Monday, July 16, 2012

Manipulating the Numbers: Motive, Means and Opportunity

I have been thinking about some important parallels between the Libor scandal and the failure of ratings for structured securities. The parallels arise from the existence, in both cases, of the motive, means, and opportunity to manipulate the numbers for private gain.

The means and opportunity arise from the fact that neither Libor nor ratings are objective reports of past events. Compare them, in that respect, to numbers like exchange rates or the outcome of T-bill auctions. The latter are reports of transactions that someone actually carried out. Libor and ratings, instead, are reports of transactions that someone thinks could possibly take place—borrowing that a bank could undertake if it wanted to in the case of Libor or future fulfillment of a financial obligations in the case of ratings. If someone says that something could happen, is that what they really think, or not? There is no way to check.

The motive for manipulation comes from two sources. One is the fact that the values of many financial instruments are pegged to these numbers. For example, the value of an interest rate swap changes as Libor changes, and the value of a credit default swap changes as ratings change. The other motive comes from regulatory reliance on the numbers. Banks (Barclays in particular) thought they could relieve regulatory pressure if they reported an ability to borrow cheaply. In other cases, financial institutions knowingly bought overrated securities so that they could indulge their appetite for risk while minimizing regulatory capital requirements.

With motive, means, and opportunity aligned in so tidy a fashion, is there any wonder that fraud took place?

All this makes one wonder about a different category of numbers that move markets, numbers like unemployment and inflation rates. Those, in principle, are objective reports of past events, unlike Libor or securities ratings. However, unlike exchange rates or T-bill prices, outsiders cannot easily verify them. The raw data are not easily accessible; and even if the raw data could be verified, the methodology of deriving the final numbers could be subject to manipulation.

Observers of emerging market economies, such as Argentina or China, frequently express doubts about reported inflation rates and other data. So do many people in the United States. Anyone who has ever blogged about inflation or unemployment rates is used to a hailstorm of comments along the lines of “Gummint lyin’ to us again!”

So far, though, most economists think our government number crunchers are doing an honest job, as best they can, given the inherent methodological difficulties and the budget constraints they face. (In support of that view, check out various bits of Congressional testimony by Keith Hall, posted here. Hall was Chief Economist of the Council of Economic Advisers from 2005 to 2008 and Commissioner of Labor Statistics from 2008 until earlier this year.)

I hope this professionalism lasts. If we learned from some whistleblower that the BLS had falsified a jobs report on the eve of an election, we would have a scandal that would make the Libor manipulation look like a tempest in a teapot.

Originally posted at


  1. This is an interesting analogy. Isn't it the case though that ratings are problematic only where there is essentially a single buyer, namely govt regulators, as for mortgage securities, but that in other instances where the supply and demand works, as in the case of ratings of countries, that they are more reliable?

  2. Yes, the ratings would presumably be more reliable where the buyer of the securities is not subject to a bias toward risk taking. Regulatory reliance on ratings is one source of the problem, but in other cases, agency problems, limited liability, or other features might create a willingness knowingly to buy overrated securities.