Jacked Up Ratings: Problems with Quantifying Trust


Contributor / Bruce Carruthers

John D. Macarthur chair and professor of sociology
Director, Buffett Institute for Global Studies
Weinberg College of Arts & Sciences / Sociology

Part 1 / Jacked Up Ratings: Problems with Quantifying Trust (0:00)
Relational trust emerges incrementally as one person extends trust and another reciprocates, creating a virtuous cycle. Modern economies are dependent on a very different model, defined by contracts, qualitative data, and ratings. This system is useful but also vulnerable as people exploit its loopholes and ambiguities.

Part 2 / Re-Engineering Trust with Peer-to-Peer Lending (1:49)
Peer to peer lending, which connects lenders and sellers at a distance, relies heavily on quantitative data while experimenting with a simulation of relational trust. By rearticulating the personal and impersonal, this model can provide transparency in geographically disconnected interactations.


BUMPER: Jacked Up Ratings: Problems with Quantifying Trust

One of the things that happens is, as we rely on quantitative measure of underlying features, like a quantitative measure or score that measures somebody’s creditworthiness or how trustworthy they are.

As these scores become consequential, as people start to take them seriously, and as they are used in actual important decision-making, there’s an incentive for people to corrupt them, to game them, to stop paying attention to what it is they’re measuring and instead focus on the measure.

And there are a couple of really clear examples of this becoming a big problem.

And one was, in 2008, people realized that the bond rating scores that had been attached by Moody’s and S&P and Fitch and whatnot, and that were attached to asset-backed securities based on subprime mortgages.

That the investment bankers and the rating agencies worked together to try to jack up the ratings as high as possible so that whenever outside investors looked at a security, they saw, “Oh, it’s AAA. Well, that’s great.”

Had they been savvy (and now they’re very savvy because they know what the problem is), they might have realized that, in fact, that AAA rating was a score that was kind of jacked up.

And so, I think in a world in which the quantitative information becomes increasingly important, what you have to do is be a sophisticated consumer of numbers and always be mindful of their limits and vulnerabilities. And you simply cannot take them too seriously.

BUMPER: Re-Engineering Trust with Peer-to-Peer Lending

Peer-to-peer lending is a very interesting experiment that, again, takes advantage of the IT revolution.

It used to be that if you were going to do peer-to-peer lending, it was going to happen in your small hometown. Those were your peers.

Those were the people who could trust you, who knew about your business, who might be willing to lend to you or whose business you knew about and to whom you might be willing to lend.

And what we’ve done is, we’ve sort of disconnected what used to be the high correlation between social knowledge and geographic concentration.

And so, peer-to-peer and similar models are re-engineering some of the differences between relational lending and relational trust and generalized trust in very interesting ways.

It’s a re-articulation of the connection between personal and impersonal. It’s a way of saying, “We can personalize what would otherwise be a default impersonal situation. We can create peers out of people that aren’t even in the same country.

Keep up with the latest insights on trust

Related Videos

Healthcare providers must recognize the importance of trust and communication in building stable relationships with patients.

Trusting Healthcare Providers and Institutions: Key Findings

Contributor / Kelly Michelson
Kelly Michelson Pediatrics Communication,Distrust,Healthcare,Measurement Research about trust in the healthcare setting has generally taken two approaches: the first is to look at it in a qualitative fashion, so to hear personal anecdotes and learn what we can from that; and the other is to look at it in a more quantitative fashion, using scales and measures to see how trust relates with specific outcomes or specific variables.

From the qualitative research, we know that things like developing partnerships, developing relationships, demonstrating competence are all very important components of establishing trust in the relationship.

Most of the quantitative data related to trust in a healthcare setting use trust scales to compare a measure of trust to a particular variable — looking at things like, do women tend to be more trusting than men of their healthcare provider? Are there racial differences related to trust? Are there differences in providers’ and healthcare settings’ relationship to trust?

And these concepts help us to think about how we act — in a clinical setting, for example — and what we teach trainees about how to build trusting relationships with their patients or with others in the healthcare team.

BUMPER: Trust and Communication

Much of what we know about trust in the pediatric intensive care unit comes from literature that looks at communication and how communication unfolds in this particular setting.

We know from some qualitative work, from Carnevale et al., that trust is a really important part of communication in the pediatric intensive care unit.

These authors interviewed physicians, nurses, and parents about communication, identified three different components of communication. And of note, one of them was relational communication. And one key factor in developing relational communication that they identified was fostering trust.

And in another work by Ames et al., we find that trust is not only an important component for healthcare providers to focus on but also for parents.

In this work, the authors interviewed parents of children who were in the pediatric intensive care unit and asked them about their roles. And one of the three roles that they identify was actually that the parents should be trying to establish a trusting relationship with the healthcare providers in the PICU.

In another study done by Vivian et al., we learned about the importance of communication among staff members in the pediatric intensive care unit.

In that study, staff members were interviewed, and we found that poor communication among caregivers within in the intensive care unit can impact trust and therefore impact how they care for patients.

So, again, we’re seeing the importance of trust between providers and patients (or, in my case, parents) but also among providers.

BUMPER: Trust in Critical Decision-Making

In terms of decision-making in the intensive care unit, much of the literature has focused on issues related to pretty challenging decisions for children who are very sick, things like withdrawing or withholding life-sustaining efforts if a child was seriously ill — some pretty serious decisions.

Some of the research I’ve done, for example, has looked at what kind of influencers contribute to a parent’s deciding whether or not to withhold or withdraw life-sustaining therapies if their child was so sick that that became something to consider.

And what I found was that distrust was one of nine important factors that parents are weighing in terms of making that kind of decision.

In another study, Meert et al. interviewed parents of children who had died in the pediatric intensive care unit to find out more about their experiences. And they found that parents who felt that clinicians were withholding information also had a sense of betrayal or distrust towards those physicians.

BUMPER: Enhancing Trust in the Intensive Care Unit

But it’s really important not just to know what happens in the intensive care unit and where trust fits into communication and decision-making; now that we have all that information, we really want to try to impact trust and to enhance better trust and better communication and hence better decision-making in the intensive care unit.

For example, Curtis et al. looked at an intervention where he tried to change multiple components of what was going on in the intensive care unit, including identifying champions for this work, providing feedback to clinicians.

Interestingly, he didn’t find that that intervention changed his primary outcome.

In another effort done by Lautrette et al., they actually educated clinicians about how to conduct a family conference.

And they came up with this mnemonic called VALUE, and each of the letters stand for something different — specifically that you should value and appreciate what the family is saying during a meeting, acknowledge their feelings and emotions, that you should listen to what they say, that you should try to understand their situation and their values, and that you should elicit questions from them.

So, they actually taught clinicians a little bit about how to focus their communication during family conferences in the intensive care unit. And they actually did find a difference.

They found that for families who had family conferences with clinicians who were trained in this manner, those surrogates to the patients in adult ICUs had less anxiety and depression after their loved one had died and less symptoms of post-traumatic stress.
Judging trust and trustworthiness is important in all aspects of our lives even grocery shopping.

Differentiating Trust and Trustworthiness: A Sociologist’s Perspective

Contributor / Bruce Carruthers
Bruce Carruthers Sociology Government,Institutions and Context,Legal Guarantees,Reputation Management,Social Psychology,Vulnerability As a social scientist, I’m very interested in trust as it concerns people because really society, human cooperation, human coordination, all of the activities that we do together, really depend on trust.

And you might kind of wonder, well, trust sounds a bit like faith? You know, we’re just going to take people on faith.

And don’t we have a bunch of ways of coordinating our activity and making sure that the left hand knows what the right hand is doing, and I can figure out what’s going on vis-a-vis my employer or all the people that I interact with — and don’t we have a bunch of formal coordination devices like contracts or instructions or standard operating procedures?

Why aren’t those good enough? Why do we have to go beyond that and trust people? And there’s a couple of reasons for this. And one of them is that, as wonderful as these formal devices are, they really do have limits.

And one of the reasons is that contracts and other instructions, lists, standard operating procedures, all of these devices — they’re always incomplete; that is, the world is more complicated and unpredictable than we can anticipate.

And so, stuff will happen that will effect whatever it is you’re doing with these other people. It will have an impact on your ability to execute whatever it is you’re trying to do, and it’s not going to be in the contract what’s up.

You might think to yourself, “Well, maybe trust isn’t such an issue if we go to the marketplace.” Let’s think about markets and capitalism and self-interest and competition — maybe that’s a world in which contracts and other formal devices will be sufficient.

And once you’ve got an airtight contract, you don’t have to worry about the personal character or the trustworthiness of the people you’re dealing with, because you’ve got a good contract and you hired a good lawyer.

So, the most famous person who thought about this sphere, of course, was Adam Smith in his famous book The Wealth of Nations, which really did talk about the virtues of capitalist production and competitive markets and so forth.

And I think it’s very telling that before he wrote The Wealth of Nations, Adam Smith wrote a book on The Theory of Moral Sentiments.

And in this previous book, he posited that people are linked through strong bonds of sympathy and empathy and trust, and that on top of this, we’re able to have markets and capitalism and all that kind of fun stuff.

It was clear to me (and I think clear to Smith) that some measure of some baseline, some foundation of trust is very important even in social settings where we might think that the issue of trust can be solved or avoided.

You might want to ask, when does trust arise? I’ve talked about it as kind of ever-present — it’s all over the place. But really, there’s two elements that drive trust situations.

One of them is uncertainty, and the other is vulnerability — that is, people are uncertain about what others are going to do (they don’t know what’s going to happen in the future), and they’re vulnerable to what those other people do to the extent that their interests and those other people’s actions are intertwined.

So, the trick for dealing with a trust situation is really addressing these two key elements: either trying to deal with uncertainty by acquiring more information and learning (or trying to figure out) what is likely to happen in the future.

…Or by managing your vulnerabilities and thinking about ways to mitigate or reduce the impact (or the potential harm) that others’ actions, future actions, could have on your interests.

So, this kind of sets up a generic recipe book for how to deal with trust situations. How do people trust?

People rely on a lot of heuristics, rules of thumb, to decide who is trustworthy and who is not, and that distinction is really important because you can’t go through the world trusting everyone, and you can’t function in the world if you trust no one.

And so, what you have to do at the simplest level, is kind of put everyone into two bins: there’s people that are trustworthy; there is people who are not. And you want to be able to trust the trustworthy and avoid those who are not trustworthy

I’m going to offer a couple of distinctions that help clarify the discussion of trust. And one of them is the difference between trust and trustworthiness. And this really speaks to who is doing the trusting and who is being trusted.

One party trusts the other, and the other party may or may not be trustworthy — that is, they deserve the trust. But someone who is trustworthy may not be trusted, and someone who is trusting may end up trusting someone who is not trustworthy.

So, these two things have to be kept separate. Another distinction is the distinction between generalized and relational trust.

Generalized trust really speaks to the question of how you deal with strangers. Do you trust abstract institutions? Do you trust the average citizen that you might run into on the street?

That kind of a thing — where you’re really dealing with someone with whom you have no relationship and about whom you have no prior information. What kind of ambient or generic level of trust do you have?

Relational trust is, what happens after you start to get to know someone? What happens after you start to develop a social relationship?

You have a history together; you have contracts; you have prior transactions. That is a very particular and non-anonymous form of trust, and it is really driven by the nature of the interaction that you have with that individual.
Contracting is an important step in building an international trust economy.

The Limits of Contracting and Trust: Key Findings

Contributor / Niko Matouschek
Niko Matouschek Economics Building Brands,Economics,Government,Legal Guarantees,Long Term Focus,Regulation The economic literature on trust really builds on and grows out of the neoclassical models of the economy that people like Arrow and Debreu worked on in the 1950s.

Now, in these kinds of models, by design, there’s no role for trust, because either the market participants can verify the quality of the product on the spot or they can costlessly and perfectly contract with each other.

Now, even though there’s no role for trust in these models, it’s still a useful starting point for us because if what you’re trying to do is understand the role of trust in an economy and to what extent it limits the efficiency of the economy, you need to have some comparison point — because if what you’re trying to do is understand to what extent the lack of trust limits the efficiency of markets, we need to compare that to something.

And a natural comparison point is an economy in which there is no lack of trust.

The other reason for why these neoclassical models are a useful starting point for the literature on trust is that they highlight the fact that there’s no role for trust if contracting is perfect. And so, to understand trust, we first have to understand the limits of contracting.

BUMPER: Economics of Contracts

That’s what brings me to the literature — the economic literature — on contracts. And one way to think about this literature is in terms of the different types of impediments to contracting that you might be concerned about.

One of them — one that’s probably less well known in the literature but I think important to the current context — are problems with enforcement. Suppose that the state, the government, doesn’t enforce contracts or doesn’t do so efficiently, then what do we do?

Here, Diego Gambetta has this fascinating book about the origins of the Sicilian mafia, in which he argues that the core function, at least initially, of the mafia was to serve as essentially a contract enforcement agency that regular businesspeople would go to when they’re trying to write a contract in which they’re committing themselves to some future action.

Other strands of the literature look, for instance, at just the cost of writing contracts. Suppose there’s no problem with having contracts enforced, but writing contracts itself is costly. My colleague Ron Dye, here at Kellogg, wrote one of the first, if not the first, paper on costly contracting, essentially.

Then there’s the enormous literature on, essentially, endogenous contracting costs, where contracting’s costly because the parties can either take hidden action or there’s a problem with hidden information.

And then finally, there’s a strand that looks at what we call “incomplete contracting” — that is, it’s trying to capture the idea that sometimes we’re simply not able to describe in words the kind of product or service that we’re trying to trade later on.

BUMPER: Economics of Trust

Once we understand the limits to contracting, we can start talking and thinking about trust and the role of trust in the economy. Here, there are really two literatures.

The first one essentially assumes that there’s at least a small number of sellers who, for whatever reason, are committed to selling a good product. They’re committed to keeping their promises.

And then, the question is, what’s the implication of the presence of these good-type sellers, if you want, for the functioning of markets?

Now, one set of papers, which is called the “Gang of Four” papers, is concerned with the incentives of the bad types, if you want — the ones who are not committed to always doing the good thing — to mimic the good agents.

So, we’re trying to understand, to what extent is this mimicking behavior possible, and how does it affect the functioning of markets?

If the literature on reputation games is about good agents, the literature on repeated games is about good incentives. Here, the starting point is to not assume that there’re good agents: everybody just does what’s in their own best interest.

Now, the question is, in that kind of situation, to what extent can repeated interaction, repeated transactions between these agents, allow them to commit to not break their promises?

Now, the key issues here turn out to be to what extent the agents are what we call “patient” — to what extent do they care about future business versus current business?

And the other is transparency — to what extent can current agents observe what has happened in the past? Whether, for instance, I really have broken my promise in the past or not.

BUMPER: Applications and Empirical Evidence

Finally, we can talk about applications and empirical evidence. One application that people have looked at is to what extent reputations are attached to corporate names and can then be traded in the market for corporate control.

The goal is, here, to try to understand to what extent this motive of trading reputation can explain real-world mergers and acquisitions.

Another application that people have studied is the market for illegal drugs, which is exactly the kind of market in which trust is important because if I’m buying drugs from you, A) I can’t verify on the spot the quality of the product, and B) we can’t write contracts.

And so, this is exactly the kind of market which only works if there’s trust between the buyer and the seller.

But in contrast to other markets, this is a market that the government may want to undermine. And it may try to undermine it by undermining the trust between buyers and sellers.

And so, then you can ask to what extent you can do that — for instance, by designing sentencing guidelines appropriately.

In terms of historical studies, the most famous are probably by Avner Greif, who looked at trade in the 11th century, a time when the trading partners couldn’t rely on the governments to enforce contracts.

He studied community-enforcement mechanisms that allowed these parties to transact nevertheless. Essentially, if you cheated on a transaction, then you were going to be punished not just by the counterparty of that transaction but by the community at large. And that provided you with incentives not to cheat in the first place.

The most recent studies use big data sets to try to understand reputational incentives in the marketplace.

My office neighbor, Tom Hubbard, for instance, has this really interesting paper about incentives in the market for car repairs — which, again, is a market in which trust is really important because if I, the car repair shop, tell you that you need a repair, you don’t really know whether you need it or you don’t.

Other pages in Videos:

Pages in The Trust Project at Northwestern University