Buying Gasoline for Hybrid Cars: 
Evidence from Daily Gasoline Prices 

Jannett Highfill
Michael McAsey
Bradley University

     The decision to buy and drive a hybrid (gasoline-electric) car has many dimensions; this paper will focus on one of them—gasoline purchases.  Hybrid owners face many of the same problems buying gasoline as any other driver.  They must refuel before they run out of gas; the tank has a maximum capacity which cannot be exceeded at any given refueling.  Time is valuable for the driver of any car; the present paper assumes one or two refuelings during a given planning period.  And gasoline prices are often quite changeable.  The present paper treats this problem by assuming the consumer knows (has perfect foresight) about gasoline prices in the present planning period, but prices in the next planning period are completely unknown.  The primary difference between hybrid and regular vehicles for the purposes of this paper is that hybrids have a sufficiently higher mileage per gallon than regular cars so that even though they have a smaller tank size, they can go much further on a given tank of gasoline.  Thus differences in miles per gallon cause the main differences between hybrid and regular vehicles in the analysis of this paper.  In particular, the present paper defines the planning period as the length of time a vehicle can be driven with only one or two refuelings.  It will be seen shortly that the planning period for hybrids is six or eight days longer than for other vehicles—thus hybrid owners have to refuel less often, other things being equal.
     The strategy of the present paper is to use data on vehicles from cars.com, miles driven per year from the Federal Highway Administration, and daily gasoline prices collected by the authors from November 2003 to November 2005 to address two questions.  It should be noted that in 2004, the main year of our gasoline price data, only two hybrids had significant sales (Lipman and Delucchi [2006, p. 116]), the Honda Civic Hybrid and the Toyota Prius, both of which are classified as compacts.  In 2004 there were a total 83,153 hybrids registered in the United States (Lipman and Delucchi [2006, p. 116]).  The first question is how gasoline purchases differ between an “optimizing consumer” and a “random buyer.”  That is, how much can the consumer save in gasoline purchases by buying whichever gas is cheapest as compared to another consumer who simply buys gasoline at random.  The analysis of this paper suggests that the optimizing consumer can save around 4.39% of the annual gasoline bill, or about $20.68 a year.  The second question the paper addresses is how much a hybrid owner saves in gasoline costs as compared to an owner of one of our other vehicle types.  Our results suggest that comparing hybrid owners to owners of other compacts, hybrid owners save about $388 a year.  Comparing hybrids to large cars and SUV/Vans gives savings of about $456 and $665 respectively.
     Although the larger problem of consumer demand for gasoline is well studied (gasoline demand being inelastic) by Nicol [2003], Cheung and Thomson [2004], Ramanathan and Subramanian [2003], and  Espey [1998], as far as we know, the consumer’s problem of this paper has only been considered in Highfill and McAsey [2007].  As far as we know, the only other paper to address the costs associated with hybrid vehicles is Lipman and Delucchi [2006].   Their analysis does not consider the Honda and Toyota hybrids although these are the most common for 2004.  Their results are very hard to compare with the results of this paper in that they include costs we do not (e.g. lifetime maintenance costs) but hold gasoline prices constant ($1.46 in 2000 prices, Lipman and Delucchi [2006, p. 123]) while we allow gas prices to vary and do not consider other costs.  Further, their results are reported on a per gallon basis rather than per time period as ours are.  Their final conclusion is that some hybrid designs are “competitive with gasoline vehicles … when viewed from a social cost perspective” Lipman and Delucchi [2006, p. 131]. 

Data and Parameters

     The gasoline prices were obtained from the (lowest grade) postings outside four stations in Peoria Illinois each day that an author was able to view them for the study period of November 18, 2003 to October 25, 2005.  The stations, along a three mile length of a single major thoroughfare, included major brands, local brands, and stations with and without repair facilities.   The overall average gasoline price per gallon of the daily minima for all four stations is $1.94.  The maximum and minimum price per planning period is shown for the 33 18-day (non-overlapping) planning periods for which full data exists.  The average range per planning period is about $0.24, and the range (of ranges) is $0.12 to $0.52.
     The “micro-micro” problem requires consumer data on daily fuel use, tank size, and the length of the planning period—see Table 1.
     Vehicle classifications are derived from cars.com. The parameter values are averages for the 2004 model year (the middle year of the gas price data).  Although in 2007 there are 10 hybrid cars/SUVs available, in 2004 there were mainly two hybrid cars:  the Honda Civic Hybrid and the Toyota Prius.  Honda also produced the Insight and Ford began production of its SUV Escape Hybrid but the Civic and the Prius were more mainstream at the time than either of the other two vehicles.  Both the Civic and  Prius feature a small gasoline engine and an electric motor operating in parallel.  The other vehicles are provided for comparison purposes.  That data includes domestic and foreign, low priced and high priced cars (“large cars” being midsize and full size). 
     Daily fuel usage is calculated by
daily fuel use = [(miles per year)/(miles per gallon)]/(365 days per year)=gallons/day
where miles per year is from Federal Highway Administration data and average miles per gallon is from cars.com.  Average tank size is from cars.com as well.  The planning period of 18 days (12 and 10 days for the comparison vehicles) are the longest feasible given the preceding parameters and the assumption that consumers must refuel at least once but will not refuel more than twice per planning period.  The number of planning periods is determined by the number of (non-overlapping) periods of 18 consecutive days (12 and 10 days for the comparison vehicles) in the daily gasoline price data.  
     Each of the 33 (non-overlapping) 18-day planning periods for which price data exists is treated as a separate experiment for Hybrids (and similarly for the comparison vehicles).  The results reported below are the average over all relevant experiments.  Recall that it is assumed that the consumer refuels either once or twice, choosing the number of refuelings which yields the lowest planning period cost of gasoline.
     It might be noted that in the analysis that follows each station is considered separately.  But we also consider the case when the consumer treats the stations as interchangeable.  That is, a “homogeneous” solution is when the consumer chooses the station with the lowest price on any given day and without regard to brand or services available.

Annual Gasoline Costs of Driving a Hybrid

     This section investigates how much an idealized consumer would pay for gasoline on an annual basis under various assumptions of consumer behavior.  It is assumed that the consumer knows the gas prices during the current planning period, but knows nothing about prices in the next planning period; thus he or she begins and ends each planning period with a half a tank of gas.  A “feasible” gasoline buying strategy requires refueling before running out of gas, and having enough gasoline in the tank by the last refueling to have half a tank of gasoline at the end of the planning period.  The solution of the optimization problem is to compute all feasible refueling possibilities, find the cost of each, and then pick the minimum.  (See Highfill and McAsey (2007) for further discussion of the consumer’s problem and some examples.)
     The “optimizing” consumer buys gasoline to minimize gasoline expenditures per planning period (given the feasibility constraints outlined above).  The costs for such a consumer are shown as “Min” in Table 2.  The “worst-case” consumer is the unluckiest buyer in the world (whose buying habits still satisfy the constraints) who just happens to buy from the most expensive station on any given day.  The solution for this consumer is shown by “Max Cost” in Table 2.  This maximum cost (given one or two refuelings), is calculated by selecting the prices and the feasible quantities that give the highest total cost in the various cases.  The third type of consumer is the “average non-optimizer” or “random buyer” whose costs are the “Random Cost” in Table 2.  The Random Cost is calculated as follows.  In the homogeneous case, for each planning period, randomly select:  a feasible pair of days, two stations, whether to buy more gas on the first day of the feasible pair or on the second.  This information then produces a total cost for the particular planning period.  This process was repeated 1000 times for each planning period as if 1000 consumers were randomly purchasing fuel from these stations during each period.  The same procedure was followed for the other columns, except that the station was fixed.
     The first section in Table 2 is “Planning Period Average Costs” which gives the minimum cost of buying gasoline in the various cases.  For example, for “Homogeneous Stations”, the average refueling cost per 18-day planning period is $22.21 for the optimizer.  The same average for the random purchaser is a dollar (and two cents) more.  The planning period costs for the unlucky consumer are only slightly more than a dollar more than for the random buyer.  Comparing the single station results, station 2 is the cheapest, followed by closely by station 4, then station 1, with station 3 being the most expensive.  Notice that on average station 2 only costs a penny more per planning period as compared to the time and trouble it requires the consumer to search for the daily cheapest station.  We note that station 3 is a “full service station” and so might speculate that station 3 offers valuable services since it is on average $.85 more expensive per planning period.   The next section of the table simply aggregates the first section to annual numbers.  Our estimate of the average annual gasoline cost for the optimizer is $450.37.  There are slightly more than twenty 18-day planning periods a year; annual savings between the optimizer and the random buyer are $20.68.   This calculation is shown in the third section of the table.
     The last section of the table gives the savings in percentage terms.  The optimizer saves between 2.37% and 8.83% of his or her annual gasoline bill compared to the other possibilities.  Notice as well that the percentage savings for Min v Random and for Max v Random are rather similar, as was suggested by the dollar savings discussed above. 
     It remains to compare the experience of hybrid owners to that of owners of other vehicles. 

Comparison to Other Vehicles

     The hybrids in this data set are compacts.  Our analysis suggests that buying a hybrid as compared to another compact will save the consumer about $388.21 a year in gasoline expenses.  There are also some “time and trouble” expenses saved in that the hybrid owner refuels less often.  Sometimes “Green” advocates argue that consumers should trade in their gas guzzlers for a compact hybrid.  Our results suggest that an SUV/Van owner who did so would save $665.53 annually and a large car owner who did the same would save $456.34 annually.
     The yearly average savings of optimizers owning hybrids as compared to random buyers or unlucky ones is smaller in dollar terms for any of the comparison vehicles—not surprisingly, since the amount spent on gasoline is much less.  In percentage terms, however, optimizing is slightly more important for hybrid owners than it is for owners of the other vehicle types.  Comparing the Min v Max, 8.83% is a little larger than 8.66%, 8.65%, or 8.33%.  Comparing hybrid compacts to other compacts (8.83% - 8.33%), hybrids owners gain a half of a percent more by optimizing. Comparing the Min v Random, 4.39% is a little larger than 4.73%, 4.72%, or 4.59%.  In this case, hybrids gain slightly less than a half of a percent more by optimizing. 

Conclusion

     From the cars.com site, in 2004 the hybrid Honda Civic has a list price about $4,240 more than the regular Civic.  Comparing the Toyota Prius with the Corolla, the list price of the Prius is about $5,680 higher.  If other costs of owning the vehicles are about the same, our results suggest that an owner would need to keep their cars upward of ten years to breakeven financially.  But it goes without saying the individual experiences differ, and it is by no means clear that the other costs of car ownership are the same between the hybrid and the regular car.  In the final analysis, people who buy hybrids may do so for non-economic reasons; but the present paper suggests there is a benefit to both behaving optimally and buying a hybrid in terms of gasoline expense.    

References

Cheung, Kui-yin and Thomson, Elspeth.  (2004). The Demand for Gasoline in China: A Cointegration Analysis.  Journal of
       Applied Statistics, 31(5), 533-544.
Espey, Molly. (1998).  Gasoline Demand Revisited: An International Meta-analysis of Elasticities. Energy Economics, 20(3),
       273-295.
Nicol, C. J. (2003).  Elasticities of Demand for Gasoline in Canada and the United States. Energy Economics, 25(2), 201-214.
Highfill, Jannett and McAsey, Michael.  (2007)  The Consumer’s Micro-Micro Gasoline Buying Decision.  International Advances
       in Economic Research, 13, 433-442.
Lipman, Timothy E. and Delucchi, Mark A.  (2006)  A Retail and Lifecycle Cost Analysis of Hybrid Vehicles.  Transportation
       Research Part D, 11, 115-132.
Ramanathan, Ramakrishnan and Subramanian, Geetha. (2003). Elasticities of Gasoline Demand in the Sultanate of Oman.
        Pacific and Asian Journal of Energy, 13(2), 105-113.

Figure 1
Gasoline Price Range per Planning Period

Table 1
Parameter Values