 Thank you very much for coming here. I want to thank Susan for arranging all of this. I've always wanted to come to New Zealand to give a talk, and this is the first time ever in my life that I've been invited. I don't know what that means, but it certainly means something to me. Honestly, it's a little bit far away from France, but it's wonderful to be here, and I hope that this is the beginning of a very productive relationship. I wanted to talk to you today a little bit about where our team is going in terms of thinking about the connection between law institutions and firm productivity, which links back into the economy. That's what I want to talk a little bit about. And I think that there's no better place to talk about this than in a law school, because in fact, it all is very much rooted. A big part of what we have shown in the last ten or fifteen years is that a big part of the differences that exist in terms of success at the firm level, but also at the economy level has to do a lot with legal institutions. But that's not it. That's not all of it. That's a big chunk. So the second part of what I will talk about is precisely about human capital of managers and how actually that makes playing a huge part of the unexplained productivity differences across firms once we control for all of these institutions that we've worked on. So that may link a little bit into some of the issues about how we move forward once we have gone through the first wave of shocks to institutions. So I thought that I would talk a little bit about two parts. I would first start talking about what I think we know. What I think we know, of course, there's an intellectual debate about everything that I would say, and I'm sure that some of you maybe on that side or on this side, whatever you want, there's an intellectual debate, and I think that's a very interesting way to move forward. But I think I'm going to talk first. I'll give you a brief summary about the impact of institutions on firms. And then I would say that beyond all of what I think we know, there's this new wave of thinking that we're working on, what else is missing, which has to do with COD education. So I will go through a couple of papers that I have with my usual co-authors, Laporta, Schleifer, and Vishni on these two topics. And it all boils down to this idea that it seems like this large, unexplained differences in productivity among firms within the same industry cannot really be rationalized from the differences in institutions, so we need something else. And this something else is going to be, we might like it or not, but it has to do a lot with the education of the CEO. And then I would try to tell you why this education of the CEO might actually be a factor. So it allows the firm to exploit its resources. So that's what I will talk about. If you want to ask me questions as we move along or if you want to just say anything that you think it's wrong from the very beginning, just please feel free to do that as soon as you want. Good. So I will give you a very brief view if you will about the impact of institutions and firms what I think we know. And then we'll move on to managerial education. So in the last, this is just a brief introduction. So in the last, now it's 20 or 25 years, people have been thinking about, since the fall of the wall with former communist countries and a lot of the liberalization of the economists, people are thinking about, so what are good institutions? So how can we promote good institutions, good regulations, good laws across countries? It was fairly easy to have this discussion back in the 80s and the 70s and in the early 90s because there was this capitalism versus communism thing and then it was very clear that some people were in the good side, some people were on the bad side. Wherever you lived, that depends on where you are. So then the wall fell down and then it's sort of like now everybody's kind of like capitalist but in fact there's a lot of different forms of capitalism. And so we saw that the institutions within capitalist countries actually vary quite widely. So that was kind of the idea. And you might say that all of these institutions have the goal of either economic or political, have the goal of promoting safer, better property rights and investment and higher investment and development and so on when some of them do and some of them don't. So what I will review is I will review the evidence that suggests that in fact within capitalist economies the institutions are very different. I will try to figure out where these difference come from. I will tell you where we are. So it comes from legal systems, differences in legal systems. So it comes back to law. So we are at the root of the problem in this school. And then I will try to tell you, show you some very pictures that shows that this actually matters quite a bit. And then we will move to what is missing. OK, fine. OK, so step back. So at some point we wrote a few papers about where institutions come from. And we say, well, think about an economy and as an economy is developing, it starts to figure out how it should regulate itself, some form of another. And as it's trying to regulate itself, you might think that those that are in power are trying to set up the institutions that make the rules for business to thrive. Now you would think that people would choose efficient institutions based on what they are, but there may be the case that the institutions that are chosen are not optimal or efficient from the beginning. And these are two big channels, theoretically two big channels that might explain why the institutions that we observe today across countries may not be efficient or the optimal ones for each country because sometimes politicians mess them up. And politicians choose or manipulate institutions because they're going to pursue other goals, their own goals rather than the goals of society, either they protect their friends or whatever you want. So this involves some inefficiency away from efficient institutions and actually leads towards some, empirically we're going to see some excessive regulation. The second channel which is the one that my co-authors and I have been exploring is this channel of colonial transplantation that is forget about politicians. On top of politicians, not everybody gets to choose what are the institutions that they have because at some point somebody got conquered. So this is good or bad, whatever you want. I'm not going to make any statements. We're not going to say that colonial powers were terrible and we should repent for the rest of our lives. The fact is for ourselves is that through colonial transplantations as the British land with their ships, they bring their common law. And as Napoleon moves around, he brings his codes. So a couple of hundred years later, it so happens that you really didn't choose the institutions that you had. It actually was the Brits that brought them to you or Napoleon that brought them to you or a couple of other guys floating around that didn't take them. So this is the big map that we economists, not very smart, would just discover like 20 or 25 years ago, which is in fact your lawyers must have known this for a long time. We just discovered it a few years ago. And we said, well, in fact, not everybody gets to choose the institutions. This is another inefficient source of laws, not efficient laws, because whatever was optimal in Britain need not be optimal in New Zealand. Whatever was optimal in France need not be optimal in Indonesia. Just because Napoleon took over the Netherlands and the Netherlands happens to own Indonesia and then banned 200 years later, Indonesia has French Napoleonic codes. It doesn't mean that Indonesia developed its best set of institutions, in particular, what we're going to talk about, commercial institutions. So this map gives you the origins of the laws. So we see that a lot of countries in the world have other border laws from common law or from French civil law, which is everything that is in orange. So you see that the biggest French export is not really wine, but French Napoleonic codes. It's not L'Oréal, but French Napoleonic codes. So what we have here is that those guys basically cover about 75% or 80% of the world. Then there is the Germans that never really liked the French. So the Bismarck then develops its codes. And then by the fact that the Japanese wanted to copy the last thing in life, which was the German codes through the major restoration, then they bring the German codes. That boom has German Napoleonic codes by accident 200 years later. And that's what explains who gets the laws. So you see how China gets these very random institutions just because a few things happened in the last 200 years. Then there's the weird guys in Scandinavia freezing themselves up, Scandinavian codes, and then the three red weirdos are about to disappear or blow up. God knows. But mostly these are the places that this is the origins of the laws. And this happened to be a very powerful map that we economists did not have a clue about until the late 1990s. So with this map then we said, this is very important because we can break the chicken and the egg problem, which in econometrics we call indigeneity. That is, is it the case that institutions developed to protect markets? Or is it that markets were very big and therefore they developed law to protect the big money? After all, you might as well know we're a law school. Law is a luxury good. So you only have it when you can pay for it. What is worthwhile having it? So this chicken and egg problem gets broken up by the fact that you didn't decide to develop your level of institutions is the Brits gave it to you. They landed in your shores and boom, this is what you have. So for us in economics this is very powerful because it's called an exogenous instrument. That is, we break the chicken and the egg cycle and then before we can say, now we know it goes from law to markets. So this was very powerful in that sense and this is what will allow me to say that all of the graphs that I will show you next which summarize like 10 or 12 years of evidence that we have actually have some econometric validity if you will. That is, we're not just talking about correlations. This is linked to that but I don't know what comes first. We're talking about causality in some sense. So we have this thing. So this is I think the three things that we know, I think that we know. And as I said, there's obviously discussion about this. But I think that even the most furious opponents of this view would actually have to agree with this slide. So the slide first says, it so happens that it all started in finance by the way. So we started with law and finance how law affects finance. And it basically says, the first thing that we know is that legal rules of investor protection in these financial markets can be measured. I know that in law school, lawyers think that nothing can be measured. I understand that. But because everything depends on something and depends on something and depends on something, fine, maybe we didn't do the best measurement that we can but we have done several measurements and they all seem to be very close. It seems that if you measure it, maybe not be perfect, a loss can be measured then and coded across countries. And it showed that some countries have stronger protections to investors than others. So some protect investors and some do not. That's a fact if you code them. And that these legal rules which show different levels of investor protection actually vary systematically among legal traditions. The slide that I showed you before where the French legal origin countries in particular are not very keen on protected investors and the British common law countries seem to be the ones that protect investors the most. Moving beyond finance, the next decade pretty much has shown that in fact civil law and in particular French civil law but you can generalize it to civil law in general that is non-common law which actually shows heavy hand of if you will government involvement in the control of capital markets but also in other areas of the economy. So we can see that civil law shows exhibits a much heavier hand of the government in all of the areas that basically regulate business, the working of business. So what I will try to show you is that imagine that you looked at the regulation of government to take this question. Yes. Because this is where it started this is what we measured but in fact contract law also when we talked about contract law it was a little bit later that has to do mostly where we looked when we talked about contract law was mostly embedded into labor regulation. So the approach that people have about how you have a labor relationship. But if you looked at contract law in general you would observe those differences but we started with I'm telling you what sort of world we started we all started with what you might call it with shareholder rights and creditor rights that's how it all started and then it moved into other areas of law which have to do with labor law which have to do with regulation of social security which has to do with the way you open up a business which just has moved on into the way a lot of the areas of business and embedded is contract law so civil procedure will also like that absolutely but I'm just telling you what I think so how it started absolutely but if you looked at contract law you would observe also very marked differences around that good so let me just try to show you a little bit of where we started and then we'll just jump to the missing part so this is the positive part so you see that how so this is more or less so from the different legal origins so I'm not going to focus only on contract law so you see these are all of the areas of law that have been codified in some form of another and you see that they map nicely into some specific market outcomes that summarizes all of the evidence ok if you put all of these you're organizing these three buckets ok let's say that you're organizing in three buckets that is financial regulation you know quarter proceedings and then other kinds of regulation, government involvement you would observe the very big differences among buckets ok so this would be let's say that this is bucket number one so this graph just tells you various measurements ok what you have the lower index is each of these names is a different paper that measures different types of protections of investors and you can see you know imagine that common law is zero you can see this comes from regressions ok controlling for everything else so you can see how all of the other civil law systems which are the French, the German and Scandinavia tend to be lower in investor protection and much more involvement of government in banking so there is more government involvement in banking so you can see that it seems that in terms of financial regulation common law countries are less invasive and it turns out that this actually impacts very heavily this comes from regressions controlling for everything else that you can think of ok it says that for example if you increase the protection of investors that go that anti-self-dealing then you have bigger stock markets as a proportion of the economy you know what people control premiums that is the value of controlling a firm because you can extract less so it also shows you for example that if you protect your creditors further you have bigger credit markets or bigger financial institutions yeah just on the previous slide New Zealand is always an outlier geographically and in many other ways I mean it's not really an outlier question about that stuff I mean it does I mean look at Hong Kong it looks much worse in Switzerland ok no this is given but this goes back to the conversation that I just had precisely with Susan so Susan was telling me that here we are sitting in New Zealand and we've done a lot of the stuff so we have a lot of these rights and we're not up there we're down there so what else is going on New Zealand is not down there all the time but this is a very big question and I would love to work on that this is what I was telling Susan maybe we should start thinking about trying to figure out how these outliers may come to the line absolutely ok now not everybody needs to fit the line so there's differences so this is exactly how can we make New Zealand move up down there it doesn't look as bad in that other one ok so which is about the control premiums that I were talking about ok so yes there is some idea valid to the idea that New Zealand is an outlier not just geographically but you know we have to we have to think about that ok I'm sorry most agricultural economies oh no but this is you know but this is what this is precisely saying that there's more agricultural countries we don't need to explain every point ok but the point of what it's saying is like if you protect investors more and you raise capital ok can you access capital markets more ok this is what this is saying ok and I don't see why any reason why you know agricultural firms may also list in stock markets ok you know I understand that farms, pop and mountain farms don't ok but they may you know there's very very big agricultural firms on list too ok just like small mama pop shops don't list ok I don't understand what these farms agriculture might do you know you think about it you know think about this one in terms of creditors you know New Zealand protects creditors quite a bit so this is credit to rights and in fact if you look at the credit market so a lot of you know part of it is like a lot of the capital that is coming to New Zealand firms that would seem to be right on the credit market it doesn't seem to be right on the stock market but it seems to be right ok but it also you know in terms of government ownership of banks you see that you know it also shows that you know the more government ownership of banks you have the least inefficient banking systems that you have ok but its interest rates spreads are bigger ok so this is you know we have several previous the last one the ones that I'm showing you in these graphs is from 2005 to 2013 or 14 I cannot remember ok so it's a decade because this is something that we did for the paper that we did the chapter that we have in this handbook of came out 2016 so probably we stopped in 2014 ok yes but if you did this graph from 1995 to 2005 you will find more or less the same this is not I mean this is not this is not really the period doesn't seem to be long lasting institutions ok that was bucket number one ok now giving the time that I have I'm going to talk very briefly about bucket number two and bucket number three otherwise I'm never going to get to Monagirial so that was the idea that you know institutions of financial markets model for the development of financial markets ok second bucket of regulation ok let's keep that regulators are thieves so that doesn't explain much ok so here's the second one which is government regulation in other states of the economy so regulation of how you open up a business regulation of entry that's regulation of labor how invasive you know how difficult it is to fire and hire people ok the involvement of government in the press and you know even the involvement of government in the army you see how the French or the civil law is much more invasive than common law across and that has impacts also in the labor market ok when I show so this shows you that if you make it very hard for firms to open up ok what you have is a lot more corruption in fact ok and then you have a much bigger you know unofficial economy so this suggests that you know what you're not buying all of this protection you know doesn't buy you much and you regulate labor is a lot more labor markets if you regulate the more you make it stiffier ok sort of like you see that how you know so labor force participation goes down and unemployment rates go up you know particularly employment of the poor ok of the poor of the young ok which then become poor ok this is the last area which is judicial institutions you know there is you know this is another you know sort of like courts and this is various measures that we have about courts you know you see that you know formalism in court so this so this is like civil procedure measuring civil procedure on how you go about collecting a check or a victim this is some stuff that we did in the past or how free are the judges from the supreme court from government influence let's say so you can see that common law seems to be less formalistic and gives more freedom to courts ok and you see how that actually fits into you know less formalism more formalism sorry fits into longer times in court and in the end it buys you nothing in terms of you know what people think about you know contract enforcement ok yes oh no thank you for telling me no I don't exclude them I'm measuring everybody else against the British against the common law origin ok so common law would be here I know he doesn't like me to do that because I'm in this so common law would be here ok so that I'm trying to show that you know there's differences versus common law ok that's what I'm trying to say ok absolutely we do not want to exclude the common law this is you know absolutely not you know the French would like us to exclude common law but we're not gonna do that ok good so this basically says there's three buckets map into something real ok and so if we were to summarize this you know it says that you know maybe as we think about that there seems to be a lot you know institutions are not very effective ok we saw that the inefficiencies come from politics and from colonial transplantation so we may be moving into inefficiently high levels of regulation and we also see that you know a lot of even more hires of inefficient regulation would be probably in French or in civil law countries that in common law countries which started with less invasive regulation ok fine you know we're going this way fine so that's the part that I think we know ok so we think that there is this very large component of the efficiency of firms and the productivity of firms and that comes out of the impact of institutions ok and the way we measure that was by comparing across countries the differences in legal institutions other people have started to go into legal institutions within countries ok so measuring differences within countries about different institutions ok so some countries perhaps some provinces have different institutions and others ok so some there's a very large literature on that and also we're going to get all of those numbers when we move into the next phase which is moving into firms and regions ok so the first part of the talk was to try to say you know listen what explains some differences in productivity but there's still you know the big possible that we observe there is that there's still very big differences in productivity among firms within very narrow industries ok within the same country which has the same institutions ok so this is not data from me ok so this is data from the kings of the guys of measuring total factor productivity ok so this is a summary of this is a summary of what people have found out ok if you look at for example this is the US you look at the graph below this is you know total factor productivity tfpq ok and you can see like let's say you know one is sort of like what we have at the center ok you can see that it can go all the way up to four or five and you produce one versus four most productive firms produce four or five the least productive firms you know produce one 64th of you know apples ok so you can see that there is humongous productivity differences within even within economies that are supposedly have very large arches to capital markets like the US think about the US very large arches to capital market hopefully very good institutions blah blah blah very large productivity differences if you move to developing countries you know so there's something like China or India you can see that these productivity differences maybe even wider ok so therefore the talk of today so it's like what is it that may explain this other part I mean I what I have just talked to you about is I said you know we can explain a lot of the stuff that we have in the world from here this difference between this these three it's all of what I said ok now I'm interested in explaining here or here ok so we we figure that out a little bit ok we have not figured this out completely ok so that's what I want to talk about ok any other questions now ok good ok so what explains this productivity differences ok so we economists usually take us forever to do anything like you lawyers figure out good things just write them down we sort of like you know we have not figured out much of the stuff and we definitely don't like to write things down ok so you know we have long about this productivity differences that we've studied says well sounds obvious it has to do with inputs right so of course if you have more inputs or better quality of inputs then you can explain higher productivity ok so you know so we the traditional approach uses like raw materials capital labor technology blah blah blah blah blah and when you put all of that in there's still humongous differences in productivity that we cannot explain ok therefore we said well this is not going to get us very far so when I need to move to the next thing so what about human like human capital human capital of managers ok there's been some work about human capital of labor and that has actually shown that it explained a little bit but not a lot ok so workers human capital the education of human workers didn't seem to be like a humongous you know a factor in explaining the productivity differences ok so what we have is so what we did is we said why don't we try to go and get new data ok across countries across regions across firms in the world in different regions in different countries and we try to see the human capital explains a lot ok so that's what we did ok so we try to analyze directly and compare you know the impact of human capital of managers but also we're going to have workers so we have some measure into the firm productivity differences ok good ok so fine so let me skip all of these guys that say that there's just to say that there's many factors that it makes plain productivity differences and I better start by saying how do we start so we start by actually looking at we start by looking at differences in productivity within countries that is within regions and then we look at differences in productivity of firms within each region ok because if there are differences in productivity within countries that come from differences in regions then we're going to be attaching too much to the differences of just by firms in fact there may be regional factors within the country that explain part of the difference and then there will be firm level factors so we have country level factors which is what I talked about so we throw that into upside now we're going to work within countries among different regions and first I will show you that there are very large differences you know in human capital across regions and that they have an impact in productivity within the region and then we're going to see at the third level within those regions differences in productivity in firms and then we see that is the human capital of the entrepreneur that explains a bunch ok is that clear ok fine so but first I need to get with from step one to step two which is I'm going to work with regions now ok so in order to do that we're going to work like crazy for five years ok and we got this data ok so we said we're going to shut the world in 2005 ok and we're going to see now we have a 2010 ok so we're going to see all a ton of data I'm going to show you what kind of data for all of these regions in blue so we don't have country data we have regional level data ok we have a few regions missing in Africa because they don't collect statistics and we couldn't force all of them to collect them ok but we forced a few ok and then so we can see that we have all of the differences in regions so within the US we're going to be talking about data at the state level ok so within Australia we're going to be talking about data about those guys in New Zealand we're going to be talking about data within those within the different regions ok is that clear so it's like a ton of data so we have you know about 1569 regions in 110 countries ok that's more or less what we have and so we're collecting data of all of that stuff sorry we're collecting data on geography that is endowments so we have now there's all of this you know text stuff like maps and so on that are collected from satellites and so on and then we have temperatures distance from the capital of the region to the coast resources like oil blah blah blah blah then we have institutions as we said institutions are going to marry and there's differences in institutions within regions of the same country so we have all of those like corruption and you know access to finance blah blah blah then we have infrastructure so we know about the city we know about a ton of other stuff ok and then we have other values like there may be differences cultural differences we have culture trust blah blah blah ethno-linguistic fractionalization and you know religions and so on of that stuff fractionalization and all of that stuff so that's a humongous amount of data which people hopefully can use and are using and then we also have obviously population differences and so on ok so we collect all of that data ok and this is what comes out of that as the beginning so I'm just going to summarize imagine that you took countries now what I'm going to try to show you is what explains differences of you know income per capita which we region which could be called productivity of the region ok so the productivity across regions within countries ok so look at you know look at Russia ok so Russia these are all of the different you know oblasts or regions if you will and controlling for other factors ok controlling like for institutions blah blah blah blah blah ok you know if you look at productivity differences among regions you can see that the average level of the school this is education of the average person in the region explains quite a bit ok of the differences of productivity that does not have to do with endowments that does not have to do with institutions that does not have to do with temperature that does not have to do with all of those theories or fractionalization or culture ethnic fractionalization or culture it's education that seems to be driving the day if you actually I just show you four countries for example but if you pull the 110 countries together ok this is the net effect so you have you know you put all the 100 countries together and you put the variation that you have from each of these countries within and then you see you know each of these dots you know for example each of these dots is this is a different province in this country so you can see different provinces ok within countries that's where we have multiple of the same ok so you can see that the net effect controlling for temperature you know what you can call geography temperature distance to coast you know resources which you know we have population et cetera et cetera and even country dummies a lot of it has to do with education in fact when you look at look at you know just streamer compared with institutions ok if you look at the years of education for you know we have it for 104 countries homogenized ok you can see that it explains first of all between countries ok education explains about 8% of you know of the r2 ok so the r2 if you just put education in a very simple one like univariate education explains 58% and you know informal opinions would be an institution corruption let's say 21% and trust would be about 80% you would say you know they're all kind of like there education is very big three times as big but you know they're there if within countries it sounds like all of those institutions like corruption and so on and so forth and trust just collapse in terms of the power to explain and it's education that seems to be the driver of the differences within yes ok so you take the average level of education of the population within the region the whole population oh ok so fine question this is what it's called the partial scatterplot so it's not the actual number this is imagine you run a regression that says you know here you have on this so like the variable that I'm trying to explain is income per capita ok log of income per capita and that's going to depend on temperature oil resources blah blah blah blah and the log of years of education ok now what I'm there is the coefficient on this and the variation around that coefficient controlling for everything else ok so the coefficient of that regression ok is 0.27 so that's controlling for everything else and that's the variation around it that's the statistic of that coefficient so it's not the actual number of years but so it's kind of like you know it's been processed through this machine so it gives the variation normalizes somewhere around to or something ok so it's not the actual number of years this is what you know Larry so it's called partial scatter plots ok which is the net effect that's why I call it the net effect ok so you can't really read the number as a number you have to just read the the slope ok yes ok so we cannot I mean this is very big we could not do that absolutely yes yes I mean I understand that you know one year you know in in Switzerland maybe better than one year in let me not say another country ok but you know in another country ok I I do yeah unfortunately we can't really do that I will show you when we move to firms ok if we move to a government ok I will show you that actually may matter you know the quality of the case may matter when we talk to about CEOs ok but you know ideally we could do this but the idea that we're going to have data for 110 countries with the quality of education with the same granularity it's very tough we have some measures ok so I will show you some you know about absenteeism like like what is this data ok so it does have an impact ok I wish I had that but you know this already was a very big effort but then very well taken ok good so that's at the regional level so that so then we said oh my god education explains a lot so what about at the firm level ok so now we get to the CEOs ok ok let me skip that this just tells you that when you put in you know education this works a lot so now we will get the firm level statistics ok let me skip that we're going to look at the firm level and what we're affirming firm level now imagine that you have firms ok so once we clean the sample you know we have this is like you know 13 or you know 20,000 depending on the data that you have we're trying to explain firm level productivity that is sales per employee clear productivity controlling for a lot of stuff ok you know country fixed effects industry fixed effects all of that stuff and then we have resources we have about you know this would be the resource like energy resources or property plant and equipment and so on and so forth and what we are trying to do is we are trying to make sure that the services are controlling for all of that and controlling for the regional education ok education of the region which is the effect that I showed you before which is significant the one that comes out you know worker education is important you know statistically significant but once you normalize and you actually do the calibration is managerial education seems to that worker education has so we're not saying that is not the educated workers do not help for the productivity we're saying that see educated CEOs have a humongous impact and the productivity of the firm ok ok so this is where like summarizing into sort of like so now we have only like 30 countries where we have all of these incredible data about expenditures or energy and expenditures as I move some of these things remove some of these things then I get into much much more countries right so this is giving you the toughest thing when you remove that it's even better so this will be like the toughest regression ok and actually the toughest toughest one the last one would be this one where we have if you really want to be very very very very concise and want to just go to the toughest sample where you have all of the data you move it all the way up to here and it's very very big ok ok so what does this mean ok so the evidence suggests that you know the we have been misestimated that returns to education to managers ok because we haven't really been measuring them and putting them in the equation for productivity we just haven't we focus on the education of workers and we have forgotten about the education of managers ok and they seem to be the education of the CEO by the way the measure that we have is the education of the CEO we don't have measures of the education of second level manager at this level so I have the years of the education of the CEO across firms ok all firms and then we're going to control for industry and whatever you want to you can do industry so it's all in manufacturing all manufacturing so we're going to throw out financials agriculture this is all manufacturing ok but it's many different you know industry manufacturing but it's a very good question manufacturing firms just so apart from education that you also have a number of years of experience of the CEOs I don't have that so it may very well be that a very highly educated person would be very health productivity but it's because that person had 20 years of experience sure it could be I mean it could be ok so what we do know and I think you know we don't show it here what we do have is how many years the guy has been in the firm ok I don't think we have it in these regressions ok because it didn't turn out once you control for that it didn't turn out to be that important what we have if you have to be working in multiple establishments we have the age of the firm which turned out to be more so older firms tend to be a little bit more productive and that seems to be very correlated that's why we threw it out in the end because that seems to be very correlated with the years of experience of the CEO and this one actually tends to be better than the other one ok so we have that for this very small sample we don't have that for this very very large thousand sample ok so when we get into this really trying to get all of the stories apart ok we still get at this actually the years of education on top of on top of that so it's I mean it's not that it's experience doesn't matter but it's sort of like it kind of gets gets subsumed ok now the age of the firm which is very correlated to the experience of the CEO it's important we know that older firms more and so on and so forth ok so that's part of it but that's a very good that's a very good point now why do we think that people actually have why do we think that these CEOs may be very good for the firms and this is just speculative this is somewhere where we actually need to go ok so it could be that firms you know that educated CEOs may be better at organizing production managing firms or adopting or developing new technologies ok that could very well be the case or it could be that you know these differences in my the education explain why not all firms are able to take on opportunities ok so good firm policies may not be successfully adopted by all firms because these firms do not or cannot afford high human capital of CEOs or highly educated managers may be able to adopt much faster something new that comes out ok so they may be faster and getting at it we do not have any clue about that yet ok we don't you know maybe some people do but you know we don't ok and not at this level so but it sounds to it's correlated to the a little bit to the CEO ok do I have time to do the next one are you serious ok so let me take five minutes ok ok just five minutes ok so then so then we said ok how can we go a little bit deeper and trying to figure out what makes the difference ok and then we said well there's this other big puzzle about there's this other big puzzle about you know why are governments different governments so why are governments some governments efficient and some governments are very inefficient ok so that's a very big question so one of the reasons is why there's political economy reasons you know there's some guys that are very corrupt there is some you know all of that stuff that the typical stories ok and then we said well another reason is maybe ok it's you know some governments have bad managers ok and these bad managers are just bad managers and they just cannot organize production just like a firm ok so this is when we decide so this is the idea that low productivity of government service may be similar to the low productivity in the private sector ok so there is inferior inputs human and physical capital and technology and there may be poor management that comes out of it ok but now we need to design an experiment that allows us to you know remove all of these stories from the equation so take them out of the room that says these stories cannot explain this productivity differences ok it can only be about human capital and initial resources ok so that's when we design this paper that's called letters ok so what we said is like obviously measuring government efficient is too big ok so we're going to focus on one single service which is very simple which is mail every country has mail ok post office ok and we're going to focus on mailing a simple letter ok like a very simple letter ok so what we did is you know we sat you know and then we mail 10 letters to 159 countries we mail 10 letters ok 2 to each of the 5 largest cities in each country and the letter basically had a piece of paper that said hi I am you know Rafaela Porta you know I wanted you know to get in touch with you to carry out our you know research project ok that's all that the letter says more or less do I have the letter here no I have the envelopes ok so but what we know we kind of like trick the post office because we put the address we included had the right city and the zip code but there was the wrong street and business name ok so so this is then we send them you know we send them into waves because then you can instrument the second wave with the first wave all of that stuff that econometricians like to do ok so the letter was one page business letter in English requesting a response from the recipient ok and then we waited for you know I don't know ungodly amount of months like 40 months or something 420 420 days so Lapo you know Laporta would go every day to his mail you know at Harvard and he would pick up his mail and then he would mark down ok I got the letter on you know the third January of 2012 ok so we knew when we posted the letter and we knew when Lapo got it ok and this is the kind of envelopes that we had so we waited for 420 days ok notice that the beauty of the experiment is that this is a letter ok so it's like you know it's not like you want to steal the letter ok so it's not corruption ok so the fishes is not someone wants to steal it was not a $20 bell ok it doesn't take really a genius to sort of like figure out that the letter doesn't go to the right others you just put it on the different you know beam ok so it takes away a lot of the political economy stories so we just focusing on resources ok so that allows us to build a production function that says well the productivity this is the productivity is how many letters returned back to us at Harvard within the period there is a convention we didn't know about this but you learned so much about stuff that you really don't care so there is this convention ok of the postal office around the world dating back from 1969 that basically says that 159 countries which is the ones that we send the letters to signed this convention that says that when the letter is wrongly addressed the country needs to resend it back within 3 months and the country it's not like the country that received the letter has to pay for it is the sending country that pays for it so if anything you would like the gringos to pay for their mistake right so the Americans have to pay back for the mailing ok so it's not like you don't want to send because you don't have resources ok so you can see this is the kind of letters that we got for example this was you know we picked also names of economists ok from the country ok and then we have a firm that doesn't exist and you can see that this is the address and then you know this is what we get ok fine so how did that come out ok so we're going to get into the quality of education here also too so this is what this is two countries for example and I will promise I'll finish in 5 minutes ok so this is two countries we send 10 letters to the Czech Republic which seems to be a fantastic country so you have here we send the 10 letters this is the name of the guys the address the postcard and the city when we send the letter ok and when the letter came back and this is the date of the limit when we stopped you know when Lapo said ok enough I cannot go to the post anymore so you see that the Czech Republic send all of the letters back completely and all of the letters within 3 months all of the letters within 3 months which is what everybody agreed on so they get so this is the number of days in the last column so within 52 days on average they send all the letters back Russia ok so this is the letters in Russia we got no letter back in 418 days 20 days ok we got no letter back and then this is very funny because I think a TV station got a hold of this thing and then they run a TV show you know which had you know grandmas babuchka's crying that they never got the mail cats roam this is from the show in the TV you know TV chain number one ok cats roaming through the mail eating the letters and so on and so forth so it was quite a show ok so if you put all of the countries together you know this is what you get you get that on average ok we had we got about 59% of the letters back ok about 35% of them came within 3 months and the average time that it took to get back was 228 days ok now you can see that high income countries did a lot better because they had 84% of the letters back 60% within 3 months and already took them 125 days and after that it just falls down for low income countries ok you only got a third of the letters back less than 10% of the letters within 3 months and it took them 336 days so there's an income component and there is an educational component you can see that in countries with low education also got you know only 46% of the letters back in 281 days ok now with all of this fantastic data or whatever maybe you think crappy data or interesting data whatever what explains you know if we got the letters back so we created these variables which is productivity ok which is there and we also have like incredible data that you would never think you would have like because these guys in the post code convention they've been collecting data of postal offices since the 1970s I mean it's like an insane amount of data like we know the number of letter boxes that exist the number of staff the technology of the sorting machine of the letter that reads the letter and so on and so on I mean it's insane the amount of data that you can get out of these guys so we have all of those variables which are inputs for production and then we put all of these measures which have to do with education or quality of education ok and you can see that controlling for everything else ok measures that people associate with you know some systems which have you know higher managerial education if you will for example here's quality of management skills we should have one that says quality of law schools but it doesn't get collected that much ok and anything that has to do with management practices which is you know better monitoring practices you know managers that set targets to their employees managers that set incentives and so on and so forth although they have only been collected for about 20 countries ok these ones are more for wider countries but they're very correlated you can see that management quality kind of like explains a lot of the differences in productivity ok so when you put that in regressions ok you can see that controlling for everything else you can see that you know the proportion of letters that we got back is much much higher if you know there's higher you know management education as we can measure it for this for this kind of like this country data ok so it says that the quality may also come in ok now I wish we had better measures of that stuff ok so let me finish with this so taken at first value ok you know so I think that what we've said is institutions such as law matter but they're not all education matters quite a bit ok and it actually explains a lot of the differences so if you want to understand the differences in variation of income across regions you kind of need to get to human capital if you want to understand the differences in productivity across firms you have to get to managerial education and even for the differences in productivity in governments ok there is this you know managerial component that turns out to be pretty good so maybe some answer for New Zealand some countries have very similar institutions that they inherit through legal transportation some of them are possibly not the best or some of them are you think they're the best but maybe ok there's something about human capital that may explain something now the puzzle of New Zealand may be bigger because human capital is not so low ok but I don't know about you know human capital of managers ok so maybe I don't know we should ask that to the law school to the business school ok but it sounds that you know people with right human capital ok maybe detonators of productivity ok so I think this is where we are kind of going this may not be very interesting but I think this is where sort of where we are ok good thank you very much