Yglesias is fond of suggesting Silicon Valley’s compensation model as a good alternative to the one that led the finance industry into distaster. The idea is that by using stock options to connect compensation to the long-term performance of a company rather than to its quarterly or yearly performance, we can make short-term risk-taking less lucrative than intelligent stewardship of a firm.
This sounds like a pretty good idea, and very well may be. But it’s worth pointing out a couple of things: first, this may just be an apples and oranges sort of situation. It may not be the case that the sort of compensation structures available to a penniless startup are practically applicable to an enormous financial behemoth. I’m admittedly no expert, but I can imagine there may be problems.
Second — and this I can say with somewhat more confidence — the performance of the software industry is not such an inspirational success story that emulating it should be assumed to be a good idea. Spend three months reading TechCrunch or its equivalent; if you can make it through that time without killing yourself, you’ll realize that Silicon Valley is incredibly inefficient, wasting vast sums of money on overhyped ideas that are stupid, unnecessary or just commercially impractical. There are very, very few firms that have created genuinely original technologies — technologies that Harry Turtledove would like to write about, technologies that may conceivably never have been invented if their creators’ parents hadn’t met. The bulk of the industry is made up of the proverbial million monkeys sitting at a million MacBooks; occasionally some Javascript comes out.
Some friends of mine have created a fake web startup. On their homepage (or twitter feed) you can find a simple mad-lib that changes every time you reload the page, and which goes like this: “[VAGUE, POSITIVE GERUND] [TECHNOLOGY A] with [TECHNOLOGY B].” This really is how the industry operates: through the mind-numbing combinatorial exhaustion of whatever technologies Google, Amazon, Adobe, Sun and a very few brilliant open-source developers can come up with. Some of this is valuable, even necessary economic activity. But a lot of it is just speculative activity which benefits no one other than the people directly involved — and which is ultimately a waste of resources. This should sound wearyingly familiar by now.
Now, it’s not all their fault. The real problem is that software development is really easy — you need almost no capital, and there’s an incredible wealth of existing technology that can be utilized. The bottlenecks to innovative commercial activity in the software realm are frequently external. These limits can be technological but not software-related (phone cameras needed to get good enough to read barcodes), cultural (Facebook couldn’t exist until college students were wired enough to adopt it) or political (only a firm of Google’s size and import could start scanning books and comfortably expect to find a way out of the orphan works problem without being sued into oblivion). And of course it’s just generally tough to start a successful business (I think I can confidently say that the average internet startup is based on a somewhat stupider idea than the average non-internet startup, but I have no idea which is actually more likely to fail). I’m sympathetic to these guys: certainly, I can understand the impulse to paint yourself a visionary who creates fundamentally new possibilities, rather than as mere skilled craftsman using tools handed down from others.
But still, it’s hard to look at the amount of investor money wasted on the web industry and conclude that its compensation practices are ones that should be emulated — particularly given that those practices are being abandoned now that the accounting gimmick that enabled them has been ended. Maybe it’s preferable to have a lot of middle-class programmers blowing through investor money instead of a relatively few upper class finance executives doing the same — I suppose it is a more progressive transfer — but that’s all that Silicon Valley’s recent history seems to promise. The way to neutralize the villains of this bubble may not be to make them more like the villains of the last bubble.
This post absolutely fails to refute the value of the ‘Silicon Valley’ compensation scheme compared to the ‘Wall Street’ version.
You spent 2/3 of the post criticizing the product development pattern of the Tech industry, which is inherently speculative (and revisionist); the compensation scheme has virtually nothing to do with this development pattern.
An appropriate comparison/refutation would address the incentives produced under each scheme for a set of scenarios. Keep your eye on the ball.
I agree that you can make a case for the incentive structure of the Valley working for finance — but that’s what others have done in posts like the one I cited. My interest was in pointing out the way in which the software industry operates on a practical basis, to illustrate that its compensation structure doesn’t necessarily lead to an efficient allocation of resources.
Put another way: the elimination of an identified moral hazard may not be sufficient to produce a system that genuinely works well.