PARIS: Three types of recessions exist. The first type are those caused by major shocks, say an outbreak of war or a sudden, sharp increase in the price of petroleum. Recall that the OPEC oil shocks of the 1970's incited two world recessions. The second category of recessions arise by chance, for example when consumer confidence dips, or businesses become uneasy and cut back on investment and/or inventory. This was the cause of America’s recession in the early 1990’s.

The third type of recession occurs when imbalances in an economy build up to unsustainable levels before exploding. This form of recession is sometimes characterized by vast increases in debt (corporate or consumer), or by dizzying stock market or capital asset speculations that eventually come crashing down. The “popping” of such an asset bubble is what happened in Japan 10 years ago, an event from which that country has not yet recovered.

Recessions of the first type are, almost by definition, largely unpredictable. Those of the second type are minor and relatively easy to repair if not to avoid. All that they usually require is a reduction in interest rates or a bit of reflation. Recessions of the third kind are the most worrying. Is America facing this third type of recession?

What, indeed, will be the nature of America’s next recession, that is, if America has a recession? Optimists predict a recession of the second type; pessimists suspect that a third variety recession is lurking in the economic shadows. Who is right? The answer depends almost entirely on one single number: the growth rate of American productivity in the future.

After adjusting for inflation, America’s stock exchanges roughly tripled in value in the 1990’s. Indeed, stock prices doubled compared to profits. The savings rate for households has gone from 5% of American GDP in 1990 to zero in 2000. America’s current account balance – the surplus of imports over exports – is now in deficit to the tune of 3% of GDP.

Do these figures reflect alarming imbalances or are they the usual signs of a boom? If America’s rate of growth in the coming decades remains what it has been over the past thirty years, the answer is simple, and terrifying. Stock market prices would be much too high. Households would be behaving in a totally irresponsible manner. The exterior imbalance would be unjustifiable, indeed, unsustainable.

If America’s annual rate of improvement in productivity turns out to be higher in the next thirty years than they have been over the past three decades, the country’s worries will be few. High stock market valuations would reflect a radiant future, households would be behaving in a reasonable way, and America would be wise to run a trade deficit in order to invest. Under such conditions, a mere nudge in interest rates by the Federal Reserve should set right any residual problem; if not, a tax cut will do the trick. Any recession would likely be short and limited.

What level of improved productivity can America expect in future years? Productivity growth will probably be higher than before 1995: the soaring growth rates of recent years are not accidents; the high tech explosion is real. All the same, I still fear that improvements in productivity will turn out to be lower than the prophets of the New Economy promise. This prospect literally gives me the shivers. Not only may the future be less radiant, but the adjustments that will be required once reality sets in could set off a nastier recession and a more persistent slowing of the economy than anything predicted today.