On April 20, 2010, the US Department of Transport released a report titled ‘Transportation’s Role in Reducing U.S. Greenhouse Gas Emissions”. The report ranked Pay as you drive car insurance as being relatively effective in reducing emissions and extremely cost effective in reducing these emissions. However, Todd Litman’s response to the article explains why the impacts and benefits of PAYD car insurance are probably much higher.
The USDOT report assumes that US federal policy would require that insurance companies at least offer PAYD policies. The report goes on to assume that this requirement would result in at most, 75% of policies having PAYD pricing. Litman takes this requirement further, appealing for a 100% shift to PAYD policies. He justifies this appeal on safety grounds, that PAYD insurance is predicted to reduce about 5,000 annual traffic related fatalities.
Litman explains that another benefit of PAYD assumed by the report is that “PAYD could reduce VMT and emissions by up to 3%, but other published studies (Litman’s and those by Bordoff and Pascal, and Edlin, to name a few) indicate much higher potential impacts (typically 8-10% VMT reductions) and benefits.”
While the USDOT report assumes many benefits of PAYD car insurance, it also raises vehicle mileage instrumentation costs as a challenge to PAYD implementation, assuming a $131 billion cost. Litman offers a much more cost-effective alternative, giving the example of Milemeter:
“Milemeter already uses (vehicle odometers) for calculating insurance premiums (they require policy holders to email a digital photo of their odometer reading). Modern odometers are extremely reliable - they must be because they are the basis of vehicle warrantee, lease and resale transactions, which often involve far larger financial burdens on vehicle manufacturers and dealers than a typical annual insurance transaction. I therefore recommend that the analysis include a cheaper implementation option, based on odometer audits or digital photos.”
Milemeter’s odometer set up is of course very similar to that of Real Insurance’ Pay As You Drive car insurance product. Real Insurance use a trust-based system: Customers tell us their initial odometer reading and pre-pay for a certain amount of kilometres. If they need to make a claim, we verify that they are within the odometer range they have pre-paid for. If they are outside their stated range then their claim will not be covered.
While the opinions and findings of Litman and the USDOT report differ, both agree that there are numerous benefits in supporting PAYD car insurance in the USA. Hopefully these findings will also encourage more investigation into the benefits of PAYD insurance in Australia.
Litman’s response to the report is available in full below:
PAYD Benefits Probably Larger Than Indicated In The USDOT GHG Reduction Report
Yesterday I sent you a message about the USDOT's new report, "Transportation's Role in Reducing U.S. Greenhouse Gas Emissions," USDOT Report to Congress ( http://ntl.bts.gov/lib/32000/32700/32779/DOT_Climate_Change_Report_-_April_2010_-_Volume_1_and_2.pdf ). Although its analysis ranks PAYD insurance as relatively effective (total emissions reduced) and extremely cost effective (net benefits per ton of reduced emission), actual potential impacts and benefits are probably much higher.
PAYD analysis is on pages 499-501. The analysis assumes that, at most, federal policy would require insurance companies to offer PAYD policies. With this approach PAYD would be chosen for lower-annual-mileage vehicles, and so would only apply to a minority of vehicle travel. Apparently, the analysis assumed that at most, 75% of policies representing 43% of VMT would have PAYD pricing. Of course, it is possible to require a total shift to PAYD. This is justified on traffic safety grounds - it is predicted to reduce about 5,000 annual fatalities - just as federal government policies have required states to implement seatbelt laws, DWI enforcement standards and motorcycle helmet laws. I therefore recommend that the analysis should included 100% PAYD as the upper-bound range.
The report assumes that PAYD would require the same vehicle mileage instrumentation as for VMT fees. The analysis assumes a $131 billion cost for mechanical "hubodometers," and so estimate a cost of $30 to $90 per ton GHG reduced. However, vehicles already have an instrument that tracks mileage, called in "odometer." Milemeter ( www.milemeter.com) already uses this for calculating insurance premiums (they require policy holders to email a digital photo of their odometer reading). Modern odometers are extremely reliable - they must be because they are the basis of vehicle warrantee, lease and resale transactions, which often involve far larger financial burdens on vehicle manufacturers and dealers than a typical annual insurance transaction. I therefore recommend that the analysis include a cheaper implementation option, based on odometer audits or digital photos.
The report assumes that PAYD would result in an average additional marginal cost of 4 to 6 cents per mile. This is a little low for fully prorated insurance ($1,100 average insurance divided by 12,000 annual is about 9 cents per mile).
The study uses elasticity values based on studies which indicated that U.S. fuel price elasticities had declined significantly over the last half-century. In particular, they used results of studies by Hughes, Knittel and Sperling (2006), which found long-run fuel price elasticities from -0.21 during 1975-80, but only -0.057 during 2001-06. Small and Van Dender (2005 and 2007) found the gasoline price elasticities were -0.09 in the short run and -0.40% in the long run during the 1997 to 2001 period, about half the values observed from 1966 to 1996. However, those results likely reflect unique factors during those years, including declining real fuel prices, rising female employment, peak Baby Boom driving years, rising real incomes, and sprawl-encouraging highway building and development policies.
More recent research indicates that price elasticities have increased back to normal levels, due in part to demographic and economic trends, including aging populations, stagnant real incomes, increased urbanization, investments in alternative modes, and changing consumer preferences. Komanoff (2008) found that the short-run U.S. fuel price elasticity reached a low of -0.04 in 2004, but increased to -0.08 in 2005, -0.12 in 2006 and -0.16 in 2007, and long-run effects are typically about three times short-run, suggesting long-run effects as high as -0.48. Brand (2009) found that the 20% U.S. fuel price increase between 2007 and 2008 caused a 4.0% reduction in fuel consumption, indicating a short-run price elasticity of -0.13. However, accounting for the basic growth trends in population and GDP, Brand estimates that this 10-month fuel consumption price elasticity increases to about -0.17.
For more information on these factors see my report "Transportation Elasticities" (http://www.vtpi.org/elasticities.pdf ), in particular, the sections on "Vehicle Operating (Out-of-Pocket) Expenses" and "Vehicle Travel With Respect to Fuel Price". Most these studies show long-run elasticities of -0.2 to 0.4. For example, Oum, Waters, and Yong (1992) estimated the elasticity of vehicle travel with respect to out-of-pocket expenses to be -0.23 in the short run and -0.28 in the long run. Button (1993) find elasticities for urban shopping of -2.7 to -3.2, and for urban commuting of -0.3 to - 2.9. Schimek (1997) finds the elasticity of vehicle travel with respect to fuel price in the U.S. to be -0.26. These results are consistent with international research (Johansson and Schipper 1997). INFRAS (2000) cites estimates of the long-term elasticity of vehicle use with respect to fuel price to typically average about –0.3.
For modeling purposes the study used the relatively low elasticity of VMT with respect to fuel price (-0.21), which they increase to reflect total average vehicle expenses based on IRS vehicle cost estimates. As a result, the analysis applies an elasticity value of -0.45 to total average vehicle costs of $0.68 per vehicle mile. I don't think these adjustments are either necessary or accurate. It would be far better to apply reasonable long-run elasticity values of -0.2 to -0.4 to the insurance price value (5 to 9 cents per vehicle-mile) relative to average fuel prices (about 15 cents per vehicle-mile, based on $3.00 per gallon fuel and 20 mpg fuel efficiency), which implies that PAYD would reduce vehicle travel 6% (-0.2 x 5/15) to 24% (-0.4 x 9/15) per affected vehicle with PAYD policies with Gold or Platinum ratings ( www.ceres.org/Page.aspx?pid=1157 ).
This study concludes that PAYD could reduce VMT and emissions by up to 3%, but other published studies (my own and those by Bordoff and Pascal, and Edlin, to name a few) indicate much higher potential impacts (typically 8-10% VMT reductions) and benefits. Of course, achieving these maximum benefits is not necessarily easy, but it is perfectly possible and cost effective, and so I think it is important that decision-makers understand this when evaluating options.
References
Dan Brand (2009), Impacts of Higher Fuel Costs, Federal Highway Administration, (www.fhwa.dot.gov); at www.fhwa.dot.gov/policy/otps/innovation/issue1/impacts.htm.
Jonathan E. Hughes, Christopher R. Knittel and Daniel Sperling (2006), Evidence of a Shift in the Short-Run Price Elasticity of Gasoline Demand, National Bureau of Economic Research, Working Paper No. 12530 ( http://papers.nber.org/papers/W12530).
Charles Komanoff (2008), Gasoline Price-Elasticity Spreadsheet, Komanoff Energy Consulting ( www.komanoff.net/oil_9_11/Gasoline_Price_Elasticity.xls). This spreadsheet uses U.S. gasoline price and consumption data from 2004 through the most recent available monthly data to estimate the short-term price-elasticity of demand.
Todd Litman (2008), Transportation Elasticities: How Prices and Other Factors Affect Travel Behavior, Victoria Transport Policy Institute (www.vtpi.org); at www.vtpi.org/elasticities.pdf.
Kenneth A. Small and Kurt Van Dender (2007), “Fuel Efficiency and Motor Vehicle Travel: The Declining Rebound Effect,” Energy Journal, Vol. 28, No. 1, pp. 25-51; at www.econ.uci.edu/docs/2005-06/Small-03.pdf.