## Leadership in the lab

## Incorpora video

# Leadership in the lab

What is experimental economics? Can the relevant methods be put to the use of understanding leadership? The lecture will attempt to provide answers.

okay we start it's a real pleasure for me to introduce Professor Martin doofen Berger who is not only a brilliant researcher but also a very kind person so I'm very happy that you are here with us he is spending a period at the Bocconi University and I think it's a quite long-standing time that he will spend there he is an experimental economist the behavioural economist and the field of research covers is a quite wide one recently well he's a game theory theoreticians as well and the main part of his scientific production is a sort of crossroad between economic game theory and behavioral economics he will I will I will I will be very short so I can I leave you the time to to introduce your your talk I just spend a couple of words to say that what we are speaking today here is about a very innovative and a challenging way to study economics which is to do it by using a laboratory laboratory in social sciences is something recently recent relatively recent because they started and many years ago already we have a laboratory in Trent as well which is a deliberate the cognitive and experimental economics laboratory we started about twenty two years ago we were the first laboratory of this kind in Italy and for this reason I think that you will find a group of people here following you very very very warmly thank you very much thanks a lot Luigi for this introduction it's really spectacular and fun to be here thank you all for coming I know I had a tough competition with the material Renzi being in jacent room somewhere I want to say thanks also to Federico for now sorry who is Regis PhD student who who has helped me settle in here in in in Trento and told me about his interesting work about risk preferences and social preferences I will talk about other themes but also related to experimental methods all right do we do you said something about 30 minutes 30 seconds okay it's been okay okay very good so I think I'm gonna stand up because I need to point to my slides and so on let's see can you hear me well if I talk like this yes even like this okay okay try to go like this okay the topic of the day is leadership in the lab so as Reggie mentioned you know I do work on game theory on on experimental economics on incorporating psychology into economic analysis using game theory and the experimental method I hope you will get a little glimpse of all these things but it will be a more specific topic today namely leadership in the lab and now I don't know how to switch the slide how do I do that okay before I come to the topic of leadership let me say something broadly about experimental economics the topic some of you may some of you may not have heard about and visuals indicate is a little bit its concerns using the experimental method to learn about economic issues one way or another and there are different ways you can do this and in in a one-hour talk like this one or one hour plus talk like this I will not have a opportunity to mention all of them but leading instance would be was I've written about here on on my slides and it's related to economic theory okay so what do economic theorists do economic theories tell stories about how the world works they look at thought of worlds and sort of try to maybe no thought of words they may be inspired by the real economy and trying to get sort of assess the anatomy of how things hang together and you know a large part of economic scholarship is is concerned with posture dating assumptions and sort of telling a theoretical story about how the economy might work okay that's interesting and valuable in its own right and many scholars will will be satisfied with doing only this you have lots of publications to make you make sense or just economic theories however you know these stories about how the world might work over of course even more important to the extent that they are empirically relevant and a big chunk of experimental economics is concerned with taking economic theory as a starting point and then asking if this theory credible does it come alive you know is it empirically relevant and this is the specific perspective I will be having in mind when I talk now before I come to the specific theory I will look at we've mentioned lab experiments you may wonder what the lab looks like this is our lab at pokorny it's called the bikuni experimental laboratory in the social sciences and maybe I'll hold it like this and you see a picture here you can picture it slightly blurry but you can get a sense there are cubicles you know students will be invited to the lab each student will be seated by in one of these cubicles and then they have a computer screen in front of them they will be getting into instructions and opportunities to make decisions you know games that they might be involved in will be described to them markets they might interact in will be presented to them and they will be making decisions may be buying and selling decisions may be bidding decisions in an auction ins of and at the end of the session there will be real payoff consequences well they are they are paid in accordance to the economic outcomes generated in the labs so you know the theories will be telling stories about economic interaction and then in the lab you actually provide settings that match the theory including paying subjects right so you know you really try to replicate the conditions described in the theory in the labs I think I hope I'm not jumping in and out of the microphone too much or is it okay the way I'm talking now this was a little bit by way of introduction about economic theory in general and and the experimental method now I will be more specific so when Tito Bowie approached me and asked if I could speak at the conference you know he mentioned that the topic was the ruling clause and leadership right so I agreed to talk about leadership in the lab I did that whilst having done no work on leadership were almost nowhere or come leadership myself okay so I warn you that I'm not an expert on leadership but when I got the task I set myself to explore a little bit to learn about the literature on leadership from an economics point of view there is a lot of literature in in neighboring fields in management in organizational psychology okay but I'm an economist and economists have written it turns out when I checked surprisingly little about leadership I mean you could also say there is almost nothing that is however something and the little race tends to tell snapshot stories to highlight some key issues that may be relevant to leadership okay and I will in this lecture highlight two such snapshot stories and mention related experiments so there will be sort of two parts to each story first where I invite you to reflect on and try to appreciate the the theoretical storytelling if you will as interesting in its own right and then you should get curious about okay we just heard a story about how the world may work is it empirically relevant okay the first story is going to be based on theory which I had no part of it's done by Monica my mark Stegman and pin Harmelin who published a paper in the American Economic Review in 2007 I put in a picture of mana and and and her co-authors to the right she was also involved in an experiment with Phil Grossman and Travis theatres so Travis deters is missing from the pictures but the other authors are there alright so we start with theory so let me now give you a simplified account of the theory of coma Stegman and her Moline so they consider a setting with teamwork okay so imagine that you have a group of people who work in a team and we stylize this team work very much and say that each individual in the team has a choice a binary choice the individual can either exhort effort work hard or not do so okay and imagine that there is a cost of working hard you know it's easier to relax and not not put into much much effort so there will be a cost of working hard and I'm gonna para meet your eyes and have numbers okay to generate payoffs so let's say that there is a cost and we'll give it a number it's gonna be three so if you exhort cost at work sorry work hard you have to pay a working hard cost of three on the other hand you cannot work hard and then you don't have a cost it's zero okay think of that as shirking or not working hard the good thing with a fort however is that it benefits the whole team everyone benefits the organization benefits right and I'm gonna have a parameter for how much that benefit is and I'll call it X so X is going to be a number and we look at different numbers of X so think of it as for each individual who puts in effort everyone in the team gets X right so for any given person exerting effort everyone in the team gets X okay now I'm going to simplify further and assume that the team actually consists of only two individuals okay everything I say and you know the author's actually look at ritual settings with more individuals and more complications but I'm gonna look at a setting with just two players and then it turns out that for a given X you get what game theorists will be cool would call a game so I think as much as learning about maybe even more than learning about the experimental method in this talk you will learn about game theory okay because my theories will be game theory based and and I'll have to introduce T game theoretic concepts to you and I hope you will enjoy that so here we go okay what I just said on the previous slide can be described again using this matrix this is what game theorist would call a game matrix okay it depicts the strategic situation I just described there are three things to note to this to learn okay in line with what I had on the previous slide first of all you have two individuals interacting they are called players in game theory okay al use of term player so one of the players is choosing a row there are two rows here okay and let's call that player play one let's say play one chooses either e4 exerting effort or in for not exerting effort okay if I'm unclear you know please raise your hand if you want to ask a clarifying question but but you know you have the first player choosing you're in a row then you have the other player at the same time choosing also you're in for a photo no effort that corresponds to columns right so one of the players choosing throws the other one chooses columns that means since you have two rows and two columns you have two times two equals four combinations corresponding to the boxes in the matrix okay they correspond to outcomes so there are four possibilities both of them choose a high effort neither of them choose n that would be the outcome and then the two other outcomes were one of them chooses effort and the other one not chooses effort right okay in line with what I said on the previous slide we I'm not done two of the three things I wanted to say I wanted to tell you who the players are and they are two and one chooses rows one chooses columns the other thing was to describe what they could do and that was the effort under know the thing the East and the ends here right therefore internal effort the thing that remains is the payoffs both do they earn okay and these are the numbers written in the cells remember the cell so the boxes correspond to outcomes and in each box I've written two numbers okay in an attempt to be pedagogical I've written the payoff of play one the one choosing rows to the left and below okay and the one to the other player to the right and the bow so for instance you know if you take the case where none of them makes our safe what well they're not gonna have any any any any nice X to enjoy and you know but they won't pay any cost so they get zero each right and that's what I've written in the cell of course if both of them let's say one of them exerts a fort let's suppose play one exert effort and the other one does not then we end up in the Pope right cell okay what are the payoffs going to get B well since there is one unit of effort provided in this case by play one both of them are gonna get X because that's what I said about X right for each unit of effort everyone gets X but the person exerting effort is going to have to pay a cost not the other one right and the cost is three so it's going to be X minus three for the first player and X for the second player of course if you have the opposite outcome or player 1 does not exert effort in player 2 thus you get payoffs X and X minus 3 okay so this was my attempt to explain this game a trace okay now think of X that's you know I'm keeping the cost of effort fixed at 3 but I'm going to start very and playing around with X of course X is what everyone gets if you have a unit of effort right and oh I forgot to say that of course if both of them exert effort we go to the top left corner since there are two units of effort provided one by each person they each get 2x but they also each have to pay a cost of 3 so 2x minus 3 I hope you followed this presentation of the matrix now I'm going to look at different at games that differ in terms of the X so think of X as a productivity parameter the higher it is the more gains are all from exerting effort did I have a question yes okay let me so the question concerns the top left box okay thanks for the question so why are the payoffs 2x minus 3 okay here is the answer for each unit of a thought okay each person gets X how many units of effort order there are 2 because there is one coming from day one and there is another one coming from player 2 so we have 2 units of effort okay one provided by each of the two players yes so I'm saying okay what happens when look at the one individual when he exerts effort on the one hand he has to pay three okay anyone exerting effort pays three but there are benefits okay namely you good you get an X and you get an X from each and everyone who provided effort so in the top left corner everyone has to pay three because everyone is exerting effort on the other hand three is the cost of three is not the only thing accepting the payoffs you also get to enjoy two x's one for each person exerting effort so the the payoff is going to be the difference between the two X yarn and the three you have to pay so that's why it's two X minus three okay now I'm going to look at games with specific X's I'm going to look at low axis y axis and intermediate axis okay once I have those three down I'm going to be able to move to the theory of leadership of kamae yes that would be wonderful thank you okay hello that's okay you can hear me okay this is gonna help me a lot I think so now I'm gonna plug in different values of X in an attempt to be pedagogical I put the old game tree up in the four right corner so you can see that one so you have the two X minus three and the zeros and everything up there and then in the slightly larger game I'm doing the same thing again but I'm plugging in a specific value of x so let's assume that x equals three okay if you have x equals three then two X minus three here it's gonna be two times six minus three so you get a nine same for the other player X minus three is six minus three so you get three X is six and zero is zero so you get this matrix okay now let's think about what players will do in this game okay and it should be pretty clear take it from a viewpoint of player one player one doesn't know what player two is doing but he doesn't really care because nine is bigger than six and three is bigger than zero so whatever the other player is doing is better off choosing high effort because the rewards are so great okay so he will the other player it's all symmetric the same logic applies you come he will compare nine and six here and 9 is bigger than six so if if he knew that the other player was choosing II he would choose he himself of course of course if the other player is choosing n he would also choose hehehe so because three is bigger than zero three is bigger than zero nine is bigger than six he doesn't have to worry he can choose high effort yes easy and a nice outcome great outcome as I wrote now I'm gonna look at the let's think of you know the situation as involving different exes right so now we have a situation which is less productive there is still yes question yes well I think what you are now saying is that you like to modify the game to the one coming on the next slide so bear with me two seconds okay I think this is the game you had in mind but tell me if I'm wrong when I'm done with the presentation okay so here is a game where X is lower okay you don't get the full kick of of of six okay you only get the kick of two x is 2 now again if I take the the game we talked about I'm not going to repeat it but you know where the payoffs came from and I'm plugging in the different values of X here okay then here instead of a six on the lost side I get to two here I get 2 minus 3 is minus 1/2 X minus 3 is 2 times 2 minus 3 is 1 so you get this game okay think about how you would play this game okay now you know if you exert effort effort gives 2 to everyone it gives 2 to you but the cost is 3 so you don't want to do it ok and you can see that in the figure because you know 2 is bigger than 1 and 1 0 is bigger than minus 1 so whatever player 2 is doing you're better off choosing no a Fortran high effort okay and similarly for the other player no effort is better than high effort and they end up here getting zero each and that's actually a bad outcome in this case because if both of them had exerted effort this is may look paradoxical to you okay they would have been better off so what I've just been saying is that if you look at the situation from the individuals point of view it's not rational to exert effort when X is so low you want to do no effort but when everyone is doing it it produces an outcome which is bad for everyone in the sense that it would have been better had both of them chosen high effort okay am i right that this is what you had in mind yeah well right my distribution function is is taking a stand I'm saying I'm really saying that for each of your units of effort there will be 2x produced because there are two players and you get the next to everyone so you get half of it okay and you're saying often I get less than half fine yes anyway this illustrates that there is a conflict between individual rationality and what's good for everyone this is the most famous game in game theory okay many of you will have heard about the prisoner's dilemma this is a prisoner's dilemma it usually stoled with a different story which I won't tell about prisoners being held in custody and the police interrogating them but you know the structure of prisoner's dilemma is much more general than that it can come up in different settings here it came up in my organization setting okay right why well because the individual has to pay the cost of his effort but the benefits of the effort is shared by everyone so it only gets a small slice that's the intuition okay so here you have actually an inefficient outcome in the sense that both of them would prefer to be up here as opposed to being down there now I go to a new game where X is even lower X is gonna be equal to one I mean in this case it should be clear that there is no benefit to doing effort and in fact not to anyone because you know if I exert a for the cost is three and we get X for each of the two so that's 2 times 1 is 2 is smaller than 3 right and if you plug in the numbers here plug in X you get 1 minus 2 minus 1 minus 1 you should do no way forth because 1 is bigger than minus 1 and 0 is bigger than minus 2 for the other player so here there is no inefficiency it's right not X can be so low that it actually doesn't it's not a good idea to exhort high effort ok unlike the previous situation where it was in principle a good idea but it wouldn't happen did I have another question okay so to summarize depending on X I put the game with x equals 6 to the left x equals 2 in the middle x equals 1 to the right I had good outcomes in the games on the sides but I had bad one in the middle okay now let's assume that you know you walk in a for more so X's come sometimes you get great opportunities and then you have high X's sometimes you have bad opportunities and you have low X's right let's suppose just for the sake of argument that you get one of these three X's with equal probability you with probability 1 over 3 you get a low execute equals 1 1/3 of the time you get a high X X equal six and one-third of the time it's this intermediate x equals 2 case right if everyone absorbed the X and chose according to our proceeding analysis well we know they would get 9 here they would get 0 in the middle and 0 over there so in expectation the payoff would be 1/3 times 9 plus 1/3 times 0 plus 1/3 times 0 so on average they get 3 yes ok now we're going to compare it to a new situation we're gonna change the nature of the organization okay so now you come to the heart of learning about the contribution the key idea of this paper I mentioned by Kumai Stegman and Harmelin okay so they say hmm you know this information about what makes a productive project is we can imagine that we can decide how to hand it out you know suppose we make one of the two players to get this information only okay and we call that fellow a leader right so one player gets two let's say that the one choosing the row okay gets private access to information about X so instead of everyone knowing that X is either 1 or 2 or 6 everyone knows it will be one of the three with equal probability but here we just tell the player 1 okay and then instead of them simultaneously choosing higher photo low effort we let the guy who has the information about X play 1 the row player make his choice before the other player so he lets the other fellow see the choice okay but it doesn't tell what X is it just leaves for the other one to make its own inferences and then choose high or low effort okay then you can improve on the previous situation and the smart idea is to do this the leader play one row player will choose to exhort high effort if X takes either the high or the intermediate value X is 6 or 2 whereas he will choose no effort if X goes the low one ok let's see what happens when you do this ok so here are the three games and I'm proposing that the leader does has indicated in purple it uses e if it's x equals 6 or x equals 2 otherwise it uses n ok now of course if the follower understands that this is what the leader is doing then of course when n is choose chosen the response you know the the follow will will figure out that the leader has chosen this row so the follower will say 0 is bigger than minus 2 I want 0 so we would choose in as well and they end up here hey that's unproblematic the interesting thing is what happens over here suppose the follower sees that the leader has chosen the high effort but he hasn't told the follower what X is ok then the follower knows it's really either here or here ok so we have to think it's equally likely it's either right so what's the choice of the follower the children choice of the follower is to say hey I can choose e and then I'm either here or here so I get either 9 or 1 9 or 1 9 or 1 the average of 9 + 1 is 5 so he will value this choice of e at 5 on the other hand he could choose no effort and he would value that at 6 or 2 6 or 2 6 or - the average is 4 right 5 is bigger than for what you want five or four you would want five right so you choose the high effort so you the the the idea of keeping the information about X private to play one and calling him a leader is that they managed to coordinate on the good outcome in the middle game you solve the prisoner's dilemma part of the situation okay right so you see instead of having if you remember well I can show you oops wrong direction I used to have efficient inefficient efficient oops going too far again here what am I to do oops sorry I went here and I got all of them efficient and if you look at the expected payoff it's 1/3 times 9 plus 1 terms no longer 0 but 1 times 1/3 times 0 so you get 10 thirds on average which is higher than 3 so you gain something yes okay now you learnt the essence of kamae Stegman and our Merlin's theory I wrote bow in the margin okay now of course as I said this was a simplified account you know their theory is more complicated more rich on nuances and so on but the essence is really what I told you the theory really tells a story you know it's a thought up world which is much more abstract than the the world you probably know and you may think why so simple ok this kind of snapshot character trying to highlight some feature of reality rather than incorporating everything but at this point you know you should think I think that hey Martin has told me this piece of science fiction basically you know is it empirically relevant right and here lab experiments may help ok and the study by Kumai Grossman and dieter's recently published in managerial and decision economics they run an experimental test and it's essentially concerned with you know where as the real world is more rich than the theory in the lab you can create the world which looks just like the theory ok so they what they do is they invite students to the lab ok they describe these games and they say things tantamount to saying with probability 1/3 X will be this and this and that right the world that the students in the lab encounter will look exactly as the mathematics but translated into lab setting where you can make choices between e and then and you get paid euros isn't that way I guess it was dollars in that case okay and then they do they have what's called treatments so when you do experiments you know you run different settings and if you change some specific detail you get two different treatments and then you can compare because you know that the only difference between the two treatments is the little thing that you changed so what do they change well they run the game the way I told you first with X known and announced and then they get certain outcomes and they do it the other way with a leader where they say hey we're appointing player 1 the leader he gets to absorb X he's not gonna tell the other fellow but he will make his choice before the other one follows up right and then I'm not gonna give you the details you can find the paper if you want to but you know basically what happens is that the theory is largely supported okay so now you still have to worry that maybe this fear is too much of a slap shot of reality but as a metaphor for some capturing something important you have a stronger belief in it than if you didn't have the support in the experimental test yes okay that'll study number one study number two coming up are you ready new example when I need to ask me you know I said I didn't do anything on leadership but I started thinking yes Oh Christian yeah you know I I think the question concerns how large X is okay you know the how large it's being confirmed it's okay my answer will be this okay when you're doing lab experiments of this sort it's hard to draw quantitative comparisons between the real world and the lab what you can get is qualitative insights here the qualitative inside would come in the form of is it the case that it's a good idea in principle to a point a leader or not and you would taste that simply with statistical significance test you know of whether let's say something like all the payoffs in the second session statistically significantly higher than the payoffs in the first session and you know you look at the data you see that what actually happens is that people play most of the time the efficient outcomes okay in the other case they don't so that's that's what I can say okay now new example when t2 asked me to give this lecture I asked myself hmm did I do any experiment on leadership and I had to honestly at first say no I didn't but then in retrospect I started thinking hey I have this paper that I did that I can actually interpret in retrospect as concerning a feature of leadership so I decided that's what I will tell you about okay so I have a paper in Econometrica 2006 with Gary Jonas which is an experimental test of theory which you know related to theory that's developed by me and Parvati gali who's my colleague at pukani and you can see my co-authors on the picture there and also me okay so now part two of the game theory lecture you learnt about game matrices now you will learn about game trees there are basically two ways that game theorists describe strategic situations where people interact and influence each other you've learnt about one you will now learn about the other so game trees game trees are sort of explicit on the order in which players move so here is a situation which you can interpret as an organization and I'm gonna loosely interpret player one as a leader okay so think of a as sort of the firm or the boss or someone who's running on organization okay and the question for the this person is whether to hire or not a worker okay so we look at interaction between this boss and the worker okay and the boss starts the game so the game starts at the top game trees at least people draw them differently I like my game trees to grow from top down sorry about that okay so they are Babylonian game trees you know if you don't hire or higher now the numbers at the end of the game tree are gonna depict payoffs okay so let's assume that if you don't hire then they get five each okay if the boss hires then the outcome actually depends on the worker or the hired person B B can again choose whether to put in effort or not to put in effort okay and this time we're gonna have an effort cost of four so the payoff is going to be fourteen if it doesn't exert effort and ten otherwise so it's costly to exert effort why would you want to do it well because it's actually beneficial to the boss okay so if you don't I'm writing always the bosses pay off on top and the workers pay off below so if you if you are hired and you exert no I thought you get a fine pay of a fourteen but the boss boss gets nothing on the other hand if you're willing to spend your four dollars worth of effort yes it costs you four but the other guys gains ten so you get ten each you know consider this compare that to what happens if you don't hire the guy then you get five each there are gains from trade this is the simplest model you could imagine to illustrate a situation with potential gains and trade between a boss and his worker right but there is tension much like in the prisoner's dilemma except now is played out sequentially because you know clearly you have to rely on the guy to - to exert a fort right and if it doesn't you you're in trouble right okay now you wonder what is this store going on okay and I could if I had hours I could talk about the store so they're probably about the galley and I we have this so a large part of my research is concerned with the following okay so economists throughout the decades have used maybe let's say more they've been more inclined to use mathematical methods and many other folks in other neighboring social sciences but they have done it at the cost of making rather simplified assumptions about human nature like you know everyone is maximizing profit ISA standard assumption where you just look at the dollars so much of what I do is trying to incorporate more richer human element into the analysis so emotions caring about you know getting even being kind to people who are kind to you and things like that right and the store I put in here to indicate something about that so imagine that you don't as a player you wouldn't only maximize you as a worker you wouldn't only look at how much you gain but you think of the other fellow then you know here is his boss he's offering me a contract and he's saying look Martin we can both gain and get 10 instead of 5 right I would feel awful if I if I don't work hard right and get the 14 and he gets zero so the the store is my bad conscience okay now the quake I wrote the store because I don't have time to to give you details for the store might be but I want you to it's read because it's sort of emotional okay and what care about the Galleon I do is we develop a seer a general theory of how people might feel guilt okay and it follows from the scholarship of many good psychologists that people tend to feel guilt when they hurt relationship partners and that depends on the relationship partners believes so if someone you know imagine that these fellows look one another in the eye and and you know you know I promise that I'll exert effort right so he manages to shape the beliefs of the other guy then he would feel terrible for not exerting the effort okay so in the theory if I were to discuss it in detail we would have this would be a mathematical formula and it depends it describes the beliefs of player B the second player about the belief of player a about the probability that player B will exert effort okay okay so that's the store now add to this situation the fact that people in organizations talk to each other right they do things like what I said they shake hands they look one another in the eye they say look I realize that there are trust issues here where you know you might think as both that I would not exert effort but hey buddy I promise you you hire me and I'll work hard right and maybe that breeds a self-fulfilling circle of beliefs about belief so that people actually do this yes promises in particular promises in particular may play a role in the psychologists have documented this in experiments way back but you know they didn't connect it specifically that we like to operate with okay so here is the idea you take the situation and you know it may be good in itself just maybe you have a big store to start with but the store will become bigger if you let people to work okay now I told you another story so again an experiment might shed light on it let's see how I'm doing on time I think I'm still doing okay right yes not that many slides left okay so the idea is now to run an experiment so we invited subjects to the lab he described exactly this drug this structure where they move in sequence the first fellow gets to play the boss and he can hire or not and the other fellow gets to subsequently choose no way for Tory effort if he's hired okay we had treatments and the treatments concerned whether or not we had communication in the picture so in one case you know we did this game without communication in the other one we had this pre play message opportunity where actually be could send the message to a okay so in particular B could send a promise you know before they started playing about how he would behave okay then we measured the store I'm not going to go into details but remember I said mathematically the store is describing beliefs about beliefs about behavior okay in the lab what we did was operate with guesses so we did things like tailing player a hey you matched with one of the fellows in the other room as your fellow B now we asked you to guess you know what percentage of the fellows in the other game in the other room choose effort right so we get data if you will and we rewarded them monetarily for accuracy in guesswork so we get data on player s beliefs then we went to play a B and said hey we just asked the other fellow to guess what you are doing now we asked you to guess the guests and the again we asked we reward according to accuracy in the guesswork so the guest guesses become data that we can use to test the theory okay and then we have research hypothesis which event essentially say that if you have words in the picture then the stars will become higher people will become more emotional and you know worry more about not living up to the other fella with expectations and that will boost hiring and effort so you get good outcomes matter well promises right so we can sort of tell a consistent story by cooperation okay so here is just to give you a little bit of flavor of its sample message so they got a sheet play B got a sheet of paper and it says you may print the message to a below if you wish let's look only at the second half of the message you know this fellow writes my country tis of thee sweet land of liberty of thee I sing land where my fathers died land of the pilgrims pride on every mountainside let freedom ring george w bush wants you to go in bin laden's as out okay i write in and out cheated a bit on the slides just to make it pedagogical i say a fourteen no effort in the experiment and we had higher and not higher in the experiment these things were described in different words they were out and in the choices but otherwise it was essentially the same game okay so you can see he's talking about the game he's trying to make the other fellow go in and further up he's making some kind of promise regarding how he would behave we had many funny messages like this anyway about the results so the hypothesis again amounts to words boosting the store boosting hiring an effort choices and largely it was poor supported again okay so just to give you a snapshot of the data if i if i make this little figure here in the no communication treatment you know this histogram would show you the blue bars or the frequency of hiring decisions the green bars or the frequency of effort decisions high effort and the red bars or the frequencies of outcomes all the way down here the good one right okay and then i do the same thing for the the treatment with the messages for b communicate and the point is that everything shoots up okay for the words help foster trust and cooperation okay so we're coming to the words the close and you know then you have an opportunity to ask me questions and you can stick to the theories we said or you can ask more generally about anything about game theory experiments or psychology and economics but let me wrap up first so in this second story you know the leadership spin would be that the leader may foster trust and cooperation by offering an employee in this case player be a communication of tunity once you give people voice promises will be issued and kept issued and kept so that helps foster trust and cooperation so the story involves words moving beliefs and the a guilt is influences choices I'll end on a on a cautionary note here because you know so I told you one story why you might get people to be truthful and trustworthy yes and not to lie and so forth so can we can we imagine some other reason why people tell the truth to foster trust and cooperation it turns out I found an answer in Rome when I was there recently and I can't deny that to you it's gonna be my last slide so it's coming here okay thanks for your attention okay there are three persons requiring the word yes first is a professor beer NER penultimate slide says promises issued and kept and of course the kept it's a large field of problems did you do experiments on a repetition of the game and if you did so to what extent did the results differ from let's say the more traditional repeated prisoner's dilemma approach okay so I'll repeat the question because I think maybe the microphone wasn't super did everyone hear the question or so the the question concerns whether we repeated this experiment whether it was done once and if repeated how did you know that matter the thing is we did not repeat okay so economists have asked in many ways the general question why do people cooperate and you know things that have been highlighted concerned well you can write a binding contract so that you will pie punished if you don't cooperate but you know that has downsides because maybe it's costly or economy is another common line which is implicit in your question is well many of the relationships we have all repeated so you wouldn't want to take advantage of someone for short-run gain because they would punish you in the future right and they'll established theories about both these things when we came to this we came from the viewpoint of thinking hey these feelings of guilt and the psychology might be let's say a third reason right and in order to to highlight that we actually wanted to distract not to disregard the other two but to abstract away from them so the only way to abstract away from them is to have a one-shot setting so that's what we did right and then we told our story for the one-shot case so I don't have repeated game data for that reason next question thank you very inspiring I'm not an economist so just to put in a practical way when you talk about emotion and behavior economics I would like to know whether how you're going to shift from the theory of games into the practical behavior of people managing and leading teams so my question is how you're going to quantify the behavior and the beliefs in the first place and the second is Maura you talk a lot about the efficiency so we know now that emotions are drivers of motivations so the question is how are you going to measure in the future the impact of emotions and the growth of efficiency okay so I'll start with the second question because I think I have better on more I have an answer to that one so I started out talking about efficiency yes and I was sort of highlighting what was and what was not efficient and implicitly I was referring only to the monetary payoff right and which is what economists traditionally have done right you're very right in pointing out that now if I'm starting to play around with what actually motivates people okay then you know we should measure also efficiency differently because the new ways to motivate people might be part of what is efficient and what is not okay and I think it's fair to say that this is a tricky issue which I don't have full answers to what to do exactly although I feel very strongly in specific examples I mean in certainly in this example you know there is no tension between the two they attend each right certainly 10 each would be considered a better outcome also taking emotions into account than five each and probably the outcome where one of the fellow gets 14 and the other one gets 0 is certainly a terrible outcome for the one getting 0 you would imagine so I think it's it's an important question and there may be other situations where it can be hard to to look at - these are tricky issues ok I think you can do two steps you should first look at the sort of the standard economists way of doing it because that will be give you some insight if everyone is better off in terms of monetary outcomes that says something it doesn't say the whole picture so you need to be mindful yes and the other question how to quantify and put this into practice you know I I guess I'm not sure I I mean I took away a little bit here I said you know give your employees an opportunity to work I mean I was trying to take something away you have to know I think it's fair to say that you know attempts to economists used to be skeptical to incorporating psychology and economic analysis they would say things like yes it matters but it's second-order effects and you know we are fine to abstract away from it it's probably noise I think recently people have started to argue that taking emotions into account will actually change outcome systematically but it's a new trend ok and much work lies in the future I really appreciate your work and I think is a very important work so don't don't take my question is something of septa sysm but if you use only student in doing this because the question is what about if you don't use student and what about if you use student from Asia or from Latin America because the question is there very often this sort of theory confirmation and by the way the question of probability is very important every time you do this kind of experiments because significance is very important and and variability is very large so you have to be sure that then you have a sort of reasonable quantification but again the most important question for me is if you have introduced enough cultural and social variability in your experiments to say that the the trend is the trend of your theory yeah so I'm going to answer by first apologizing and then defending ok the the apologizing part is that you know imagine researchers who sit in universities and want to taste stuff in the lab you know it's very convenient to use students okay students are first of all you know if you worry about stakes in experiments it's pretty good you students they don't have super much money so you know if you pay them a ten fifteen euros so that's probably more to them than it would be to to sort another subject so you get fairly motivated students which is nice you know if you're testing theories that imply sort of complicated reasoning you know students are pretty smart so you know it's it's it's you don't worry so much about well you know for many purposes they are very good subjects okay for some purposes of course they are not representative and and one needs to be mindful of that however then you also have to knowledge that there is a big chunk of work done outside the D lab so you know using experimental game similar to what I've described you know for instance anthropologists in the last ten to fifteen years have really starting to take the tools of experimental economies to heart and go out in the different parts of the world and and compare cultural differences you know and you can take a game like the one I have in the lost for instance which is called the trust game you know because player a has to trust player B to do things and you can really compare trust levels in different cultures so some of this work is done let me say one more thing like for instance there is a literature which I haven't talked about but on pricing in financial markets okay financial economists debate about how we should understand price movements in in in in the real world you know should we think of manias and panics or could we have more sort of fundamental theory to develop by economist to describe it there is a heavy debate and it will never finish okay because no one observes the fundamental values in reality well in the lab you can actually control fundamental values because you can define the dividend structure and have a market and so on so you can test these theories in a good way and people have done it on students and you get certain results now in this connection many scholars or commentators wondered you know okay you get these results with students would you get them in with professionals okay so Vernon Smith you may have heard his name if you know one experimental economist it should be he may go to Nobel Prize the moral price in 2002 for his work in experimental economics and he was starting this series of experiments testing where the bubbles occur so you know excessive prices relative to fundamentals in in in markets and they do with students you get huge bubbles and at that point someone said okay but these are students would it happen with professional traders so he actually took the effort and went down to some Pacific Stock Exchange whatever it was and invited traders to his lab right and cranked up the stakes I think to make them happy and played and they made even bigger bubbles so you know that's one example of so people are taking these concerns into question and sometimes they are more important than others but yes are there other questions if not can I add my went one question from my side I think that you are right when you say that economists did not care too much about leadership in the literature you started by saying this at the beginning I think that I think that you're right totally right but at the same time I believe that some aspects of leadership are in some way nested in some economic phenomena studied also by game theory for example you cannot have a leadership without having some form of your article structure you know you need a boss but the boss is something you know at the top of the hierarchy even if it is not a formal hierarchy but nevertheless to have leadership you need this okay now if you think to some games very common commonly used in behavior in experimental economics like the ultimatum game or the investment game where there is some form of sequentiality in the game itself well in that case you can see some form of hidden hierarchy meaning that for example in the Ewell team Adam game the ultimatum game is a very simple game it works in this way there are two players and the one is the proposer and the other one is the receiver the proposer isn't down at by getting money he receives money for example 10 euros and he has to decide how much to send to the to the receiver of this amount of money some X and then what the receiver can do is only to accept or reject if he or she rejects then both the players lose the money not nothing they get nothing zero if he assets of course they share that the pine the way that the proposal decided to share okay this kind of game with which is a sequential can in some form of hierarchical relationship so you can imagine the proposer like a leader in a sense you understand what I mean do you think that this could be a way to look at the leadership issue at least from this perspective of the use of power and yerkey within games or it is not something that can be studied of the top of my head I'm first not sure how to answer but then I get the following Association so you tell me if it's related to what you said okay so I'll come back to these Anthropology experiments I mentioned it used to be that the game Luigi mentions the ultimatum bargaining game people used to say you can and this is coming also to the student question okay you pick students anywhere in the world and you can play with large stakes or small stakes you get the same outcome basically lowball offers get rejected and that's actually anticipated to people over a fair chunk of money often half of the pie okay they don't try to take advantage of the others and they shouldn't because they would be punished by rejections okay that used to be the case that people thought this was a universal pattern until the anthropologists came and started playing the game in different cultures okay then you do get very different outcomes actually that somehow from an anthropologist point of view I think they argue are very you know useful for understanding the culture now of course in this here comes leadership connection and what I'm shooting from the hip okay you would imagine that maybe you have different leaders in this societies maybe they are matriarchy call or patriarchal in different cases and if you know maybe suppose you compare matriarchy : patriarchal societies maybe what you said would have implications for different outcomes depending on whether the first mover is a man or a woman in these societies but I don't know is that a useful angle Thank You Lou geez thanks everyone

# {{section.title}}

##### {{ item.title }}

{{ item.subtitle }}