It’s time to hoist Henri Tournyol Du Clos from the comments; he’s always insightful, but what he said about Libor last night is particularly spot-on:
The problem is that unsecured fixed rate interbank lending is an activity that firmly belongs to the past, not the present and – most probably – not to the future of financial markets. Its volume, respective to other financial instruments, has been dwindling for the last 20 years and there is not a sufficient level of activity any more for prices to reflect anything except, well, the level at which bank treasurers do not want to trade.
Unsecured interbank fixed rate lending is a remnant of a world where derivatives markets were small and where cash financial markets were roamed by thousands of mid-sized commercial banks with a small exposure to financial markets. Thanks to the banking sector’s consolidation, we now have
a) liquid derivatives markets, traded by everyone (bank ALMs, market-makers, asset managers, insurers, commercial firms, Tallahassee widows, you, me, my dog if I had one, etc)
b)a few big financial supermarkets, like Citi, JP Morgan Bear Chase, etc, that do not need to borrow or lend unsecured from each other. In fact, they try to collateralize as much as possible a portion of what they trade together, whatever the product, and make it subject to margin calls.
The only unsecured lending that banks have to do is overnight reserve management: effective Fed funds in the US, Eonia in Europe and Sonia somewhere in-between. So forget about those 4 or 11 month Euribor or Libor or Whereverbor rates that have been only a fiction kept alive everyday like some ancient custom; base all derivatives on the effective overnight rate now; and let us stop wasting time on trying to measure precisely something that is not traded and does not exist in the real, transaction-based, financial world.
This is certainly true of fictions like 6-month Libor and 1-year Libor, which are often used as benchmarks but which are all but unused IRL. Take an adjustable-rate mortgage which resets every year, at, say, Libor plus 400bp. The fiction underlying the mortgage is that the bank in question can borrow at Libor, lend at 400bp more than that, and lock in a very nice spread. But that hasn’t happened in years.
Does that mean that we should stop using Libor as a benchmark, and start pegging mortgages and corporate loans to overnight interest rates, a bit like we already do with credit cards? That seems dangerous to me. But I do think that it’s time to stop talking about "fixing" Libor, as though there’s some Platonic ideal of "the" interbank lending rate which a well-designed survey could somehow reveal. There isn’t.