Statistics Canada
www.statcan.gc.ca
Common menu bar links
Main article
- Introduction
- Canadian Vehicle Survey 2007: personal or business vehicle use and fuel consumption data
- Comparing the CVS to other data sources on fuel consumption
- Conclusion
1 Introduction
The Canadian Vehicle Survey (CVS) is a voluntary, vehicle-based survey that provides quarterly and annual estimates of road vehicle activity. This includes vehicle-kilometres and passenger-kilometres as well as a number of other elements related to the trip such as gender of driver, time of day and season (see "References" section no. 2, Statistics Canada, 2008).
In 2007, the sample size of the CVS was increased in order to address a data gap regarding consumption of fuel for personal use. The CVS was seen as a possible solution to getting better insight into the household component of fuel consumption. By differentiating between types of vehicle use the CVS can provide estimates of fuel consumed for personal and business purposes.
The distinction between the type of vehicle use could potentially pave the way for a number of improvements to existing statistics that use household fuel consumption data:
- total household energy use and total greenhouse gas emissions;
- household use of transportation fuels;
- analysis of trends in the household component of total greenhouse gas emissions.
This report aims to present a national profile of annual vehicle fuel consumption by type of use –personal or business. More importantly, the analysis focuses on the comparison of the fuel quantities estimated via the CVS and other known data sources that produce similar types of information, particularly other data sources within Statistics Canada. The data comparison will provide further validation and foundation for future work using the CVS datasets on fuel consumption by type of vehicle use.
2 Canadian Vehicle Survey 2007: personal or business vehicle use and fuel consumption data
2.1 Survey overview and additional questions for 2007
The CVS is a cross-sectional sample based survey. The survey consists of two steps. The first step is a computer assisted telephone interview (CATI) with the registered owners of the sampled vehicles. The interview is used to collect some general information on the usage of the vehicle as well as to ask the respondent to complete a trip log specific to his or her vehicle type. The vehicle log asks detailed information about vehicle trips, distance driven, vehicle type, fuel consumption, as well as demographic profile of drivers and other related information. 1
In 2007, respondents driving vehicles less than 4.5 tonnes (light vehicle category) were asked if the vehicle was for personal or business use. This was consistent with the general notion that personal use vehicles fall primarily in the light vehicle category and that only a very small portion of overall fuel consumption would be attributable to people driving very large vehicles for personal use. The following question was added to the CATI portion of the survey in order to determine if trips were for personal or business use:
Is this vehicle used for business only, personal use only or a combination of business and personal use?
Respondents who stated that the vehicle was used for a combination of personal and business use were asked a follow-up question to assign a percentage of use.
Driving for business purposes was defined to include the following activities: service calls, sales calls, deliveries and pick ups, as well as business trips. Taxis are also in this group. 2 If the trip was part of the driver’s job, then the trip was understood to be classified under business use. Therefore, driving for personal use could be interpreted as any type of driving that cannot be connected directly with work purposes. Vehicle use to commute to and from work is not considered business use (driving as part of the job).
2.2 Scope of report and survey highlights
In 2007, the CVS had a sample of 26,987 vehicles drawn from ten provinces. Another 11,693 vehicles were included in the sample for the three territories. Provincial response rates ranged from 53% in Alberta to 69% in Prince Edward Island (see "References" section no. 2, Statistics Canada, 2008, pages 32 and 37). All motor vehicles except buses, motorcycles, off-road vehicles and special equipment were included in the CVS sample.
The 2007 CVS estimated that 0.34% of the total vehicle population is fuelled by something other than gasoline or diesel. 3 Given such a small proportion, the CVS estimates for the quantity of fuel consumed by these types of vehicles are not reliable. The focus of the comparison between the CVS and other sources will, therefore, concentrate only on vehicles fuelled by gasoline and diesel.
Table 1 illustrates the overall estimates of fuel consumption by vehicle type. In 2007, vehicles consumed 42,694 megalitres (ML) of gasoline and diesel with 76% of this total coming from light vehicles (less than 4.5 tonnes) and the rest from heavy vehicles (4.5 tonnes and over).
Canadians driving light vehicles used 32,597 ML of fuel, 75% of which was identified as personal use while the remaining 25% was for business use (Table 2).
Across the country, Ontario had the highest share of total transportation fuel consumption (38%) for light vehicles, followed by Quebec (20%), Alberta (14%) and British Columbia (12%). This is to be expected, as these were the provinces with the greatest numbers of vehicles (see "References" section no. 2, Statistics Canada, 2008).
3 Comparing the CVS to other data sources on fuel consumption
In order to use this new data effectively going forward, comparing the CVS data to other sources that collect and publish similar types of data was deemed to be important. This comparison will set the stage for future analytical work that will be conducted using the CVS, including the personal or business split, as a data source.
Aside from the retail pump sales data from the Monthly Refined Petroleum Products Survey, 4 a common limitation on the comparability of the CVS data with other data used in this report is that CVS data uses 2007 as the reference period. For the other sources against which the CVS data are compared, only 2006 or earlier data are available. However, it should be noted that there has been relatively small changes in fuel consumption over the years. For example, from 2006 to 2007, the CVS showed a 1.5% increase in gasoline consumption by light vehicles to 31,305 ML. Since 2004, there has only been a 2% overall increase in gasoline consumption by light vehicles.
3.1 Retail pump sales, 2007
The Monthly Refined Petroleum Products Survey 5 publishes annual estimates of provincial retail pump sales in Canada. When comparing these data to the CVS, it is important to note that this survey reports fuels distributed at retail locations.
The overall totals for diesel and gasoline produced by the two surveys appears to be very close. Table 3 shows that the overall estimate for the consumption of gasoline and diesel for the retail pump sales figures is 1 percent more than the total volume of gasoline and diesel produced by the CVS. Although not perfectly comparable, as explained below, the comparison seems to indicate promising results in terms of the CVS estimates.
For the purpose of comparing the CVS with respect to personal fuel use, it is necessary to look only at light vehicles. This type of vehicle cannot readily be excluded from the retail pump sales data. There are vehicles such as heavy vehicles, buses, motorcycles and off-road vehicles that use these retail pumps. There are also many examples of other types of special equipment that would fill up at retail locations including: boats, lawnmowers, snow blowers, gas lawn trimmers etc. These vehicles and implements are not in the population of interest for the CVS.
Because of some basic differences in scope, when the heavy vehicles are excluded from the CVS, it is expected that retail pump sales will report substantially higher volumes of fuel than those estimated by the CVS where only light vehicles are included. Excluding heavy vehicles, the difference between CVS fuel estimates and the retail pump sales data shows that the retail pump sales data are 33% higher than the CVS.
3.2 Comparing CVS data to the Environment Accounts and Statistics Division (EASD) fuel estimates
Environment Accounts and Statistics Division produces data on energy use by sector and fuel type. The most recent year of available data is 2004. It contains values on gasoline and diesel consumption by household, business and non-business sectors.
An attempt was made to compare the personal use CVS data to the household component of the EASD figures. While there is a definite spread between the years of data available, this comparison does provide a useful perspective for looking at the personal component of transportation fuel.
To produce energy consumption estimates for the household sector, EASD applies the personal or business fuel consumption allocation from the Input-Output Accounts. This allocation is applied to the retail pump sales data in order to assign an appropriate amount of motor gasoline and diesel fuel to the household sector. 6
The business component of transportation fuel consumption is not readily available through these data because of non-transportation uses of fuel that are captured by the data set. For example, agricultural production activities such as harvesting grain are captured by EASD but not by the CVS.
The values produced by EASD are for all sectors regardless of how the fuel is being consumed while the CVS is concerned only with fuel consumption for transportation purposes. For instance, the consumption of diesel by a diesel generator or in agricultural production would be captured by the EASD figures, whereas the CVS would not measure this number. Because few people have diesel generators, and most fuel consumption at the household level is for transportation purposes, the data sources should be comparable for the household sector.
Table 4 illustrates that in 2004 the household sector, was responsible for 28,860 ML of gasoline and diesel as measured by EASD. This total is 11% lower than the 2007 CVS total quantity for gas and diesel of 32,597 ML.
For a more accurate comparison, it is the value for the 2007 CVS personal component of light duty gasoline and diesel consumption (24,452 ML) which more closely aligns with the concept that is being captured by the EASD value. This means the values produced by EASD are 18% higher than the comparable 2007 CVS values. A simple linear regression to project the values that EASD would likely produce for household fuel consumed based on the past trend showed the EASD value to be estimated at 22% higher (29,832 ML) than the CVS, if the projected values are accurate.
Taking into account the difference in reference years and the yearly increases in fuel consumption since 2004, would narrow down the difference. Conceptually, these two sources are the closest to one another, of all those compared in this study. The main difference is the amount of gasoline consumed by households operating private boats, off-road vehicles and gas-powered equipment which are captured in the EASD estimates but not in the CVS. However, these amounts represent a relatively small proportion of overall total, hence the estimates from EASD’s energy use accounts would be slightly higher than the CVS estimates for total gas and diesel use.
3.3 Comparison of CVS with non-fuel quantity data: Survey of Household Spending and fuel price data
Another method used to compare the CVS data was to estimate fuel consumption through average provincial fuel prices and reported expenditures on fuel at the household level.
Average household level expenditures on gasoline and other fuels for owned and leased automobiles and trucks/vans is available yearly through the Survey of Household Spending (SHS survey no. 3508). In the SHS, people are asked for their expenses in relation to automobile expenditures. They are asked to report annual expenditures for transportation fuel.
The monthly price for regular unleaded gasoline and diesel fuel at full service and self-service fuel stations by urban area is available through CANSIM Table 326-0009 (Average retail prices for gasoline and fuel oil, by urban centre, monthly).
In order to calculate a provincial average yearly fuel cost, monthly fuel price data was used to calculate the yearly average for each urban centre available. This yearly average was then converted into a provincial average based on the urban areas reporting in each province.
Total litres of fuel was then calculated for each province by using the following formula:
Gasoline and diesel weights were calculated for each province using provincial vehicle counts for the 2007 CVS. These are composed of the ratio of the number of gasoline vehicles to the total vehicle count and the number of diesel vehicles to the total vehicle count. The weight was then used as a proxy to estimate the proportion of households that spent money on gasoline vis-à-vis diesel.
This method provides an estimate of fuel consumption at the household level, separate from business consumption. The estimates produced are 10 % lower at the national level than the personal use estimates produced by the CVS with half of the difference between the two data sets coming from Ontario (Table 5).
3.4 Comparison of CVS to National Inventory Report 1990 – 2006: Greenhouse Gas Sources and Sinks in Canada
The National Inventory Report (NIR) is produced by Environment Canada and was submitted to the United Nations Framework Convention on Climate Change in April, 2008. The report contains a component that estimates greenhouse gas emissions (C02) in Canada by vehicle type and province. This estimate is produced by applying a conversion factor 7 to a fuel consumption value by vehicle type in order to estimate greenhouse gas emissions.
Fuel consumption is calculated using the following formula:
Fuel consumption = Population * Vehicle-kilometres-traveled * Fuel consumption 8
Fuel consumption – Gasoline and diesel consumption is estimated first by using the population, vehicle kilometres and fuel consumption ratios. This amount is then adjusted based on top-down estimates from Statistics Canada’s Report on energy supply and demand (RESD) 9 . If the amount from the model calculation is greater than the amount of fuel reported in the RESD, then the two amounts are averaged.
Population – Light duty vehicle and truck populations for 1990 to 2002 were obtained from the Canadian Vehicles in Operation Census, maintained by DesRosiers Automotive Consultants Inc. From 2003 to 2006 the number of light duty vehicles and trucks were estimated based on observed trends.
Vehicle-kilometres traveled – This is an annual measure of kilometres traveled for a vehicle. Odometer readings are estimated for the country based on emissions monitoring tests from Ontario in 2004.
Fuel consumption ratio (FCR) – Average provincial FCRs by vehicle class and model year are estimated using provincial vehicle sales. These fuel consumption ratios are adjusted to be 25% higher than the laboratory ratings figures to account for real life driving conditions.
The main difference between the two data sources is that the CVS is a sample based survey whereas the NIR compiles information from several different sources in order to model emissions based on fuel consumption values.
Another difference between these two sources is that each defines light vehicles differently. The CVS defines light vehicles as those vehicles with a gross vehicle weight rating (GVWR) below 4.5 tonnes, whereas in the NIR, light vehicles are those with a GVWR below or equal to 3.9 tonnes. Given that the CVS has a higher threshold for defining light vehicles, it may be expected that there is also a corresponding higher estimate for fuel consumed by light vehicles in the CVS. This is not the case, as will be shown later.
Data tables received from Environment Canada show that the overall total is 11% higher than what is produced by the CVS (Table 6).
4 Conclusion
Each of the data sets used for comparison had differences stemming from various definitions, methods, and scope of the source. The direction of the differences between estimates is generally consistent across the provinces for all of the data comparison methods chosen.
Larger differences are evident between the CVS, excluding heavy vehicles, and the retail pump sales data. In no case is there a higher fuel quantity reported by the CVS than the retail pump sales. This is to be expected because of the broader population coverage by the retail pump sales.
Given the different population definitions and the differences in methodology used by the various sources, the data collected through the CVS do not appear to be unreasonably out of line and could be considered to be at least as accurate as any other estimate.
