The spread of the global financial crisis to Latin America in general, and Brazil in particular, has changed the complexion of the problem in several ways. Obviously the stakes are now even higher than before, especially for the United States; obviously, also, the fact that contagion can take place in this way makes it even harder than before to blame the whole crisis on "Asian values" and "crony capitalism". Beyond this, Latin America’s troubles bring out in an even clearer fashion than before the unpleasant dilemmas that all developing countries (and perhaps even smaller developed economies) now face.

 A good way to illustrate these dilemmas is to consider a classic analysis of macroeconomic policy in an open economy: Trevor Swan’s 1955 discussion of the difficulty of reconciling "internal balance" (i.e., more or less full employment) with "external balance" (an acceptable current account deficit) . The "Swan diagram" analyzes the economic effects of two kinds of policies: those that affect the overall level of domestic expenditure, such as the fiscal deficit; and those that affect the relative demand for domestic and foreign goods.

 The figure shows a standard Swan diagram. We imagine a country with a pegged exchange rate and high capital mobility, so that interest rates are determined by the need to avoid rapid depletion of reserves, and in effect monetary policy is removed as a tool of stabilization. Thus the "expenditure level" policy variable on the horizontal axis is fiscal; glossing over many complications, we can simply think of it as the budget deficit. On the other axis we show an "expenditure composition" variable, the cost of production in our country relative to that abroad.

 What Swan pointed out was that the nature of the difficulties facing a country depend on where in this space it resides. To see this, we draw two curves. One curve represents conditions under which the country has "internal balance"; as drawn, it is upward-sloping. The reason is that any rise in the country’s relative costs would tend to reduce exports, increase imports, and thus reduce employment; to compensate, to keep employment constant, the country would need to have a fiscal stimulus – a larger budget deficit. At any point to the right or below this internal balance curve, the economy will suffer from too much demand for its goods, and will experience inflationary pressures. At any point above or to the left, it will suffer from unemployment.

 The other curve shows conditions under which the country has "external balance". It slopes downward, because an increase in spending would other things equal increase the current account deficit; to offset this the relative cost of production in this country would have to fall. At any point below or to the left of the external balance curve, the country will have a current account surplus (or at least a deficit below what is really appropriate), at any point above or to the right an unacceptably high current account deficit.

 These two curves define four "zones of economic unhappiness", shown in the figure; only at the point where the two curves cross is the economy without problems internal or external. (All happy economies are alike; each unhappy economy is unhappy in its own way). A country may have an external surplus or deficit; it may have unemployment or inflation. And by considering what kind of economic problems a country has, one gets a clue as to what kind of policy action is appropriate.

 Which brings us to Latin America. Focus on Brazil, although Argentina’s situation is similar in some respects. Brazil is clearly in the top zone of unhappiness: there is little inflation, but high and rising unemployment, together with a worrying external deficit. The textbook analysis clearly indicates, then, that Brazil’s relative costs are too high. In principle, at least as far as macroeconomic policy is concerned, it is not clear which way the budget deficit should go (although realistically the size of that deficit raises enough concerns about long-run solvency that it surely must be reduced whatever the diagram says).

 How can relative costs be reduced? Well, the obvious, textbook answer is a devaluation, which brings about an immediate reduction in costs measured in terms of other countries’ currencies. And five years ago many economists – myself included – might have cheerfully recommended a one-time devaluation, or a temporary period of floating, as a way to get Brazil’s relative costs into the right place. After all, Britain and Sweden did it in 1992, and nothing bad happened. Indeed, by any measure the devaluing economies did better than the hard-currency countries.

 But nobody who looks at the terrible experiences of Mexico in 1995 or Thailand in 1997 can remain a cheerful advocate of exchange rate flexibility. It seems that there is a double standard on these things: when a Western country lets its currency drop, the market in effect says "Good, that’s over" and money flows in. But when a Mexico or Thailand does the same, the market in effect says "Oh my God, they have no credibility" and launches a massive speculative attack. So the question for Brazil is, do you think that the market will treat you like Britain, or do you think it will treat you like Mexico? And this is not an experiment that any responsible policymaker wants to try.

 And yet the economic unhappiness remains. Indeed, even if the current speculative pressure on Brazil lets up a bit, the country will clearly be forced both to retain high interest rates and to make massive cuts in its budget deficit – pushing it much further from internal balance, and probably causing a sharp recession. In effect, fear of speculative attack has paralyzed macroeconomic policy, and even forced it into acting perversely.

 So what are the options? Brazil – and many other developing countries – now have three possible courses of action, all extremely dangerous. They are:

  1. Hold the line on the exchange rate, and rely on gradual reductions in relative costs – via productivity increases and deflation relative to the rest of the world – to restore internal balance. In principle this should eventually work. However, the operative word is "eventually": all experience (Britain in the 1920s; France since 1987) suggests that this is an extremely protracted process. Even aside from the sheer economic cost, can the social and political fabric stand the strain?
  2. Hold your breath, cross your fingers, and devalue or let the exchange rate float – accompanying this with market-friendly policies like sharp fiscal contraction and privatizations, in the hope that the markets will treat you like Britain instead of Thailand. But they probably won’t.
  3. Impose temporary currency controls to prevent a speculative attack, then use the breathing space to engage in a one-time devaluation-cum-fiscal-stabilization, in the hope that after a little while the markets will calm down and let you return to business as usual. But currency controls are hard to implement and enforce, they disrupt normal trading relations, and they may impair confidence for a long time to come.

 What is the right answer? Honestly, I don’t really know – but neither does anyone else. What is clear is that something has gone tragically wrong with the whole system, if these are the available alternatives.