiments, have a greater voice in public policies. They're asked for help when setting up public auctions, when creating systems to assign students
00:42:14
to colleges or universities... So... if we want to be useful in our profession... when it comes to these types of problems, then we have to be ready to look closely at these details, which are actually crucial, but often overlooked. And economists tend to neglect them, almost through a sort... of professional deformation.
00:42:41
We're accustomed to overlooking details, since we learn to create models that remove the details and focus solely on reasoning, which leads us to form a habit of neglecting details. And they are often neglected by political decision-makers, too, who, in the end, are also... who often succumb to the temptation to want to change people by using models that they... create themselves,
00:43:10
as James Scott stipulated in Seeing Like a State. So, what we need to do is... Something that I hadn't realized a few years ago, and eventually discovered over time, is that... approaching... political decision-makers on all levels, whether that means a bureaucrat, a mayor, or a minister, with a very open attitude, and detailing each of the components
00:43:43
of a program or a policy is alone very useful, no matter the results of an annual evaluation. This means taking on the attitude of a plumber. So, we can't know the parameters of the problem beforehand and can only do our best by adjusting things in the process. Nothing like a scientist who... knows the limits of a problem, or an engineer who can control their environment,
00:44:13
at least within a laboratory. A plumber-economist or a plumber-decision-maker starts... with a model or on a hunch... but can keep their minds open to trying out something else if their hunch doesn't work out. Since we're talking about plumbing I of course chose a plumbing example.
00:44:40
This program was... I led this project, along with Véolia, in Morocco. The government of Morocco wanted to provide all citizens, even the poorest, with running water. So, they created subsidized, social connections... and asked everyone involved... all the companies that wanted a concession to use the sewage
00:45:07
and drinking water networks in a large city to plan on also including these subsidized connections. In Tangier, all the plumbing work was completed by Véolia. It involved... digging trenches, placing tubes, and so on, so lots of work involved. Financial plumbing was also a factor, since there was a loan system in place for each beneficiary.
00:45:34
Everything was in place, the house had been built, so to speak, the pipes had been installed... but there was very little demand. When we met with Véolia, their main question was "Why? "Why is demand so low?" And by visiting households we eventually came to see that many people had simply been scared off by a complicated process. They had to go to their city hall several times in person,
00:46:02
with all their documents in order, to prove that they were indeed the owners or renters of a household that could receive a social water connection. So, what we suggested to Véolia and the government was to take over the last stage, by going to visit... potential beneficiaries in person, to ask if they were interested, and if so by taking photos of the documents in question, which would've been copied once they'd gone to their city hall.
00:46:27
Everyone agreed, so we began the experiment. We used... a randomized approach. In the villages... In the control households... the demand remained at 10%, but rose to 70% for households where we intervened. So after conducting the experiment for about six months we were able to see the impact of suddenly having running water at home, and the people in... some households were absolutely thrilled
00:46:56
to have water. There was a lot less daily tension, an improvement in mental health, and so on... so that we were then able to return to the control households and extend that simple procedure to all samples. So this goes to show you that, even though everything was in place for a project that would have been very beneficial, even just that one small, last detail, which had been overlooked,
00:47:23
had managed to drastically reduce the demand for water. Since 2009, this type of experiment has become widespread. We've gone from... small, tightly-controlled experiments to experiments on a massive scale, in conjunction with governments, on matters as varied as the flow of finances, the organization of environmental... audits,
00:47:51
information technology, respect for user rights, and so on. When all the plumbing... is in place, the benefit is also... These projects are always done with tightly-controlled criteria, in conjunction with an administration that wants it to be successful, since their desire is to learn how to implement better what they had intended to do in any case, beforehand, so expanding the project's scope happens a lot faster.
00:48:19
There's no need to convince anyone, since they're already convinced. What was the impact, then, of these experiments, beyond just economic policy itself? There has been a rise in empirical economy experiments. The database at the American Economic Association now contains 7,000 experiments in 163 countries. The field has evolved, so, as I said, more large-scale experiments
00:48:51
primarily with governments, sometimes without... To give you an example, I took part in experiments during the COVID crisis related to information that... directly concerned several tens of millions of people. At the other extreme there are very controlled types of experiments, in which the details are very crucial, in order to link field experiments
00:49:18
with a theoretical test. So there are the advantages of laboratory experiments, as well as those of field experiments. That influence also goes beyond... empirical experimentation. I think we've sort of gone from... The fact that experimentation exists as an actually feasible objective, and not just as a dream, has sort of forced the rest of the economy
00:49:47
to pay more attention to the issue of identification, and to what it means only to make causal arguments instead of a correlated arguments. Even in modern macroeconomics... experiments are being conducted, for example, on the way that people react to the news... of inflation, the way people understand how prices are formed, and so on.
00:50:19
In regards to the theoretical economy... the multiplication of... Or, going back to Seurat's painting, it's just one, but an accumulation of experiments, that finally end up obliterating a few fundamental hunches that many of our models and institutions are based upon. An example of that would be the effect of income,
00:50:44
a rather... fundamental idea in economics: If you are sufficiently wealthy you don't need to work as much, so if social benefits are also sufficiently generous people work less. But, empirically speaking, we just don't get that result. Ultimately, the accumulation of experiments ends up being reflected in economic policy, to a certain extent,
00:51:12
as well as in modeling. There's still quite a bit of work to be done, like on this topic in particular, it seems. Following a study at Yale University, it was shown that people who had received benefits... financial assistance during COVID, continued to work if they found a job. The Wall Street Journal titled its article
00:51:40
"Economists vs. Common Sense". By common sense, they meant, "if you pay people not to work, "fewer will work. "Except at Yale, it seems." So there's still some progress to be made here. HAVE YOU SPENT THE LAST 15 YEARS LIVING BENEATH A ROCK? Some of you might now be asking yourselves if I've spent the last 15 years living beneath a rock, with my blissful enthusiasm
00:52:06
about the way things are happening. Just within the last two years we've had COVID-19, but over the last 15, there's been a rise in inequality, the breakdown of democracy, watching as robots... begin to replace us, as well as climate change, of course. I have not been living under a rock for the last 15 years, since the title of our book in 2020 was about these "hard times" that we're living through.
00:52:35
That said, those "hard times" are only hard for rich people. Those hard times... are what you can see here, in this "elephant curve" chart, which was copied by... Thomas Piketty and his group, showing us that... from now until-- Since 1980, the wealthiest 1% of the population captured 27% of overall growth...
00:53:06
to the detriment of the 90%... to the detriment of 90% of the people in the United States and Western Europe. But the other 50% of the poorest captured 12% of global growth, so the world's poorest people did see a rise... in their standard of living. So... we can definitely contrast my enthusiasm here
00:53:34
with the reasons to feel despair shown in this graph. That said, there is still one problem that isn't only for the rich, and actually ties rich countries to poor countries, and that... is climate change. Here you can see some of the illustrations... in the books you... heard about earlier on, published by Olivier.
00:54:01
The children are very hot. And there's no more water. An unbalanced climate could change everything. Climate change could... very quickly, over the next ten or 20 years, change everything and cause us to slip backwards and lose much of the progress that I've spoken about tonight. And for two reasons... and in poor countries, in particular, of course.
00:54:31
Poor countries have the misfortune to be located in places where it's already hot. However, the effect of heat... isn't linear. The body cannot handle heat beyond... more or less 35 degrees C. The same goes for cultures. Beyond 35 degrees C, something happens that is obviously much worse
00:54:59
than at 20 degrees C. So we're hearing about global warming, by an additional two to three degrees. But two or three degrees is not at all the same thing in Sweden where a summer day is something like 25 degrees C, as in India, which already has many hot days in a year. This map shows you the number of days from now until 2100 with temperatures over 35 degrees C.
00:55:24
Darker red means more days. As you can see, the Sahel, of course... the north... of India and all of Pakistan, and so on, will be places where temperatures get hotter and hotter. So, if we... Given the models that we have of global warming, places that are already very hot as of today, many of which are very poor, are going to be hit harder. The second problem: the impact of the same increase in temperatures,
00:55:53
the impact of the same temperature, is more lethal and more serious in a poor country, since poor countries have... fewer ways of protecting themselves. On this map, you can see Pakistan and... Saudi Arabia. Both countries are expected to have about the same temperatures. But Pakistan is much poorer than Saudi Arabia, a country that... can provide...
00:56:21
People work outside less and there's air conditioning, so the mortality rate that we can predict for the future by looking back at what's happened in the past is expected to be much higher, so the mortality will rise faster in Pakistan than in Saudi Arabia, despite the rise in temperatures being the same. In general, the impact on... mortality rates
00:56:45
from hot days... This shows... the temperatures-- the impact that temperatures have had... on mortality rates. It's higher in colder countries, since they aren't used to the heat, and in poor countries, since they have fewer means of protecting themselves, since they work outside, there's no air conditioning, etc. So, when you combine those two elements,
00:57:12
the fact that it will be hotter in poor countries and that poor countries suffer more from the same rises in temperatures, you get a result showing that mortality... the rise in mortality because of the climate, will be... One more day a year above 35 degrees C will be most common in Africa, so when you factor in both elements together...
00:57:40
mortality rates because of the climate will be much higher in the poorest countries, and in Africa, in particular, than anywhere else. But in Sweden global warming will actually save lives. It will be less cold in winter, and fewer people will die from cold. But in Africa, that obviously will not be the case. This study by Michael Greenstone and others in Chicago concluded that by 2100
00:58:05
73 million people per year could die every year from the effects of global warming. That's a lot - more than have died from AIDS and tuberculosis combined. That's the first problem we face. The second is that the reason the effects will be stronger in poor countries is mostly because of the world's rich countries, since the world's wealthiest citizens both consume more and produce more emissions.
00:58:33
This work was done by Lucas Chancel... who shows quite succinctly how... the 10% of the world's wealthiest citizens emit, or are responsible, for about 48% of all global emissions through their consumer habits... and how the 50% of the poorest are responsible for 10% of emissions. That is called the 10-50 rule, and it works both ways.
00:59:01
Since the world's richest-- Since the wealthiest citizens in the world are responsible for more emissions through their behaviors, we're also talking about the people who live in those rich countries. So in North America for example, one person consumes the equivalent of 20 tons of CO2 per year,
00:59:28
whereas one person in Africa consumes 1.6 tons. So there's a very critical problem affecting poor countries that is being caused by wealthy countries. This is an unprecedented problem, in the sense that all the problems we faced before, like infant death, could be solved through actions in poor countries, but the climate crisis cannot be solved in Africa, since the people in Africa are not causing the emissions.
00:59:56
It's a very thorny political issue, since the Western world made such a... marvelous display, during the COVID-19 crisis, of how little solidarity there is around financial aid, as well as around vaccines, and that, under pressure, we show no solidarity... no steps towards solidarity with poor countries. So poor countries are now in an impossible situation, since the problem already is and will be theirs,
01:00:24
but they didn't create it. And the paradox is that becoming rich as fast as possible is now the best adaptation strategy, since you know nobody is coming to help you, so why not get rich too? That was India's justification... towards... the end of the COP 27 conference... when it suddenly abandoned, at the very last minute, its plan to... gradually do away with coal,
01:00:53
by saying, "Well, we need to burn coal for now, so..." What can be done? Well, define margins for action in order to enter into binding agreements. At the COP 28 an initial fund was proposed for losses and damages, but there is unfortunately no specific mechanism for contribution. So the fear is that the same thing that is... happening
01:01:20
with the COVAX initiative, when the 100 billion euros that had been promised over and over... were never actually paid out. The United States refuses to agree to pay reparations, because it feels, if it allows one reparation to go through, then... it will be facing infinite criminal responsibility. But we don't really have to consider this idea of reparations.
01:01:46
We can just look at today. That's not the past. It is still the case that the citizens of rich countries emit more, and the wealthiest people among them in particular. We could also implement a very gradual international tax on very high incomes and on multi-nationals, as Thomas Piketty proposed. That idea does not appear to be that far-fetched, because the responsibility in question... isn't infinite
01:02:14
for the projections. That is associated to the... to the idea of a carbon tax, but not directly to emissions, and is more indirect, since the wealthy emit more. That ties into carbon consumption. Then the idea of a fund that was proposed, contributing to a fund meant solely for the poor... to be distributed as donations, not loans, for damages,
01:02:41
as well as for adaptation and mitigation in poor countries. To me that is an essential need, because nothing like it has been done as of yet. So does the climate emergency mean we should abandon the concept of moderation? Should we just forget it all... except for political activism? I don't think so. Perhaps because of my temperament... There are still millions of poor people in the world today,
01:03:14
whose lives are endangered by many different factors, not just global warming. Ethics and our instincts force us to be concerned about them, in the here and now, and not just as abstract people in the future. That's the first reason. That's why I will still focus on topics like education, nutrition, and health, but today, even though I know it's important to think of the climate too. Incidentally, my second reason
01:03:41
for which I do not think we should give up on moderation: The climate ecosystem now somewhat resembles the fight against poverty 15 years ago. There's this desire... for win-win solutions, for some sort of miracle, which can simultaneously put us on the path towards growth that is even better than its predecessor but respects the planet. That is possible, but not a given.
01:04:09
There is also this unrestrained trust in technology and the market, as we saw at the COP 27... which, out in the field, are often the source of rude awakenings. So, in order to find answers for the climate, and to know what's really going to work, we all need the same sort of patient, evaluative work, very time-consuming work, that was done in the fight against poverty.
01:04:37
If that fund existed today and were backed by contributions, I don't think anyone would be sure how to use it the best way possible today. So there's definitely a need, a need for evaluation, for experimentation, for plumbing, and it's always there. We have to continue tending our gardens, choose a specific topic... even if it seems modest, and sort of clear the way around us
01:05:04
to do the best we can. We might make significant progress, and we might not, but at least... we'll have tried. I'll end with this image of tending a garden... with a quote from Keynes at the top of the slide: He could have said "plumbers" or even "gardeners"... Thank you! Giuseppe Ilisco Subtitling: Hiventy by TransPerfect
End of transcript