Three is a nice long article in the New York Times by David Leonhardt, in which we learn, among other things, that Tim Geithner is actually an old China hand. Which should come in handy as he tries to right the bilateral imbalance between the two countries that has clearly become unsustainable.
Most writing on this subject, the Leonhardt piece included, concludes that the Chinese end of the solution lies in stimulating consumption demand in China, and reducing that country's ridiculously high saving rate. But the discussion typically ignores an important issue: the secret of China's growth is that it has made a rapid transition into producing tradables (mainly manufactures). If the correction in China's trade balance comes at the expense of slowing down this process of structural change, it will also result in a reduction in the economy's longer-term growth rate. This would be a bad bargain for China, and it is an important reason why the Chinese authorities have resisted a significant appreciation of their currency (which, in the absence of other compensating policies, would have the effect of reducing the profitability of investment in tradables).
So if growth in China relies on continued structural change in the direction of tradables, but the international context no longer allows a large external surplus, must the world have to choose between global macro stability and Chinese growth? Is there any way out of the conundrum?
Actually, there is, but we need to think a bit out of the box. The economics of the situation is actually pretty simple. If China wants a larger supply of tradables than the market equilibrium produces, it can achieve this by directly subsidizing the domestic production of tradables (through tax incentives or rebates or reduction in the cost of inputs), while letting the real exchange rate appreciate to equilibrate the external balance. Direct subsidies don't tax the domestic consumption of tradables the way that currency undervaluation does. Presto! We have both the structural change China needs, and the external balance the world economy requires.
Put differently, if China gives up the exchange rate as an instrument of industrial policy, it will need a substitute. Explicit industrial policies are the obvious substitute. Across the board spending by China on infrastructure and other things doesn't quite achieve the required goal, unless the spending is targeted on projects that disproportionately reduce the costs of producing tradables.
There are of course small details to worry about, such as the WTO's Agreement on Subsidies. But we do need to think creatively about squaring the circle in question, and I cannot come up with any other alternative.
If you want to read more about this, here's a paper that explains the argument in greater detail.