The recent kerfuffle over LIBOR (or perhaps, Lie-bor) has led many, including those living in developing countries but reliant on the global interest rate benchmark, wondering how the apparently flawed system was allowed to become such an major global institution. Blame for the entire fiasco has, predictably, come from the usual suspects on the usual suspects: apportioned equally to bankers' moral shortcomings and/or regulator complicity. These strike one as lacking imagination: after all, people (bankers and regulators inclusive) respond to incentives, and it is the institutional setting that governs the types of incentives individuals embedded in the system face. Presumably, if the existing framework for obtaining reported LIBOR rates did not involve substantial submitter discretion,[*] then both traders and regulators would react very differently to potential exploits of the system. After all, both groups would see little benefit from attempting to game the system if it offered little chance of success.
Through all the spilled ink (or pixels)---especially with regard to the now-conventional proposal to simply base LIBOR on actual transactions data, rather than self-reported numbers---little has been said about whether moving to such a system would yield actual efficiency improvements. In thinking about such issues, it is useful to keep two aspects of economic theory in mind.
First, in a standard double-auction setting such as LIBOR, theory and experiment suggest that efficient price discovery can occur with a remarkably small number of participants; with markets fiding equilibrium with as few as 6 to 8 agents (JSTOR subscription required). The argument that LIBOR is inherently inefficient just because it has only a small number of participating banks is therefore a weak one, in light of this fact. By the same token, however, the ease of attaining efficient outcomes with a small number of agents means that a setting where only actual transactions data are used to establish LIBOR need not be threatened by the fear of potentially thin markets.
Indeed, the natural advantage that using actual transaction data offers is the possibility that rates adapt to changing market conditions. In calm times when markets are thick, large institutions with superior economies of scale will tend to possess a funding cost advantage. Consequently, the transacted market LIBOR rate will be lower, reflecting the overall positive spillovers that result from having larger players engaged in the wholesale funding market. In turbulent times, however, the size and relative opacity of large bank balance sheets may lead to rational concerns by market participants over counterparty risk. In such circumstances, smaller banks with balance sheets less exposed to market volatility could find themselves being able to secure funds at a lower cost than the larger banks can. These smaller banks will then be the ones setting LIBOR at the margin. Again, the transacted rates will ultimately reflect the best available rates, even in such thin markets.
Second, even if we want to move away from a straightforward market mechanism based on realized transactions, we are aware of mechanisms that elicit truthful, incentive-compatible bids from market participants in alternative auction settings. Although such mechanisms alone cannot, alas, guarantee the absence of collusion (the problem with LIBOR), we also know that collusion becomes harder to sustain with a large number of players and free entry. So a system that combines both of these elements---preference revealing bids and ease of entry---could be the basis for an alternative market structure for generating LIBOR. Note that even in this setup, it would still be the case that LIBOR should be set on the basis of actual transactions, rather than imagined ones.
The bottom line is that there are good reasons to rely on actual transacted data to establish LIBOR, subject to some potential tweaking of the mechanism. LIBOR is widely considered one of the world's most important benchmark rates, and the erosion of trust in the system due to recent events suggests that it is high time to move away from the traditional, "gentleman's agreement" approach to LIBOR rate-setting. This would yield clear benefits not just for the millions of people around the world who implicitly rely on LIBOR in their financial contracts, but also for rebuilding trust in the international financial system, more generally.
*. Submitters repond to the question: "At what rate could you borrow funds, were you to do so by asking for and then accepting inter-bank offers in a reasonable market size just prior to 11 am?" Even setting aside deliberate manipulation, the scope of the question offers significant flexibility to the submitter, both morally and legally: the language includes terms such as could and reasonable, which leave substantial scope for individual judgement and discretion. To the extent that this was by design---maybe the BBA fears thin markets rendering LIBOR too volatile on a day-to-day basis---it is now clear that such a degree of flexibility is, on net, a negative for the system as a whole.