The way Washington works, something is true today if we thought it was true yesterday and the day before.
This is great news for bad ideas. Once a career government official hard-wires a belief into the bureaucracy, it is there forever.
This helps explain the mess we are in. An ossified regulator sets up a stiff and inflexible regime of rules, confident it will be effective. Those being regulated adjust their behavior and get away with economic murder.
Likewise in fiscal policy, the tyranny of dead ideas funnels taxpayer dollars toward ineffective programs. The best example is the $862 billion economic stimulus.
Last week the Congressional Budget Office reported that the stimulus enacted a year ago created 1.4 million to 3 million full-time-equivalent jobs in the fourth quarter. The CBO drew heavily on commercial forecasting models and the macroeconomic model used by the Federal Reserve.
Both of these old-school Keynesian models, by the CBO's own admission, "tend to predict greater economic effects" from legislation like last year's stimulus.
As the CBO put it, the range of estimated job creation was "intended to encompass most economists' views and thereby reflect the uncertainty involved in such estimates."
Let's pause for a short history.
Era of Models
Back in the 1960s, economists thought that large Keynesian models like those relied upon by the CBO could enable them to simulate with great precision the response of the economy to different policies. After an explosion of model building, activity in this area ground to halt, not because the perfect model was developed, but because the effort was convincingly discredited.
I received my Ph.D. in the 1980s from the University of Pennsylvania, which was a center of macro model building in the 1960s and 1970s. By my time there, such modeling had become no more than a footnote in the graduate curriculum.
The large-scale Keynesian forecasting models were discarded by most of the profession because they didn't work. One of the first to demonstrate this was Charles R. Nelson of the University of Chicago, who in 1972 showed that forecasts based on simple extrapolations significantly outperformed theory-laden macroeconomic models in competitions.
About a decade later, Virginia Tech economist Richard Ashley performed a similar exercise that found the big macro models "so inaccurate that simple extrapolation of historical trends is superior for forecasts more than a couple of quarters ahead." To paraphrase Ashley, all you need to outperform the fancy models is a ruler and a pencil.
Economists went on to refine techniques that correlate the way things move with each other over time, relying on hard evidence rather than Keynesian theory. These techniques generally find that policies like last year's stimulus have much smaller effects than the macro models would suggest.
Yet the analysis of job creation of the stimulus bill essentially ignores this vast literature.
Here is the best sign of how bad it is. Last month, one of the leaders of the macroeconomics profession, Robert Barro of Harvard University, wrote a piece for the Wall Street Journal that drew on his own experience studying the response of the economy to fiscal policy.
Barro has been one of the primary contributors to the macroeconomic time-series literature that has tried to estimate effects from observed economic data, rather than assume effects, as Keynesian models do. He is a lock to win a Nobel Prize someday. Barro wrote that his work suggests that the stimulus increased output last year by about 0.8 percent--an estimate below the range that the CBO claimed would characterize the breadth of views of economists.
If the stimulus gave a smaller boost in gross domestic product, it also created fewer jobs.
Rather than rely on large-scale computer models that are filled with tenuous assumptions to simulate what might have happened, a better approach is simply to let the data speak. Many countries, including the U.S., have tried economic stimulus measures. We can study the evidence of what happened.
The CBO's estimate range, which would have been correct in 1969, is silly today. Its analysis is based almost exclusively on speculation within the context of old-fashioned Keynesian models. The fact that these effects are inconsistent with actual experience, as captured by analysis like Barro's, is inexcusably ignored.
CBO's Rosy View
When political leaders ask the CBO what will happen if a big stimulus bill is adopted, they get an intellectually indefensible, rosy scenario. If you wonder why the stimulus was so large, perhaps it's because the CBO told Democrats that it would be so good. The fact is, our giant, ineffective stimulus was designed for economic models that were discarded when Richard Nixon was president.
It's a safe bet that if 8-year-old Sasha Obama ever becomes president like her father, the bureaucracy will still be using the same models.
Kevin A. Hassett is a senior fellow and director of Economic Policy Studies at AEI.