Matthew Yglesias is somehow offended by the Antiplanner’s recent post on Cato’s blog about the huge decline in the productivity of our socialized transit industry since 1970. He never addresses or even acknowledges any of the arguments made in my article. Instead, his problem is that the article “fails to acknowledge any government role in promoting the usage of private automobiles.” Since my article was about transit, not automobiles, I don’t see why I need to acknowledge government’s role in driving any more than I should acknowledge government’s role in our failed education system or any other government failing.
It could be that Yglesias is arguing that I am somehow inconsistent because object to socialized transit without objecting to socialized highways. If so, he would be wrong: In books, papers, and policy statements I have argued that highways should be funded out of user fees, not taxes; that states should encourage private highway construction; and that the federal government should get out of the highway business. That isn’t in any way inconsistent with my article on transit.
You need to consult your medical doctor initial if you are to acquire this drug specifically if you are having several clinically cialis 40mg related sicknesses because this substance may be contraindicated for your problem. Two features cialis online online that set Columbia River Knife & Tool has is the assist Opening mechanism. So levitra generika this was taken as a rock solid vote for hyphens. The nation’s five largest populated cities constitute just 6% of the heritageihc.com buy cheap levitra electorate. It could be that Yglesias is going further and arguing that “government’s role in promoting the usage of private automobiles” somehow contributed to the huge decline in transit productivity since 1970. If so, he would be wrong. Since 1970, government’s role in transportation has mainly been to steal money from the users of roads that carry 85 percent of American passenger travel and spend that money supporting transit systems that move barely 1 percent of all passenger travel. In 2008 alone, more than $15 billion in federal, state, and local highway user fees were spent on transit, which did highway users little good (see cell O17)
Yglesias says that government land-use policies have caused “regulatory impediments” to creating transit-friendly cities. I’ve previously invited Yglesias to join me in calling for the elimination of all government land-use policies, but he has remained silent about that. Instead, he would rather rant about things I didn’t say.
Since Yglesias is so fond of attacking me for things I haven’t said, let me return the favor by pointing out that one of the most common arguments that people like Yglesias use against highways (and he himself may not have made this argument) is what they call “induced demand.” That is, they say we shouldn’t build more roads (even if paid for entirely with user fees) because they will just lead people to drive more. This is supposed to be a bad thing because driving is supposed to be bad. But really, they are admitting that Americans want to use and are willing to pay for more roads, whereas in their demands for more subsidies to transit they are admitting that Americans are not willing to pay for their transit-oriented utopias.
The real problem with the Cato report is that the data on which the conclusions are based is not consistent.
If you look at the actual tables, there is a substantial discontinuity in 1983-1984, as the reported number of employees rose by well over 30% and the reported costs rose by 45% in that one year alone, compared with changes of 0.5% and 5%, respectively, in the year just prior. Of course, that’s because the reported data set changed, so that all sorts of new modes were included that hadn’t been included before.
If you look instead at just “Operating Employees by Mode” (Table 11), and choose a more recent and reliable data set, you’ll see that operational personnel for light and heavy rail combined increased by just 19% in the 24 years from 1984 through 2008, while operational personnel for buses increased by just 25% over the same period. This is a far, far cry from the 180% and more increase in personnel as reported over 40 years that you see in the Cato article.
Going back to Table 38 to assess operating costs, I note that the data is only reliable by mode from 1988 on. In that period, the cost of light and heavy rail operations combined increased by 98%. Bus operational expenses, on the other hand, increased by 129%. Inflation increased consumer prices by 82% in this same period. Meanwhile, Table 2 shows us that, during this same span, passenger miles increased by 60% for light and heavy rail combined, but only by 4.8% for buses.
In other words, between 1988 and 2008, light and heavy rail operational expenses increased by just 16% more than the CPI – a far cry from the 195% increase reported by Cato. Moreover, this 16% increase bought 60% more passenger miles. Obviously this is a substantial INCREASE in productivity, not the decrease in productivity reported in the Cato article.
Meanwhile, bus operational expenses increased by 47% more than the CPI – still well below the multi-modal 195% increase Cato tried to convince us of. However, THAT increase bought just 4.8% more passenger miles. Clearly, this is a substantial decrease in productivity. Yet we repeatedly find this blog advocating in favor of buses rather than rail.
I’m very pleased that Mr. O’Toole likes to cite his data sources. With the actual data in hand, one can clearly show that he has no idea what he is doing.
Anyone reading Yglesias’ post will understand the argument immediately. Presumably Randal isn’t so dumb to not understand it, so Randal is using a rhetorical device.
And speaking of arguing things people don’t say, this absurd statement But really, they are admitting that Americans want to use and are willing to pay for more roads, approaches buffoonery. And thanks for the laugh.
DS
Antiplanner,
I would like refer you to the Wikipedia page on the deleterious effects of socialized transit in Havana, Cuba. Here’s the money quote:
“Havana was renowned for an excellent network of public transportation by bus and taxi. A subway system modeled after that of New York City was even proposed in 1921.[56][57] In 1959, Havana’s buses carried out over 29,000 daily bus trips across a dense layout of routes that connected the 600,000 inhabitants of Havana. After the Socialist Revolution, all business were nationalised, and public transport was assigned to the Ministerio del Transporte (MITRANS). In the Province of the City of Havana, Provincial Transport Authority functions are carried out by 11 divisions. But this bureaucratic, complex system of central control produces today only 8,000 trips per day, for a population that triples that of 1959”
Amazing, triple the population but service is cut to one third! I guess that’s ok because transit employees live well in Cuba and Cubans don’t suffer as many of those evil cars on the road either.
It doesn’t take more than a causal observer to see you rail against transit waayyy more than you do against highways. I’m not sure why you’re stunned by this. It’s kind of hard to be against 15 billion for transit when cars got more than 60 billion (probably more if we knew how those bonds were paid off) — including 8 billion from the feds.
http://www.fhwa.dot.gov/policyinformation/statistics/2008/hf10.cfm
I’m not sure why Yglesias would want to eliminate government land-use conditions with you or other zoning that benefits urban environments. I assume he is liberal, afterall. He’s just saying a Libertarian cannot point to this mythical magic ferry “free-market” beast that Libertarians like to point to when in fact that very market and supposed “choice” is about as predictable as the sun coming up every morning.
Expressways right now make up less than 2% of roads in the US, even if they were all private they would still depend on 98% of property taxpayer funded roads for traffic.
Mr.O’Toole might live in a house with out windows, but he still needs a door.
I’ve given this matter some thought, and I think there are some definitional issues that need to be addressed.
Randal’s definition of productivity, as employed in the Cato post is basically a measure of consumed output (ridership) divided by an input (number of laborers). One problem with this is that the output can be influenced by exogenous factors (e.g. income growth, changing urban structure), as well as factors that are under the control of public transit agencies themselves (fares, service levels). Also the input measure looks only at labor and ignores others (energy, capital). Perhaps that was by choice and the point was to demonstrate the decline in labor productivity as others have. But I sense there was more to that post.
Then Adam argues that there are inconsistencies in the (APTA) data and that different measures should be used. He not only throws out the pre-1984 data, but also uses passenger-miles as a measure of service consumption, while Randal used boardings, which are consistently available going back beyond 1970. This is an important omission, since it ignores the disastrous experiment with federal operating subsidies during the 1970s and early 1980s which caused a run-up in both expenses and employment levels.
The discontinuity issue is not as severe as Adam makes it out to be, and certainly does not warrant throwing out all employment and cost data prior to the change in reporting conventions used in the APTA data. While limited data availability on some topics might limit modal comparisons, there is still enough data to review Randal’s claims in his previous post.
There is no reason to look only at operating employees when evaluating transit systems. When Randal referred to the number of employees needed to produce transit service, I take that to mean all workers that are employed by a transit agency. Perhaps there was some confusion here, but I think a consistent measure needs to applied. It is important to count total workers, since it is not possible to have a non-trivial increase in service without also increasing staffing levels in non-operations divisions (maintenance, administrations, etc.).
Let’s look at total employment then. We can account for the change in reporting conventions by assuming that the number of employees increased from 1983 to 1984 at the same rate that it did from 1970-1983 (about 2.7% per year), then subtracting the remaining difference (about 55,000 workers) from the totals in later periods. This would mean that, all else equal, there were about 332,000 employees in 2008. That represents a 140% increase over 1970 levels. Not quite a 180% increase, but much closer to 180 than to 25, and based on a more careful method of measurement.
@MJ: I agree that a consistent measure needs to be applied, which is why I used only the more recent and more reliable data. If you are going to change the methodology to include numbers all the way back to 1970, then you have to somehow account for why your analysis comes up with a 140% increase over 40 years for all personnel, when I have shown that the last 20 of those years show an increase of 25% or less for operating personnel. Where are all these extra people working, since it is clear they are not working in operations? Or is your point only that there was a devastating drop in productivity from 1970 through 1984, but that this trend has since reversed itself?
As my analysis should make clear, I don’t think your calculated 140% increase is anywhere close to a realistic figure. The data that far back is simply not consistent with the data being collected today. You are not comparing sufficiently similar measures to be able to be sure of the quantitative trend, which is what you need in order to justify a claim of a drop in productivity. Without consistent total employee data, you can’t possibly get meaningful results based on that data, which was the entire point of my post.
Also, I have no idea why you think boardings are a better measure of productivity than passenger-miles. If two rail lines have the same operating expenses and each has 1,000 boardings per day, but one line carries an average of 5,000 passenger-miles, and the other carries an average of 20,000 passenger-miles, can we not conclusively conclude that the second system is substantially more productive? It is completely obvious to me that passenger-miles are by far the more accurate measure. Since the data are available for a longer period, you may still be able to draw some useful conclusions from boarding data alone, but a quantitative measure of productivity would not be one of them.
Using flawed data leads to flawed conclusions. Your attempt to “repair” the data is simply insufficient.
It’s expensive to have drivers & so much capital for public transit. Only paid about 1/3 by user. Such a deal, if the routes & times fit.
“Expressways right now make up less than 2% of roads in the US,” -Highwayman
Source? Lane miles? Miles driven?
“In other words, between 1988 and 2008, light and heavy rail operational expenses increased by just 16% more than the CPI – a far cry from the 195% increase reported by Cato. Moreover, this 16% increase bought 60% more passenger miles. Obviously this is a substantial INCREASE in productivity, not the decrease in productivity reported in the Cato article.” – Adam
Did those 60% more passenger include new service that was built during that time? Or are you saying all lines that existed in 1988 are carrying 60% more passengers today?
prk166: “Did those 60% more passenger include new service that was built during that time? Or are you saying all lines that existed in 1988 are carrying 60% more passengers today?”
Of course it includes new service. Operational productivity still increased, though, whether the service was new or old, since the passenger miles per operational expense increased. I did not amortize capital expenses for new construction, since adequate data just wasn’t there, but Cato didn’t attempt that, either.
As my analysis should make clear, I don’t think your calculated 140% increase is anywhere close to a realistic figure. The data that far back is simply not consistent with the data being collected today. You are not comparing sufficiently similar measures to be able to be sure of the quantitative trend, which is what you need in order to justify a claim of a drop in productivity. Without consistent total employee data, you can’t possibly get meaningful results based on that data, which was the entire point of my post.
The same set of data were used from 1970 to 1984. Thus, there should be no reason to dismiss it out of hand. It is consistent over that time period. By focusing only on 1984 to the present, you are missing the huge increase in funding levels that occurred during the 1970s, which allowed for much higher staffing levels. This included large numbers of administrative and maintenance personnel, all of which were on the payroll of public transit authorities. I made some reasonable assumptions to account for the discontinuity in your data, which should allow for some meaningful conclusions to be drawn on the experience going back to 1970, which was the original focus of this post.
If two rail lines have the same operating expenses and each has 1,000 boardings per day, but one line carries an average of 5,000 passenger-miles, and the other carries an average of 20,000 passenger-miles, can we not conclusively conclude that the second system is substantially more productive?
No. The only thing those data would indicate is that one is serving a different market (i.e. longer trips). It says nothing about productivity. Your example is also unrealistic, since if the number of boardings is the same and one service is providing more passenger-miles, operating expenses would almost certainly not be equal. Trip length is highly correlated with operating expense.
Since the 1970s, much the of increase in transit service has come in the form of trying to expand service to new markets, often with things like longer-distance express bus service and commuter rail. Since these markets tend to have longer trips, average trip lengths should increase, and passenger-miles should increase faster than boardings. This would be even more evident if transit providers actually measured linked trips instead of unlinked boardings. Of course, the tradeoff has been that transit providers have been spent considerably more on (federally-funded) capital improvements to allow these extensions, with much of the money going to new rail systems.