After shrinking away from the international economic news for a brief spell, the USD/RMB exchange rate is the hot topic of the day again. Part of this renewed interest is due to the imminent semiannual report  by the Treasury to Congress on exchange rate policies (due mid-April), part of it relates to the relevance of a competitive dollar for U.S. President Obama's national export initiative , and part of it is due to Chinese Premier Wen Jiabao's spirited defense  of the Chinese position. And, of course, part of it is the renewed comovement of the two rates since the global financial crisis, after a brief interruption between July 2005 and 2008 (see figure).
Source: World Bank staff calculations, from IMF IFS 
Many illustrious commentators have already thrown in their lot, including a dean of economics columnists , a distinguished think tank guru , a respected investment banker  and, of course, a Nobel prize winner . The World Bank, for its part, has also suggested greater exchange rate flexibility, in the quarterly update  issued by the Beijing office.
With such distinguished analysis already on the table, it would seem hazardous to wade into a debate where each side seems to hold strong views on the subject. But rather than take a position, I'll exposit a little on what academic research has had to contribute to the topic, organizing my thoughts along the three broad themes (some of which were alluded to  by Dan Drezner).
- On equilibrium exchange rates
While its entirely possible to claim  that the undervaluation of the yuan is indisputable on the basis of observed capital flows, the extent of a currency's over- or undervaluation must ultimately be considered against some theoretical reference rate. Of course capital export must, as an accounting matter, accompany a current account surplus. But the identity tells us nothing of the behavioral relationships that ultimately determine the variables in the balance of payments identity. For that, we need a model. It is this model that gives us a theoretical reference rate, and determining the fashion by which capital flows enter into this model-determined rate is a matter of, among other things, taste. Now, what do different models suggest? Estimates based on naive PPP  and FEER  (fundamental equilibrium exchange rate) models typically find some degree of undervaluation, while studies using an extended PPP  or BEER  (behavioral equilibrium exchange rate) approach produce more ambiguous results, especially if statistical uncertainty is taken into account. The bottom line that we need to recognize is that an overvaluation claim does rest on a model-specific notion of what an equilibrium exchange rate should be.
- On export and import demand elasticities
It is generally understood that exchange rate depreciations will generally lead to an improvement in the trade balance (subject to some technical criteria ). Even so, the magnitude of this adjustment to the exchange rate change will be conditioned by export and import elasticities. In this regard, however, the evidence is mixed on how much bilateral trade flows will budge due to a yuan appreciation. Some authors  (PDF) have found that the effect will be minimal, while others  find a much stronger response. Regardless of whether you may be an elasticity pessimist  or not, it is likely that any adjustment to the bilateral balance would need to come from a Chinese demand response (since a relatively more dearer RMB will simply mean that U.S. consumers start buying clothes, electronics, and toys from other low-wage economies like Bangladesh, India, and Vietnam). Is this forthcoming? Hard to tell, but one must imagine a massive increase in Chinese demand for (full-price) Hollywood movies and Silicon Valley software, farm equipment, and airplanes (the goods and services that that U.S. has strongest comparative advantage).
One more complexity that is introduced with expecting real exchange rates to bear the primary burden of correcting the bilateral deficit: Since current exchange rate correlations between the major trade deficit partners  and East Asian exporters differ wildly (see figure), the total effect of an RMB revaluation is difficult to pin down in global general equilibrium. While this is likely to reduce the U.S. trade deficit overall, engineering a successful rebalancing in this fashion does require that the Marshall-Lerner condition be satisfied for every country pair with an imbalance inconsistent with fundamentals. Guaranteeing this can be especially tricky in a post-financial crisis environment .
Source: World Bank staff calculations, from IMF IFS 
- On the structural determinants of current accounts
While forcibly realigning real exchange rates may serve to reduce external deficits in the short run, medium and long-run current account balances are ultimately a reflection of the underlying structural factors that determine saving and investment. Among other things, differences in national saving patterns , financial maturity  and crises , and both economic size  and growth prospects  across economies appear especially important. There is even a thesis that demographically-driven sexual competition  is a key determinant of excess household saving. Real exchange rates then must move as a consequence of adjustments to these underlying factors, in order to bring about changes in the current account. But here, real rates are endogenously determined. What an exogenously-imposed revaluation may do to the evolution of these structural variables is less clear, and the extent to which their adjustment to new equilibrium levels will be excessively disruptive is even less clear.
Update: More heavyweights weigh in on the currency issue: Goldman's chief economist Jim O'Neill  argues, among other things, that China's current exchange rate is not overvalued, but that domestic inflationary pressures are likely to induce the Chinese to voluntarily allow the RMB to appreciate, anyway. And another Nobelist  calls for cautious restraint in order to avoid a trade war  at a particularly inconvenient time.