Forecast Errors

3/06/2013 09:30:00 AM
A lot of people have noted that economic forecasts throughout the recovery have been systematically too optimistic. Ezra Klien posted this excellent graph of successive economic forecasts since the recession began:

One question that remains to be resolved is whether this forecast error is due to bad forecasting technique (or random shocks), or is policy induced.* One theory I have is that by forecasting a recovery, economists have been, throughout the recession, implicitly signaling policy makers that they can ratchet up austerity measures. In reality, those forecasts all relied on the assumption that policy would not be substantially altered, and in particular not altered in response to the forecast. I'm not sure that assumption is satisfied. Here is one example:

Not long ago, all the major forecasts were saying that housing inventories would finally bottom in 2012, and start growing in 2013. Then we faced another debt ceiling crisis during which congress punted on the sequester cuts, ultimately allowing them to happen, while failing even to raise the debt ceiling (they offered only a three month exemption from the ceiling). Now we have tens of billions in spending contraction, on top of hundreds of billions already enacted, plus a likely total government shutdown over the budget resolutions, and another debt ceiling crisis. And all that policy-induced panic has really paid off: forecasters now thing that housing inventories will continue to fall throughout 2013. This is, in my view, a policy-induced forecast error.

Meanwhile, Congress has once again deferred the economic recovery for another year.

*Turns out there actually is research showing most forecasting efforts are basically worthless. The methodology, though, cannot discern whether this is because the techniques just don't work, or policy and other variables react endogenously with our forecasts, as theorized above.