The Achilles Heel of Cowboyeconomics

The Wall Street Journal says everyone in the world would benefit from a dose of Cowboyeconomics. You know, exporting tons of debt, to support consumption growth well in excess of income growth ...

Obviously, global equilibrium requires someone to lend the US the money it needs to run large deficits. Not everyone can follow the same strategy -- the US can only export debt if someone else is willing to import it.

There is little doubt that central banks have been the big (net) buyers of US debt exports over the past two years. Cowboy economics is underwritten by the Japanese Ministry of Finance and a bunch of Chinese Communists (Ok, a bit unfair, but the People’s Bank of China is one of the United States most important creditors)

The big question is how much, if any impact, this is having on the market.

Opinions differ -- both in the market place and on different pages of the Wall Street Journal.

Broadly speaking, those who argue the impact of the central bank bid is overstated make the following points.

Treasuries rallied in December, despite a sharp fall off in foreign purchases of Treasuries.

Foreigners (mostly foreign central banks) may hold over 50% of all outstanding Treasuries, but their holdings tend to be concentrated at the short-end of the curve, and they are primarily price takers, not price setters. The market price is set here in the US, by the actions of American institutions who buy and sell Treasuries like mad. The yield of the ten year bond is determined by the willingness of American investors to hold the ten year rather than some other asset.

The same is true of the broader market. Foreign central banks may be the biggest foreign buyers of US assets at the margin, but private foreign investors in aggregate hold more claims on the US than foreign central banks -- and US investors hold more claims on the US than anyone. Market prices are largely determined not by the price at which central banks are willing to add to their portfolio of dollar reserves, but by the price that American and foreign private investors demand to continue to hold their existing set of claims on the US.

The counter argument rests on the following set of arguments.

Foreign central bank buying influences US debt markets not just directly, but also indirectly. By keeping Asian currencies weak against the dollar, central bank intervention helps keep US import prices low -- and since US firms often get a cut of the profits from buying Asian goods cheaply to feed their distribution channels (see Walmart), it also contributes to corporate profits and corporate savings. Such "general equilibrium" effects make Americans more willing to hold Treasuries at current prices.

Official data underestimates actual central bank support for the US fixed income market, potentially quite significantly. Consequently, attempts to use the US data to formally estimate the impact of the central bank bid are likely to be off. Moreover, those models try to infer the impact of $400 b plus of dollar reserve accumulation in 2003 and 2004 from the impact of less than 100 b in dollar reserve accumulation in the 90s -- that is a hard task. The impact may not be linear.

Markets are segmented, not seemless wonders. By buying assets that neither Americans nor foreign private investors want(at current prices), central banks are changing the global equilibrium. Paul McNully of PIMCO laid out this argument in his recent commentary.

This theory is founded on the notion that investors are not homogenous, with many if not most, constrained to a "natural habitat", borne of regulation and/or heterogeneous risk appetites. Most simply, this theory is a refutation of any grand efficient markets hypothesis, resting on the simple observation that global market participants do not view debt in various currencies and at various tenors as perfect substitutes, via forces of arbitrage.

... We must also take as a given that the rest of Asia, and much of the EM world, does not want to see their currencies appreciate very much (or at all) against China, and thus shadows China’s fixed exchange rate regime with the United States. These are facts, not opinions. To wit, we must take as a fact that, for the moment, what is frequently called the "Bretton Woods II" arrangement, does exist.

Yes, America does, as custodian of the global reserve currency, have "exorbitant privileges," as de Gaulle bitterly bellyached 40 years ago. Indeed, they are even more exorbitant now than then! America presently has the necessary degrees of freedom to pursue an accommodative, domestic demand-oriented monetary policy, aimed at domestic job creation, secure in the knowledge that:

Inflationary pressures will be defused by a portion of domestic demand "leaking" out to the rest of the world via the current account deficit, and; Foreign central banks will act as a "put" under U.S. debt prices or, if you prefer, act as a ceiling on the level and term structure of U.S. interest rates.

This is reality: the market segmentation hypothesis of the level and term structure of U.S. interest rates dominates the other two hypotheses.

It is the argument that makes the most intuitive sense to me. My gut instinct is central bank flows of this magnitude -- Central bank reserve accumulation (roughly $600 b, after netting out valuation gians) is close to the total global current account deficit (around X, but only Y if you look at all of Europe as integrated economic region), so in some sense central banks are intermediating the overall (net) global flow of funds -- have a real impact on markets.

Consider the actions of Japan’s Ministry of Finance -- a huge net buyer of (relatively) short-dated Treasuries in 2003 and 2004. Most private investors, both at the US and abroad, are not thrilled by nominal yields of 2% or 3% -- hell, the big US investors buying long-term treasuries that yield 4% do so because they can gear up by borrowing at 2% to buy at 4%, multiplying their potential terms. But the MOF funds itself very, very cheaply in yen terms, and is consequently a natural buyer of low-yielding Treasuries.

To quote Fred Bergsten:

You may be amused. I talked to the Japanese vice minister for finance about it a month ago, and he actually went out of his way to say to me, "You know, we really like the exchange market intervention we’re doing." He said, "We borrow money from the Bank of Japan at zero interest, and we buy dollars and put it into your Treasuries, and they don’t pay much, but they pay 2 [percent]--and that’s better than zero--we’re making lots of money." [Laughter] And I said, "But you’ve got a lot of unrealized capital losses." And he just laughed and said, "We don’t mark to market. We’ll never mark to market. Besides, 20 years from now the yen may be back at 130--who knows? We may get big capital gains." In short--and that’s part of the problem--that’s where Nouriel’s [editor’s note: and Brad] got it wrong. It can go on for a very, very long time, blocking the adjustment process.

Put differently, without Japanese demand, the Treasury would not have been able to sell so many short-term Treasury notes (reducing the average maturity of the Treasury stock) over the past two years, at least not without offering a higher interest rate to entice other investors into the market. It probably would have been forced both to pay more on its two and three year notes, and to issue more ten year notes. And since these assets are not really perfect substitutes, that would have an impact on the market.

I spent a bit of time digging around to come up with some statistics to try to flesh out this world view, in very broad strokes.

In broad terms, I estimate that central banks added $460-70b to their dollar reserves. Suppose $60-70b was used to build up their bank accounts (we know $40b went into the banks during the first part of the year). That leaves $400b to purchases debt securities. US data records purchases of $33b of short-term treasuries, around $203b in purchases of longer-term treasuries and $22b of other long-term US debt securities. That leaves $142b unaccounted for -- assuming that the overall estimate for dollar reserve accumulation is right. I would assume that some private purchases of US long-term securities were really central bank purchases.

Let’s look at how this impacted different markets.

Equilibrium in the Treasury market required that the market in 2004 absorb $380 billion of new issuance, and to find investors willing to hold the existing stock of $4044 billion. Central banks -- through recorded channels bought $267b, and foreigners bought 410b billion -- more than absorbing the increase in the stock (It would not shock me if central banks were behind say $50b of the private buying, bringing their total buying to around $320b).

Equilibrium in the foreign exchange market came because foreigners bought $900 billion of US long-term debt, plus at least $33 billion of short-term debt. That $930b financed a current account deficit of $660 billion, $100 b of net US purchases of foreign securities, and an additional $170 b net outflow (both a net FDI outflow and probably a net outflow from the banking sector -- think US banks loaning money to US owned hedged funds domiciled in the Caribbean). Central banks provided a net inflow of $400 billion or so through their purchases of US debt securities, and if they added say $65 b to their total (dollar) deposits in the the global banking system in 2004 (they added $41 b in the first half of 2004).

Equilibrium in broader US fixed income market required that the market absorb net new issuance of roughly $1400 billion (extrapolating from the bond market association’s data through q3). Foreigners absorbed $900 b of that, and central banks maybe $370 b of that -- almost as much as domestic US investors absorbed.

Foreign central bank demand for $320 b plus of treasuries (assuming some unrecorded buying) was a three-fer. It supports equilibrium in the treasury market, in the foreign exchange market and in the broader fixed income market. But its overal impact on key prices hinges -- thinking in a market segmentation kind of way -- on answering three questions:

a) What would the impact of the end of Central bank intervention be on the Treasury market -- someone, either at home or abroad, would have to be willing to absorb the roughly $300 b of the $400 b annual increase in the debt stock associated with ongoing deficits now being absorbed by central banks. There presumably would be some impact on price -- and, since that would feed back into the deficit through higher interest rates, the total size of the needed issuance would go up. To the extent that most of the additional issuance would be absorbed domestically, the US would need to find other ways of funding its external deficit.

b) What would the impact on the foreign exchange market be? The dollar would have to fall and interest rates would have to rise to either generated the flows from foreign private investors needed to replace the (estimated) $460-70 b in financing central banks provided in 200, or to reduce the United States need for borrowing (hard in the short-run).

c) What would the impact on the fixed income market be? Someone would need to absorb the $360 b in long-term securities (mostly treasuries) now being absorbed by foreign central banks, or the scale of issuance would need to fall to reduce the need for absorbtion. There would need to be some adjustments in price and the volume of issuance to make private investors, both at home and abroad, willing to absorb the total new issuance.

My point is simple: there fairly clearly would be some adjustment -- though there certainly is room to debate the size of the adjustment; and a lot depends on how various private actors react. If -- as Nouriel and I suspect -- the withdrawal of the exchange rate insurance provided by expectations of central bank intervention would tend to reduce, in the first instance, private inflows into the US market from, say, Japanese pension and insurance companies, that would tend to increase the overall size of the required market shifts. On the hand, if private investors would step in and fill the gap left by the withdrawal of central banks with only small changes in US interest rates and exchange rates, then the overall impact would be muted.