The danger of repo

By Felix Salmon
August 6, 2012

Remember how there’s a very good chance that Treasury’s new floating-rate notes are going to be linked to some kind of repo benchmark? Well, here’s another reason that’s a bad idea: the repo market is shrinking fast, at least in Europe — and if it can shrink in Europe, it can do so in the US, as well.

What’s more, we want the repo market to shrink. Gillian Tett, in her latest column, explains that as rules about collateral tighten up, they create what Manmohan Singh calls a “second deleveraging”. But weirdly, Tett thinks that an increased focus and reliance on the repo markets is a good thing in the long term: “a financial system in which transactions are secured on assets is likely to be a healthier system than one which is largely – or patchily – unsecured,” she says, “particularly if that collateral is valued in a regular, disciplined basis”.

This is exactly wrong. Repos are a form of informationally-insensitive asset: they epitomize the paradoxical and ultimately destructive desire on the part of people with money to lend out money but to take no credit risk while doing so. Informationally-insensitive assets are a bad idea in general, for reasons which are probably familiar at this point to most readers of this blog: they breed complacency, tail risk, and deluded, magical thinking. But repos are a particularly bad species of the genus, because they are a direct replacement for old-fashioned unsecured credit.

Lending money in return for interest on that money is a form of investing: one entity, with money to spare, invests that money in a venture which can put it to good use and profit from it. If all goes according to plan, both win. The borrower might be poor but has ideas, and the ability to make money in the future; the investor makes such profits possible.

When you move from a credit-based system to a repo-based system, however, all that changes. At that point, future profitability isn’t enough to get you cash: instead, you need to be rich already, and you need to be able to hypothecate your existing assets to some lender. If we’re talking about the banking system, here, we’re talking about a world in which banks simply cease to trust each other at all, and the answer to all interbank credit questions is “no”. The only way for banks to lend to each other is to either go through some central counterparty, hub-and-spoke style, or else to retreat to the world of repo, where banking prowess counts for nothing and all that matters is collateral quality.

The implications of such a world are already being seen: Tett says that “collateral arbitrage” has now become a profit center at some banks. Far from trying to lend out money to creditworthy borrowers, banks are beginning to make money by gaming inconsistent repo rules. No good can come of this.

And in times of crisis, a reliance on repo markets makes all banks incredibly fragile, and vastly increases the risk to taxpayers should a bank fail. Once upon a time, banks had equity, they had debt, and then they had deposits. If a bank failed, the bank’s equity would be wiped out first, and then its debt. The depositors were senior, which meant there was relatively little chance that the FDIC would have to bail them out.

Now, however, bank debts are shrinking, replaced with repo operations. As a result, when a bank fails, the equity gets wiped out first — and then there’s no cushion any more before the depositors start losing money and need to be bailed out. The rest of the finance world is senior to depositors: they have repo collateral, which makes them secured creditors, and secured creditors are senior to unsecured creditors, even when the unsecured creditors are just mom-and-pop depositors.

The more that the world of finance relies upon repo, the less it relies upon relationships and trust and underwriting and all the other ties which bind. The financial sector can’t afford those ties to be severed: the cost of breaking them, in terms of foregone growth and profit, is far too great. But we seem to be doing exactly that.

Update: See also Carolyn Sissoko, from February,  a great post. h/t Waldman.

13 comments

Comments are closed.