Kieran’s piece on kids being driven to school reminded me of a post I’ve been planning for a while. One of the issues debated at length on my blog is that of speeding and law-enforcement measures such as speed cameras. I’ve argued against speeding and in favor of rigorous law-enforcement. Not surprisingly, and perhaps reflecting the fact that more than 80 per cent of drivers regard themselves as above-average, this has been very controversial. You can read some instalments in the debate here and here or use the search facility for “speeding”. Unfortunately most of the extensive and interesting comments were lost in a database failure.
In the course of this debate I discovered the fact, surprising to me, that, although the rate of road deaths per person in the United States is nearly twice that in Australia and the United Kingdom, much of this difference can be accounted for by the fact that distances travelled in the United States are a lot higher and are rising (there are problems with the numbers and biases in the measure, but I’ll leave that to one side for now). The differences between US and UK are plausible given differences in population density and well-developed public transport in London at least, but the differences between the US and Australia certainly surprised me. Australia is every bit as car-dependent as the US and has much lower population density.
All of this is a prelude to the fact that, in economic terms, time spent travelling is a really big deal. In their book Time for Life, based on the 1985 US Time Use Study, Robinson and Godbey estimate that the average adult American spends 30 hours a week in paid employment and 10 hours a week travelling (they also, controversially, argue that working time has been falling, not rising). It’s pretty clear that distances and times spent travelling have increased since 1985 in the US (in both the US and Australia, driving is by far the dominant mode of travel).
If, as I’ll argue below, most travel should be regarded as being in the same economic category as working and if, as the stats linked above imply, Americans spend about twice as much time travelling as Australians, then reducing travel times to the Australian level would be equivalent to a productivity improvement of between 12 and 15 per cent. As it happens, combined with the relatively small difference in hours of paid work, adjusting for hours of work and travel would just about eliminate the gap between Australian and US GDP per capita (about 20 per cent on standard PPP estimates).
his is also important because quite a few commentators have argued that one of the factors promoting productivity growth in the US has been the rise of “big-box” edge of town stores like Walmart and Costco in place of small inefficient local shops (here, for example is Robert Gordon, cited by Brad deLong). As Steve Sailer points out, this apparent productivity growth has been achieved by transferring costs to shoppers.
Bigger stores mean fewer stores of each type, and that means longer drives, with larger parking lots and longer aisles to trudge through.(I’m less impressed by Sailer’s argument that more diversity pushes the responsibility for choice onto consumers, but that’s by the bye).
Now let me justify my claim that travel time should be added to work time in deriving economic measures of productivity. On the output side, at least two-thirds is associated with the basic business of getting and spending (commuting, childcare and shopping). The remainder is associated with free-time activities but, except in the case where the travel itself is part of the activity (‘a drive in the country’ and so forth) should, I would argue, be classified as work.
On the input side, driving is more stressful and unpleasant than most paid employment activities. One way of thinking about this is to look at full-time jobs that mostly involve driving (cabbie, courier etc) – these are generally considered high-stress unpleasant jobs.
Driving is more dangerous than work in general. Australian data suggests that there are about one and a half times as many work related deaths than road fatalities, but when account is taken of the number of hours spent at work and on the road, this means driving is several times more dangerous.
I should add that, in all of the above, I’ve treated distance travelled by car as a proxy for time spent travelling, on the assumption that average speeds are about the same. This might seem inconsistent with my emphasis on the enforcement of speed limits, but the reduction in average speeds associated with enforced limits is quite low – most of the time the binding constraint is congestion. My casual observation suggests that urban traffic doesn’t move any faster in the US than in Australia.
As far as I can tell, this issue has been almost completely neglected by economists, except for the special purpose of evaluating savings in travel times associated with improved roads. The standard practice appears to be to value travel time at about half the average wage. This is consistent with the economic analysis I’ve proposed on standard assumptions about the disutility of work (if the disutility of work starts at zero and rises linearly until the marginal disutlity equals the wage, then average disutility will be equal to half the wage).
The most likely reason for this neglect is the assumption that travel represents a more-or-less constant overhead cost associated with getting to work, shopping and so on. If so, it can be ignored without changing anything of interest. But the evidence seems to be that this is not true. Travel times (or at least distances) differ greatly between apparently similar countries and vary significantly over time. This means that analyses of productivity and living standards that disregard travel costs are likely to get wrong answers.