BoE research shows big banks were too big to fail | FT Alphaville

BoE research shows big banks were too big to fail

Brace yourself. We have a formula:

P(eventi) = α + βXi + εi

If only that formula had been available previously.

Academics Andrew Rose from the Haas School of Business and Tomasz Wieladek from the Bank of England have discovered, using multinomial logit regression techniques, that there’s a link between the size of a bank and the likelihood of it being bailed out in a crisis.

The evidence can be found in Working Paper No. 460, entitled “Too big to fail: some empirical evidence on the causes and consequences of public banking interventions in the United Kingdom.”  An extract (emphasis ours):

Our results suggest that the size of a bank is an important determinant of key public British banking interventions: capital injections, nationalisations, and government funding or central bank liquidity insurance schemes.  In particular, the size of a bank relative to that of the entire banking system increases the probability of an intervention, suggesting that large banks are more likely to receive public sector assistance.

And here’s some detail around that crucial formula…

We use a reduced-form model to examine which bank characteristics affect a bank’s ex ante probability of receiving government assistance:

P(eventi) = α + βXi + εi

where: eventi is a dummy variable that takes a value of one if bank i has been affected by a given public sector intervention following the onset of the global financial crisis, and is zero otherwise; X is a vector of factors that drive the probability of public intervention; ε is a wellbehaved residual that represents omitted factors such as ‘management quality’ or ‘interconnectedness with the rest of the global banking system’ and {α, β} are coefficients to be estimated.

Related links:
Bigger breasts make for bigger tips — Springer Link
Abstinence-only education does not lead to abstinent behavior — ScienceDaily
False confessions may lead to more errors in evidence — Association for Psychological Science