If forecasters admit that their forecasts are based on coincident indicators released with a lag, then why do we need them at all?
Economists were clearly stunned by the strength in consumer spending evident in the recent retail sales report for January.
Non-auto sales rose 1.2%. Excluding gas station sales, which tend to be driven by changes in price, they posted the fourth consecutive solid monthly increase.
'Our projection for total real consumer spending growth this quarter (-1.5% annualised) looks much too low now in the wake of this report,' wrote economists at Goldman, Sachs & Co.
'Consumption growth could be stronger than the 2.5% decline we have been assuming,' wrote the folks at Deutsche Bank, in a model of understatement.
'The consumer seems to be ignoring the bad news in the labour market,' Lehman Brothers' economists noted before raising their first-quarter real GDP forecast by one percentage point to 1.5%.
'This is the second time this year that incoming information has suggested an upward revision to our GDP growth estimate for the first quarter, which in itself is an economic recovery signal,' wrote Bear, Stearns & Co. (Is it a recovery signal because they missed the cues?)
The above snippets from the daily economic missives are not meant to suggest that Goldman, Deutsche, Lehman and Bear are bad forecasters, although some of these folks need to understand the difference between a demand-driven and supply-induced rise in market rates.
If these are what we call forecasters, and they themselves admit that their forecasts are based on coincident indicators released with a lag, then why do we need them at all?
'What you see out the window is what the forecast is,' says Ram Bhagavatula, chief economist at Royal Bank of Scotland Capital Markets, who had the second-highest forecast for first-quarter real GDP (4%) in the semi-annual Wall Street Journal survey in January.
'There are leading indicators of the economy,' Bhagavatula says. 'Policy changes lead to an economic response. The lags are not precise but when you cut the fed funds rate by 475 basis points in 12 months, historically that qualifies as an enormous easing.'
What strikes Bhagavatula as odd is that none of this is either high-tech or proprietary.
'I had lunch with Mickey Levy (chief economist at Bank of America Securities) the other day, and we laughed at the idea that we are two old dogs and do not even have to learn new tricks,' Bhagavatula says.
The first step to forecasting is believing in something: in this case, the ability of monetary policy or interest-rate changes to stimulate the economy. While it is fashionable in each business cycle to assert that this time is different and monetary policy is impotent, it is hard to find a historical example to support it.
Before you point to the Great Depression in the US and the lost decade of the 1990s in Japan as examples of monetary policy not working, the level of interest rates is not always representative of the stance of monetary policy. Low interest rates with slow or no money growth (in the case of the Great Depression, a huge contraction in broad money) does not represent a stimulative policy.
If you do not believe that monetary policy has an effect ' if you believe that the economy sort of ebbs and flows of its own accord ' then I cannot help you. You will have to look out the window for your forecast.
Once you have a bias that the economy will improve, then it is a question of figuring out when, Bhagavatula says. Policy lags can be long and variable, yet historical relationships are often useful.
Take 1998, for example. The prevailing view was that the world was coming to an end following Russia's default and the near-collapse of hedge fund Long-Term Capital Management.
The Fed cut interest rate three times, which had the effect of calming the financial markets and goosing an economy that was already motoring. The monetary stimulus drove the stock market bubble and the economy.
The current cycle poses certain forecasting challenges, to be sure. The bursting of an asset bubble is not an everyday occurrence, providing little guidance for economists comfortable with a business cycle in which the central bank fights inflation, causes a recession and then fights the recession.
Still, monetary policy affects aggregate demand, which sends a signal to producers. This time around, demand never collapsed.
Final sales to domestic purchasers, a good measure of final domestic demand (it excludes inventories), never went negative on a year-over-year basis as it did in previous recessions. The only quarterly decline was a 0.3% fall in the third quarter, which to date is the only negative quarter for real GDP.
The strength in demand led Bhagavatula to look at the leading indicators of production, which had started to improve before 11 September.
'The consumer was affected less than people expected (by 11/9) while producers were affected as much as people thought,' he says. 'Production was cut while consumers kept spending.'
After steep cutbacks and a record inventory draw down in the fourth quarter, 'production was the only piece that could adjust,' Bhagavatula says.
It sounds almost too simple to be a forecast.
New York, Bloomberg newsroom
Duo start roles on 1 October
Where true value lies
Economy to thrive despite global risks
Behaviours, animals or something else?