Simon Wren-Lewis’s post has sparked vigorous discussion   (and rejoinder) of whether the teaching of Mundell-Fleming, omitting uncovered interest parity, to our undergraduate students is defensible. In addition, he wonders whether it makes sense to use an LM curve, given the absence of a stable money demand curve, and the fact that most advanced country central banks target a policy rate. I think these are good questions for teachers, as well as those who provide advice regarding policy measures.
Pedagogical Issues: the Intuition/Rigor Tradeoff
When I teach Mundell-Fleming, I try to convey the intuition in a model without expected depreciation (and without an exchange risk premium), either under fixed or flexible exchange rate regimes. Then, I add expected depreciation, which is consistent with uncovered interest parity. Once one adds expected depreciation, it’s interesting to note that the results of fiscal and monetary expansions do not change qualitatively.
In other cases, I’ll try to incorporate expectations directly. This can be done using the IS-LM framework, or as Paul Krugman has. I use the Blanchard exposition, notes here. What I like in particular is the incorporation of IS-LM as well as the interest rate parity condtion.
Figure 1: Fiscal expansion. Note: Rise in E denotes home currency appreciation. Source: Blanchard.
Figure 2: Monetary
expansion contraction. Note: Rise in E denotes home currency appreciation. Source: Blanchard.
Fiscal expansions lead to a rise in GDP, a rise in interest rates, and a currency appreciation, in the absence of an exchange risk premium and default risk premium. A monetary
expansion contraction leads to a rise decline in GDP and a currency appreciation, when not at the zero interest rate bound. The same results would obtain assuming capital flows respond to the interest differential.
How’s the model work? Here are the graphs I usually provide to explain “the dazzling dollar” of the mid-1980’s:
Figure 3:Federal structural budget balance to potential GDP (solid line), and Fed funds interest rate minus lagged one year CPI inflation. Source: CBO, St. Louis Fed FRED, NBER, and author’s calculations.
Figure 4: Fed funds rate minus lagged one year inflation, and log real value of the US dollar against a broad basket of currencies. Source: St. Louis Fed FRED, Federal Reserve Board, and NBER.
Real World Issues
Uncovered interest parity (UIP) is one of the most important concepts in international finance. It is also one of the most difficult ones to verify. Uncovered interest parity is the no-arbitrage condition that one cannot obtain in expectation a higher rate of return in one currency than another. Since expectations are unobservable, tests of UIP are necessarily a joint hypothesis, usually incorporating the rational expectations hypothesis, in which case it is almost universally rejected at short horizons (Thanks to Simon for the cite of my recent work). I guess it’s a question with no particularly good answer whether to incorporate an assumption that is seemingly at variance with so much data (although I certainly object to short run/instantaneous purchasing power parity, the other central tenet of international finance.)
Reflecting my focus on emerging markets, I am somewhat less enthusiastic about assuming UIP in cases where capital mobility is less than infinite. Then “textbook Mundell-Fleming” might not be too bad at conveying the basic intuition.
Simon Wren-Lewis’s larger concern regards the inclusion of the LM curve. Here, I can see the merits of the argument –- in advanced economies, and many emerging market economies, the policy instrument is a policy rate, rather than some sort of aggregate money stock. Moreover, the stable money demand function which underpins the conception of the LM curve is an elusive creature.
That being said, the linkage of the money base to the external accounts is important for less developed countries, under fixed or managed exchange rates – assuming less than perfect capital mobility. When there is a balance of payments surplus, then money base increases, thereby increasing the money supply in the absence of sterilization procedures.
Now, it might be true that the monetary authority still targets a policy rate; but the resulting flat LM curve or MP curve would obscure the fact that sterilization procedures would have to be in place to keep the respective curves in place.
To be honest, I’m not sure what framework works best for conveying intuition when the economy being examined is at the zero lower bound (not relevant for most of the emerging market economies and the LDCs).
Teaching approaches differ; otherwise, we’d all use the same the textbook. What gets stressed will, I think, necessarily differ depending upon the economies analyzed, and phenomena being explained (US 1985 calls for a slightly different framework than China 2007).
In any case, I think that the conversation represents the best of what the blogosphere can provide –- an informed exchange of views of what works best in conveying economic intuition. (Note: So far, no name-calling, and no admonitions for a rewrite of theory in favor of an Austrian weltanschauung, or a Kondratieff wave approach to the universe.)
A great exposition of Mundell-Fleming with rational expectations and UIP by Maurice Obstfeld here. And a simple model to explain apparent deviations from UIP using an McCallum-like “exchange market shock” and central bank reaction function a.k.a. Taylor rule, here.