What Would Really Bring about a Dollar Dive?
One of the things about reading the op-eds and various articles in the blogosphere is the tendency to hype the possibility of the collapse in this, or the collapse in that. The most recent "bubble" in this type of writing involved hyper-inflation, commodities (silver, anyone?) and the dollar. Now I read things like QEIII would bring about a collapse in the dollar  (as if anybody really thought QEIII was politically likely, even if it were advisable on economic grounds), or easy monetary policy. Here's a choice quote from Jim Rogers:
"I would expect to see some serious problems in the foreseeable future.... By 2011, 2012, 2013, 2013, I don't know when, we’re going to have an economic slowdown again," he said. "This time it’s going to be a real disaster because the US cannot quadruple its debt again. Dr Bernanke cannot print staggering amounts of money again."
"How much more can they print without a serious collapse of the US dollar?" he said.
I can't figure out where that cited "quadruple" comes from. Debt held by the public (in current dollars) has not even tripled since G.W. Bush came into office, and has not even doubled since Obama came into office (see FREDII if you don't believe me). As a share of GDP, it rose from 0.49 to 0.63 from 2009Q1 to 2010Q4. Oh, well. Time to drop the hyperbole, and look at some data.
To begin with, I think it useful to ask whether we've actually printed a lot of money, where money is defined as currency plus checking deposits, or currency plus savings deposits (what we who teach macroeconomics, or teach money and banking, call M1 and M2).
Figure 1: Log ratio of M1 to real GDP (blue) and M2 to real GDP. NBER defined recession dates shaded gray. Source: Federal Reserve Board via FREDII, and BEA, 2011Q1 advance release, NBER, and author’s calculations.
As I've noted before, the Fed has more than doubled the money base, but this is not the same as doubling the money stock. That could happen if the banks lend out the excess reserves, but that hasn't happened as of yet. For a discussion of the correlation of money base growth and inflation, see this post.
Thinking about Stimulus, Monetary Policy, the Dollar, and Timing
I've read a lot about how deficits threaten the dollar. And certainly they can, but I think one has to have a more nuanced view than "reduce the deficit immediately, or the dollar crashes today". To think about this in an organized fashion, it's useful to consider the dollar's value in the context of several models. First, think about the dollar in a portfolio balance model  . In a static model version of the model, the current stock of US government debt relative to that of other countries’ stocks of government debt drives the risk premium on dollar assets. A stable relationship obtains if (1) preferences for government debt are constant, and (2) the correlation of relative returns are constant. But the story of the 2000's, and in some sense the "saving glut", is that the preference for US government debt has not been constant (e.g., [Caballero et al.]). Rather central banks have exhibited an upward demand shock for US government debt.
It is entirely possible that there will be a negative demand shock for US government debt, going forward. But the story of the last three years has been a series of shocks that have impelled flight to, not away from, US government debt. So, while it is conceivable that there will be no further sovereign debt shocks in the rest of the world (in which case the increase in US debt might very well induce a weakening of the currency), right now, with 2011 US net debt at 72.4% of GDP (IMF WEO April 2011), it is unclear why we should have a dllar crisis when the Euro area has a 66.9%, UK at 75.1% and Japan at 127.8%.This is not to deny that the debt trajectory is worrisome, with net debt approaching 85.7% in 2016. But that means one needs to do serious entitlement spending restraint and tax revenue increases that will take effect in the future, not nickel and dime-ing the one-fifth of total current budget accounted for by discretionary non-defense spending today.
That leads to the second point, that spending restraint and the early withdrawal of monetary stimulus now would exert a drag on growth, which would also weaken the dollar. A relapse in growth would also worsen the depression in tax revenues which accounts for a large part of the accumulation of debt during the Great Recession; hence, even in a portfolio balance model, too-early fiscal contraction could actually instigate a dollar dive.
There are other drivers for the dollar's value. Figure 2 highlights the correlation between GDP and the dollar’s strength, over the last 28 years.
Figure 2: Log real value of USD against broad basket of currencies (blue) and log ratio of US real GDP against RoW (export weighted) real GDP, detrended with quadratic in time. NBER defined recession dates shaded gray. Source: Federal Reserve Board via FREDII, and BEA, 2011Q1 advance release, NBER, and author's calculations.
Clearly, the strength of the dollar is correlated with the strength of the US economy relative to the rest of the world. (A DOLS(2,4) regression of the real dollar on relative GDP with a quadratic on time yields an elasticity of 1.82; however, the series do not appear to be cointegrated except at about the 20% msl.) There are a variety of reasons this correlation arises. The first is from the monetary approach  -- higher income induces a greater demand for money, and in the monetary approach to the exchange rate, that leads to a stronger currency (when prices are sticky, as in the Dornbusch-Frankel model). Contractionary monetary and fiscal policy implemented now which leads to a lower GDP in the future would mean ultimately a weaker currency.
Second, the higher income is typically associated with a higher real interest rate differential, partly due to Taylor rule fundamentals (see    ) and partly because higher income leads to greater investment in physical plant and equipment (that's how government spending could "crowd in" investment ), and hence demand for credit. That draws in capital leading to an appreciated currency.
The foregoing analysis places into context the proposals for fiscal retrenchment and monetary tightening now. The latter might induce a short term boost in the dollar's strength, but both would almost surely induce a medium-term weakening.
It must be remembered that the dollar in and of itself is not usually thought of as a target variable. The dollar is a relative price that is key to re-allocating aggregate demand, and at the same time allocating capital, across borders. Drastic moves in the dollar's value could indeed destabilize the financial system. But policy should not single-mindedly focus on the dollar's value. That's key for remembering why we want to stabilize the debt-to-GDP ratio -- it's to establish a sustainable path for output per capita over time.
How to Engineer a Dollar Crash
For certain, what would be key to causing a crash in the dollar's value would be a failure to raise the debt ceiling in a timely fashion. In almost any model I can think of, that would either cause a flight from US government debt, or -- even if we only go to the brink -- elevating the risk premium, and hence total interest payments, on US Treasury debt indefinitely. Thus, it's the height of irresponsibility to make unrealistic demands for deficit reduction based solely on spending cuts, thereby risking a crisis. [E.Klein] [M.Thoma]