Exchange Rate Expectations and Interest Rates
Since I’m (unhappily) short US Treasury bonds at the moment, I should be glad so many people think that uncovered interest parity means US bond yields have to rise when investors lose confidence in the dollar. (See the comments section of this post from Brad DeLong.) But it just ain’t so! Unquestionably, there are reasons to expect interest rates (bond yields) to rise when the dollar weakens, but those reasons are more subtle, and have different economic implications, than a brute force application of uncovered interest parity would suggest.
The theory seems simple enough: if investors expect the value of the dollar to decline, they will require higher yields on their dollar-denominated bonds, to compensate for the lower exchange value of the dollars in which they expect to be paid. The problem is that people take the clause “investors expect the value of the dollar to decline” to refer to a condition independent of how investors respond to that condition. In most models, what actually happens when “investors expect the value of the dollar to decline” is that the value of the dollar immediately declines, and then investors no longer expect the value of the dollar to decline. And most of those models include uncovered interest parity.
Suppose international investors, holding US bonds, suddenly come to believe that the value of the dollar will decline. What do they do? First they sell their bonds, which does cause yields to rise temporarily, but… What do they do with the proceeds? Do they sit on those dollars and wait for yields to rise enough to entice them to buy bonds again? I would think not, if they expect the value of the dollar to decline: instead, investors convert those dollars into (for example) euros and sit on those euros (or invest them in European bonds) until some combination of interest rate and exchange rate movements entices them back to the US bond market. But they don’t have to wait very long, because, in the process of converting those dollars to euros, they have pushed down the value of the dollar sufficiently that it no longer needs to decline. Suddenly, with the dollar no longer expected to decline, US bond yields become very attractive. They become so attractive that investors bid yields back down to their old levels.
In the paragraph above, I’ve snuck in some elasticity assumptions to assure that the value of the dollar doesn’t rise again when investors convert their euros back into dollars to buy back the US bonds. In principle, interest rates and exchange rates should be determined simultaneously. But if we believe in the expectation-based world of uncovered interest parity, then we must also believe in a similar expectational parity between short-term and long-term interest rates. Short-term interest rates are set by monetary policy, and long-term interest rates depend on expectations of future short-term interest rates. So if we hold expected monetary policy constant, all interest rates will be constant. If long-term interest rates rose, investors would try to exchange all their short-term bonds for long-term bonds, but since monetary policy assures a perfectly elastic supply of short-term bonds, this process would not stop until long-term rates came back down. Under these circumstances, the only way the market can adjust to a loss of confidence in the dollar is by bidding down the exchange value of the dollar immediately, until it is no longer expected to fall.
Now, you might point out that expected monetary policy will likely respond to the declining value of the dollar by raising interest rates. That’s true, since the cheaper dollar makes US goods more attractive and thereby produces a stimulus that monetary policy needs to offset. But that is a consequence of the monetary policy reaction function and the goods market equilibrium, not a consequence of uncovered interest parity. And it doesn’t allow one to make the argument, “When investors lose confidence in the dollar, US interest rates will rise, causing the economy to weaken,” because the Fed is only expected to raise interest rates enough to prevent the economy from strengthening, not to weaken it relative to its original condition. There are other, more subtle arguments you could make about why interest rates might rise further (the inflationary impact of the J-curve, behavior of foreign central banks, etc.), and the economy might actually weaken, but we have already wandered far afield from uncovered interest parity. If you want to argue that a loss of confidence in the dollar will weaken the US economy, you’ve got a lot of explaining to do.
The theory seems simple enough: if investors expect the value of the dollar to decline, they will require higher yields on their dollar-denominated bonds, to compensate for the lower exchange value of the dollars in which they expect to be paid. The problem is that people take the clause “investors expect the value of the dollar to decline” to refer to a condition independent of how investors respond to that condition. In most models, what actually happens when “investors expect the value of the dollar to decline” is that the value of the dollar immediately declines, and then investors no longer expect the value of the dollar to decline. And most of those models include uncovered interest parity.
Suppose international investors, holding US bonds, suddenly come to believe that the value of the dollar will decline. What do they do? First they sell their bonds, which does cause yields to rise temporarily, but… What do they do with the proceeds? Do they sit on those dollars and wait for yields to rise enough to entice them to buy bonds again? I would think not, if they expect the value of the dollar to decline: instead, investors convert those dollars into (for example) euros and sit on those euros (or invest them in European bonds) until some combination of interest rate and exchange rate movements entices them back to the US bond market. But they don’t have to wait very long, because, in the process of converting those dollars to euros, they have pushed down the value of the dollar sufficiently that it no longer needs to decline. Suddenly, with the dollar no longer expected to decline, US bond yields become very attractive. They become so attractive that investors bid yields back down to their old levels.
In the paragraph above, I’ve snuck in some elasticity assumptions to assure that the value of the dollar doesn’t rise again when investors convert their euros back into dollars to buy back the US bonds. In principle, interest rates and exchange rates should be determined simultaneously. But if we believe in the expectation-based world of uncovered interest parity, then we must also believe in a similar expectational parity between short-term and long-term interest rates. Short-term interest rates are set by monetary policy, and long-term interest rates depend on expectations of future short-term interest rates. So if we hold expected monetary policy constant, all interest rates will be constant. If long-term interest rates rose, investors would try to exchange all their short-term bonds for long-term bonds, but since monetary policy assures a perfectly elastic supply of short-term bonds, this process would not stop until long-term rates came back down. Under these circumstances, the only way the market can adjust to a loss of confidence in the dollar is by bidding down the exchange value of the dollar immediately, until it is no longer expected to fall.
Now, you might point out that expected monetary policy will likely respond to the declining value of the dollar by raising interest rates. That’s true, since the cheaper dollar makes US goods more attractive and thereby produces a stimulus that monetary policy needs to offset. But that is a consequence of the monetary policy reaction function and the goods market equilibrium, not a consequence of uncovered interest parity. And it doesn’t allow one to make the argument, “When investors lose confidence in the dollar, US interest rates will rise, causing the economy to weaken,” because the Fed is only expected to raise interest rates enough to prevent the economy from strengthening, not to weaken it relative to its original condition. There are other, more subtle arguments you could make about why interest rates might rise further (the inflationary impact of the J-curve, behavior of foreign central banks, etc.), and the economy might actually weaken, but we have already wandered far afield from uncovered interest parity. If you want to argue that a loss of confidence in the dollar will weaken the US economy, you’ve got a lot of explaining to do.
Labels: economics, exchange rates, interest rates, macroeconomics
posted by knzn at 11/21/2007 12:44:00 PM
2 Comments:
A few issues:
1) Equities. Movements in equities effect both the dollar and the exchange rate.
An expectation that equities will fall could create a falling dollar and falling interest rates at the same time.
You could say "Why not hold German bonds instead of US as the until the exchange rate stabilizes?"
Well, if you expect that at some point US equities will recover and at that point the dollar will surge , US Treasuries might be a better parking lot than German bonds.
2) Foreign government holdings. I read quickly somewhere that in Oct government purchases of treasuries exceeded net foreign purchases, implying perhaps that US treasuries are supported by China and OPEC even as the Euro investors are pulling out.
Suppose that investors believe that the dollar will decline because the Fed adopts inflationary policy and starts to expand the money supply. It would appear at the first sight that expanding the money supply at 10%/year would translate into gradual decline of value of the dollar at 10%/year. Consequently, yields on US treasury bonds would have to rise by 10% wrt say Euro-denominated bonds.
Right?
Looks like a paradox to me. Standard theory is that the Fed expands the money supply by buying bonds on the open market and that has the effect of lowering yields. But if you also have to take ECB into account, the effect is opposite.
Post a Comment
<< Home