Science and technology

(1) Economists are often puzzled when a technology that is plainly powerful fails, for years, to show up in measures of productivity. A common explanation is that the technology is simply less useful than its enthusiasts claim. An alternative account holds that the most valuable general-purpose technologies, those that can be applied across many industries, deliver their largest gains only after the organizations using them have been redesigned around them, and that such redesign is slow.

(2) The electrification of manufacturing is the case most often cited in support of the second account. Nineteenth-century factories were typically powered by a single steam engine whose motion was carried to individual machines through a system of rotating overhead shafts and belts. Because power was lost along every length of shafting, machines had to be clustered close to the engine, which pushed factories into cramped multi-storey buildings. The first factory owners to adopt electricity typically replaced the steam engine with one large electric motor and left the shafts in place. Their power bills sometimes fell, but the way work flowed through the building did not change, and neither, to any great degree, did output per worker.

(3) The substantial gains came later, when manufacturers gave each machine its own small motor. Freed from the shafts, machines could be arranged in the order in which a product was actually assembled; buildings could spread out into single-storey plants with better light; machines could be added or moved without rebuilding the power system; and a breakdown in one machine no longer halted every other. None of these benefits required a better motor; they required a different factory. And a different factory was costly: it usually meant new buildings, new ways of supervising work and a workforce trained for different tasks. Many firms therefore waited until their existing plants wore out before rebuilding.

(4) Proponents of the redesign account argue that the lesson applies well beyond electricity. When a firm installs new software but keeps the approval procedures and job boundaries that its old paper-based system required, they contend, it is reproducing the single large motor attached to the old shafts. Skeptics reply that the analogy can be used to excuse almost any disappointing technology, since its benefits can always be said to await a reorganization that has not yet happened. The redesign account is most persuasive, then, not as a general promise but when it identifies specific constraints that an old arrangement imposed and that a new one could remove.