Other than the weather, economic projections are probably the most common prediction seen in the mainstream media. And, like the weather, they are generally wrong.

The IMF puts out its World Economic Outlook twice a year -- once in the spring and again in the autumn. A case study for the lack of utility in economic forecasts can be shown by considering how the IMF's predictions for real annual GDP growth in the United States have held up since 1999.

The way to read this table is to take the year you are interested in on the vertical axis, and then move across the horizontal axis to see how the various annual IMF predictions have changed over time for the year in question. The diagonal of yellow shaded cells are the predictions for a particular year made in the autumn World Economic Outlook for that year.

Overall, we see a bit of a coin toss in accuracy. Just look at the year 2000. In late 1999, the IMF predicted U.S. real GDP would grow 2.6 percent in 2000. Then, late in the year 2000 itself, the IMF revised its projection for that year up to 5.2 percent. The next year (i.e., 2001), the economic growth for 2000 was revised down to 4.1 percent, then subsequently revised down further to 3.7 percent in the mid-2000s before being revised back up to 4.1 percent during the past few years.

The financial crisis -- and particularly, its magnitude -- was not predicted. In 2007, the IMF predicted a U.S. real GDP growth rate for 2008 at 1.9 percent. In late 2008, the growth rate for that year was thought to be 1.6 percent -- just a slight decline from the prediction made the year before. But over the next few years, the growth rate for 2008 was progressively revised downwards until reaching -0.3 percent.

Same goes with 2009. In 2008, the growth rate for 2009 was projected at 0.1 percent. Subsequent years saw the growth rate revised down to between -2.6 and -3.5 percent.

Also in 2008, the IMF predicted a growth rate of 3.1 percent for 2011. Now it is looking like the actual growth rate for 2011 was somewhere in the neighborhood of 1.5 percent.

Some years appear to have had reasonably predictivity, others not so much -- but there is no general predictable pattern about the level of predictivity (i.e., we have no idea in advance which years are more likely to be more incorrect than others), which leads to an unproductive tail-chasing exercise. Obviously there can also be strong political and other vested interest motivations behind economic predictions, further bringing them into disrepute beyond the simple quality (or lack thereof) assessment.

Like weather predictions, economic projections are unreliable. It is perhaps time to devote less public funds to both endeavors, and instead leave the crystal ball gazing to the private sector -- if there is a demand for it.