Years ago, I did some work with the Department of Energy touching on inter alia technical progress and energy intensity.  At the time, our taxonomy spoke of fundamental technical changes, which might (or not) involve substantial reductions in the energy intensity of industrial processes, contrasted with incremental technical changes, which would bring some reductions in energy intensity.  (And we struggled with whether that distinction was best, as it was hard to come up with a priori criteria for

The term of art these days appears to be disruptive innovations, although the concept is spawning another destructive business fad.  (Harvard's Clayton Christensen appears to be the principal sponsor of the expression, although I seem to recall passages in Jean Tirole's Theory of Industrial Organization providing a game-theoretic framework for disruptive innovation.)  The fad, as is common with many business fads, involves attempting to achieve by force that which can sometimes be achieved by serendipity, and more generally involves great expense to achieve worse than nothing.
Disruptive innovation is a theory about why businesses fail. It’s not more than that. It doesn’t explain change. It’s not a law of nature. It’s an artifact of history, an idea, forged in time; it’s the manufacture of a moment of upsetting and edgy uncertainty. Transfixed by change, it’s blind to continuity. It makes a very poor prophet.

The upstarts who work at startups don’t often stay at any one place for very long. (Three out of four startups fail. More than nine out of ten never earn a return.) They work a year here, a few months there—zany hours everywhere. They wear jeans and sneakers and ride scooters and share offices and sprawl on couches like Great Danes. Their coffee machines look like dollhouse-size factories.
It's worth reading the entire New Yorker essay from which I extracted the quote, as it lays out, in some detail, the same problem we faced years ago in attempting to define a technical change as fundamental.  You can only identify the successful creative destruction after the fact, and sometimes the creator blunders into a successful formula.
Christensen has compared the theory of disruptive innovation to a theory of nature: the theory of evolution. But among the many differences between disruption and evolution is that the advocates of disruption have an affinity for circular arguments. If an established company doesn’t disrupt, it will fail, and if it fails it must be because it didn’t disrupt. When a startup fails, that’s a success, since epidemic failure is a hallmark of disruptive innovation. (“Stop being afraid of failure and start embracing it,” the organizers of FailCon, an annual conference, implore, suggesting that, in the era of disruption, innovators face unprecedented challenges. For instance: maybe you made the wrong hires?) When an established company succeeds, that’s only because it hasn’t yet failed. And, when any of these things happen, all of them are only further evidence of disruption.

The handpicked case study, which is Christensen’s method, is a notoriously weak foundation on which to build a theory. But, if the handpicked case study is the approved approach, it would seem that efforts at embracing disruptive innovation are often fatal. Morrison-Knudsen, an engineering and construction firm, got its start in 1905 and helped build more than a hundred and fifty dams all over the world, including the Hoover. Beginning in 1988, a new C.E.O., William Agee, looked to new products and new markets, and, after Bill Clinton’s election, in 1992, bet on mass transit, turning to the construction of both commuter and long-distance train cars through two subsidiaries, MK Transit and MK Rail. These disruptive businesses proved to be a disaster.
In no particular order: stressing the organization simply to stress it, which is a related fad, sometimes breaks it. And Mr Agee's behavior might better be understood as that of a rent-seeker: prior experience in a Democratic administration plus Democrat majorities with transit-dependent constituents equals spending on trains.  (The Allen-Bradley Company makes a cameo appearance.  Yes, the company and its successor Rockwell Automation have made great strides miniaturizing industrial control.  But they had to joust with Square D and Cutler-Hammer, to think of two competitors in Milwaukee, and the advantages of solid state control were more about securing market share than about disrupting other industries.)  And in steel ...  there are reasons I'm still hanging onto my steel industry data base in retirement.
In his discussion of the steel industry, in which he argues that established companies were disrupted by the technology of minimilling (melting down scrap metal to make cheaper, lower-quality sheet metal), Christensen writes that U.S. Steel, founded in 1901, lowered the cost of steel production from “nine labor-hours per ton of steel produced in 1980 to just under three hours per ton in 1991,” which he attributes to the company’s “ferociously attacking the size of its workforce, paring it from more than 93,000 in 1980 to fewer than 23,000 in 1991,” in order to point out that even this accomplishment could not stop the coming disruption. Christensen tends to ignore factors that don’t support his theory. Factors having effects on both production and profitability that Christensen does not mention are that, between 1986 and 1987, twenty-two thousand workers at U.S. Steel did not go to work, as part of a labor action, and that U.S. Steel’s workers are unionized and have been for generations, while minimill manufacturers, with their newer workforces, are generally non-union. Christensen’s logic here seems to be that the industry’s labor arrangements can have played no role in U.S. Steel’s struggles—and are not even worth mentioning—because U.S. Steel’s struggles must be a function of its having failed to build minimills. U.S. Steel’s struggles have been and remain grave, but its failure is by no means a matter of historical record. Today, the largest U.S. producer of steel is—U.S. Steel.
We'll return to the disruption of American Steel at the coda. First, though, the economics, courtesy Joshua Gans in Digitopoly.
Take his prescription that established firms ‘disrupt themselves.’ This is crazy talk to an economist (which is one reason he doesn’t like us). Suppose you take resources and invest in your own disruptor. If disruption occurs, you still lose the entire value of your existing business. All that has happened is that you have kept your name alive. The retort may be that something can be preserved but remember, Christensen is essentially saying firms need to act as if nothing can be preserved. I don’t mind the idea that established firms should not be complacent but hastening their demise on speculation seems weird when there is no upside.

Instead, the focus on the doomed incumbent leads Christensen away from the obvious alternative. The incumbent should ‘wait and see.’ They will see all manner of potentially disruptive technologies being deployed and instead of removing them from their radar as irrelevant, they should continue to monitor them to see what happens. Because, when the one in ten or a hundred or whatever turns out to be successful, they can then move to acquire them and realise a more ‘orderly transition’ to the new technology. Indeed, as I read Lepore, I got the sense that even with Christensen’s iconic examples, the end result was incumbent preservation through acquisition. And this is not just theorising. My own recent paper with Matt Marx and David Hsu demonstrates just that: disruptive technologies (identified after the fact) are associated with start-ups competing and then being acquired as much as they are associated with those start-ups growing as independent firms.
Indeed. We can only identify disruptive innovations after the fact. And in addition to faddishly tearing up the company every six months, there are two other things that can go wrong. You can wait too long to identify the disruption. Or you can identify the wrong disruption.  Here's Lynne Kiesling.
But the genesis of innovation is in uncertainty, not risk; if truly disruptive, innovation may break those historical relationships (pace the Gans observation about having to satisfy the incumbent value propositions). And we won’t know if that’s the case until after the innovators have unleashed the process. Some aspects of what leads to success or failure will indeed be unknowable. My epistemic/knowledge problem take on the innovator’s dilemma is that both risk and uncertainty are at play in the dynamics of innovation, and they are hard to disentangle, both epistemologically and as a matter of strategy. Successful innovation will arise from combining awareness of profit opportunities and taking action along with the disruption (the Schumpeter-Knight-Kirzner synthesis).
And I was in this fight a long time ago.
The most intriguing implication of this research is that the "lethargic" steel producers who postponed the replacement of their open-hearth furnaces or only recently replaced them with large electric furnaces made an optimal choice. Early adoption of a (too small) basic oxygen converter may have been a mistake.
But a coherent model of inventive behavior that combines economic hysteresis (waiting, as Mr Gans notes) with joining the fad (that's called cascade behavior) still eludes me.  There are, however, sharp pencils and stacks of note-paper at hand, and the prospect of thinking time undisturbed by electronic mail, conference calls, office hours, or the obligation to meet any classes.

No comments: