Jacked Up Ratings: Problems with Quantifying Trust

Applications

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.

Transcript

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.