IN 1996 Alan Greenspan began asking why the flashy information technology spreading across America seemed not to be lifting productivity. He was not the first to wonder. A decade earlier Robert Solow, a Nobel prizewinner, famously remarked that computers were everywhere but in the statistics. But Mr Greenspan was uniquely positioned, as the chairman of the Federal Reserve, to experiment on the American economy. As the unemployment rate dropped to levels that might normally trigger a phalanx of interest-rate rises, Mr Greenspan’s Fed moved cautiously, betting that efficiencies from new IT would keep price pressures in check. The result was the longest period of rapid growth since the early 1960s. Despite his success, few central bankers seem eager to repeat the experiment and many remain blinkered to issues other than inflation and employment. That is unfortunate. A little faith in technology could go a long way.

Central bankers are not known to be a visionary bunch. Turning new ideas into more efficient ways of doing things is the job of firms. The capacity of an economy to produce—the supply side—is primarily shaped by things such as technological progress, population growth and the skill level of the workforce. Monetary policy is typically thought not to influence this process. Its responsibility is the demand side of the economy, or people’s willingness to spend. Central bankers typically see themselves as drivers who press on a vehicle’s accelerator and brakes. The state of the engine is someone else’s bailiwick.

Not all economists have seen so sharp a delineation between supply and demand. In 1973 Arthur Okun mused that in an economy with very low unemployment firms would coax more output out of their workers. More efficient firms would outbid less efficient ones for scarce labour, boosting productivity. By letting spending grow rapidly and unemployment tumble, a central bank might induce productivity to grow faster. In the 1980s Olivier Blanchard and Larry Summers further developed this notion in their work on “hysteresis”. They reasoned that, if weak demand led to a long period of joblessness, workers might find their skills becoming obsolete and their connections to the labour market eroding. A short-run monetary failure could create a long-run drop in supply. Correspondingly, a central bank that responded to recession by allowing unemployment to fall to inflation-stoking levels might find that this overheating lures discouraged workers back into the labour force, and pushes firms to give them the training and equipment they need to thrive. Demand, in such cases, might create its own supply.

In fact, the role of a central bank in managing productivity is even more fundamental than these theories suggest. Good monetary policy is essential to capturing the full benefits of new technologies. Suppose, for example, that a tech firm creates a cheap, AI-powered, wearable doodah as good in monitoring health and diagnosing ailments as going to the GP. Deploying it takes some capital investment and hiring, but also leads to much larger reductions in spending on conventional practices. In other words, this magical innovation leads to a rise in the productivity of health services. Hurrah for that! But the need to shift resources around in response to this disruptive new technology creates some difficulties. Spending on health care is a reliable source of growth in employment and in demand. A sudden drop in such growth might push an economy into a slump. The cost savings that consumers, health insurers and governments enjoy thanks to the new technology would help; perhaps some people would plough their newly saved cash into elective procedures like plastic surgery, at clinics which might then have to expand and hire new workers. But there is no guarantee that lost spending on doctors and related equipment will be offset by increases elsewhere.

Indeed, in a paper published in 2006, Susantu Basu, John Fernald and Miles Kimball concluded that advances in technology are usually contractionary, tending to nudge economies towards slump conditions. They estimated that technological improvements tend to depress the use of capital and labour (think, in this example, stethoscopes and doctors) and business investment (new clinics) for up to two years. To those living through such periods, this depressing effect would show up in lower inflation and wage rises. That, in turn, suggests that an alert central bank with an inflation target ought to swing into action to provide more monetary stimulus and keep price and wage growth on track. That stimulus should spur more investment in growing parts of the economy, helping them to absorb quickly the resources freed up by the new, doctor-displacing technology and thus averting a slump.

Two obstacles usually get in the way of such a benign outcome. First, these steps unfold with a lag. The slowdown in price and wage growth will be gradual, as displaced workers tighten their belts and compete with other jobseekers for new employment. Central banks might then wait to see whether low inflation reflects a genuine economic trend or is merely a statistical blip. Even after they act, their tools take time to have an effect.

What is not seen

The greater difficulty may be the trouble that central bankers have in imagining that dizzying technological change is possible, let alone imminent. And the risks they face are asymmetric. Had Mr Greenspan been wrong, the high inflation that resulted would have been there for all to see; had he played it safe, no one would have known that a boom had been achievable. Such possibilities can only be guessed at; they are not found in the data. Sober technocrats are not given to leaps of faith. But to risk a bit of inflation for a chance at a productivity-powered windfall is a wager more central bankers should make.

Visit our Free exchange economics blog