What's less clear is whether a computer would have done better than central bankers during the global financial crisis. The deep downturn in economic activity required them to take creative and dramatic action to save the global economy from freefall. Unprecedented forms of monetary support like Ben Bernanke's "credit easing" and the return of Operation Twist could hardly have come from a computer programmed to buy and sell the same government securities over and over again.
A computer would also have fallen short at the European Central Bank, despite its narrow focus on inflation. Last year, the growth rates of the 17 economies in the eurozone ranged from -7 percent in Greece to +7 percent in Estonia, adjusted for inflation. How could you set a single monetary policy for all of these countries? It would be impossible for anyone, human or machine.
With these experiences in mind, the right way to think about monetary policy may be like driving a car on the highway. Today's cars can be driven manually or using cruise control. Under normal conditions, cruise control works just fine, maintaining a constant speed by adjusting the amount of fuel going to the engine. When the driving gets a little hairy, you want humans in charge. Their reactions may not be perfect, especially in retrospect, but the cruise control computer doesn't have the same analytical ability or range of available actions.
The driving has certainly become pretty hairy in the United Kingdom, and so it's understandable that the Bank of England would want to draft a foreigner widely seen as at the top of his field. To his credit, Carney is nothing like Greenspan -- his statements are sharp and transparent, using a combination of plain language and data to convey his conclusions. But the days of stardom for people like Carney may be numbered.
Put simply, the computers are catching up. Within a decade, self-driving cars will likely become commonplace on American highways. Just like central bankers, the cars will have to process a lot of information quickly and use a limited number of tools to choose a safe path forward despite constant uncertainty.
Of course, self-driving cars will have to perform as well or better than humans, even in the most difficult situations. Yet if a rule as simple as Taylor's would already have performed as well or better than Greenspan did throughout his entire tenure at the Fed, then surely a computer to replace Bernanke or Carney is within the capacity of current technology. The only questions left are who will program it, and who will be the first to give it a test drive?