Attention T-Mart Shoppers

by

Ross M. Miller
Miller Risk Advisors
www.millerrisk.com
March 12, 2007

[This commentary  appears in the March/April 2007 issue of Financial Engineering News. This is the raw, unedited version of that commentary.]

Now that Henry Paulson has settled in as U.S. Treasury Secretary, I have a suggestion for him. He should visit LEGOLAND in Billund, Denmark. I have no doubt that the majority owners, the Blackstone Group, will be happy to give him and his Secret Service agents the run of the place.

Secretary Paulson should make DUPLO Land his first stop. I must explain to the more culturally deprived of my readership that DUPLOs are the preschool version of LEGOs, which (for the even more culturally deprived readers) are a construction system that allows one to build up amazing structures from a variety of multicolored bricks and other useful pieces. DUPLOs differ from full-fledged LEGOs in that the set of available pieces is more restricted and the pieces that are available are not as versatile. DUPLO's deficiencies are not a bug of the product, but rather a feature to make them manageable for less dexterous young hands (such as those reaching out from my "Hedge Funds: Corruptors of Youth?" column, which provided Ben Bernanke with free advice.)

DUPLO Land crudely represents the current state of the world's financial system. The U.S. Treasury serves as the provider of the big, clunky colored blocks that directly or indirectly serve as the fundamental pieces that are directly or indirectly behind virtually all financial instruments. DUPLO Land is an impressive place until one visits the rest of the theme park. The elaborate structures built from the smaller and more versatile LEGO blocks represent what the world financial system could become if only the Treasury would get its act together, discover modern marketing, and transform itself into T-Mart. (Note that the general LEGO metaphor has been kicking around financial engineering for some time, dating back at a very minimum to a 1987 paper by Charles Smithson.)

Consider the lowly T-bill. As I write this, the Treasury provides the global economy with 26 varieties of them—one maturing every week for the coming six months. You would have to add all 31 Baskin Robbins flavors to them just to get the full 57 Heinz varieties. As T-Mart, the Treasury could increase the number of available Bills by nearly a factor of 10 by providing bills with maturities for each business day up to the legal limit of a one-year maturity.

But who would want so many different maturities? Plenty of people, apparently. Filling in the missing maturities is a major reason that there is an immense market for T-bill-secured repurchase agreements (repos). Great, you might think, the market has already filled in the hole in the market, why does T-Mart have to do anything? The answer is that a repoed T-bill is not financially equivalent to a real one—the counterparty risk and illiquidity that tag along are not always welcome.

The holes at the short end of the yield curve, however, are small potatoes compared with those at the long end. While daily maturities for notes and bonds would never fly, a good start would be to have T-Mart commit to auctioning a new 30-year bond on a fixed periodic schedule. Stretching the longest maturity date out to at least 40 years would not hurt either.

I learned the importance of the long bond early in the 1990s when I was building models to price some of the exotic insurance products of the day. Most products were priced off Treasury STRIPS, whose maturities could only go out as far as the principal payment of the on-the-run 30-year bond. Fortunately, a new long bond arrived at least once a year through 2002. Then, it went into hibernation only to awaken last February.

The ostensible reason for long bond's hiatus was that with the Treasury was running a "surplus" so it had no need to borrow long. This argument would hold some water if the Treasury had planned to get out of the borrowing business entirely by 2020 or so. But how likely was that?

Moreover, the Treasury could have argued that there was simply no demand for the long bond—indeed, last August's reopening of the February 2036 bond went badly—but there the fault lies squarely with the Treasury itself. For the long bond to be a truly useful LEGO block, a new one must arrive like clockwork, even if that clock ticks just once or twice a year.

The Treasury's stated mission is to finance the government debt at the lowest possible cost to the taxpayers. One could argue that as all Treasury debt is issued at the risk-free rate for that maturity, the NPV of all its debt is zero regardless of how it is laddered. The problem what that argument is that liquidity matters and the Treasury is in the liquidity business. The better the Treasury is at providing liquidity, the cheaper it can borrow. In a world teeming with financial engineers, a full set of intertemporal LEGO would be inherently more liquid both collectively and individually than a hodgepodge of pieces issued in a clearly myopic manner. There may be individual Treasury securities that will only trade as frequently as Britney Spears does something silly that makes the news, but that is still frequent enough to provide fresh prices for the financial system to exploit.

Filling in the holes in the yield curve should just be a point of departure for the Treasury on a grander mission. Forward contracts, floating-rate notes, and swaps are just a few of the products one should find at a well-stocked T-Mart. The Treasury has already successfully dabbled in derivative securities beyond the rudimentary STRIPS, auctioning off "Y2K options" to provide emergency fin-de-siècle liquidity in 1999 under the aegis of Secretary Lawrence Summers. (Many thanks to George Zachar for pointing these options out to me.)

If Henry Paulson is like Robert Rubin, his predecessor in the treasury secretary's slot and fellow alum of Goldman Sachs, he might wish to return to Wall Street some day. Therefore, he would not want T-Mart to cannibalize one of the Street's valuable franchises. The historical precedent of Treasury STRIPS, however, demonstrates that there is no need for him to worry. Merrill Lynch beat the Treasury to the business of stripping coupon and principal payments and then giving them clever acronyms. But any business Merrill lost in the short run to the Treasury entering its market, it likely made back many times over on all the new products that STRIPS made possible.

A more fundamental problem with Treasury securities is that while they are exempt from state and local taxes, they are remain subject to federal taxation. In the case of STRIPS, that the tax is levied in such a punitive way that their use is largely restricted to tax-sheltered environments. The federal tax on Treasuries simply ends up back in the Treasury, so why not simplify things by making the paper completely tax-free from the start? This way, the true LEGO nature of Treasuries can shine through.

Finally, as any LEGO devotee can tell you, those blocks can become smart with the addition of programmable microprocessors to the MINDSTORMS line of robots. Treasury securities themselves may never be intelligent, but the markets in which they trade certainly could be a lot smarter than they are now. Just how that might work will be the subject of a future commentary.