Peak car?

From David Metz’s Limits to Travel blog: Peak Car?

I read with interest Phil Goodwin’s Comment piece in Local Transport Today of 25 June in which he introduced the concept of ‘peak car’, and look forward to his promised further exposition. In the meantime, let me observe that as far as London is concerned, peak car use came and went at least fifteen years ago, when none of us noticed. Transport for London’s most recent ‘Travel in London’ report records a steady decline in private transport’s share of trips since at least 1993 (then 50%, 41% in 2008). Correspondingly, public transport’s mode share has risen from 24% to 33%, while walking and cycling have been steady at about 25%.

I read the beginning of Goodwin’s piece, but it is behind a registration wall.
It begins:

Rail, bus and tram use all peaked and then declined, so why do so many people assume that car use will either keep rising indefinitely or reach saturation and a ‘steady state’ condition?

I agree with the general point. But it must reach saturation before it drops. Whether it drops quickly after the peak, or slowly, depends on specific conditions. London suggests the drop may be faster rather than slower.
As gas prices drop and if the economy recovers, I expect we will see somewhat more travel by car in the US than the past couple of years, but the rate of growth from the 1990s and before is a thing of the past (unless travel by car gets much faster).

Review of “Bay Area/California High-Speed Rail Ridership and Revenue Forecasting Study

Review of “Bay Area/California High-Speed Rail Ridership and Revenue Forecasting Study” (pdf) Final Report Prepared at the request of the California Senate Transportation and Housing Committee, Submitted to the California High Speed Rail Authority, Prepared by: David Brownstone, Mark Hansen and Samer Madanat, Institute of Transportation Studies, University of California, June 30, 2010
The Executive Summary of this report is below. In short, the demand forecasts likely mis-stated demand (and thus profitability) of the proposed California HSR for a variety of reasons. The error bands are likely to be much larger than reported.

Executive Summary
“We have reviewed the key components of the California High Speed Rail Ridership
Studies. The primary contractor for these studies, Cambridge Systematics (CS), has followed generally accepted professional standards in carrying out the demand modeling and analysis. Nevertheless we have found some significant problems that render the keydemand forecasting models unreliable for policy analysis. This Executive Summary describes the most serious problems. The body of this report elaborates on these problems and describes additional concerns we have.
In broad terms, the approach taken by CS includes a model development phase and a model validation phase. In the model development phase, both historical data and survey data were employed to develop a mathematical model of interregional travel. The individuals surveyed were interregional trip makers. However, the mode choices of the individuals surveyed were not representative of California interregional travelers. For example, nearly 90% of long distance (over 100-mile) business passenger trips are made by car, while 78% of the long distance business travelers sampled for the study were traveling by air.
The travelers in the sample were asked a series of questions concerning the mode choices they would make for the interregional trip that they were making at the time they were surveyed, under various hypothetical values of travel time, cost, service frequency, and service reliability for each modal alternative (auto, air, conventional rail, and high speed rail). In analyzing the data, the fact that the mode shares actually used by the travelers were not representative of traveler population was not taken into account. Since it is likely that travelers on different modes attach different degrees of importance to different service attributes (e.g. air travelers care more about travel time than auto travelers), it is likely that the resulting model gives a distorted view of the tastes of the average California traveler.
CS attempted to adjust for this problem in the validation phase, by making sure that the model accurately replicated the observed market shares for the existing travel modes in the year 2000. Model predictions of trips by mode were compared with observed values. Parameters obtained in the model development phase were adjusted in order to obtain good agreement between predicted and observed values.
Unfortunately, the methodology employed by CS for adjusting the model parameters has been shown to be incorrect for the type of model they employed. The parameters are therefore invalid and the forecasts based on them, in particular of high speed rail mode shares, are unreliable. (It should be noted that at the time CS performed the study the incorrectness of their adjustment method was not known.)
We found other problems in model development and validation. CS changed key
parameter values after the model development phase because the resulting estimates did not accord with the modelers’ a priori expectations. While this is frequently done in this type of work, it is important that the a-priori expectations be based on experience with like contexts. Unfortunately, some of the a-priori expectations used by CS are valid for intra-regional, but not for inter-regional ridership models. Specifically, the modelers increased the parameter for headway (the time between successive aircraft or train departures) and set it to a value typically found in intra-regional travel demand models. This adjustment made the predicted shares of the travel modes very sensitive to changes in frequency.
Another problem was that CS employed a model structure that does not allow for
travelers to choose between high speed rail stations, thereby exaggerating the importance of having frequent service at the single station that is judged to be “best” for a given trip.
Together with the inflated value of the headway parameter described above, this
unrealistically favors alignments that avoid dividing services onto branch routes, such as Pacheco. Correcting this deficiency would almost certainly reduce, although probably not eliminate, the ridership difference between the Pacheco and Altamont alignments found in the CS study.
In the model validation phase, several other parameters of the mathematical model were adjusted. As a result of this process, many of the model parameters were assigned values that were considerably different from those obtained in the model development phase. In some instances changes to the model parameters were informed by professional judgments of the consulting team as well as the goal of replicating observed behavior.
The resulting “validated” model, which is used to generate subsequent high speed rail
ridership forecasts, provides reasonably accurate “backcasts” for the year 2000, reflects
certain patterns of behavior observed in the traveler surveys, and accords with
professional judgments of the consultant. However, the combination of problems in the
development phase and subsequent changes made to model parameters in the validation phase implies that the forecasts of high speed rail demand-and hence of the profitability
of the proposed high speed rail system-have very large error bounds. These bounds,
which were not quantified by CS, may be large enough to include the possibility that the
California HSR may achieve healthy profits and the possibility that it may incur
significant revenue shortfalls. We believe that further work to both assess and reduce these bounds should be a high priority.”