Carmen Medina (director of the CIA's Center for the Study of Intelligence) and Rebecca Fisher have an interesting article examining why economists had trouble forecasting the housing crisis / financial meltdown / etc. that we're living through, and what it suggests for other kinds of intelligence-gathering and analysis:
“Why did economists not do a better job anticipating the crisis?” was the question everyone seemed to be asking as the global economy began to unravel last fall. The consensus seems to be that most economists not only failed to see the crisis coming but also were downright hostile to the few who argued that The Great Moderation—the era of economic stability brought about by modern banking system controls—wasn’t so great after all....
The fact is that most economists and business experts did not anticipate this economic regression, or its particular timing, with any great degree of specificity, despite the astute analysis of Larry Summers and a few other highly regarded theorists. Economist James Galbraith estimated that, out of thousands of economists, perhaps only eight or 10 individuals really saw the crisis coming....
Leaving behind the issues of bias on the part of economists (which has already been discussed among intelligence officers, along many dimensions) and “group-think” because, again, we are deeply familiar with these pitfalls, six lessons from the economists’ experience seem to have unique applicability to what we, as intelligence professionals, do.
The six lessons are:
- There are no easy, obvious, straightforward policy responses to the economic crisis.
- We are overly sanguine about how close our information and intelligence sources approximate reality.
- Traditional economic analysis has trouble dealing with human irrationality.
- Timing is very different from analysis.
- How we think about causality in the world has great bearing on the priorities we set as an intelligence service and as a nation.
- The complexity of the modern world is overwhelming our existing intellectual and informational models.