Letter to the FT: Three ways the US Treasury could repair the bonds market

As has been reported in the FT, the US Treasury believes that Treasury prices are open to some manipulation, and is trying to deter even the appearance of such. But the fault lies, at least in part, with the Treasury itself, who could and should improve the functioning of the market in three separate ways.

1. Bonds should be much larger. The current 30-year, the 4½% Feb 2036, has $26,397 million outstanding. This is the long bond of the world’s pre-eminent economic power, yet there are five long gilts that are larger, as is Germany’s long bond.

   On 15th May 2008 US bonds and notes totaling $87bn mature, and between the date of this cashflow and 2036, US tax revenues will grow by something like inflation plus real economic growth. So, the 30-year could be $250bn, and at maturity still be a smaller proportion of tax revenues. Reducing the number and increasing the size of bonds would improve liquidity, and make it more difficult for a player to acquire a monopoly position in any one security.

2. Indeed, US Treasuries could be of de facto infinite size. The UK DMO has a facility whereby one can borrow any gilt, which would be created for this purpose, in return for lending other gilts and paying a punitive fee. (This facility was suggested by me in 1996 and introduced in 2000.) The UK’s fee is too punitive — it should be 25% not 90% of the BoE policy rate — but even so it has done much to deter dealers from acquiring monopoly positions in individual securities. Happily, the US Treasury is consulting about starting such a facility.

3. The auction mechanisms used by government borrowers are descended from pre-computer days, even though modern technology allows a much better system. Consider an auction in which the government is selling $10bn of a bond. This bond has a fair price, reasonably well known to market participants. If I bid an edge more than that fair price, I can buy much of the bond (though the rules prohibit my acquiring the whole auction). If my friends and I bid an edge more, we can buy the bulk of the auction and be in a position of considerable market power.

   Even if we don’t do this, other dealers should fear that we might. So instead consider splitting this auction into fifty ‘auctionettes’, each of $200mn, conducted electronically at a pace of one per minute. Yes, my friends and I could still buy the first few auctionettes, but doing so would raise the price of subsequent auctionettes. Paying this higher price would further raise the price of later auctionettes. So acquiring a monopoly position would cost a lot more than “an edge”, because other dealers could respond as the position is built.

Of course, any reform entails an implicit admission that things previously could have been better, but surely fixing the system would be better than involving the lawyers?