See more vehicle-related facts in AUTOSTAT >>
Vehicle Greenhouse Gases to 2030 in Metro Vancouver and the Naked Emperor
(To download a pdf version of the study, go to bottom of this report)
Once again the Federal Government has announced a new commitment to reducing GHGs, this time by 30% by 2030 from a 2005 baseline, after failing to meet its 2009 commitment of a 17% decline by 2020 (now on track for less than a 1% reduction). Similarly, in 2007 the BC Government passed legislation committing the Province to reducing GHGs by 33% by 2020 from a 2007 baseline (18% by 2016), a goal that is acknowledged by most knowlegeable people as being impossible.
This Report examines how well Metro Vancouver likely will do in achieving these reductions in terms of vehicle GHG emissions. Projections using our VKT/GHG Forecasting Model and based on detailed assumptions from the US Energy Information Administration, the Metro Vancouver Regional Growth Strategy and forecasts of transit availability and options, suggests that, far from declining, vehicle GHGs will likely increase between 2007 and 2030. This despite a robust improvement in new vehicle fuel consumption rates over the period and a decline, at least after 2020, in average vehicle kilometres travelled.
Using conservative (e.g., emission-saving) assumptions, the Report details how the stock of vehicles will likely change given population changes, how fuel efficiencies for the entire stock is expected to change based on new vehicle consumption rates, and how the average kilometres travelled each year is expected to alter based on a variety of economic conditions (fuel prices, average incomes, etc.) and policy initiatives (transit availability, fare card prices, regional growth, etc.).
The Report also examines what would be the impact of an alternate fuel price forecast using the USEIA “High Oil” case. What is found is that, although emissions would be lower, even with a very high price of oil, emissions will still not decline to the levels anywhere near the 33% goal committed to by the Federal Government.
VEHICLE GREENHOUSE GASES TO 2030 IN METRO VANCOUVER AND THE NAKED EMPEROR
The Federal Government recently agreed with other G7 partners to eliminate GHG emissions by the turn of the century and, as an initial step to achieving this goal, announced that Canada would reduce GHG emissions by 30% by 2030 from a 2005 baseline. This goal in emission reductions is similar to the BC Government’s own objective of reducing total provincial GHGs by 33%, although that aim is to be accomplished by 2020, a mere five years from now.
In British Columbia, vehicle emissions represent approximately 25% - 30% of all GHG emissions and therefore a reduction in vehicle emissions ought to form an essential component in any strategy to lower overall emissions. Similarly, since vehicles represent roughly 35% - 40% of all emissions in Metro Vancouver, one should expect that a strategy for lowing vehicle emissions in the Lower Mainland is also essential.
With the recent defeat of the transit referendum in Metro Vancouver, an interesting question is raised: how, in the absence of major investment in transit infrastructure and facilities, will vehicle GHGs change over the next 15 years in Metro Vancouver, and will this change meet provincial and/or federal goals. This Report assesses that question using our VKT/GHG Forecasting Model and input data for the Metro Vancouver Regional District. The result, which may not be surprising to some, show that GHGs will continue to rise in the medium term and, far from declining over the next decade and a half, vehicle GHG emissions are likely to be some 1.3% higher in 2030 than in 2014 (8.5% higher than in the baseline year 2005).
Given that outcome, one must question the ability of Metro Vancouver or the Province to achieve the stated GHG reduction goals. Indeed, are existing “strategies” to reduce GHGs conflating assuming an outcome with a plan; that rather than assessing the actual impacts of GHG-reducing initiatives, the “plan” becomes a collection of ideas, assuming those initiatives will suffice. If so, then perhaps it’s time to expose the naked emperor.
Any forecasting model is highly dependent on the input assumptions, and our Model is no exception. The main economic and vehicle-related assumptions are taken from the US Energy Information Administration (USEIA) whose own internal model provides the best systematic and objective set of official US government data on expected new vehicle attributes and changing driver behaviour. The USEIA model provides internally consistent forecasts for the following:
US Growth and Energy Price Data: The model incorporates forecasts of US economic growth, inflation, and crude oil prices which, in turn, drive expectations for new vehicle efficiencies and new vehicle purchase patterns. For the purposes of this analysis we have used the USEIA 2015 Reference Case, although later in this Report we do provide an alternative scenario based on the USEIA “High Oil” scenario.
Figure 1: Energy Prices, 2010-2030
Source: USEIA (2015 Reference Case) and VKT/GHG Forecasting Model
Figure 1 above displays the estimates of real ($2013) crude oil prices (WTI $US/bbl) which we have incorporated into our Model as well as the resultant real ($2002) gasoline and diesel prices in Metro Vancouver.
Fuel Efficiencies: USEIA projections for expected new model fuel efficiencies by vehicle type and fuel type are generated based on US Environmental Protection Agency (USEPA) regulations and, to a small extent, on expectations of US growth and future crude oil prices. Figure 2 below displays the projected trends in new vehicle fuel consumption rates over the study period.
Figure 2: New Vehicle Efficiencies 2010 – 2030
Source: USEIA (2015 Reference Case) and VKT/GHG Forecasting Model
* Efficiencies for all vehicles except Medium-Duty are for gasoline vehicles; Medium-Duty is the weighted average of gasoline and diesel;
Incorporating those trends into our Model suggests that the weighted-average new passenger vehicle fuel consumption rates will improve by some 28% over the forecast period, falling from a weighted average of 10.7 litres/100 kms for model year 2010 vehicles to 7.7 litres/100 kms for model year 2030 vehicles. Medium-duty vehicles are projected to experience a 10% decline between 2010 and 2030, falling from a weighted average of 16.9 litres/100 kms in 2010 to 15.2 litres/100 kms by 2030. Other vehicles (e.g., heavy trucks, buses – not shown on graph) also are expected to experience significant improvements to fuel consumption rates of approximately 25% over the forecast period.
New Vehicle Purchase Patterns: Over the last decade and a half in the US, in British Columbia and in Metro Vancouver, there has been a general move away from the purchase of small and large cars toward pickups (light-duty and medium-duty) as well as minivans. The USEIA data suggests a continuation of these trends in vehicle purchase patterns as displayed in Figure 3.
Figure 3: New Vehicle Purchase Patterns
Source: USEIA (2015 Reference Case) and VKT/GHG Forecasting Model
The USEIA also projects new purchases by fuel type, specifically an increase in the share of electric (both pure electric and plug-in electric-hybrid) vehicles. Integrating these trends in our Model, the percentage of new electric passenger vehicles (excluding pickups) purchases is estimated at 4.8% by 2030. Light-duty trucks are projected to have a lower penetration rate at around 1%. By 2030, then, approximately 3.5% of the passenger vehicle total stock is projected to be either pure electric (slightly greater than 1%) or gasoline-electric hybrids (with roughly 2.5% of the total stock).
Besides data from the USEIA on the general economy and vehicle characteristics, the Model also incorporates data on a number of other variables:
Real Per Capita Average Incomes: Growth in Metro Vancouver for this variable averaged 1.1% over the 2001-2013 period (the latest for which there are data); accordingly, in keeping with our policy of using conservative assumptions, we have used a growth of .75% for the forecast period (USEIA forecasts per capita GDP in the US will grow at ~1.5% over the forecast period);
Population Growth by Age and Sex: these data are based on Metro Vancouver’s Regional Growth Strategy which identifies population growth by small area (Transit Area Zone – TAZ) over the forecast period;
Density and Driving Distance variables: These data are based on the Metro Vancouver Regional Growth Strategy (density equals population divided by area in hectares for each TAZ) and weighted population average of the distance of each TAZ from the Central Business District (downtown Vancouver) and from each Regional Town Centre);
Transit policy variables: These data include forecasts of transit access, number of bus hours, the number of Skytrain hours, and the real dollar price of the monthly fare card.For the purposes of this forecast, we have assumed that transit access follows the Regional Growth strategy expectations, that bus hours increase with population growth, that Skytrain hours increase with the completion of the Evergreen Line beginning in 2018 and population growth, and that real dollar fare card price stays constant over the forecast period.
Given the above assumptions, we are now in the position to estimate the trends in GHG emissions over the forecast period 2015-2030. In an earlier paper, we used our VKT/GHG Forecasting Model to examine the trends in vehicular GHG emissions in Metro Vancouver between 2007 and 2014. The analysis examined the trends in the three factors that influence GHG emissions and found that emissions increased to 2009 and then began falling, only to begin rising again starting in the middle of 2012 and has continued to increase through to the end of 2014.
This paper again examines the same three factors that contribute directly to changes in future GHG emissions: changes in the vehicle stock to 2030, changes in the average fuel efficiency over the same time period, and changes in the average number of vehicle kilometres travelled (VKT) in the coming 15 years.
Passenger vehicle stock is driven by the vehicle-to-population ratio. While the stock ratio varies over time, there has been a general decline from an average of .677 vehicles per person in 2001 to .649 vehicles in 2014. We have extended the time trends for each sub-group over the forecast period, with the result that by 2030, the vehicle-to-population ratio is projected to fall to .620 vehicles (this effectively constitutes an 8% decline in the total stock per capita.
Figure 4 below highlights the trend in vehicle stock over the forecast period suggesting that, even with a falling ratio of vehicles to population, the stock of passenger vehicles is projected to rise from approximately 1.26 million in 2010 to 1.56 million by 2030. Recreational vehicles historical have relatively constant stock ratios (motorcycles increasing slightly; motorhomes falling slightly).
The other vehicle groups (Medium-Duty and Other) are driven by a combination of stock-to-population ratios and economic growth. The Medium-Duty stock ratio is projected to rise over the forecast period while Other has had a relatively stable ratio over the last decade or so and is continued to remain stable. The effect is that the stocks of these vehicle groups show a slightly greater increase in overall stock than passenger vehicles.
Figure 4: Total Vehicle Stock, 2010-2030
Source: VKT/GHG Forecasting Model
Fuel Consumption Rates
Earlier in this paper we described the trends in new vehicle fuel consumption rates. But in any year, the overall fuel consumption rate is a weighted average of new vehicles and the stock of older vehicles.
Our VKT/GHG Forecasting Model automatically generates this weighted average based on the aforementioned new vehicle consumption rates, new vehicle purchases by vehicle class and fuel type and the stock and efficiencies of older vehicles after accounting for vehicle attrition.
Figure 5 on the following page highlights the trend in aggregate fuel consumption rates from 2010 to 2030.
Figure 5: Fuel Consumption Rates 2010 - 2030
Source: VKT/GHG Forecasting Model
Vehicle Kilometres Travelled (VKT)
Average VKTs per vehicle are estimated econometrically for each separate vehicle class, fuel type and license type (e.g., a separate equation for small gasoline cars driven for pleasure only) as explained in detail in a previous report, but are heavily influenced by fuel prices, incomes, age of the vehicle and regional and transit policy variables. While VKTs are estimated for each vehicle type and model year, the aggregate average follows the pattern highlighted in Figure 6 below.
Average VKT fell during the recessionary period 2009 to mid-2012 (the historical VKT/GHG Model is a quarterly model) but by 2014, VKTs returned to averages slightly higher than those recorded prior to the recession. Average VKTs are projected to rise again in 2015 largely due to the fall in fuel prices at the pump and a return to higher levels of economic growth. After that, average VKTs flatten out and then begin to decline after the 2021 period.
Figure 6: Average VKT per Vehicle, 2010- 2030
Source: VKT/GHG Forecasting Model
Total Fuel Consumption
Total fuel consumption in litres is calculated as a simple identify: Total Stock X average Fuel Consumption Rate X average VKT divided by 100.
Figure 7: Fuel Consumption in Litres, 2010 - 2030
Source: VKT/GHG Forecasting Model
As outlined in an earlier paper (Vehicular GHGs in Metro Vancouver), fuel consumption dropped during the period beginning in 2009 but began in mid-2012 and has continued to increase. Passenger vehicle fuel consumption is projected to remain more-or-less stable over the next five years and then begin a slow decline. By the end of the forecast period, 2030, passenger vehicle fuel consumption is expected to be about 4.8% below levels recorded in 2014 and about 0.6% below the baseline 2005 values.
Fuel consumption for all vehicles, in contrast, continues to climb over the near term, and by the year 2030 is still 0.8% higher than total consumption in 2014 (8.6% higher than in the base year 2005).
While fuel consumption and GHG emissions are closely linked, the link is not quite exact. A litre of gasoline burned will always produce 2,289 grams of CO2 while burning a litre of diesel always produces 2,263 grams and propane 1,510 grams. Consequently, a small difference in total emissions may occur if the proportion of diesel-to-gasoline-to-propane vehicles changes over time. At the same time, however, newer vehicles generally emit much less CH4 and N2O emissions (gases that have much higher CO2 equivalency) and therefore as the vehicle stock renews itself over time and older, higher emitting vehicles drop out of the vehicle stock and/or drive fewer kilometres, emissions will fall. Figure 8 below highlights the impacts on emissions over time.
Figure 8: GHG Emissions, 2010-2030
Source: VKT/GHG Forecasting Model
Emissions for passenger vehicles plateau between 2015 and 2022 and then begin to fall thereafter, with levels in 2030 roughly 4.9% below 2014 levels (and 1.7% below base year 2005 values).
Emissions for all vehicles will continue to increase between 2014 and 2021, only beginning to fall in the year 2022 onward. Despite this fall, by the end of the forecast period, emissions for the entire vehicle stock will still be higher than those experienced in 2014 by about 1.1% (8.7% higher than base year 2005 emissions). Given that vehicles represent around 35% of all GHGs in Metro Vancouver, it is understandable that one questions the idea that overall GHGs could reach a level 30% lower than 2005 levels by 2030.
“High Oil” Case
The analysis in this paper is based primarily on the USEIA “Reference Case” forecast of crude oil prices, projected fuel efficiency improvements and expectations of new vehicle purchase patterns. In this section, we examine the impacts on emissions of using the USEIA “High Oil” case.
Figure 9: “High Oil” Emissions, 2010-2030
Source: USEIA (2015 Reference Case and 2015 High Oil Case) and VKT/GHG Forecasting Model
Figure 9 on the previous page highlights the difference in vehicular GHG emissions between an economy with Reference Case crude oil prices and vehicle characteristics versus an economy with “High Oil” crude prices and “High Oil” vehicle characteristics. Under the Reference Case, nominal crude prices remain under US$100 for the next ten years while under the “High Oil” Case, nominal crude prices jump up to approximately US$130 in 2016 and continue to rise over the forecast period reaching over US$250/bbl by 2030.
The interesting result is that, while these extremely high oil prices do affect overall emissions, particularly in 2016 when oil prices experience a very large jump, GHGs would still continue to increase from 2017 to 2022 and even by the end of the forecast period (2030), emissions would only be 5.5% lower than actual emissions in 2014 (versus an increase of 1.1% under the Reference Case).
The overwhelming conclusion is that one cannot rely on higher fuel prices to provide a solution to vehicular GHG emissions in Metro Vancouver and that to achieve the stated government goals for vehicles, other policy initiatives will need to be considered. It is important to point out that this also means that a reliance on increasing the Carbon Tax is equally inadequate as a solution to reducing GHGs. A recent paper (The Impacts of the Carbon Tax on Vehicle Fuel Use in Mero Vancouver) provides a much more detailed examination of how the carbon tax does and doesn’t affect overall driving behaviour and the resultant impacts on fuel consumption and GHG emissions.
The power of modelling GHG emissions properly is that one can examine in detail the assumptions used and ask what would be the impacts of differing expectations. In this example, we used a different crude oil price forecast, but we also could examine the impacts of a different projection of average incomes, or transit availability, or population growth and density, or uptake of electric vehicles or any number of other assumptions.
For this analysis we have used documented assumptions from a variety of credible sources (e.g., the US Energy Information Administration, Metro Vancouver Regional Growth Strategy) and when assumptions are ours alone, we have endeavoured to be conservative in our outlook – that is, using assumptions that generally are emissions-reducing. These assumptions include:
A declining stock-to-population ratio leading to a lower total stock than if a simple stable stock-to-population ratio were used;
A refinery margin that is much higher than historic levels recorded between 2001 and 2010 leading to a higher fuel price assumption and a downward influence on average VKT and consequently lower emissions;
A growth in average per capita incomes lower than the average recorded over the 2001-2013 period resulting in a downward influence on average VKT and consequently lower expected emissions;
Transit availability and access: bus and Skytrain hours (after the opening of the Evergreen Line) increase with the population, transit access improves by 25% (following TransLink assumptions prior to the Transit Referendum), and Fare Card costs remain constant in real dollars.Given the failure of the Transit Referendum, these are fairly optimistic assumptions, especially in the near term, and will tend to depress any projections of GHG emissions.
Based on these conservative assumptions, the high-level conclusion of this analysis is that under the likely conditions we will experience over the next 15 years there is virtually no chance that vehicular GHG emissions will fall to any significant degree in Metro Vancouver, and indeed are likely to rise unless there are major additions to transit availability and options and/or substantive changes to driving behaviour. As such, the likelihood that the Federal goal of a 30% decline in emissions by 2030 over 2005 levels (or clearly, the Provincial goal of a 33% decline by 2020) can be met must be considered suspect.
But a further conclusion is that the “plans” governments have put forward to counter GHGs appear to be more sets of assumptions posing as plans rather than concrete initiatives with explicit expectations and, perhaps as important, explicit tracking of how well these initiatives have been implemented and how well they have actually affected GHG emissions. At present, with the exception of a number of papers prepared by Pacific Analytics, there has been little or no examination of what is impacting vehicular GHG emissions and how various policy variables (e.g., fuel efficiency regulations, the Carbon Tax, Fare Card pricing and other transit attributes, local density regulations, etc.) have, and likely will, affect emission trends. Indeed, the latest “official” estimation from the Provincial Government of community vehicular GHG emissions is for the year 2010. Those estimates had no underlying analysis and, given that community emissions decreased between 2010 and 2007 (mainly because of the recession), this is the reason why many people claim that vehicular GHGs are falling when clearly they are not.
When a proper comprehensive analysis of GHGs between 2007 and 2013 was done this past spring, it showed that increasing fuel prices was the most important determinant of lower GHGs, a factor totally unrelated to any GHG “strategy”. In contrast, it also found that the increase in Bus Hours was also important to falling emissions as was the increase in Skytrain Hours, at least for the city of Richmond.
A strategy without a comprehensive method for tracking and assessing the impacts of each initiative such as above makes it difficult to evaluate whether the initiatives themselves are responsible for GHG changes or whether other factors are responsible. A case in point is the supposed “success” of the carbon tax in reducing vehicular GHGs. A proper analysis of the carbon tax found that, while vehicular GHG emissions did decline in the 2009 – 2012 period roughly coinciding with the implementation of the carbon tax, the carbon tax itself had very little impact on GHGs. Yet, because there has little effort to assess these impacts, we are still hearing how the carbon tax is the solution to reducing GHGs.
Whether GHGs can be reduced as much as projected by governments within the timeframe allotted, is almost a moot point. An equally important point is that thus far most government targets comprise a series of assumptions, but assuming an outcome is not a plan, a point that governments and government planning departments everywhere should heed. At least in the case of strategies for reducing vehicular GHG emissions, the emperor has no clothes.
 Our VKT/GHG Forecasting Model was commissioned and used initially by the BC Ministry of Transportation to examine fuel tax implications. Recently the Model has been used by the Ministry of Environment for generating Community Energy and Emissions Inventory (CEEI) data and by TransLink. A complete description of the Model along with all data sources can be found at http://pacificanalytics.ca under the AutoStat tab.
 The USEIA model is a comprehensive econometric model in which the price of crude oil directly impacts economic growth, new vehicle purchasing behavior, and the impetus (slightly) to improvements in vehicle efficiencies. Higher fuel prices result in somewhat larger uptake in more fuel efficient vehicles, especially diesel and electric, and leads to somewhat greater purchases of smaller vehicles.
 Metro Vancouver fuel price projections in our Model are based on crude prices plus inflation-adjusted refinery markups, retailer markups and various fuel taxes. Refinery markups, however, have increased from an historical average of approximately 13.4 cents per litre over the 2001 – 2010 period to the present 30 cents per litre (one of the main reasons that retail prices have not followed crude prices downward). If refinery markups were to return to historical averages (adjusted for inflation) then retail prices would be roughly 15 cents lower than projected which would lead to even higher fuel consumption and hence greater emissions.
 The USEIA model generates forecasts of fuel efficiencies in term of miles per US gallon. For our purposes, these are converted to fuel consumption rates – litres per 100 kms.
 The decline represents a composite of reductions in vehicle-specific consumption rates, changes in the types of new vehicles purchased (e.g., movement toward larger vehicles), and changes in new vehicle fuel types (e.g., movement toward a greater proportion of electric and diesel vehicles).
 See our Report “Vehicle GHG Emissions in MetroVan” (February 2015) for a detailed analysis of the changes in vehicle purchase patterns over the last decade. It can be downloaded from the website: http://pacificanalytics.ca under the tab AutoStat.
 Some readers may think that a penetration rate of just less than 3.5% for electric vehicles is low, but it should be considered that electric vehicles are still a very new technology and still demand a very large purchase price premium. Acceptance of new technologies generally is slow; take for instance hybrids. In the early 2000s, hybrids were the newest up-and-coming technology that was going to replace the standard gasoline vehicle. Yet in 2014, less than 0.3% of new passenger vehicles sold were hybrids and of the total stock, only 1% were hybrids.
 All of these factors are estimated at a detailed level: 14 vehicle classes (small cars, large cars, SUVs, pickups, minivans, medium-duty pickups, medium-duty vans, heavy-duty city, heavy-duty highway, small and large transit buses, small and large regular buses, and taxis), five fuel types (gasoline, diesel, hybrid, other, and electric), 2 license types (for pleasure only and to/for work purposes), and by model year 1990 (and earlier) to last forecast year.
 The Model actually incorporates vehicle-to-population ratios for six sub-groups: males 16-24, 25-64 and 65+ and females 16-24, 25-64 and 65+. These ratios vary substantially from each other – the male 25-64 ratio in 2014 was .921 while the female 16-24 was just .168. The forecast for each of these ratios are input independently in the Model.
 See the Report “VKT Forecasting Model Description” found under the tab AutoStat – Models in the website http://pacificanalytics.ca
 An analysis of the impacts of electric vehicles on emissions (and tax receipts – and what the impacts could be on both TransLiink revenues and monies allocated to the Ministry of Environment for highway maintenance and capital spending) using our VKT/GHG Forecasting Model is presently in progress.
 Some Community Energy and Emissions Plans (CEEPs) include such assumptions, unfortunately with little background as to how these measures will be met. Delta, for example, in its CEEP projections for 2030 assumes (see page 25 of CEEP) that there will be a 2% reduction in emissions due to decreased idling not due to traffic congestion (or a reduction of ~ 25 minutes of idling per week per vehicle). One must question how many people could actually reduce their idling that much on average per week. It should also be noted that idling due to traffic congestion is already accounted for in the projected improvements in fuel efficiencies of new vehicles through cylinder deactivation and micro-hybridization technologies.
Similarly, there is an assumption (undocumented) of a 25% improvement in fuel consumption rates for the passenger vehicle stock between 2016 and 2030 (page 16 of CEEP). In contrast, the USEIA is forecasting a 25% improvement in new passenger vehicle efficiencies, corresponding to an improvement in efficiency for the total stock of just over 15%, at least in Metro Vancouver. This difference of 10% equates to 10% higher emissions from that assumption alone.
At the same time, there are no references to the expectations for future fuel prices or income levels (other than the catch-all reference: Business As Usual – BAU). Does BAU imply that fuel prices and incomes will remain flat (is that constant or nominal dollars) or are they saying that fuel prices and incomes do not affect driving behavior???
The issue here is not that we disagree with the vehicle GHG projections that Delta has generated; it’s that the projections are based generally on a set of assumptions that in the most part are not well defined, and further, there is no testing of the implications if one or more of these assumptions are not met.
 See “Vehicle GHG Emissions in MetroVan” (February 2015) for a detailed analysis of GHG emissions in Metro Vancouver over the 2007 – 2013 period. It can be downloaded from the website: http://pacificanalytics.ca under the tab AutoStat.
 See, for example, “The Shocking Truth About BC’s Carbon Tax: It Works”; http://www.theglobeandmail.com/globe-debate/the-insidious-truth-about-bcs-carbon-tax-it-works/article19512237/
A copy of this report can be downloaded from the website: http://pacificanalytics.ca