Score One for the Commons (or maybe not)
I was just listening to Dave Levine’s interview of Mike Madison, which touched on his work with Strandburg and Frischmann on constructing commons. Here’s one more possible case study for the group, suggested by Michael Lewis via Felix Salmon:
[T]here are smart [high frequency trading] HFT shops, and then there’s Goldman Sachs. The smart shops execute their strategies using lightweight, open-source, flexible code. Goldman, by contrast, considers its enormous, clunky, proprietary codebase to be a source of competitive advantage — it has to, in order to justify the bonuses it gives to the people in charge of that codebase. Goldman knew that Aleynikov was its best programmer, but it never really grokked why he was good: he was an expert at replacing clunky Goldman code with much simpler and more elegant open-source solutions.
So while Aleynikov thought he was streamlining Goldman’s technology, Aleynikov’s bosses got million-dollar bonuses by claiming that he was adding to a proprietary codebase in which they placed enormous value. And when Aleynikov thought that he was simply emailing his own notes to himself, Goldman decided that he was stealing proprietary information of enormous value — and that, since it was enormously valuable, of course Aleynikov intended to use that code against Goldman in his new job.
Three cheers for open-source! Or maybe 2…or 1. Because the ultimate endeavor here–HFT–is not exactly a boon to the economy. As Wallace Turbeville has demonstrated, “HFT siphons value from the pipeline of capital intermediation, impeding the long-term investments the economy needs for sustained job growth.” I make the case against HFT here; let’s just say that it’s hard to argue that a queueing rule keyed to thousandths or millionths of a second is any better than one that simply allocates trades that happen to come in at the same hundredth, tenth, or (horrors!) half of a second at random, or in (roughly) equal lots.
When it comes to HFT, Goldman’s hapless effort to propertize a codebase was a great example of the “upside of IP’s downside.” The enterprise itself is socially useless at best, pernicious and destabilizing at worst. Here’s to a decline in the commons in HFT code, and to the megabank bonus games that doomed it at Goldman.