Tim claims to be an economist but his blog summary of economic forecasters and their resultant consensus forecasts suggests he may be of little more value than the choir he is critical of. Im sure his real job is about quantifying data to enable better economic research and forecasting, but he doesnt seem to incorporate the necessity of his work into his over simplified analysis of economic forecasting.

In good faith, I was formally trained as an macro-economist and now work as an analyst on Wall Street. Moreover, I have not read his formal research; only this blog post. Bias in hand, here is what I think is wrong with Tim's analysis and the common-man view of economics.

First....."uh, duh!". Economic forecasting is difficult and even the state-of-the-art is "relatively" inaccurate. Amazingly, Tim's research found what is widely accepted as common sense....ECONOMIC FORECASTING IS REALLY FRICKIN HARD!!!!! The data sets are increasingly massive. Data is imperfect. Quantitative methods and predictive models are developmental and imperfect. People rarely behave consistently. Given all this, its amazing the forecasters are ever even close.

Second....It's actually all "relative". Tim's criticism appears to have essentially two points: That economic forecasts are either correct or incorrect; and that forecasters consistently make similar mistakes. Well, his data isnt too useful because it doesnt account for relativity. Economics makes relative comparisons. Measuring an economic forecast in binary terms (hit/miss) is like scoring golf by "holes-in-one" (yes/no). This measurement isnt fair or helpful because it equally values huge inaccuracies and near-misses. Well, what if all those forecasts were within a fraction of a percent of the true economic measurement? What if they were no more accurate than random? For example, predicting 5% growth before a 4.5% measurement is MORE ACCURATE than predicting a 5% contraction (negative growth) before the 4.5% growth. Moreover, consistently coming within 10% of the mark is clearly useful because it helps reduce risk, especially relative to random. Eliminating risk may be impossible, but reducing risk is not.

Hind-sight is 20-20, Tim. Economics is about predicting future events. Statistics is about reporting historical results. Which endeavor would be easier to accomplish accurately, Tim? Tims work appears to be hind-sighted. I'd describe him as a statistician or historian, rather than an economist.