Wednesday, 30 March 2011

Path Dependence

Can an economic fallacy validate itself?

Through the 80s and 90s, Paul David made a minor career of shouting that path dependence made for pervasive market failure. His favourite example: the QWERTY keyboard. Tyler Cowen and I looked at path dependence in our 2002 edited volume on market failure. I'm going to self-plagiarise fairly liberally here from that book's introduction. Here's the summary of David's lock-in argument:
The basic argument behind lock-in is simple: collective action problems may cause individuals to end up stuck in a technology that is less than optimal. When compatibility plays a role in determining value, switching from an established standard to a superior standard may become very difficult. The market may fail to abandon the inefficient standard because it has no mechanism to coordinate the move to a new standard. When compatibility effects are strong, the private gains of an individual switch would be lower than the social gains of the switch. Each individual fails to take into account that a successful collective switch would yield significant benefits for others as well. This makes individuals more reluctant to switch than they should be, and the market may become “locked in” to an inefficient standard. Absent a centrally coordinated move, individuals and firms would stay with the old standard.
Liebowitz and Margolis examined the various cases David put forward for inefficient lock-in and found them wanting. DVORAK was only found superior to QWERTY because the tests were conducted by Lt. Cmdr. Dvorak himself; later tests found no difference. VHS beat Betamax not because of lock-in but because customers valued long recording time over image quality. Most businesses chose DOS over Apple because it taxed the operating system less and afforded more system resources for business applications. WordPerfect dominated word processing software only until Microsoft Word started providing a superior product, despite strong potential lock-in from users having memorized arcane function key arrangements - anybody remember the cardboard cutouts folks would place over their keyboards to help remember what the WP function keys did?

Perhaps the best case for inefficient lock-in and market failure is that people keep pointing to QWERTY as an inefficient market failure despite its having been refuted ages ago. Here's David Brooks:
John McWhorter, a linguist at Columbia University, wrote that people should be more aware of path dependence. This refers to the notion that often “something that seems normal or inevitable today began with a choice that made sense at a particular time in the past, but survived despite the eclipse of the justification for that choice.”

For instance, typewriters used to jam if people typed too fast, so the manufacturers designed a keyboard that would slow typists. We no longer have typewriters, but we are stuck with the letter arrangements of the qwerty keyboard.
Path dependence in stories about path dependence. It seems the QWERTY story can't be killed.

A reasonable case can be made that nuclear power generation technology exhibited lock-in where US Navy designs for submarine reactors were ported over into domestic power generation because they were proven to have worked; other designs may have been better, but the submarine engines had a head start.

But again, we have to ask the questions Liebowitz and Margolis highlighted:
  • Was the choice suboptimal ex ante? Likely not. The US government was in an awful hurry to get civilian reactors going, not least of all to help produce plutonium. Delaying things a decade while deciding which design was best would not have met the political goals sought.
  • Was the choice suboptimal when viewed ex post? Possibly - we now have rather better reactor designs. But nothing requires that new reactors follow old designs. Legacy reactors get mothballed when they reach end of life; if new ones are built, they can use the updated designs.
And it would be a bit rich to point to nuclear technology lock-in as having anything to do with market choices and market failure.

As Tyler and I argued in the book introduction, there are two possible policies for mitigating inefficient lock-in.

First, government can impose delays on standard adoption in hopes that delay ensures that the best choice is made. But we have to weigh both the losses that accrue in the interval in which no choice is made and the possibility that it's only through market experimentation that we can discover which standard really is best. The techies would have chosen BetaMax over VHS; it was user choice that showed that what mattered most was recording duration, not picture quality. Delay in that case probably would have enforced the inferior product choice.

A second policy option would be helping in coordinating ex post moves away from inefficient standards. But even this is perilous: just look at how many people would still want to force us all to flip from QWERTY to DVORAK. And do we really want to open up that avenue for lobbying?


  1. Eric, I'm wondering what you think about the problem Japan is facing with blackouts. The east and west islands are on different systems (one German 50hz, the other American 60 hz). There's a bottleneck is transmitting across the systems. NPR suggests that the systems would have been reconciled years ago except for economic interests protecting monopolies.

    Wikipedia gives several examples of systems that have been reconciled over the years.

  2. @John: I really don't know much about that case beyond what's in the links. I'd want to know a lot more about regulation in the Japanese power industry before saying much. On the face of it, it seems utterly absurd to have the two standards. Suppose on the other hand that a whole lot of industrial machinery is geared to either running on 50hz or 60hz. Switching costs then would be substantial.

  3. I'm currently learning to use Latex. As far as I can tell, there is very little that Latex can do that the current version of Microsoft Word (with a few add-ins like Mathtype) can't do. My higher-ups tell me that there's a wee bit of snobbery going on with the ubiquity of Latex use, and it's likely that journal-article submissions in Word could get rejected for not having been delivered in Latex.

    I've also been told that Word has previously had nothing on Latex, and it's only recently that Microsoft upped its game. Maybe good-enough versions of Word (and other programs) haven't been around long enough for this to count, but does widespread, unwavering Latex use seem like a bit of path-depdence to you?

  4. @James: LaTex makes things look like science. I'd say more signalling than path dependence.

  5. What about driving on a different side of the road to neighbouring countries? It seems that many countries have overcome this lock-in with government coordinated effort (your policy option 2).

    I also feel that network services are like that - eg, electricity standards, and competition between rail and roads (where a history of road subsidies has made the road network more competitive for haulage than rail, even though there may be long term benefits from a coordinated shift in investment to rail).

    More here

  6. @Cameron: I'm not sure what other entity could coordinate a change from one side of the road to the other; anybody else trying it would be arrested.

    I don't know about the Aussie case, but at least in NZ, the legacy of rail subsidy / subsidy through punitive regulation on trucking has left us with a too-extensive rail network. A few lines are viable; lots of others should be shut down. Road user charges and petrol taxes largely cover roading costs, barring a few boondoggles like Auckland proposed tunnels.