People are too obsessed with measuring things perfectly.
Dr. Roberto Rigobon is Professor of Applied Economics at MIT Sloan. We discussed his theory of how we measure the wrong things at the wrong time.
Podcast Name: Masters of Data
Episode Name: How to Measure Ethical Behavior (Guest: Roberto Rigobon)
Ben: Welcome to the Masters of Data podcast. The podcast where we talk about how data affects our businesses and our lives, and we talk to the people on the front lines of the data revolution. And I'm your host Ben Newton. Our guest today was a ton of fun to interview. Doctor Roberto Rigobon is a professor of applied economics at the MIT Sloan School of Management. He has been teaching and advising students at Sloan for over 20 years and he has won both the Teacher of the Year award and the Excellence in Teaching award at MIT three times. He has an exuberant personality, and made an interview effortless with his stories and analogies. We sit down in his office and talked about his theory of aggregate confusion, and how we measure the wrong things at the wrong time. So without any further ado, lets get started.
Ben: Thanks everybody for tuning in to the Masters of Data podcast. And I'm excited to be here with Doctor Roberto Rigobon, thanks for coming on this show.
Roberto: Oh no, thank you. Thank you for the invitation.
Ben: Doctor Rigobon is a Professor of Applied Economics at the Sloan School of Management at MIT. So I'm sitting here with him in his office and I like to start just talking about peoples background, just to humanize you a little bit. I mean from what I've read, you have a really interesting background. So how'd you end up at MIT at Sloan?
Roberto: First, so I was born in Venezuela, and I studied there. And I went to the university to study chemistry.
Ben: Oh really?
Roberto: Yup. Yup. But my best friend decided that he wanted to study electrical engineer, and then I didn't want to lose my best friend, so I just went and I studied electrical engineer. When I graduated I actually worked for only one day, literally one day. I said, “Crap, I cannot do this anymore.” So as a typical engineer-
Ben: Its good you figured that out quickly.
Roberto: Yeah, it was super quick. Actually after one day I knew I had to find another job. So I moved immediately to management, as almost all the engineers do. Then after several years I did an MBA, in Venezuela, and that's when I met economics. So I fell in love relatively late, I mean most people, when they do PhD's they tend to do that right after college. In my case, I worked for five years in the pharmaceutical industry, I did an MBA, and that's when I decided to do a PhD. I came as a student to MIT, and then I stayed in the building. I think I stayed at MIT now 25 years.
Roberto: And even my sabbaticals, I did them here. This is how much I love the institution. I think the Dean is thinking about reducing my salary now because he knows there's no way I'm leaving this place.
Ben: That's really interesting. So you teach economics here at the Sloan school, so what are you focused on in terms of your teaching and your research?
Roberto: I teach MBA, so I teach kind of a basic macro, which is very policy oriented.
Roberto: So try to think about how political events and economic outcomes, how they interact with each other, what are the limitations? I mean sometimes something that we would like to do on the economics is very hard to implement socially and politically. It's unacceptable for example, that's very typical. And what is happening now in the developed world is that we're becoming more and more like an emerging market if you think about it. The problems that emerging markets used to have about 20 years ago, all the developed nations are having it right now.
Ben: Yeah, yeah.
Roberto: So populism and things like that, that are effecting tremendously all developed nations. So it has become an interesting world for me. Because I used to differentiate them, and I call this period the emergisation of the world. So that's what I teach, and my research is complimentary to that because my research is about measurement. I try to improve how we measure all our life, and this is true for everything from personal aspects all the way to discrimination, to GDP, to drop consumption. So my work is in trying to think about how to measure things better.
Ben: That's really interesting 'cause I, I've talked to a couple of people already about bias in data, and data is so deeply tied into our world today and how we make decisions.
Ben: So how you measure it, and being aware of that, is massively important.
Roberto: Well, and actually I think we are lousy at measuring. I mean I think, I mean think about all the important events in life, we tend to measure super late. We only concentrate on extreme behavior. We tend to measure very infrequently, so it's hard to make any decisions. We measure a lot of perception, not facts. And then we concentrate on the wrong statistics. So five for five, this is the worst that you can imagine. So to me it's actually a miracle that we make a good decision.
Ben: When you and I were on the phone earlier, you had talked about some stuff you're looking at recently. You called it aggregate confusion.
Roberto: Confusion. Yes.
Ben: Yes. Talk to me a little bit more about that. What does that mean?
Roberto: So the project of aggregate confusion is to try to think about how we measure ethical behavior. There are many data providers in the world, about, I would say more than 200.
Ben: And when you say ethical behavior, what does that mean?
Roberto: Okay, so this goes from how do you, do you pollute or not in the environment? Do you treat women correctly in the labor force? Do you treat your labor force correctly? Do you use child labor?
Ben: Got it.
Roberto: I mean so firms make many, many decisions and what they're trying to do is to understand dimensions about your ethical behavior. And what I mean ethical behavior, is because we're trying to measure things that we know that are morally incorrect. To pollute the environment we think is morally incorrect, to overuse water is morally incorrect, to lie about your marketing. But this is really difficult to measure or prove, no?
Roberto: So I mean its profits we measure relatively well compared to how you treat your workers.
Roberto: I mean, so that is why they separate the two. For example your willingness to pay, well that would be a credit rating agency, and that's about an economic outcome. And we think that we have a pretty good idea how to do that, but your willingness to pollute, we have no idea how to do that. So we group all of this type of measures in what is called ESG, Environment Social and Corporate social responsibility. I call them ethical.
Ben: Okay, got it.
Roberto: You see?
Roberto: That's why, kind of I put, but it's actually kind of my personalized. By the way, very few people-
Ben: Oh I like that.
Roberto: Agree with me.
Ben: About calling it ethical?
Roberto: Yeah. Although when I look at the categories that you are measuring, they're all about ethical behavior and all of them-
Ben: No it makes sense.
Roberto: Have some moral aspect. So I mean, you should not pollute. I mean you're not violating any law in particular, I mean you should pollute less.
Ben: Yeah, yeah.
Roberto: So you should use less plastic. I mean I'm not violating any laws. So this is the problem of these measurements that you are trying to think about a way that you should behave, without any particular regulation. So I mean, you should treat women nicely. There should be no reason why you have only 20% women in management for example.
Roberto: Or even less. But you're not violating any law by not having them. Only there's a proven discrimination. You see?
Ben: Yeah, yeah.
Roberto: So what occurs here is that because it's such a gray area, what we have is that we have a lot of moral judgments without any evidence, and then we have no idea how to react to whatever data we produce. And therefore we tend to, that's what I mean that we tend to measure very late. I mean this happens in may areas, from drug consumption to discrimination. All of that will have the same feature.
Ben: Okay. And then the aggregate confusion research is like you're researching how to correct that? Like how to-
Roberto: So there are two parts. One is to try to understand why the measures that exist are so inconsistent.
Roberto: One is to try to understand, can we improve some of the measurement by using what we're already doing? That's kind of the first part. The second part is in those areas where we think that we are not measuring that correctly, that we actually provide new measures. Let me give you an example about a measure that we do badly that I think that most of your listeners will understand in 10 seconds.
Roberto: So in the United States we have an opioid crisis. A massive opioid crisis all over the place. How do we measure that? Well we only count ... And by the way we love counting. I mean it's so interesting that we just love counting.
Ben: You can understand it.
Roberto: Yeah, I think so. And we started doing that seven thousand years ago, and we have not stopped counting.
Ben: Its like, "Wow I can use my fingers."
Roberto: Exactly. So we count the number of people that either have an overdose, have to be revived, or enter into the Emergency Room. So that's more or less how we measure the drug consumption. So how do we know we have an opioid crisis? Because we have too many people entering the Emergency Room with an overdose. Now if you think about it, first you're measuring, you're not measuring the consumption of opioids, you're measuring the excessive consumption.
Ben: Mm-hmm (affirmative).
Roberto: So in other words, you're only looking at the people that are consuming in excess, not necessarily everybody that is consuming. And you're making an assumption that the people that are consuming in excess somehow behave the same as the people that are just consuming moderate amounts. And that's not necessarily right. So one is, you're measuring extreme behavior in the sense that the guys overdose. Second, this is by definition late, the guy just had an overdoes.
Ben: Mm-hmm (affirmative).
Roberto: So I mean like, I mean so-
Ben: You didn't prevent anything?
Roberto: You prevented ... Exactly. You think because you measure that relatively well, the number of people that enter an Emergency Room, I mean it's a very well established, understood statistic.
Roberto: Now we start concentrating on that, as opposed to concentrate on the whole distribution of behavior. I mean if I want to understand the opioid consumption, I should start with the guy that is consuming a little bit. I should not wait until the person has an extreme event to actually make an action, take an action or do something else.
Ben: I expect that's a lot harder.
Roberto: Yes, of course. Of course. Because counting people in an Emergency Room is easy. I mean, figuring out how people are consuming in their home is way harder. I understand this exactly, that's what I mean that we tend to concentrate on the wrong statistic. It's because the statistic is so easy to understand and to compute-
Ben: The ones that are immediately available.
Roberto: Then we started doing that, and therefore we pay attention to that as opposed to anything else. And then basically we measure not only late, but we measure very infrequently. So it's like, it's a moment where it's almost like regret, because it happens in the last three months. So it's like, I don't have the data updated fast enough so I can take an action. A lot of our reaction, and public reaction, is to the public perception of the problem.
Ben: Oh, yeah.
Roberto: So let me give, so this continues to be a problem in opioid. But let's think about a different problem like refugees from Syria in Europe. I mean have you, lately no one is even talking about it.
Ben: That's true.
Roberto: How many-
Ben: That's true, it's kind of out of the news. Yeah.
Roberto: It's all off the news. But why it made it to the news? Do you remember when these little baby was found?
Roberto: Okay, okay. Well it was found on the beach, that outraged all Europe.
Ben: They had a visual connection to it.
Roberto: Exactly. And now we are emotionally very attached to the problem, we count it, again we are counting, how many people are escaping or dying in the ocean. But after that impression stops, we stop paying attention. We are not outraged by the numbers. Today, by the way, according to the statistics, there are more people dying in the ocean than before. But it's jut marginally worse. So imagine we have a 100 per week lets say, what was it, all the statistic, it would be 110. Well it's just 10% more. I mean, so people are not outraged by the number, we just take it normally.
Roberto: And therefore we have a statistics we are measuring clearly too late. We are measuring very infrequently. But now we are not even outraged, so we are not even paying attention.
Ben: People usually don't get outraged by statistics, right?
Roberto: They don't.
Ben: They get outraged by the images-
Roberto: By the visual.
Ben: By the emotional connection.
Roberto: Yup, you're absolutely correct. You're correct. So when you think about all of that, for example about three years ago, maybe I have this wrong, we had all these scandals in collages of sexual harassment. Have you heard anything this year? I actually heard one news only, related to that.
Ben: You're right, I haven't heard as much.
Roberto: There was one news about a judge, you remember the Stanford case?
Ben: Mm-hmm (affirmative).
Roberto: That the judge was too lenient? The only news that I have heard, do we know this year about colleges and sexual harassment, is that that guy was kind of let go. That's it. In fact, I really don't know if it's true. It is like, this is how little it is. So there are two possibilities. One, we have solved the problem and there's no more sexual abuse in US campuses.
Ben: Probably not.
Roberto: Or we are not paying attention. And that's, you see, that's what I mean. It is a miracle that we make good decisions. We measure too late, too infrequently. We pay attention to the wrong statistics, we are a lot based on perception, and what has outraged us at the moment. And by the way, the Hollywood intensity has slowed down. I mean you realize?
Ben: How do you mean?
Roberto: In the sense that we were outraged, about a year ago, this was in the new all the time about sexual harassment in Hollywood.
Ben: Oh. Right, right, right. Yeah.
Roberto: Which is what I think took over after colleges, and I don't know what is happening, but a new one will take over soon.
Roberto: It will be maybe in politics, whatever. I mean, so we don't know.
Ben: Well I guess it's hard to maintain attention on these things too because-
Ben: People get, they get inured to it. They get overwhelmed and-
Roberto: But again, so by only measuring the moments when you are outraged, you make very emotional decisions, very short term decisions. And we are not actually improving the society at all.
Ben: So how do you go about changing that? It makes sense when you think, I mean how do you change it?
Roberto: So I mean, first I think that what I would like to do is to try to measure the process before we get to these extreme events.
Ben: Mm-hmm (affirmative).
Roberto: So it's a very good question. I don't know if I can, if I have an answer about how to make people aware of the problems that we have in such a way that we can all, as a society make better decisions. So I don't know that second part. I do know that if we measure for example, treatment of the labor force, before people need to go to a court to file a complaint. If we measure that earlier, we might have a chance to stop the abusive behavior.
Roberto: So in other words, think about it, I am an organization, and I think that actually most firms and organizations in the world are kind of good. They would like to be able to treat their workers nicely.
Roberto: So if they knew how costly their decisions are in the livelihood of their workers, they might be willing to change. So the way to change it is not necessarily to make, I mean the whole 360 million people in the US, aware of the problem. It's much better to make each firm aware of their own problems, so each one can take an action. And therefore a problem that we are unaware, and we cannot have the capacity to keep attention on the media by measuring individually, we might be able to change the behavior. Let me give you an example. You have an Apple watch there?
Roberto: The day you get one of those guys, you start measuring how many steps you take. You start measuring your heart rate at night.
Ben: Yeah. Sure.
Roberto: I mean you start measuring how many hours you sleep. And guess what? That automatically changes your behavior, for the better.
Ben: Yes. Yes.
Roberto: These are not measuring your health. I mean, no way. I mean no way that will tell you if you're healthy or not.
Roberto: But it is related. It's an imperfect measure of something relevant. And that's beautiful. When you think about it, you're imperfectly measuring something relevant. As opposed to the way we do it with the doctors. You go to the doctor once a year. I mean you're young, so maybe you will go once every two years. But I hope that you go once a year to your doctor.
Roberto: But when you go to the doctor only once a year, yeah, they are measuring something perfectly, which is your blood.
Ben: But only that one time.
Roberto: It's only one time. It's too late, too infrequently. And by the way, at that moment it's only regret. What do you want me to do in the previous year that I did nothing? I did not exercise enough. Yeah, I know. So what do you want me to do? I mean this, you measure so infrequently that it has no impact on your behavior. But that little watch, man that has an amazing impact. Because if your target is, lets say 10 thousand steps, and it's actually lets say, five pm and you only have walked for five thousand. Guess what, you start walking like crazy. You know?
Roberto: So I think that that should be the approach. By allowing people and organizations to measure themselves better, to make that transparent to them-
Ben: But even if it's imperfectly I guess-
Roberto: Even if it's imperfect. I think we actually can change behavior. Exactly. So the phrase will be like, if by perfectly measuring something irrelevant, is much worse than imperfectly measuring something that is relevant.
Ben: Okay. I like that. Yeah, that makes a lot of sense.
Roberto: So going back to the opioid. I think we perfectly measure the number of kids that enter an Emergency Room to be revived. Because we don't make a mistake on the diagnostic, all the hospitals report that. So all the EMT's they will report that. So we have a perfect measure, or something that I consider difficult to take an action. It's too late. No, I'm not saying that this is irrelevant, it's that from the decisions point of view, it's hard for me tot take an action. It is too late, and it's not easy for me to understand that my actions will have an impact on that behavior.
Roberto: So imagine we have a fantastic educational program in the city of Boston, fantastic. I mean that it is super effective. That changes the consumption of opioids of 90% of the population. It just happens to be that that 90 was not the extreme.
Ben: I see what you're saying.
Roberto: So guess what happens? We do our educational program, and there's the same amount of kids that enter an Emergency Room. It has not changed because that 10%, we didn't reach them. I mean we cannot have a program that is perfect.
Roberto: So we did not reach all of these kids, and therefore in the statistics we have done nothing. Although we have improved the life of 90% of-
Ben: 'Cause you don't even know if you're measuring the same thing?
Roberto: Exactly. Because I'm not measuring, I have no idea. And then we are fixated on the number of kids that enter the Emergency Room, so this program that was useful, we eliminated because we have no evidence of success.
Ben: You know one thing that comes to mind when you say that, do you think that this happens less because people are afraid of measuring imperfectly and using things that aren't, you know?
Roberto: That's a good point. It is conceivable. You're right. I never thought about that, but you're right. It is the case that politicians would be very afraid to take actions on probabilistic statements.
Roberto: So because measuring imperfectly, what it really means, is that you have a probability of making a mistake.
Roberto: We tend to prefer something that we can count, which is much easier to understand than a probability. And that we can visibly tell people, "This is my outcome."
Roberto: So we don't measure how upset people are in the traffic, which they're super outraged.
Roberto: We only measure when this behavior goes to the extreme, when now you have a dangerous situation, and you cause an accident. But again, it's because it's hard to measure, it's hard to explain how we measure, then it's hard to take an action on something that we don't understand very well.
Ben: You know, one of the things that I was thinking about too, and I've seen this come up in the business world too, is that I think there's maybe a sense that lets say that I'm the social worker dealing directly with the people that are struggling with opioid addiction. Or I'm in a business, resources, whatever. Do I feel comfortable measuring, even intuitively, what's going on and saying, “I'm seeing this.” And reporting it up the chain because it's not, I don't feel it's a perfect measurement, so I hold that back and I don't communicate it.
Roberto: That does happen, yes. I see that.
Ben: Yeah, 'cause I mean even, 'cause I would even think like in a business setting, I think when you're at the executive level you get, you don't see these things. Even the imperfect measurements. Because people don't communicate it to you, so you can be blind to what's going on in your own company.
Roberto: I agree with you entirely. So for example, I can assure you that the psychologists within firms that are dealing with coaching their own staff. Okay? They will know, for example, if someone is being a jerk in a meeting. Or you could say, “I feel that they are passing over me because I'm a lady.” Or “They are passing over a promotion where I feel that people are treating me badly because I'm gay or because I'm black.” I mean they can actually, that individual in the organization will have that piece of information. And the reason is because they are coaches and therefore they come to you.
Roberto: The question is, once that happens, how do we transfer that information to the organization? That is an assessment. It's a very imperfect assessment. Again, it's very late because the person is coming to a coach, already with a situation where they feel that it's bad enough that they are willing to say something. But my point is, that should be actionable. And the question is how do we treat those conversations? If we treat that conversation as an outlier, then we tend not to report it. Or it gets lost in the statistics. You see? And therefore you are creating a massive moral issue, and you're not realizing that this is an extreme event.
Roberto: So for someone to come and complain about some form of discrimination in the labor force, that means that there's a lot of discrimination that was not severe enough, that decided the person to go and file a complaint. So when you get to that point, the question is not to investigate the outlier. You see? It's to investigate everybody else that is not telling you anything. And I don't think we do that. I think we only report the outlier, most organization will do that. Again it's the hardest statistics, "Well I have one complaint about labor practices." And then they, hopefully they do something about that. But I think most organizations just report that and they forget it.
Ben: Yeah, you know, for whatever reason it brings to mind, yeah I've heard statistics about even customer support organizations When you think your customers are complaining about something, it's usually for every complaint you get there's dozens of people that had the same problem but didn't complain.
Roberto: Did not complain, exactly.
Ben: Because they didn't reach the level of frustration with them. So you're not actually measuring the real impact. You're just measuring the impact of the people that got frustrated enough to say something.
Roberto: Exactly. I mean, but think about your reviews on restaurants. How many times have you been in a restaurant that you're not satisfied, but you did not give them a bad review? Did you not even bother to go to Yelp and give them like minus 20.
Ben: Yeah, yeah, yeah.
Roberto: See how, minus five stars if you could. And the reason why you don't do that is because it requires, I mean it's an intrusive, it's effort on your side. You're already upset.
Roberto: Now you will go when you cross that line. That means that in your personal life, think about how many times you have not reported dissatisfaction. And I'm not saying that you should be whining all the time. I mean, okay, so.
Ben: Yelp used right.
Roberto: I'm not advocating whining, a whiny society. I'm just providing you examples that even those reviews are extreme outcomes. And therefore firms, when they look at those extreme outcomes, they should think that there's a lot of marginal outcomes that are almost equally bad, but did not cross that line to become a complaint. So that's why it's hard to measure, because you don't have any form of reporting. So you have to understand the processes, how the process of discrimination takes place, how the process of having a hostile environment takes place and so on. So this, and again, they will be very imperfect measures, which is your first point.
Ben: That reminds me of another thing I was talking to another guest that mentioned that in airlines, pilots and stewardesses and anyone that works for the airlines can report mistakes when planes fly. Anonymously. And they will never be punished for it. Like a pilot can actually say, “I fell asleep when I was flying.” When obviously he's not supposed to. But he can report that. They can also report it to their airline officially, but they can report it through this body. And I'm forgetting what the name of the body is. But there's-
Roberto: The FAA? No.
Ben: Well I think it may be related to the FAA.
Ben: But it's like-
Roberto: I don't know. Yeah.
Ben: Yeah, but it's really, it came to mind 'cause it's really interesting. What they did is they had anonymous reporting. So this has been going on for like 30 plus years now, where it's actually significantly reduced the issues with flying, and made it much safer, because what they've done is they've identified problems that are happening.
Roberto: Exactly. Exactly.
Ben: And they've done it in a way that its not, I guess to lower the threshold for reporting.
Roberto: Exactly. So have you realized how interesting it is, our social interpretation of a whistleblower versus ratting your colleagues?
Ben: I hadn't thought about that.
Roberto: Now it's the same.
Ben: What's the perception?
Roberto: Yeah, but when I said, “You're a rat.” That doesn't sound like you're a nice person. No, you're ratting your colleagues means that you're a miserable bastard. You deserve what is coming to you.
Ben: You can get shot in a back ally.
Roberto: But whistleblower's, you assume that is somebody taking a personal sacrifice.
Ben: He's heroic.
Roberto: He's making a personal sacrifice, he's heroic, exactly. He's a hero. You're right. So one way to deal with a lot of moral issues is to have a much better system of whistleblower where it's not considered ratting. I think that having a system of whistleblower, like the one that you described with the airplanes, is a good system to have. And again, you want small events that could be disturbing for one person or the other to be actually shared. I think that a lot of those systems should be used, for example, to avoid the culture of a particular organization.
Roberto: As opposed to do law enforcement and try to catch individuals. And I think that the reason why the system will work that way is you want to provide a measure for the whole organization in terms of how are we treating people? How are we treating our workers? How do we treat management versus the workers in assembly? How do we treat people in the warehouse? How do we treat, you know, drivers from Uber? And so on. So the idea of that system is just to measure treatment, and therefore you want the tiny things to be reflected in those systems. That eliminates a little bit of the snitching or the ratting the other.
Roberto: So in the sense it will be about just providing information about behavior. That also will lead, kind of the example I was saying, that two people in an organization might have very different cultures, and therefore what one person might consider inappropriate, the other one will be fine with it. But if we have no way to communicate those differences, then what do you do is you just repeat the situation until it explodes. You know?
Roberto: So either you never want to have lunch with me, which I find it insulting because as a Latin American how can you not have lunch with me? Or you are actually having lunch with me, but actually you don't want to have lunch because for you actually lunch means something different. And if you find it strange that I want to have lunch with you all the time. So it's interesting because actually I see that here in MIT. Different groups have different patterns of behavior. Some of them have lunch every Friday. Some of them have lunch every day. We are economists, we hate each other. We never have lunch. So in some sense, you see organizations should be able to adapt to those cultural differences, and if you have communication that you can highlight what bothers someone, you can make better decisions.
Roberto: Now the systems, I think they are good. And they are necessary. But I don't think they're enough. There's something else that is kind of your personal feedback. How are you behaving? So that's a tougher one, and again in a lot of this world of imperfect measures, I think that it can be done. By the way, kind of my research is truly to think about that, that way. But again, you have your Apple watch, and you interrupted a lot of people in a meeting, and the Apple watch said, "You interrupted a lot today." And I think that that's fantastic feedback. There's no judgment in the feedback, you will make your own judgment. Again, my example when I explained this to the students, I said, "I am Latin American, okay? And I cannot live without interruptions. I'm sorry. This is part of who I am."
Roberto: And I end the sentences of others, even if I don't end them the way they wanted, I think that that's the proper way to end the sentence. But we think there should be no reason why I should actually interrupt African Americans more than Whites. If I interrupt, I should be an equal interrupter. You see, I don't have a problem with interruption per se; I have a problem that this is actually biased.
Roberto: So a watch could tell me that. It doesn't matter how much you talk or interrupt, you're not interrupting equally. And therefore that would be an amazing piece of information. Because I work on this research, you know what? Now I'm very, very aware of my interruptions.
Ben: Of what you're doing. Yeah.
Roberto: Of who I am. And one thing is actually making me interrupt less, because I have realized that kind of, my culture, we interrupt too much. So I say, “Maybe I don't need that many interruptions."
Ben: It’s funny when you talk about the Apple watch, of course this is where my mind goes, because I get little rings for how much I exercise. It’d have to rethink the interface there, it’s like, “You only have five more inappropriate statements to make to complete your ring.”
Roberto: Oh I see.
Ben: So that’s interpreting it the wrong way.
Roberto: Yeah. Yes. Actually I think those rings are very non informative. I have to say that, I have my Apple watch too. I love Apple. But those rings are, is the wrong way to present the information. Don’t you think? I mean, except for the calorie one, the other ones, there just completely useless.
Ben: Oh yeah, yeah. Because I always get my, I always get my standing one, all you have to do is stand up once an hour and you get it. I see what you’re saying. But the calorie one, yeah, that seems to work pretty well.
Roberto: Actually we don’t know. That’s the beauty of imperfect measures.
Ben: Yes, yeah, exactly.
Roberto: We don’t know if it works, but the day-
Ben: But at least I have it now.
Roberto: Exactly. You have it and then when you get several days and you have not closed a ring, then suddenly you kind of work a little bit.
Ben: Yeah, I feel very guilty. [crosstalk 00:30:08] myself.
Roberto: But okay, you see. But this is actually why I think having it makes an impact. You have the same guilt when you go once a year to your doctor. Here you have your guilt every five days. The difference is that if you get your guilt every five days, you might do something on the weekends. And therefore you will do 52 times exercises. Like going 52 times to the doctor. And this changes dramatically your behavior. Now of course I will like those rings to measure something slightly relevant.
Roberto: You don’t want these just to be random data.
Ben: Right. Right. Right.
Roberto: But people are too obsessed with trying to measure things perfectly as opposed to, especially with human behavior and with things that are so difficult to measure, I think we have to settle on measuring imperfectly.
Ben: Imperfect is better than nothing.
Roberto: No, exactly.
Ben: So I guess to wrap up, I mean where do you go next with this? I mean where are you focused?
Roberto: So I’m focused, a lot of my research moves with the students that I have. So just depending on their interest, that’s more or less what we’re doing. One of the ones that we have right now, which I think is really cool one that we’re studying, is in the inconsistency between the statements that firms make and their lobbying activity.
Ben: Mm-hmm (affirmative).
Roberto: So something that does happen, in fact I have seen a couple of reports lately, that people are starting to pay attention, that you say something on your corporate social responsibility, and then you’re lobbying differently.
Ben: Oh, interesting.
Roberto: So I mean you say, “Oh I hate guns. I hate guns. We should have no guns.” This is my public statement, but then I just give massive amount of money to a think tank that actually supports guns. Okay? I’m just giving you an example of what I actually saw in the newspaper. These examples I gave you, actually I saw in the newspaper.
Roberto: Oh you said, imagine that you are saying, “Oh I just want to reduce CO2 emissions.” But then on the other side, you are working really hard to make sure that there’s no incentives to buy electric cars. Or solar panels. You are trying to stop solar panels but on one side you’re saying, “Oh we want to become a green producer of energy.” So one thing that we’re trying to do is to measure your statements. So when you make a statement on your corporate social responsibility report, you make statements about treatment of workers, I mean most firms make statements about many different areas.
Roberto: We have classified this in 32 aspects that go from emissions, to resources, to labor, to discrimination, to criminal behavior, to corruption, things like that. And then we look at where your money in the US is going. And the beauty, it’s hard to tell because sometimes when you give money to a particular bill, it could be because you want a yes or you want a no. So how can you tell? No?
Roberto: But the problem is that we can see patterns. I mean, so we can see how the money that is going for lobbying has a particular objective.
Roberto: And therefore we can infer that objective. In other words, you give money for a particular bill, you don’t know if it’s yes or no, but then you observe that you’re supporting a think tank that supports a no. Okay? So then you can start making, again, probabilistic statements, but what is aligning you that that money goes to a particular. And what we want to do is to measure the disconnect between what you say and what your lobbying activity is.
Ben: Yeah, that could be really interesting.
Roberto: It really could, no? So this is an example of if I would have told you about this example, you would have said, “How can you measure that?” I say, “Well this is the way to measure.” Now it’s gonna be totally imperfect, but whatever we do I want to make it totally public and totally transparent. That people go to a webpage and they just look at the company, and we have a measure of the inconsistency between the statements and what they do. But also for example, one of the things that we want to do is to have a very clear appeal process for that. In other words, if the company, if I made a mistake, I want the company to come and say, “Hey, you made a mistake. There is evidence here that we actually were supporting the yes not the no. And therefore that amount of money is incorrectly allocated.” Beautiful.
Ben: I just wrote the check 15 minutes ago.
Roberto: Here is the evidence. Well we just … By the way, the beauty of also doing this is that I am not, I don’t have a judgment of which bill you should support. Actually I don’t care.
Roberto: I do need to be honest, that is a decision that the political system have to do. I have views about how the US should be organizing society, but that’s not what I want to impose here. I want the companies to be truthful, that what you say on the report is what you do with your lobbying money. So one is very non reported by the way. They’re trying to change some of the rules about reporting on lobbying activity now. So, which might complicate a little bit, our analysis in the future, further.
Roberto: But anyway, again, I don’t mind producing something imperfect. But the beauty of also having an appeal process is that we will try to be very transparent and honest on the fact that, “Yes, we know it’s imperfect so it can be fixed.” If you provide the information then we can fix it. I have no problem actually including that piece of information into the analysis. And you see, that’s almost like creating a whistleblower system, or a reporting system, by shaming your behavior.
Ben: Mm-hmm (affirmative).
Roberto: You see?
Ben: Yeah, that’s fascinating. I’m looking forward to seeing that.
Roberto: So anyway, so we will do that. So I don’t know what the name is going to be, but something like, inconsistency will be in the name somewhere. So it would become a part of a project soon.
Ben: Inconsistent consistencies.
Roberto: Yeah. Or Consistent Inconsistencies. That’s what … No, because that’s what we are gonna be detecting. So I don’t know that, actually that name doesn’t sound that bad. But it might be too general.
Ben: But it’s gotta be catchy.
Roberto: It has to, yeah exactly. And the students do all the work, I make sure of all the entertainment value. So the catchy names and, I’m in charge of that.
Ben: I like that.
Roberto: Yeah, yeah, yeah. It’s very clear separation of work here.
Ben: Well I really enjoyed sitting down and talking to you. This was a lot of fun.
Roberto: It was great. Thank you.
Ben: And I appreciate your time.
Roberto: Me too. Me too. I have a good time. And thank you so much for this opportunity, and I hope that our paths cross again.
Ben: Absolutely. I would like that.
Ben: And thanks everybody for listening in. And that’s a wrap.
Speaker 3: Masters of Data is brought to you by Sumo Logic. Sumo Logic is a cloud native machine data analytics platform delivering real time continuous intelligence as a service to build, run, and secure modern applications. Sumo Logic empowers the people who power modern business. For more information go to sumologic.com. For more on Masters of Data, go to mastersofdata.com and subscribe. And spread the word by rating us on iTunes or your favorite podcast app.