Well, let me welcome all of you here. My name is Bill Colglazier, the Science Technology Advisor to the Secretary here at the State Department. It is my great pleasure to initiate the first lecture by the current crop of Jefferson Science Fellows here at the Department. For those of you who are not familiar with the program, it was created in 2003. It is for distinguished scientists, engineers and scientists of all types to come and work in the State Department for a year and then to be available as consultants for several years after that, and I guess this is the ninth class if I've gotten that correctly. The thirteen this year I think, the three working at USAID and 10 working at the State Department. Scientists who come in from university, it takes a while to adjust to this new culture of the State Department but they learn pretty fast and they figure out how they can contribute and they all have a terrific experience learning both about diplomacy here as well as trying to contribute from their scientific knowledge.
And actually to introduce our first Jefferson Fellow I am going to introduce the introducer, who's Jonathan Shrier. He is the Special Representative in the State Department for Global Food Security. He is a distinguished Senior Foreign Service Officer. In his current job he is responsible for coordinating all aspects of U.S. diplomacy related to food security and nutrition. He is also the Deputy Coordinator for Diplomacy for Feed the Future. He came to this recent assignment from the Secretary's Policy Planning Staff. He has also served the Department of Energy, the Principal Deputy Secretary and Acting Assistant Secretary for Policy and International Affairs. At DOE he also served both in the National Security Council and the National Economic Council. As a Foreign Service Officer he has had a tour in Beijing, he also worked through the Under Secretary for Economic Business and Agricultural Affairs where he handled international trade and investment issues. He has degrees from the National Defense University, the University of London, and undergraduate at Dartmouth and he also speaks Mandarin Chinese as well as Arabic. So we are very glad – and our speaker today is actually in his office, we are very glad to have Jonathan do the introduction.
Thanks, Bill, for that kind introduction, and do people want to do this in Arabic or?
We will stick with English for today. So, thanks for coming to here Clyde's remarks on “Insuring the Subsistence Farmer.” I want to just say a quick word of background on the U.S. Global Food Security effort which is primarily centered around a presidential initiative launched by President Obama called Feed the Future. And this grew up out of a recognition that there had been this period of incredible progress in global agriculture and improvements in food security, that is, in reducing hunger, during the 60's and 70's, and that story was really largely one that was based in science and technology. The story centers around Dr. Norman Borlaug, plant scientist, who is responsible for introducing major innovations in agriculture that include dwarf wheat varieties and improved corn varieties and techniques of preventing plant diseases that led to major increases in agricultural productivity in Mexico, in South Asia, as India took on board many of these innovations. It also involved introduction of new fertilizer techniques and so on, that led to major reductions in hunger, and major increases in the availability of food around the world. And there was this sentiment that the job had sort of been done and there were other priorities to attend to, and the result was that the global effort to invest in spreading these scientific and technological innovations further sort of fell off of the agenda. And there was a decades-long decline in spending in this arena.
With the food price spikes that occurred in 2008 which threw tens of millions of people back into poverty, adding to the world's hunger burden and poverty burden, there was a recognition by world leaders that we had to get back to these basics. We had to get back to insuring that the world's people had enough to eat and enough of the right foods to eat, that they have sound nutrition. And so the effort that has been launched both in the U.S. through Feed the Future, and globally through the Lockwell Food Security Initiative has devoted a lot of attention to the economics of improved agricultural productivity and improved earnings for rural households and certainly that dismal science is one key part of the solution. Other parts draw from other scientific disciplines, including further advances in agronomy, plant sciences, animal husbandry and so on, and certainly the genetic engineering revolution has been a key part of the recent improvements, further improvements in productivity.
And of course all of these things depend on numbers, and this turns us to Clyde's area of specialty which is mathematics and statistics, and the statistics, by the way, also comes into play in our work in food security because part of the new spirit of work on food security has been a dedication to achieving real world results, and that means tracking our progress through effective monitoring and evaluation systems, and again this requires thinking through what the right techniques are, and what the right analytical procedures are for looking at those results. In our work we also try to take a comprehensive approach and now I am getting to the real connection to Clyde's lecture today. We've wanted to look at food security more comprehensively than the world may have done in the period of the green revolution of the 60's and 70's where really it really was very much about what farmers do in their fields and by extension what livestock herders do and so on with their animals, or fisherfolk do with fisheries. But we really need to look at complete food systems and complete agricultural systems and that means, among other things, thinking about risks that those systems face. And when you have risks the natural inclination of anyone who is intent on reducing those risks is to turn to insurance to try to address those risks, to mitigate those risks, to manage those risks and so that is what we will be hearing about today form Clyde, which is the topic of insurance, especially for small holder farmers.
This is a topic that was seen as important enough to be a component of the latest bit of progress in food security on the U.S. and Group of Eight countries' agenda which is the new alliance for food security and nutrition launched by President Obama at the time of the G-8 summit, the Camp David summit which was held in May of this year. And the new alliance includes a commitment by G-8 members working with African governments and institutions and international organizations to launch a global action network on risk insurance, and to help key African countries that we are working with to conduct risk assessments and develop risk mitigation strategies, national strategies for addressing these risks. And so the work that Clyde is going to be talking about is directly relevant to presidential and prime ministerial-level agendas and so that's why it's important.
To just give you a little bit of background about Clyde, he -- we are very pleased to have him working in our office. He is the P.W. Horn Professor of Mathematics and Statistics at Texas Tech University. He received his PhD from the University of Wyoming in 1971 and Clyde's research interests include control theory and the development and analysis of mathematical and statistical models in agriculture, the environment, and medicine. He has collaborated with engineers and scientists in a number of areas, including aeronautics, bio-engineering, economics, plant science, soil physics, epidemiology, and chemical engineering on a variety of scientific topics. He's a Fellow of the Institute of Electrical and Electronic Engineers, a Fellow of the American Statistics Association, and an elected member of the International Statistics Institute. In November of 2001, he received an honorary doctorate for his work in engineering from the Royal Institute of Technology in Stockholm, Sweden. And before I turn the floor over to Clyde to talk to us about insuring the subsistence farmer, I just want to read the one line that really stood out to me in the abstract of his presentation which is to say that, “Very elementary recursive models of assets will be shown, but not belabored.” And so I will be interested to hear from folks afterwards whether they think he has accomplished that. With that, Clyde, let me turn it over to you.
Thank you both for the very nice introductions. Yes, I started here in August and have learned a lot and still have an awful lot to learn about getting things done here, you know, when it takes two days to write a letter saying that somebody's not coming, you know that you've got a lot to learn. But anyway, I have been working on this, I started because I went to a talk and found it interesting and so I began working on it. The standard disclaimer that these are my ideas and are not necessarily those of the State Department, and I would also like to thank Xavier Gine for showing me through the literature of indexed insurance, and of course Robert Martin for lots and lots of hours talking about the economics of insurance and economics in general. This sort of thing is, you know, you don't do by yourself, you do it in conjunction with lots of other people and that is certainly what has happened with this.
So who is the subsistence farmer? That is the key question, I guess. You know, we used to have subsistence farmers in the United States quite regularly, I think you could call my family subsistence farming in Arkansas, we certainly didn't make enough money on the farm to live there. So the farm it could be a traditional family, husband, children, grandparents living in a village or living in conjunction with a clan. Often now it turns out it's a wife with children and grandparents, but the husband has gone to some urban area to try to make some money. Rarely, it's a husband with children and grandparents, see that mentioned very, very seldom in the literature and now too often it's grandparents and children. This is partly because, largely because of the HIV/AIDS problems in many parts of Africa and this is something that comes into this and needs to be studied much, much more. What do they farm, what's the farm like? Well, they are small and that's the key point. There in 10's of acres at most. What they farm depends on what country they are actually in, what the climate is, what the environment is, but at the basic level they grow some food crops. You would hope that they would grow enough food crops to provide food for themselves, but they seldom do. According to one expert, in the lower levels of the subsistence farmers the farmers are providing about enough food for two months of support. So that we'll come back to. They grow corn, maize as it's called in the literature, ground nuts, which I gather are peanuts, although there is a ground nut in the United States that's not a peanut, but I think what they are referring to in Africa is actually peanuts, some vegetables and then perhaps some cash crops. One of the things that is being encouraged is cash crops, cotton especially. They may have a draft animal if they are wealthy. Probably some goats a few cattle, perhaps, depending on where they are at, chickens and perhaps a few other animals. But basically they don't have much, that's the whole point.
The area that we are looking at is the Sahel area of Africa so it's the area directly under the Sahara Desert, the north part of it is, I would say, extremely arid and as you go farther south it becomes more tropical. In lots of ways it’s a lot like west Texas where I am from, in fact it is in some ways it is very, very similar to west Texas. The important thing for the food security is that it has been subject to a series of droughts over the last few years, in fact there were three successive droughts which we will come back to a little later. So what happens in a drought? They're barely surviving in good times, they are barely having enough food, and so the crops fail and they have to sell assets. The assets could be the cattle if they have it, the family is dependent on food aid. Men go to the urban areas to try to find work and one thing that is really important is that often the children drop out of school and try to find work. So in some sense, the children dropping out of school is like that the family is readying the children and when they go away then they are not having to be fed at home and perhaps they are actually bringing in some money. So the children can be an asset in times like this, or they can be a liability. I had thought that perhaps there was some problem with children being sold and I have been told that that is not a problem in this area of Africa, not a major problem.
So the whole point of what we are doing is what can be done to increase the resilience in the Sahel. Bio-technology, better seeds, availability of fertilizer, irrigation systems, mechanization, farming techniques, all of the things that have gone into moving the United States from basically subsistence farming in the 1920's to what we have now. A big thing is infrastructure, roads are a major problem. It's hard to get crops to the markets and it's hard to get people to the farms. Communication that is becoming better with the cellphones. Irrigation systems are basically non-existent, there in some places there are. There are storage systems that are rodent and insect-proof that are not there, post-harvest losses are huge.
And then the area that I'm interested in and for the purposes of this talk, at least, is financial systems. Microcredit, micro-insurance and the idea on this is, you know, bio-technology is great but if you don't have any money then how are you going to buy the better seeds, the fertilizer, the pesticides and et cetera? So you borrow money and right now this is one of the things in particular, USAID is working diligently on is providing loan systems for these people. Crop insurance, and this is primarily what I am looking at here, the U.S. model won't work in Africa. The U.S. model, you buy insurance on your wheat crop and a big hailstorm comes and wipes out your wheat crop and the adjustor comes out and says, “Okay, you've lost 90 percent of your crop and so we will pay at the 90 percent level.” But these farms are so small that if you had to send out an adjustor to every farm, you simply would not be able to make the insurance pay.
So something has been developed called index crop insurance and this has become quite the thing in the last few years, but it is actually not particularly new. There was a PhD dissertation at the University of Chicago in the 1940's on indexed insurance, so the topic goes back a long ways. I think I stole this from Gine's definition, “Index crop insurance uses weather observation as proxies for losses in production or quality and does not require loss verification.” So the way it works: you have weather stations and if the weather station says that you have had less than n-millimeters of rain, then it pays. If you have had more than n-millimeters of rain, then it doesn't pay. It can be a lot more complicated than that, but it can't be extremely complicated because if it is the farmers don't understand it and this has been one of the problems with the use of indexed insurance is that it's been difficult for the farmers to understand what the insurance is doing. Farmhold assets, these are all the things that make a farm possible. The right to farm the land this is something we are discovering, that I am discovering, I think a lot of people already knew it, is a big problem in Africa. Having valid tenure to the land, you know, you are not going to invest a lot of money in improving a farm if you are pretty sure that within a year somebody is going to come and say, “This farm belongs to me instead of you.” So land tenure is really important. Tools to farm the land, you know, technically I suppose a hoe would be enough, but it's not, really. You need cash and you need crops to feed the family these are part of the assets. Household items, the television, cooking pans, et cetera, these are assets and these will be sold if conditions become bad enough. The goal of insurance, and the goal what I am looking at, is to keep enough assets to keep the family on the farm and not hungry. You know, this is one of those beautiful goals and, you know, if we could just do this then everybody would be happy but that's the whole problem is having to do it. You know, I gripe at people for writing things that say basically that things will get better, if we make things better things will be better, and that's sort of what that sentence says.
So model of farmhold assets, you know I am a mathematician, I have a right to make a model although Lawrence assured me that no one would understand me if I really did, but we'll give it a try and see. Okay, assets should increase at harvest time, that would be what one would hope. You would expect that assets would decrease leading up to harvest and so we say that the value of the assets at time 't' is x(t), so just index by time and for any given farm there is almost no hope of actually predicting what x(t) is. That's beyond what we could expect to do, but when you have large groups of things, lots of things come into play, lots of statistics comes into play that says on the average for large sets you can predict. You know, I can't predict what Jonathan is going to do in the next 30 minutes, but if I look at a large crowd of people often you can predict what the crowd is going to do, and that's sort of what I am looking at here and the goal of insurance is to protect assets from risk. Or I guess I should say the goal of insurance is to protect assets in the presence of risk. So a little mathematics.
Okay, this should take you back to college algebra, which you all had. We'll build a model that can be simulated. Now remember, models are at best approximations. You can get information from a model and the analysis of a model, but you don't get answers from models. This is something I have preached to my students and anybody that listens. I'm sure everybody hates to hear this, is that mathematics never gives answers but it does give additional information and that's the whole point of mathematics. But so then we have this nasty-looking equation, and you know, we've invented all this notation so that nobody else can understand it. It's like the, you know, the ancient priests in Rome or Greece that had a language that nobody else could understand and so that's exactly what mathematicians have done. So this is not all that complicated. Basically all it says is that the .01 that says that things are increasing but slowly, the dW is just some random noise, and the part that's really important in the model is the dP, and that we'll use to represent shocks to the system, in particular droughts although it could represent civil unrest, it could represent a lot of different things. But droughts in the main. So we'll look at the pieces of it. The .01 is just a number that says in the absence of all problems the assets will increase very slowly. And that .01 says it is indeed very slowly; dW is, we have some biology-trained people this is Brownian motion, and it's simply a random effect that can affect the assets. It can make the assets bigger, it can make the assets smaller. You know, your aunt dies and leaves you some money, that's an increase. You break your arm and can't work, that's a random decrease. 'dP' is a term that models a shock to the system and in particular a drought. And then up here we have that coefficient of dP is this 1 – x(t) and that says as x(t) gets closer and closer to 1, then the effect of that shock becomes less, so the better off you are the less damage you will receive in terms of a drought. We can argue and I am sure I will have arguments about this at least going home tonight that over just what these coefficients should be and how the coefficients could be formed, but this is a model, we get some information from it but be aware we do not get answers from that model.
Okay, so then the question is what terms can we affect, and I want you to know that my wife and I went through this whole document making sure we had affect and effect used correctly each and every time, because I invariably get them wrong. The .01 term, that was the coefficient that said how things would increase, that we can change. That we can change by giving better seeds, better cultivation methods, irrigation systems, better roads, less post-harvest loss, lots of things. There are lots of things we can do to that term that will help the assets to increase. It's hard to change the dW term, the random effects, we can't do much about that but we can change and we can affect the coefficient of the shock and this is the important thing for the talk because the coefficient of the shock then what would we like to do. We would like for them to come out of the drought, we don't want them to come out of the drought better than they went in, that's not realistic, but you would like for them to come out with assets left and that we can work on.
We did simulations – actually, I say we did simulations, I didn't do simulations. I have a graduate student that loves to do simulations, at least he says he does. This was with three shocks. This is pretty bad. We looked at the model for three consecutive droughts and the average asset we had a list, we had the assets banded between zero and one and that is just for convenience, could have been between zero and 100, or whatever, it doesn't matter. And so we started with the average being 0.5 and the end of this one period of time which roughly represents 10 years, the assets have gone down to an average of about .2. So this is pretty disastrous. We started it off with 10 percent below .1 and 10 percent above .9. That is some really poor farmers, some a lot better. For this particular run of simulations none of the lows survived. They all moved out of farming and moved into the city and became wards of the state, I guess. So that was actually with 42,000, almost 43,000 runs. The beauty of the computer is that you can do a lot of runs and without doing any work at all. You know, I called him about 11:00 one morning and I said I need to have these put in the talk and he said, “Okay, would 1:00 be soon enough?” And so he got back within two hours. Okay, now, here we have reduced the value of the shock and we have also reduced it down to one shock and we have reduced it to .1 from .3 and then we made a small increase on the coefficient. We doubled it from .01 to .02 and we see that we get something that appears much nicer. That still many of the lows went away and that's inevitable, I don't know how you can keep those that start off really low, just a random effect is going to get rid of most of them. But then, one shock, the shock had been reduced from .1 to .3 and I increased the coefficient 20-fold, so say they are going to be 20-fold more apt to keep their assets and it didn't make that much difference. That I find very interesting and simulation need to be refined, lots of things need to be done more to the simulation.
But I think this is starting to tell us something really pretty interesting, is that we can increase their level of farming during the good times and it will have some effect, but if we really want to help them we've got to do something about what happens during the droughts. And that's, I think, something very important to draw from this. So some thoughts: the largest effect is in reducing the magnitude of the shock, so this suggests that insurance of some sort is a necessity. Crop or asset insurance, indexed insurance is certainly a candidate, but there is another form of insurance which we have used in the past and would like to get away from but direct food aid is also a form of insurance during the droughts. It’s hard for me to really get my thoughts around how this would actually, should work. So indexed insurance, let's go back and look at that. Typically, the index would be given in this form; it pays full value if the rainfall is less than m millimeters, pays nothing if the rainfall is more than some level. Pays on a sliding scale if the rainfall is between those two numbers, then there's been other much more complex forms of indexed insurance proposed and tried. One of the problems is: how do you measure the rainfall? And typically what happens is in an area there may be three or four at most government controlled weather stations and these weather stations measure the rain quite accurately, but you know, if you are 20 kilometers from the weather station, you're not going to be very happy with the results. So there are other things that can be done; I am doing some work with a colleague in Japan on some ways to mitigate that.
The insurance should be actuarially fair. This is important and hard. Actuarially fair that simply means that on the average if I buy the insurance I will receive a payment just often enough to break even. This term on the average though is a real kicker on this. For Africa this means being able to predict rainfall over a multiple year period. This is tough, I mean we've got reasonable records that go back about 20 years and if you are looking at a 50 year event, lots of luck. And then we have this series of droughts that are associated with El Nino, I am told, and this makes things even more complicated. And then we have the other thing of climate change. I am quite willing to say the word climate change in a mixed audience. I am unwilling to discuss what might be causing climate change in any audience. I find that that's a good way to get fights started in a bar.
Okay, some assumptions. We assume that the insurance is fair. The crop if it makes will have a fixed value: C. This is really simplified because crops don't have a fixed value C it depends on how good the crop is and so it's variable but for the purposes of argument we'll say it has a fixed value. The price of the insurance policy is P and that's not exact either because often you can buy with this indexed insurance. You buy shares, and so you may buy one share which is like buying a box of Girl Scout cookies from your neighbor but not enough to support your family and one share is just bought out of politeness. If the index is not made, then the policy pays the value of the crop which again is a simplification and if rainfall exceeds the index then it pays zero, and for purposes of argument right now, we'll assume that there is no sliding scale, that it's fixed. So we want to know what the value of the insurance is to the farmer. So top line, if we buy insurance and it rains, the crop makes then farmer gets the value of the crop minus what he paid for the insurance. But it could happen that he buys insurance and it does rain and still there is no crop, the elephant gets into the field and destroys it, the locusts come, all these things that are not part of the index and in that case he doesn't get anything but he has already made a payment so he has a negative value minus P, the payment. It could happen that he buys the insurance, it doesn't rain, but the crop makes and that could happen for a variety of reasons, but one of the things that might happen is that he's 20 kilometers from the station and he had a shower that the rain gauge didn't get, or he might have been slipping out and irrigating some stuff by hand. So then he really comes out well: he gets 2C, twice the value of the crop minus the payment and on down through these. Notice that if he doesn't buy insurance he either gets the value of the crop or he gets nothing. That's his two choices.
And now, this set of probabilities, probability of rain plus crop .9, these are just made up numbers. The funny thing about them, farmers have a pretty good idea about what these probabilities were, whereas the guy that's selling the insurance may not, and so the farmers could probably actually use this to hedge a little bit. And so I've assumed that the probability of rain in a year is .9 and go through the arithmetic, but it’s just arithmetic. But what we find is that with insurance then his expected value is .99 times the value of the crop minus the payment, so not bad. If he doesn't have insurance under the same set of probabilities, his expected value is .905 times the value of the crop. So clearly it's to his advantage to go buy the insurance. And so, one would expect that all these farmers are rushing to the insurance office to buy insurance, but it turns out they're not.
So why aren't they rushing? Well, there is a lot of literature out there and sort of my favorite is some of Townsend's work, Gine, Fafchamps, Urdy, Gobbel. And then what you find out is that many of these villages have a sharing system in place and how does this affect insurance? So suppose that there are 20 families in the village and I buy insurance, and I'm the only one that buys insurance, and so I have to put out an amount of cash, of P, to begin with so if it rains and the crops make, then my net gain is C minus P, not bad. But if it doesn't rain and crops fail then I get my insurance but I have also had to pay for it, but now all my cousins come and say “It's time to share.” And so I end up with a net gain of C/20-P which may actually be negative and very apt to be negative. So if one person in the village buys the insurance then this is not a good deal for that person. Gine certainly recognizes this and Townsend has been pushing this for many years. So that was sort of with perfect sharing, less than perfect sharing, we all have less than perfect sharing we all have paid taxes and so your net gain is whatever the tax rate is. But if everybody in the village buys insurance then the indexed insurance works just fine. This suggests that instead of trying to sell insurance to the individual farmers that one should go to the headman in the village and then make the case that it would be a really good deal for the whole village to buy an insurance policy. And I am certainly not the first person that has thought of that, that thought occurs in the literature. One really sad thing, though, is that in many of the villages the sharing part has been disrupted by HIV/AIDS epidemic. When you only have grandparents and children in the village then there is a problem. There is nothing to share; I suppose is the big problem.
So indexed insurance, it's a good option for subsistence farmers; in fact, it's a great option for subsistence farmers. It's very important but we have to respect the village culture, we have to educate the farmers as to what the insurance can do for them and we have to, you know, talk to the village chief or the head man in the village and say, “You know, if you have a sharing system in place then it would be a really good idea for the village to buy the insurance and then tax the villagers for.” We have to have better estimates of rainfall, this is something that is critical and important and hard. One of the things that I have been experimenting with is giving every farmer a rain gauge, which they're going to cheat with -- I mean, I would if I was such a farmer. But you can do a lot by doing some averaging and some sort of truth and that and one can get some very good maps of rainfall with this. I'm doing this work with a colleague in Tokyo. We have to develop actuarially fair policies. This is a problem because we don't know what actuarially fair is in this situation. Some of the early policies that were let people really felt that they were priced too high for what you got out of it.
So, an important question. You know, this could apply to farmers but it could also apply to me. I am a farmer and at the time to buy seeds and insurance I have exactly S dollars and I'd like to invest some of it, but how much should I invest and how much should I keep as a back-up? With the money I'm investing, how much should I put into insurance, and how much should I put into upgrading the farm, and how much can I afford to borrow? Of course, now, if we borrow, we don't usually go to the bank for this we put it on a credit card but you know how much can I afford to put on the credit card for this? And so this is a question and this is a question that farmers have to answer, this is probably a question we can't answer for them, but this is quite important.
So that's the end of the slide show, it's not the end of the problem by any means. These problems are out there, they are very hard. USAID is doing a lot to work on these but there is still a lot that needs to be done and some of these things, you know, I have talked about, the sharing, but there is also many other reasons why farmers are not buying the insurance. But almost all can be answered by education of the farmers as to what insurance can do and then education of the people that are trying to sell the insurance. You know, don't try to sell individual insurance to a farmer in a village that shares. This is not going to work. Nobody is going to be happy with it. Okay, I will open the floor to questions.
[unintelligible] is lovely and I appreciate the presentation; however, one of the things that I am not quite clear about is how will the farmers be able to afford the insurance, when you are looking at an area that has on the average more than 80 percent of the farmers are women, they have very limited income, they are providing for grandchildren and taking care of themselves. Where do you propose they would get the funds to buy insurance when they will need every penny they have to buy food because most of those people are the hungry people?
That's the central question and, I mean, this last slide in some sense addresses that in that these decisions have to be made and I think governments will play a big role in this is perhaps insurance will have to be subsidized. You know, I mean, we can subsidize insurance, or we can subsidize food aid, and it’s not clear to me at the moment which is the most effective, but I agree with you 100 percent, you know, the cost of the insurance is critical. Now, there is some work being done on tying the insurance to some of the micro-financing and I don't feel competent to really address that question right now, but it'd certainly be something that I would love to talk with you about, Lily. Yes?
Clyde, when you were doing your research, did you come across any places where this is working, any success stories where this is actually having a positive impact? I am just wondering because we're seeing it more as a suggestion when we're looking at various organizations and what they think we should be doing, but it'd be nice if we had a pilot program to point to or some kind of history of progress with this.
The literature is confusing. Some of the literature suggests that it has been successful, but if you read the literature very closely it’s not so clear that it was really successful. You know, I see data that says “Oh, we sold so many million dollars’ worth of policies,” but then if you look at the number of policies they sold and these shares, it looks like people are buying one or two shares which is not enough to insure the property. What I would love to see would be a very rigorous economic experiment done in Africa. I wrote to probably the best experimental economist in the country and asked if he had some ideas and he has pointedly not answered me. So, perhaps he doesn't or perhaps he doesn't have time to think about it, would be my guess. But yes there needs to be -- the experiments need to be done, and you know one of the things that would be extremely valuable would be to have a small meeting of people who have tried to sell insurance and then ask, “What went wrong?” you know, so we can avoid the mistakes that have been made. And then also find out, you know, “When you did sell insurance, how did you do it?” and you know, “What was your method of convincing the farmers that it was a good deal?” I think that a small conference like that would be extremely valuable.
Arol Levee from the EU delegation to the U.S. I've come across a study that said that farmers in a region very close to the one you studied, farms in Kenya in this case, were reluctant to actually even buy fertilizer much less insurance. This was of course a tangible asset that seems on the face of it perhaps to be obvious to buy, but of course given the constraints of farmers and the pressures that they face, it is a difficult decision to make. I am wondering whether in fact there is a case for interdisciplinary research between behavioral economics and the studies that you are doing, and psychology and others in order to actually look at those trade-offs and to see how indeed you would persuade, educate, influence, incentivize the decisions that would lead to success? So that is my question.
I agree with you 100 percent that it’s not something that an agronomist with USDA is going to solve immediately. It is not certainly something a mathematician will solve in the next lifetime or two. But you need to bring a group of people together, and again, I think you know to ask the question, you know, “When you did sell insurance why did it work, why were you able to sell it? And when you didn't sell it, why couldn't you sell it, you know? What went wrong?” Also, this question of risk, you know, and farmers are reluctant in general to try new things until it's been demonstrated that it works. I have a graduate student from Zimbabwe and what she said is that, you know, if you want to convince people go to the headmaster of the village and get a demonstration plot and sell or give insurance to the headmaster and then let the village see that it has worked. And perhaps that's one approach and I'm sure there are others. But farmers in general, in correspondence with Gine, he pointed out that farmers don't like to buy the new seeds because they feel the new seeds are riskier than the traditional seeds. And again, this is a matter of education.
Great talk. Couple of questions, the first one is just on the basic assumption that we should be providing insurance. So obviously 50 percent of farmers are subsistence farmers, but do we actually want to prolong or encourage people to be subsistence farmers rather than moving to larger scale farms, get encouraging migration from rural areas to urban areas and actually moving to greater productivity that you see on those larger farms. And then the second question is about the indigenous sharing systems, whether those could be viewed either as a basis for an existing insurance mechanism, or to somehow co-op those mechanisms to build the kind of insurance mechanisms that we need.
The second question, that one's easy. This is what Townsend has been promoting for years is that the sharing system is a form of insurance and his beautiful little book, ” The Medieval Village Economy”, talks about this in medieval England and how the sharing and the allocation of plots were done in order to mitigate risk to individual farmers. So that one, yes, I think that sharing is indeed. Now the other question is harder and politically charged, you know, should we encourage these low level farmers to remain on the farm. I think the perfect example is to go back to the United States in 1930 and look what happened in Texas and in the South. There we had subsistence farming, the typical thing was 60 acres and a mule for the cotton farm and these people were not making a living, you know, barely staying fed, and then during the early part of the Depression they migrated. They migrated sort of in two directions. A lot went to California, but also a lot went to the northern cities and so you had this urbanization. And the problem in Africa is that there are really no jobs in the urban areas either. You know, I hate to say it out loud but what you really need is sweat shops you know so that there is some kind of employment for these people when they get to the cities. You say, “Oh, sweatshops are terrible!” Yes the sweatshops are indeed terrible but at least if you are working, you know, even for a minimal thing, you probably are going to have enough to buy a little bit of food and have a place to sleep. Whereas if you are on the farm, probably working more hours you are not guaranteed of either. A good example of that is what happened in New England around 1800, the movement of the farm women from the farms to the cities. So, I feel that it ultimately, that some of the subsistence farmers will not survive. I just don't see how we can manage that but I am not going to say, in public at least, that all subsistence farmers should move to the cities and we should go back to the big farms. It sounds suspiciously like 1880 to 1970 in Africa.
Chris Fall, Office of Naval Research. So I am wondering: if I buy insurance, how do I game it, right? Or maybe the larger question is, you know, is a guaranteed payout based on you know something like rainfall liable to shift, you know, incentivize people to do things that might not be in the larger interest. You know, maybe I plant crops that will survive the rain so I get the payment and I get the crops at the same time, and you know maybe that shifts the - I don't know. Have you thought about those sorts of issues?
Yes, I have had many discussions about these kinds of issues back and forth on the commute. That's an ever-present problem and we face it, you know, in our day to day life with insurance, you know, if I have health insurance I am probably more apt to go to the doctor than if I didn't have health insurance. So do I get sicker when I have health insurance? Well, the records would probably indicate that I do, although I probably don't really. Yes, so there is some disincentive and hopefully again with extension work and good agricultural education some of this can be mitigated. Okay, thank you very much.