After last Friday's announcement of the third consecutive drop in US employment figures, following hot on the heels of Ben Bernanke's admission that the American economy is already in recession - or more precisely that “a recession is possible ... there's a chance that for the first half as a whole, there might be a slight contraction” - I am probably the only economist left in the world who still believes that a US recession is likely to be avoided.
Obviously, I drew some encouragement for this view from the surprisingly decent ISM index, which triggered the huge rally in global stock markets on Tuesday (with some help from UBS and Lehman and the news of further regulatory moves to ease the credit crunch).
But the monthly employment figures are widely regarded as the real litmus test for the recession/soft landing debate.
So does Friday's 80,000 employment decline mean that the US is in recession? And does it matter anyway whether the technical definition of a recession is or is not satisfied? The answer to the first question is No and the answer to the second is Yes.
A drop of 80,000 in payroll employment is still not enough to signal a recession. If the US had already been in recession since the start of this year, as most economists believe, monthly payrolls would by now be falling by well over 100,000 a month, rather more than the average of 77,000 recorded in the past three months.
More important, other short-term indicators, such as the purchasing managers' indices, the industrial production figures and the quarterly consumption statistics (up a very decent 2.8 per cent in the fourth quarter), would be confirming the signs of a generalised downturn.
Instead, all these figures are still consistent with a mid-cycle slowdown similar to the ones that the American economy experienced in 1995-96 and 1986-87.
But does the distinction between a recession and a mid-cycle slowdown really matter anyway, given that the US is obviously suffering from one of the worst housing and financial crises in living memory - a fact that nobody (including me) can any longer deny? Actually, this distinction does matter quite a lot.
The difference between a recession and a slowdown is not just a matter of semantics, because there is a world of difference between a dislocation confined to one or two parts of the economy - say, housing and mortgage lending - and a generalised economic decline in which the weakness of demand in a few sectors creates a self-sustaining downward spiral of falling employment and incomes, weakening consumption and investment and further declines in activity across the economy as a whole.
It is the self-reinforcing and contagious nature of recessions that makes them different - and much more dangerous - than the sector-specific slowdowns that occur in market economies all the time.
This is why the National Bureau of Economic Research defines recessions in a very specific way: “A recession is a significant decline in activity spread across the economy, lasting more than a few months, visible in industrial production, employment, real income and wholesale-retail trade.
A recession influences the economy broadly and is not confined to one sector. Expansion is the normal state of the economy; most recessions are brief and they have been rare in recent decades.”
This definition of recession is not just a matter of semantics. The broad-based and lasting nature of the economic decline is absolutely essential because a sector-specific or transitory slowdown is not sufficient to counteract the natural expansionary momentum of a capitalist economy.
This, incidentally, is why comments frequently seen in the media about a “housing recession” or “manufacturing recession” or “retail recession” or “export recession” are always technically inaccurate.
Problems affecting only one or two parts of the economy are a continuous feature of any market economy with its constant variations in demand and supply for particular products and services.
Such fluctuations can be very painful and disruptive for the sectors involved, but at the macroeconomic level they rarely do much serious harm.
But surely the financial and housing crises in America are now so severe that there can be no hope of avoiding the sort of self-sustaining downward spiral described above?
Apart from the lack of statistical evidence (so far) of a recession already mentioned, there are three other reasons for continuing to resist the consensus view.
First, the occurrence of a severe financial crisis is not, in itself, a sufficient reason for expecting a recession.
In fact, there have been severe financial crises that did not lead to recessions in the middle of every previous economic cycle: for example, the Latin American defaults of 1984, the stock market crash of 1987 and the savings and loan and housing collapse of 1988-89 and the Mexican peso, Asian and Russian crises of 1995-98.
A second reason why an American recession will probably be avoided - or, if it does occur, will be extremely brief - is that powerful expansionary forces are about to come into play in the US economy in the months ahead.
From the second week of May onwards, every American household will receive tax rebate cheques of between $600 (£300) and $2,000. As a result, personal disposable income will grow at an annualised rate of well over 10 per cent in the third quarter of this year.
Even if only half this money is spent, US consumption will therefore be guaranteed to grow by at least 3 per cent in the third and fourth quarters. And beyond that, the lagged effects of the Fed's recent monetary easing will kick in from the start of 2009.
This is why Mr Bernanke could afford to state so confidently on Wednesday that the US economy would return to trend or above-trend growth by early 2009.
Thirdly, there is the US housing market.
Financial markets imply a further decline of about 20 per cent in US house prices, but financial markets are sometimes wrong (as must surely be obvious by now).
US house prices have already fallen almost 15 per cent from their peak and as a result property in America is no longer expensive in relation to average incomes.
In fact, the ratio of average house prices to personal disposable incomes is now 5 per cent below its 40-year average and only 9 per cent above the record low it reached in 1990 and again in 1995.
Given that disposable incomes are rising at about 6 per cent annually, the housing valuations would fall to a new all-time low within 18 months or so, even if there were no further decline in house prices from now on.
And the last time American homes were as cheap as they are today in relation to disposable incomes, interest rates were much higher.
For example, in 1990 and 1995, when house price to income ratios were last at about present levels, standard 30-year US mortgage rates were 10 per cent and 8.5 per cent respectively.
Today, the corresponding rate is 6 per cent. As a result, affordability indices that take both interest rates and incomes into account show US property to be very good value already.
For example, the National Association of Realtors' composite affordability index stands at 135, compared with a record high of 140.8 reached in 1999, at the start of the recent housing boom.
After a few more months of falling house prices or rising disposable incomes, American housing will start to look irresistibly cheap.
In sum, if America can get through the next month or two without sinking into a serious recession, the danger should be past by the second half of this year.
anatole.kaletsky@thetimes.co.uk