Willem Buiter Cries "Doom! Doom!" for the Dollar

Willem Buiter:

Our currency and our problem: In 1971, the then US Secretary  of the Treasury, John Connolly told his European counterparts: "the dollar is our currency but your problem." So far, Connolly's statement continues to be true.  Every time the dollar weakens, US exporters and US import-competing industries are gaining competitive advantage and/or increasing their profitability.  The explosive growth of US export volumes... is part of the reason that, despite the collapse of US housing construction, the US economy is still expanding.... The weaker dollar also improves the US net external investment position.... With so much of US external liabilities denominated in dollars, every time the dollar weakens, the world largest debtor feels a little wealthier. The Chinese authorities, despite moves to diversify their foreign asset holdings, still... hold over a trillion dollars... in US Treasury... The Japanese authorities have a similar exposure... have re-confirmed their reputations for being among the world's worst portfolio investors.... [T]he Chinese and Japanese authorities... presenting their tax payers with a further $200bn to $300bn capital loss... a heavy price to pay for access to US markets for your exports, especially for a poor country like China....

I fear, however, that the good news about dollar weakness for the US is about to come to an end. Sooner rather than later, the weakness of the dollar, and fear of its future weakening, will trigger a large increase in long-term US interest rates, nominal and real.... The further weakening of the US dollar will continue to boost the tradable sectors of the US economy, but any sharp increase in long-term nominal and real interest rates will hit investment spending.... It won't be pretty.  Expansionary monetary policy measures will be limited because a collapse of the dollar will have non-trivial inflationary consequences...

It is not clear to me what model Buiter is working in. In economists' default model, expectations are, in general, not of a particular rate of change of the dollar but of a future level for the dollar. If domestic interest rates are high (relative to interest rates abroad, adjusted for risk and other factors) then the value of a currency will be above its long-run expectation. If If domestic interest rates are low (relative to interest rates abroad, adjusted for risk and other factors) then the value of a currency will be below its long-run expectation. But it is not the case that expectations of decline drive up domestic interest rates--not unless a central bank is driving up domestic interest rates because it wants to keep a currency worth more than its long-run expectation. And the U.S. Federal Reserve is not in the business of pushing up domestic interest rates in order to keep the value of the dollar high.

So how then is it that Buiter expects "the weakness of the dollar, and fear of its future weakening" will "trigger a large increase in long-term US interest rates"?

One possibility is the following chain of causation:

• Past declines in the value of the dollar push up import prices.
• Rising import prices produce inflation.
• Existing inflation leads workers, managers, and savers to expect future inflation.
• The Federal Reserve has to raise interest rates to create mass unemployment to keep those expectations of future inflation from turning into actual high inflation.

But there seems to be another line of argument back there: one in which demand for dollar-denominated bonds is diminished by the mere fact that they have been a money-losing asset class in the past, and that this is a source of excess volatility in the currency markets. Such excess volatility is a bad thing for U.S. consumers of imports--they will face lousy terms of trade. It is a good thing for U.S. manufacturing companies and their workers. It is probably a small net minus for the country as a whole. However, it is not enough of a net minus to justify the Federal Reserve hitting the economy on the head with a brick--raising interest rates to recessionary levels--in order to prop up the value of the dollar.

The potential problem is only if rising import prices make people scared of rapidly-rising inflation. This makes me think that a suggestion Greg Mankiw made once--that the Federal Reserve should focus on and disseminate not core inflation--inflation ex food ex energy--but supercore inflation--inflation ex food ex energy ex imports.

Comments

[James Wimberley]

Buiter's comments make sense if he's thinking simply of stopping a run on the dollar, a panic rush for the exit which would by itself only stop at an exchange rate far below any equilibrium. To do this, a central bank has to raise rates sky-high for a while. The last major run was I think Britain's ignominious exit from the ERM in 1992; not a good analogy since the fixed ERM rate presented speculators like Soros with a one-way bet, whereas the dollar floats. But the scale of a run on the dollar today would be much larger - hundreds rather than tens of billions. Would such an event really not have longer-term consequences?

[James]

I believe he is thinking in the internationalist model, rather than the domestic model. As such, if the US continues to run extraordinary levels of debts, and the demand for dollar debt slackens relative to supply, then long term rates will rise, since most new US debt is being sold internationally.

How that translates to domestic policy and model is another matter, and of less consequence to the prospective holders of debt.

[lostoption]

"...the weakness of the dollar, and fear of its future weakening, will trigger a large increase in long-term US interest rates..."

"Long-term US interest rates". Isn't that the basis of Buiter's argument? The retort seemed to assume that the issue was whether the Fed would take action. But, can the Fed significantly influence "long-term" rates?

[lloyd667]

Well, Brad, here's the model: uncovered interest parity.

Viewed in (for example) euros, an investor is comparing a return of "r" (the interest rate on euro-denominated assets) against one of "r*" (the interest rate on dollar denominated assets) less the expected depreciation of the dollar against the euro over the holding period of the investment.

This relation is what drives, for example, the famous Dornbusch overshooting model and, I would guess, features in almost any other international macro model these days.

The level of the exchange rate is irrelevant, except to the extent that it signals a change in the rate. For instance, if the dollar is viewed as "undervalued", market participants presumably think it will rise and will therefore, all else equal (as economists are fond of saying) will accept a lower dollar nominal interest rate.

Buiter's point is, I guess, that investors now believe the dollar to be overvalued, and so think it will fall. Thus, they will want to be compensated for this by a higher nominal interest rate on dollar denominate assets. This does not mean rates have to rise in the US, but the alternative of a decline in the rest of the world seems implausible.

[Ian Whitchurch]

The key moment will be when a member of the NYSE issues bonds denominated in something other than US dollars.

Losing the extraordinary priviledge will be hard for the US to cope with.

[kevin quinn]

Well lloyd667 beat me to the punch. Still, Brad must have something up his sleeve, for he is surely quite cognizant of uncovered interest parity. Maybe we're not taking delong view. Brad?

[Matt]

The model I use is that money represents a demand on economic production reserves. Or, cash reserves are there to cover any estimation error in future production.

Money is never meant to hold value longer than the spectrum of the underling production volatility.

This is the problem I see with my dynamic yield curve, http://stockcharts.com/charts/YieldCurve.html

The tail at the three month mark keeps sticking up because some large (mostly foreign) organizations are continuing to build up short term money reserves beyond what they need.

The fed's job, under the current management scheme, would be to continually lower interest rates to keep that tail from wagging. The result will be an inflation cycle, which at this point probably cannot be stopped.

Bernanke should announce his plan to keep the yield curve shapely. Foreigners at risk of losing reserve value should be instructed to open up investment houses in the U.S. to solve their problem.

Better yet, Bernanke should publicly explore the possibility of allowing foreigners to issue their own version of high powered dollars into the economy, make the monetary system competitive.

[knzn]

I'd like to go all the way to superdupercore inflation, which would exclude food, energy, and imports and be adjusted to correct for the food, energy, and imported content of products that don't fall directly into those categories. (For example, the Fed shouldn't mind if FedEx raises its rates just to compensate for rising energy costs.) I've advocated using unit labor costs, because it's a measure that already exists. But in principle one could include capital's chunk as well.

[knzn]

Possibly Buiter has in mind the recent tendency of central banks to hold longer-term securities as reserves. For example, if China were to stop pegging to the dollar, or to be expected to stop pegging to the dollar, it could cause the yield curve to steepen, as the demand for longer-term treasuries would decline. In principle, the Fed should compensate by cutting short rates, but it's easy to imagine the Fed passively resisting the potential inflationary impulse from imports.

[knzn]

lloyd667 and kevin quinn, I think Brad is assuming uncovered interest parity, but he is saying that, in most models, the adjustment would be in the exchange rate rather than the interest rate. That is, foreign investors may (initially) insist on higher interest interest rates and sell US bonds, but when they do so, they will convert their assets to other currencies, causing the dollar to fall. Once this process is complete (which should happen immediately), the dollar has fallen enough that it is not expected to fall further, and the old interest rate becomes an equilibrium once again, so they will buy back the bonds again with cheaper dollars. Really all this should be anticipated, so no real selling of bonds needs to take place, just a cheaper dollar.

Even in the Dornbusch model (if I remember and understand it correctly), it is (roughly speaking) the interest rate changes that cause the exchange rate changes (overshooting), not the other way around. Very roughly speaking, interest rates are set by Fed policy and expectations of Fed policy (if we hold foreign rates constant), and exchange rates have to adjust so that the expected future path of exchange rates is consistent with the expected path of interest rates set by the Fed. That's why you get a currency overshoot if something unexpected happens to change Fed policy.

In this case we're talking about a change in the expected long-run equilibrium value of the dollar rather than a change in Fed policy. Fed policy should still determine interest rates and provide a constraint within which current exchange rates will adjust to make the expected future path of exchange rates consistent with expected interest rates (covered interest parity).

[josh bivens]

to answer the question about a model where expectations about future changes, not levels, of the value of the dollar can lead to interest rate movements, i think brad advertised a (very good) piece by Krugman on the dollar recently.

from that paper

"The key to this approach is arguing that the real question is not whether the dollar must eventually depreciate. It is whether the dollar must eventually depreciate
at a rate faster than investors now expect.

That is, the only reason to predict a plunge is if we believe that today’s capital flows are based on irrational expectations – that the future path of the exchange rate that investors expect is inconsistent with a feasible adjustment path for the balance of payments."

From the dollar plunge, you move pretty quickly to a spike in interest rates, as investors herd for the door.

not sure if this is what buiter has in mind, but, it could be.

[anonymous]

Core inflation excludes certain categories of goods on the grounds that their higher volatility would introduce noise into the statistics. Fair enough. But this so-called "supercore" inflation would exclude goods not based on categories, but based on their origin. Is there really any objective economic rationale for such a redefinition? If we exclude Canadian widgets from inflation calculations, why not exclude red widgets or widgets made on Thursdays?

[Measure for Measure]

We would exclude Canadian or Japanese widgets if we think that their price fluctuations are transient, like that of food and energy.

But that's not the case: if anything import prices tend to be stickier as foreign firms seek to hold on to market share.

I can't see what DeLong and Mankiw are getting at.

[christofay]

Bernanke cuts and the price for oil, long term borrowed money, food, foreign currencies, and stuff goes up. What don't you see in that trend?

Relatively we're getting poorer.

Anyways it's been the world's worst overseas investors (reading the above) the Japanese, and the world's most profitable Communists, the Chinese central banks who have been bailing out the world's most overinflated (there's an appropriate term) central planner Greenspan. Do your rosy colored models account for central planning? Of course after falling for a while at some point the xera will reach a new plateau.

[christofay]

New Name:

The Socialization of Risk for the Well Off Republic of Berkeley with affiliated Cantons in Cambridge, Mass., Boston, Mass., selected streets of Manhattan, NYC, tony NYC suburbs in Cn., you get the idea.

[knzn]

Changes in import prices may be permanent (unlike, perhaps, typical changes in food and energy prices), but changes in the inflation rate for imports will be transitory if they result from exchange rate changes that are not echoed in domestic prices. Such exchange rate changes are a one-shot deal: prices change once, but they don't become unstable. Therefore, increases in import prices can be tolerated without violating the Fed's mandate to pursue price stability. Arguably, targeting a price index that includes imports would violate the Fed's mandate to pursue high employment.

[kaleidescope]

How about: Continued weakness of the dollar -- and the expectation that there's plenty more where that came from -- means that nobody wants to buy or hold U.S. Treasury Bonds. But continued government and current account defecits means the U.S. has little choice but to import credit from abroad. The price of T-bills plummets and interest rates soar in order to attract the needed credit.

Of course, the dollar could decline to the point that U.S. manufacturing-sector workers are suddenly competitive with teenage HIV-positive girls from the Chinese hinterland. In which case there wouldn't be a great need for credit from abroad and interest rates could remain low.

[Albrt]

I advocate a super-duper-duper-duper core inflation measure, in which the only component is the price of stone disks like the ones from Yap, but produced in Indiana. Of course, if any demand ever develops for Yap-Indiana stone disks we would have to change components.

By the way, those of us who find ourselves a dollar short on the grocery bill one week, two dollars short a few weeks later, then four dollars short a few weeks after that soon begin to base our expectations on the rate of price changes rather than a future price level. Even if we can't graph it or express it as an equation. I suspect, but cannot prove, that the Chinese have also figured this out. I leave the policy implications to my betters in the academic community.

[Ginger Yellow]

"The key moment will be when a member of the NYSE issues bonds denominated in something other than US dollars."

Huh? They do this all the time, albeit usually through European funding arms. Coke did a euro deal two weeks ago.

[Nahtanoj]

Various commenters seem to get the point -- expectations of a dollar decline reduce the demand for long-term dollar-denominated bonds, requiring an increase in long-term dollar-denominated interest rates to attract them back. It is easy to picture it happening, especially when the papers are full of stories about net buyers of US assets considering diversification out of dollars, severing their link to the dollar, etc. Once this happens, a "rush to the exits" might occur, as market participants try to get out before all the other participants do. But Buiter does not explain why the marginal holder of US bonds will act this way now, instead of three years ago -- or five years from now. (But maybe he doesn't have to show that now is the time -- he's just pointing out, in light of the drumbeat of articles along these lines, that such a "dollar run" is possible.)

[wood turtle]

Why for?

[piotr]

Albert, good but an imperfect idea.

I visited on occasion the commisary of Bundesministerium Des Finanz, or German Treasury Department, and I noticed that the prices are rether low. So, the inflation index should be based on meal costs as measred, let's be broad here.... the average of the commissaries of the Treasury Department, Federal Reserve and SEC. One could also include services in the basket, namely, what the officers in the respective institutions pay for parking. As you noticed, having Yap-Indiana stone discs in the indicator leaves too much for chance.

Now, if Fed wants to justify jacking up interest rates, all they have to do is to increase prices in their comissary. In turn, the Treasury can counteract by decreasing their prices. Hopefully, we will not fall into liquidity trap when Treasury workers get food for free.

Try as we can, one cannot devise "happily ever after".

[Maurice Lanselle]

Once upon a time, a high-level (director|manager) of a (market|business unit) of a multinational corporation (I was told) would pound his fist on the table and say "I disagree". Whether it was because he didn't understand what was being proposed and wanted it restated, or because his travel schedule was such that he wasn't sure what he had heard, I don't know. Be that as it may, pounding on the table and saying "I disagree" provoked plenty of explanation. Has enough now been explained?

[albrt]

piotr:

I bet it would be easier to get journal editors to look at an article with "Yap-Indiana" in the title rather than "Treasury Commissary." That's the real criterion for judging economic theory, no?

[resim]

We would exclude Canadian or Japanese widgets if we think that their price fluctuations are transient, like that of food and energy.

But that's not the case: if anything import prices tend to be stickier as foreign firms seek to hold on to market share.

I can't see what DeLong and Mankiw are getting at.