Developing a National Vehicle Fuel Economy Database & Baseline: Kenya Case Study Climate XL March 2011
Table of Contents List of Tables...2 List of Figures...2 Acronyms and Abbreviations...3 Background...4 Purpose and Target...4 STEP 1: Set Your Objectives...6 STEP 2: Obtain Vehicle Registration Data...7 STEP 2: Data Cleaning...8 STEP 3: Restructure Your Database to conform to GFEI Methodology Guidelines...9 STEP 4: Populate National Database...11 STEP 5: Estimate National Average Fuel Economy...11 STEP 6: Report Findings on Fuel Economy and CO 2 Emissions...12 Lessons Learned...13 Partnership is Key...13 Vehicle Fuel Economy Baselines are Time-Intensive...13 A Digital Registry is Recommended...14 Endnotes...28 List of Tables Table 1: IEA Data Frame of Key Attributes (expanded list)...10 Table 2: Fuel Consumption by Year and Fuel Type...12 List of Figures Figure 1: Steps in Developing a Fuel Economy Database...5 Page 2 of 28
Acronyms and Abbreviations CO 2 GFEI GHG IEA ITF KBS Km KRA LDV NEDC UNEP Carbon dioxide Global Fuel Economy Initiative Greenhouse Gas International Energy Agency International Transport Forum Kenya Bureau of Statistics Kilometers Kenya Revenue Authority Light Duty Vehicle New European Drive Cycle United Nations Environment Programme Page 3 of 28
Background The United Nations Environment Programme (UNEP) in partnership with Climate XL Africa conducted the first-ever systematic survey and analysis of the Kenyan light duty vehicle fleet s fuel economy for baseline year 2005, and subsequent 2008. This practical approach to auto fuel economy is part of the larger work of the Global Fuel Economy Initiative (GFEI), a joint initiative of UNEP, the International Energy Agency (IEA), the International Transport Forum (ITF), and the FIA Foundation aiming for an integrated approach to improving the global car stock s efficiency by at least 50% by 2050 (i.e. the 50by50 campaign). This level of improvement is feasible using existing, cost-effective incremental fuel economy technologies. Whole 50by50 is a global target, each country s contribution will differ according to national baselines and trends. Therefore, the GFEI is working with countries around the world to determine a) what the baseline auto fuel economy is from 2005 and b) what the growth in emissions and vehicle ownership trends are. These two sets of numbers help countries to determine where they are now, where they are headed, and point to solutions for what can be done to limit greenhouse gas emissions from their vehicle stock. Purpose and Target This guide describes the process used in developing the vehicle fuel economy database and calculating the 2005 baseline for Kenya. The Kenyan automobile fuel economy database and baseline were developed: 1. To provide a basis for tracking progress in improving fuel economy, and 2. To provide evidence for policy making and development of appropriate guidelines for vehicle fuel efficiency and green house gas (GHG) standards. Specifically, the aim was to compile data on the light duty vehicle (LDV) fleet registered in Kenya since 2005 using the GFEI Baseline Methodology (see Annex I for details), and to determine the trend in fuel economy for this fleet and the corresponding carbon dioxide emissions expressed in grams per kilometer (gco 2 /km). Arising from the work done Kenya and lessons learned thereof, the purpose of this guide is disseminate the experience gained to help build the capacity of developing countries to participate in the GFEI by providing them with a tested approach to developing a fuel economy database and baseline. Page 4 of 28
Steps in Developing a Vehicle Fuel Economy Database & Baseline Figure 1: Steps in Developing a Fuel Economy Database Page 5 of 28
STEP 1: Set Your Objectives It is important to ask the question: why undertake a national fuel economy baseline calculation? Transportrelated greenhouse gas emissions account for approximately one-quarter of energy related CO2 emissions globally, and are expected to rise to one-third by 2050. The world s light duty vehicle fleet is set to at least triple by 2050, at which time two-thirds of the planet s vehicles will be found in developing countries (compared to about a quarter today). Global efforts to mitigate climate change can only succeed with improved vehicle efficiency worldwide. In addition, CO2 reductions must be paired with vehicle emission standards to reduce conventional pollutant emissions, including dangerous particulate matter. The global economy is set to quadruple while the global demand for transportation is set to more than double by 2050, the fastest increase of any economic sector. A greener transport sector is central to a low carbon economy and will help to create significant investment and employment opportunities. Cleaner, more efficient vehicle systems are a major part of the sustainability picture and a cost-effective way of improving the energy efficiency of the sector as a whole. Improved vehicle fuel efficiency: 1. is key to addressing climate change. The adoption of more efficient vehicles concurrently with cleaner fuels and stringent emission standards will significantly improve our ability to meet climate change mitigation targets. Even if vehicle kilometers driven double by 2050, vehicle fuel efficiency improvements and emission standards would cap CO2 emissions from cars at current or lower levels. 1 2. improves urban air quality by reducing conventional emissions, including particulate matter and black carbon. According to research, improving vehicle efficiency is one of the most cost effective interventions to reduce transport-related emissions (i.e. NO2, PM and black carbon). 2 Cleaner, more efficient vehicle technologies can also significantly reduce environmental damage and lost economic potential. 3. leads to foreign exchange savings and lower oil import bills. Financial savings can be achieved through reducing the energy demand and petroleum imports of a country. Improvements in auto fuel economy can result in estimated savings in annual oil import bills worth over USD 300 billion in 2025 and 600 billion in 2050. 3 4. increases fuel cost savings for consumers. For many individuals, most or all of the cost of improved technology in more fuel efficient cars could be off set by the fuel saved in the first few years of car usage, especially at higher oil prices. 1 50by50 Report. The Global Fuel Economy Initiative, 2009. 2 Pathways to a Low Carbon Economy. McKinsey, 2009. 3 Based on an oil price of USD 100/bbl. Page 6 of 28
5. is one of the best ways to reduce economic vulnerability to oil shocks and improve energy security. Many developing countries are net fuel importers and therefore vulnerable to volatile oil prices in international markets. Policies that reduce the size of the net oil import bill relative to GDP e.g. vehicle efficiency policies, fuel switch to renewables - will lessen the impact of future price shocks. The reasons underlying a decision to address automotive fuel economy and greenhouse gas emissions are particular to each country and government. In the case of Kenya, significant gains in fuel quality in the past few years (including lead-free petrol and plans for lower sulphur fuels) have meant that addressing vehicle emission standards and the efficiency of the auto stock in general are a natural progression in reducing overall road transport emissions. The objectives of calculating a fuel economy i baseline are (1) to provide a basis for tracking progress in improving fuel economy through future national policies and standards, and (2) to ensure that Kenyan policymakers have improved information to inform their choice of policy instruments for reducing vehicle emissions. In addition, the Kenyan experience will inform and refine the GFEI methodology, and also serve as a working model for similar data gathering and analysis work in Africa and other regions in which the GFEI is active (see globalfueleconomy.org). The level of detail and quality of information about vehicle markets differs from country to country. But in general, vehicle registration data of most developing countries is fragmented and incomplete. Therefore, the first task in developing a fuel economy database and calculating a baseline average is to build a comprehensive picture of the country s vehicle stock and yearly registrations. For reasons discussed below, this can be a time-consuming process. STEP 2: Obtain Vehicle Registration Data Records of car imports and/or registration are the main sources of data for establishing an automobile fuel economy baseline ii. In Kenya, the Registrar of Motor Vehicles a Department of the Kenya Revenue Authority (KRA) is the official public repository of vehicle registration data. Since 2005, when it digitized its registry, the KRA maintains a digital and searchable database of vehicle registered in the country. 4 The GFEI methodology uses 2005 as the baseline year, for ease of comparison of data from numerous countries. Once this first 2005 baseline year is established, it is recommend that the same calculations be done for 2008 and then, ideally, for every 2 years after that in order to establish a trend and to have a solid base from which to monitor the impact of auto fuel economy policy. In order to facilitate access to records in Kenya, UNEP wrote a Letter of Introduction describing the nature of the study and its importance to the Government of Kenya and the GFEI, which Climate XL used to approach various institutions it had identified as potential sources of data. It should be noted 4 In countries without digital databases, vehicle records would be extracted from manual filing systems, adding to the staff hours allocated for this work. Page 7 of 28
that agencies collect and maintain records to suit their own needs. iii Such data may or may not be compatible with the type of database sought, in which case it would need modification (see STEP 3: Restructure Your Database to conform to GFEI Methodology Guidelines). For example, vehicle registration by KRA includes all categories of vehicles motor cycles and mopeds, light duty vehicles (LDVs), large buses with gross weight in excess of 3500 kg, and heavy trucks. But for the Kenya baseline, only information on LDVs (i.e., group of vehicles with a gross weight of less than 3500 kg) was required. This meant that all other information pertaining to, for example, motor cycles, trucks and tractors was irrelevant and therefore had to be discarded. Also, KRA data is not classified in terms of the 'year of registration' and therefore it was difficult to a priori categorize vehicles on the basis of their registration date. Given that the recommended baseline year for the GFEI is 2005 (this is in order to ensure harmonized global comparisons across countries), it was crucial to establish the year of registration of all vehicle entries. Other notable constraints included absence of engine technology data. KRA only considers the year of manufacture. Information on transmission type and mileage readings of vehicles is also not recorded at registration. To overcome some of these shortcomings it was necessary to define the type and format of vehicle data we wanted. A data capture form (Annex II Table 5: Data Capture Form) was designed and sent to the Commissioner, Registrar of Motor Vehicles with a request for those data sets. However, to build a comprehensive picture of a country s vehicle stock a number of state institutions may need to be consulted. In Kenya s case this included the Kenya Bureau of Statistics (KBS), which is the custodian of all official government data and the Transport Departments of government ministries and parastatals, which keep records of their vehicle fleets and their operational costs including the mileage readings and fuel consumption. It is important to involve and inform the agencies and focal points that you are dealing with throughout the process of collecting, cleaning and analyzing data. This ensures buy-in and endorsement of results, in addition to more efficient access to data in later years. STEP 2: Data Cleaning 'Cleaning' data involves removing errors in acquired vehicle data. In the Kenya case cleaning involved: a) Removal from the data set of vehicles not classified as light duty; b) Separation of new and used vehicles at time of registration; c) Correction of data entry errors e.g. spelling mistakes; and d) Addition of other relevant fields e.g. vehicle horsepower, transmission type, axle configuration etc., to make it as comparable with the GFEI database as possible (See Annex I for required/minimum information points for a baseline and STEP 3: Restructure Your Database to conform to GFEI Methodology Guidelines. Page 8 of 28
The cleaning process included sorting out the raw data obtained from the Registrar of Motor Vehicles to fit the objectives of the exercise (including the agreed-upon data points for a GFEI baseline) and to ensure that we only carried out analysis on relevant entries. Different countries break down vehicle classes in different ways (see Annex I for guidance on vehicle categories), but typically fuel economy standards cover passenger vehicles below a certain weight (e.g., 3500 kg for LDVs); separate auto fuel economy standards are sometimes established for commercial vehicles, freight vehicles, etc. in time, so including this class of vehicles in the initial baseline may save you time down the line. However, due to resource constraints in the Kenya case all heavy duty vehicles were discarded from the database obtained from KRA but it should be noted that other countries may decide to include heavy duty vehicles in their baseline. The classification of vehicles and the scope of the baseline depend on the needs of each country undertaking a baseline exercise. Resolving the issue of new vs. used cars was also an important factor in developing the Kenyan automobile fuel economy database. Second-hand or used vehicle imports dominate the Kenyan vehicle market and there is no clear separation between new vehicle registrations from re-registrations and new vehicles from 2 nd hand vehicles being registered for the first time. This distinction is critical to establishing a globally comparable fuel economy database as well as for tracking changes over time. If vehicles are re-registered each year and are included in official databases, these should be separated out from the first time registrations. In addition, for the first time registrations, those vehicles that are 2 nd hand should be separated from new vehicles. For example, vehicles imported into the country at 3 or 5 years of age will be registered for the first time within Kenya as 'new registrations'. While these are clearly not new vehicles, in the Kenyan case they were nonetheless included in the official database of new vehicles on the grounds that a 3-5 year vehicle imported from Japan (the largest source market of used cars in Kenya) is perhaps 'as good as new'. It is therefore important to establish a cut-off point beyond which admissibility into the auto fuel economy database based on age (or mileage) is disallowed. The process of developing database should be a back-and-forth with the relevant government officials, particularly where additional fields of information would be useful in subsequent years. The baseline process may suggest additional data points for collection by the relevant agencies, making calculations for later years less time consuming. STEP 3: Restructure Your Database to conform to GFEI Methodology Guidelines Developing a fuel economy database is a collaborative exercise. In developing the Kenya automobile fuel economy database Climate XL actively collaborated with the International Energy Agency and UNEP for technical support and information on how similar exercises have been carried out in other countries (e.g. Western Europe and the U.S.). This collaboration led to a number of changes that were made on the KRA data to make it as comparable to the GFEI global database of baselines being developed by the IEA as possible. For example, terms used in the IEA database were adopted in the Kenyan database to harmonise the category body types. Thus, 'lift back' was substituted for 'station wagon' and 'estate'/'sedan' for 'saloon', etc. Page 9 of 28
The IEA s extensive dataset (Table 1: IEA Data Frame of Key Attributes) on auto fuel economy has 24 attributes that would ideally be collected if national information is detailed enough. However, realistically, national data sets are often must less exhaustive, and the KRA data only included 7 attributes for cars registered in Kenya. Climate XL expanded the KRA data by conducting research using additional sources including car dealers, manufacturers and internet searches related to specific car makes and models. With these data limitations in mind, the GFEI has identified (see Annex II) the essential/minimal attributes that are necessary for a robust and internationally-comparative auto fuel economy baseline database and calculation. Regardless of the number of categories you choose to include in your baseline database, in addition to LDV s, the absolute minimum information required for each vehicle is: 1. Vehicle make and model, and if possible configuration (this typically is labelled by the manufacturer using a sub-model number or other designation; it can indicate transmission type, trim level, optional accessories, etc.) 2. Model production year 3. Year of first registration, if different from model year 4. Fuel type 5. Engine size 6. Domestically produced or imported 7. New or second hand import 8. Rated Fuel Economy per model and test cycle basis. This can be done either by getting data from country of origin or manufacturer, or by testing of a select sample of vehicles. 9. Number of sales by model Table 1: IEA Data Frame of Key Attributes (expanded list) 1. Vehicle Type 2. Model 3. Manufacturer 4. Body type 5. Simplified Body Type 6. Segment 7. Axle configuration 8. Driven wheels 9. Engine cylinders 10. Engine ccm 11. CC Category 12. Engine kw 13. KW class 14. Engine horse power 15. Engine valves 16. Fuel type 17. Model year 18. Number of gears 19. Transmission type 20. Turbo 21. Gross vehicle weight 22. Height Page 10 of 28
23. Length 24. Number of seats STEP 4: Populate National Database Once the database has been compiled, it should contain records (likely several hundred) that show all of the different vehicle makes/models/configurations that have been registered/sold in the country for a given registration year (e.g. 2005, 2008, etc.). In most cases the registration data will lack key attributes, most significantly fuel economy. This and other attributes 5 can then be added to the database by checking manufacturer information on specific models (usually available at automobile manufacturers Websites as well as in car guides). 6 If the tested fuel economy number for the vehicle is not included in the registration data, they can be obtained elsewhere and mapped into the database. The fuel economy figures for a given make, model and year can usually be retrieved from the vehicle manufacturers (for example, via websites and direct requests), and from various international industry associations. The GFEI is compiling a list of fuel economy data into a common database for use by countries undertaking baseline-setting exercises. Please contact clean.transport[at]unep.org for more information. See Annex II, Table 3 for an example of the Kenyan data capture spreadsheet. STEP 5: Estimate National Average Fuel Economy While vehicle fuel economy may be reported according to various test drive cycles, we recommend that for the sake of comparison, all drive cycle data obtained be converted to the New European Drive Cycle cycle. Conversion factors are conveniently available in a work sheet from the International Council on Clean Transportation: downloadable from www.theicct.org/info/data/globalstdreview_conversionfactor.xlsx. Where information was not available, the corresponding CO 2 emission values were computed from the global standard review- conversion factor (see Annex II, Table 6: General Procedures - Conversions. 5 These include, for example axle configuration (i.e., number of driven wheels); driven wheels (which wheels are driven); engine cylinders (number of cylinders); engine displacement (engine size); CC capacity (engine size category for aggregation purposes); engine kw and engine horse power (engine power); engine valves (number of valves); engine type; number of gears; transmission type; turbo; trim level; wheelbase; emission; etc. 6 Examples of useful Websites include: http://www.carfolio.com/specifications/models; www.edmunds.com/toyota; http://www.carfolio.com/specifications/models/?man=4131 http://www.epa.gov/fueleconomy/gas-label-1.htm; and http://www.carfolio.com/ Page 11 of 28
Given the fact that data on all vehicle makes and models was not available to Climate XL, particularly for older vehicles, data on the next closest model was used on the assumption that there is marginal variance between one generation model and the next. For the purposes of the Kenyan baseline, used vehicles imported into the country were considered theoretically "new" in terms of CO2 ratings and the fuel economy for the vehicle taken to be that based on test cycle provided by the manufacturer. This was a reasonable alternative given the difficulty in determining the vehicle depreciation in fuel consumption as well as the relatively good condition that most vehicles imported into the country tend to be in (i.e., they are usually not more than 8 years old). STEP 6: Report Findings on Fuel Economy and CO 2 Emissions Findings from the Kenyan baseline indicate that the average fuel consumption for vehicles in Kenya in 2005 was 7.69 L/100km with a corresponding CO 2 emission of 184.23 gco 2 /Km while in 2008 fuel consumption dropped to 7.6 L/100km, with a CO 2 emission of 184.65 gco 2 /Km. Diesel engine vehicles were found to consume more fuel per 100 kilometers travelled as compared to petrol powered vehicles. New vehicles were also found to have higher consumption as compared to second hand imported vehicles as shown in Table 2 below: Table 2: Fuel Consumption by Year and Fuel Type 2005 2008 Average 7.69 7.6 (l/100km) Diesel 8.67 9.09 Petrol 7.52 7.2 Source: GFEI Kenya Baseline, 2011 Table 3: Percentage of new vehicles registered per year Year Total reg. New vehicles 2005 13,577 3085 (23%) 2006 35,749 4333 (12%) 2007 53,689 6114 (11%) 2008 46,259 11226 (24%) 2009 43,649 8554 (20%) Source: GFEI Kenya Baseline, 2011 Page 12 of 28
Table 4: Vehicle registration by fuel type The Kenyan findings were shared and discussed at a national meeting, and included the major government agencies that had contributed data to the baseline. The summary of that meeting, and implications for similar exercises in other countries, are documented at http://www.unep.org/transport/pcfv/pdf/gfeiafricasummary_30%20november2010.pdf. Lessons Learned Partnership is Key Partnerships are important in the short term when developing vehicle fuel economy databases as well as in the longer term, for tracking progress in improving fuel economy over time. Partnerships can be between national institutions as well as with international agencies. Countries embarking on development of automobile fuel efficiency baselines are likely to benefit from pursuing partnerships between public authorities such as national and or regional vehicle registration bureaus, auto clubs or associations (e.g., Automobile Association of Tanzania, Automobile Association of Uganda, Automobile Association of Zimbabwe, Automobile Association of Namibia, Rwanda Automobile Club, etc), environmental agencies; statistical bureaus, etc., to give the process the official seal necessary to get quick buy-in and speed up the process of data gathering. Vehicle Fuel Economy Baselines are Time-Intensive Developing an automobile fuel economy database is a demanding and time consuming exercise, especially in developing countries due largely to unavailability of specific data for vehicle makes and models in the country and related attributes. The scarcity of information, dispersed nature of available sources and the size of the database require careful planning and consideration when allocating resources and staff time. Page 13 of 28
A Digital Registry is Recommended For countries that have electronic vehicle registration systems, or digital registry, the process of retrieval of vehicle data is much faster and more efficient. However, digital databases are not error free and diligence should be exercised in handling such data as multiple errors are likely to be introduced during the process of capturing data into the fuel economy database. As Kenya s case shows, official databases may also be incomplete for purposes of establishing automobile efficiency baselines and may more often than not need supplemental data from multiple sources. Page 14 of 28
Annexes Page 15 of 28
Annex I Getting to know your vehicle stock: a step-by-step guide to baseline setting 1. Introduction 2. Data needs for vehicle database development 3. Sources of information 4. Estimating baseline fuel economy, fuel consumption and CO 2 emissions 5. Country examples 6. Resources 1. Introduction In order to track trends in stock-wide light-duty vehicle (LDV) emissions intensity and progress in improving fuel economy toward national and international targets (i.e. the 50:50 targets), policymakers must establish a starting point from which to measure and monitor emissions. Some countries worldwide have already undergone this important inventory, and the Global Fuel Economy Initiative (GFEI) has distilled the basic steps and considerations for your ease of use in this guide. A good understanding of the starting point, or baseline, will allow policymakers to choose the right combination of technology and policy instruments needed to meet national emission, energy security, and efficiency goals. In this guide we deal with fuel economy from newly registered LDV's. With the fuel economy information, it is also possible to estimate average CO 2 intensity. The GFEI methodology uses 2005 as the baseline year, for ease of comparison of data from numerous countries. Once this first 2005 baseline year is established, we recommend that the same calculations be done for 2008 and then, ideally, for every 2 years after that in order to establish a trend and to have a solid base from which to monitor the impact of auto fuel economy policy. The base and subsequent year measurements are taken from vehicles entering a country s vehicle stock for the first time, including new vehicles manufactured in the country, new vehicles imported and second hand vehicles imported into the country in other words, all vehicles newly registered in that year. However, it is useful to keep separate track of these three categories of vehicles, as well as creating a combined average set of information. In summary, the baseline-setting exercise consists of the following steps: Page 16 of 28
1. Establish the baseline year (e.g. the GFEI uses 2005) 2. Establish the data points you will need to collect in order to calculate a robust baseline 3. Find and evaluate available new LDV vehicle registration data sources and their quality 4. Calculate your baseline year average fuel economy and other characteristics for newly registered vehicles 5. Repeat the same exercise using uniform methodology at regular intervals. In section 5 below, we provide practical examples from countries that have undergone baseline-setting exercises 2. Data needs for vehicle database development The baseline, as mentioned above, should only consider vehicles that are new to the country s vehicle stock for that year, i.e. newly manufactured or newly imported (including second hand imports) and thus newly registered in that year; a car that is already in-country, but is re-registered because it is re-sold should not be counted. (It may also be useful to track total vehicle stock characteristics, but this would take a different approach, such as street surveys, and is not covered in this document.) Before you start gathering information for calculating the fuel intensity for the new vehicle registrations for a given baseline year, there are certain key data items that are required to fully develop a clearer picture of the vehicles in operation. These basic data items (or characteristics) should form the backbone of a database of the newly registered vehicle characteristics for that year (e.g. 2005); from this database, the country s average fuel economy for that year will be calculated. While in this guide we are primarily focused on LDV s, some countries may be interested in including additional categories of vehicles. Vehicle segmentation (or categorization of types by weight, interior volume, cc's) has no clear definition or assigned values, and differs from one region to another, although it is often based on body shape and interior volume. Vehicle fuel economy standards are typically differentiated based on vehicle category. While there is no agreed international classification scheme in place, we recommend using the following categories for guidance: Vehicle Segment A: Mini / Micro / Small town car Smallest cars, with a length between 2.50m to 3.60m Examples Citroën C1 Fiat Panda Smart Fortwo Page 17 of 28
B: Small compact Slightly more powerful than the Minis; still primarily for urban use; length between 3.60m and 4.05m Mitsubishi Colt Opel Corsa Suzuki Swift C: Compact Length between 4.05m 4.50m Mazda 3 Subaru Impreza Volvo S40 D: Family cars Designed for longer distance; fits 5-6 people; length is 4.50m to 4.80m BMW 3 series Chrysler Sebring Lexus IS Light vans Size is similar to D, but interior volume is maximised to accommodate larger families Big / Full size cars Have generous leg room; can comfortably transport 5-6 people; generally have V8 engines and are 5m or longer in length Chevrolet Uplander Ford Galaxy Volkswagen Sharan Cadillac DTS Jaguar XJ Mercedes-Benz E Class SUV / All terrain The original cars were utility cross-country vehicles with integral transmissions like the Jeep Dodge Durango Jeep Grand Cherokee Nissan X-Trail Toyota Land Cruiser Regardless of the number of categories you choose to include in your baseline database in addition to LDV s, the absolute minimum information required for each vehicle includes: Vehicle make and model, and if possible configuration (this typically is labelled by the manufacturer using a sub-model number or other designation; it can indicate transmission type, trim level, optional accessories, etc.) Page 18 of 28
Model production year Year of first registration, if different from model year Fuel type Engine size Domestically produced or imported New or second hand import Rated Fuel Economy per model and test cycle basis. This can be done either by getting data from country of origin or manufacturer, or by testing of a select sample of vehicles. Number of sales by model Additional information that would be useful for advanced analysis and should be collected, if possible, includes: Vehicle Information / Identification Number Fuel type (petrol or diesel) Injection system type Body type Transmission type and other vehicle configuration details, as available Vehicle foot print Vehicle curb weight Emissions certification level Use of vehicle (private, public, for hire, etc.) Vehicle price Second hand vehicles imported into the country will be registered for the first time within the country and thus are new registrations. Although they are clearly not new vehicles, they are new to that market and thus should still be counted, particularly as they constitute a significant proportion of new registrations in some countries; for example, around 80% of new vehicles in Kenya are second hand ex- Japan imports. These should be clearly defined as second hand, so that analysis can be done on all new registrations as well as on new versus second hand vehicles. The database should contain the initial registration of each vehicle, with the date of initial registration and the model year of the vehicle. If vehicles are re-registered each year or after a re-sale, and these are included in the government database, these re-registrations should be taken out of a baseline database so as to avoid double-counting. The image below shows an example of part of a database that 1) differentiates between vehicle types by make and model and 2) gathers information for those vehicles along the basic vehicle characteristics needed to form a picture or baseline of the average fuel economy for any given year. Page 19 of 28
Make Model Condition Body Type Engine CC Fuel Type Model Year Registration Date L/100km CO2 BMW 316I Used S.WAGON 1596 Petrol 1989 2005 7.5 176 CHEVROLET OPTRA Used SALOON 1799 Petrol 2005 2005 6.2 145 CHEVROLET NULL Used S.WAGON 1799 Petrol 2005 2005 6.2 145 NISSAN SUNNY Not Specified SALOON 1970 Diesel 1998 2005 6.6 177 MITSUBISHI LANCER Used SALOON 1600 Diesel 1998 2005 6.9 185 SKODA OCTAVIA Used SALOON 1800 Diesel 2004 2005 7.0 188 SKODA OCTAVIA Used SALOON 1800 Diesel 2005 2005 7.0 188 TOYOTA COROLLA New S.WAGON 1970 Diesel 1998 2005 7.0 188 TOYOTA COROLLA New SALOON 2000 Diesel 1998 2005 7.0 188 FORD RANGER New VAN 2500 Petrol 2005 2005 8.1 170 HONDA CR-V NULL S.WAGON 1970 Petrol 1998 2005 9.3 217 3. Sources of Information The majority of OECD countries and a few other countries currently track all vehicle sales and vehicle characteristics (via registration records) and have well-established systems for measuring average new LDV fuel economy on a yearly basis. All countries should be able to do this, but data collection and analysis systems must first be established. When vehicle registration data is available: The best approach for gathering baseline fuel economy data and developing estimates involves obtaining official records of new vehicle sales or registrations. Registration data is usually collected by national governments as part of the system of regulation and taxation of vehicle ownership within the country. For example, in Kenya, this is the Kenya Revenue Authority, and in Chile, the Ministry of Transport keeps records of new registrations. Many countries store registration records in digital databases or are starting to develop them from paper-based records. If data is available in paper form, it will have to be inputted by hand into a specialized auto fuel economy spreadsheet for use in analysis. This time-consuming exercise should be Page 20 of 28
taken into consideration when planning for the required resources to set up a database and collect information. When data is not available: It may be the case that vehicle registration data is not available (or not available to groups interested in establishing the baseline estimate). If it is determined that a full database of vehicle registrations is either not available or not easily utilized, there are other approaches to obtaining at least an indicative estimate of baseline fuel economy levels, e.g. via a representative sample of new vehicle sales. Statistical methods may be applied to determine a representative sample. One option in this case is to work with importers and manufacturers to obtain their data on vehicles imported and / or sold within the country. Such reporting may already be a requirement in many countries. If such an approach is taken, it is important to cover enough importers and manufacturers to be confident that a representative sample of different vehicle types has been collected. While including all manufacturers is clearly best, at least all major manufacturers / sellers should be included in the sample and those not covered should be checked to see if their exclusion is likely to skew the sample in some manner (e.g. if those specializing in particularly large or sporty vehicles are excluded, etc.). It may also be possible to collect vehicle information via a vehicle count sampling system essentially, a visual inspection of vehicles in use around the country, for example in parking areas. However, this approach requires careful efforts to obtain a representative sample, along with a detailed-enough inspection of each vehicle to obtain key attributes of the vehicle and to ascertain that the vehicle is new to the vehicle stock (which can be very time-consuming). Inspection of the VIN (vehicle identification number) can help in this regard since this number usually indicates the year of vehicle manufacture and key attributes of the vehicle. The registration number or licence plate also often indicates when the vehicle was first registered in the country, indicating if it is new to the stock. Additional guidance from the GFEI on this approach is forthcoming. If you would like additional advice on this survey method, please contact clean.transport[at]unep.org. 4. Estimating baseline average fuel economy, fuel consumption and CO 2 intensity Once the database has been compiled with new vehicles registered for the base year, it should contain records (probably several hundreds or thousands) that show all the different vehicle makes / models / configurations that have been sold in the country for a given model year, the number of vehicles sold and their fuel economies. Page 21 of 28
If the tested fuel economy number for the vehicle is not included in the registration data, they can be obtained elsewhere and mapped into the database. The fuel economy figures for a given make, model and year can usually be retrieved from the vehicle manufacturers (for example, via websites), and from various international organizations. The International Energy Agency, UNEP and partner organizations are compiling a list of fuel economies into a common database for use by countries undertaking baselinesetting exercise. Please contact clean.transport[at]unep.org for more information. While vehicle fuel economy and may be reported according to various test drive cycles, we recommend that for the sake of comparison, all drive cycle data obtained be converted to the NEDC cycle. Conversion factors are conveniently available in a work sheet from the International Council on Clean Transportation: downloadable from www.theicct.org/info/data/globalstdreview_conversionfactor.xlsx. Once your database is as complete as possible, and checked for errors, it can be used to develop baseline estimates. At the simplest level, taking a weighted average (by sales) of all new (including newly imported second hand) vehicles in the database will provide the average fuel economy of new vehicles sold in the country in the given year: In a similar way, average CO 2 intensity can be obtained through weighted average with the sales of each model: Page 22 of 28
If you have decided to include multiple categories of vehicles in your baseline database (as mentioned in Section 2 above, e.g. LDV, heavy duty etc), this will allow the generation of distributions of sales by vehicle class and the average fuel economy in each class. This can be very helpful in tracking how sales and fuel economy changes in the future, and for comparing to other countries. Finally, fuel economy can be compared to attributes such as vehicle size, weight, or engine power, to see how these attributes relate. This can help in establishing different vehicle standards by weight or size class. For a global comparison of your country s figures for base year 2005 and onwards, contact the GFEI for support on your baseline-setting exercise and for guidance on methodology issues at clean.transport[at]unep.org. 5. Resources The following sources of information for fuel economy baselines, conversion tools and downloads can assist in the development of your baseline database: A Test Cycle Conversion Tool: www.theicct.org/info/data/globalstdreview_conversionfactor.xlsx A global comparison of Vehicle Fuel Economy Standards: http://www.theicct.org/passengervehicles/global-pv-standards-update/ South African Comparative Passenger Car Fuel Economy AND CO2 Emissions Data: http://www.naamsa.co.za/ecelabels/ UNEP Vehicle Fuel Efficiency Baselines: Practicalities and Results Global Fuel Economy Initiative in Africa, Working Session, November 2010. Summary and Country Case Study Presentations: www.unep.org/transport/pcfv/pdf/gfeiafricasummary_30%20november2010.pdf U.S. Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends (1975 through 2010): http://www.epa.gov/oms/fetrends.htm Page 23 of 28
U.S. Fuel Economy Policy: http://www.fueleconomy.gov/ U.S. Fuel Economy Regulations: http://www.epa.gov/oms/climate/regulations.htm U.S. Auto Fuel Economy Database: http://www.fueleconomy.gov/feg/findacar.htm Page 24 of 28
Annex II Table 5: Data Capture Form Vehicle category Cars New/newl y registered Type & registration Used/newl y registered Used/not newly registere d Identificatio n number Mak e Mod el Mod el year petr ol Fuel type dies el Engin e type Engin e size (cc) Transmissio n type Bod y type privat e Use of vehicle For hir e PS V Mileage (odomet er reading) Emissions certificatio n level Rated fuel efficienc y 7 Minivans Pickup trucks Delivery vans Commerci al vehicle 7 Grams per mile (g/mi) CO2 emissions and miles per gallon (mpg) Page 25 of 28
Table 6: General Procedures - Conversions Step 1 Calculate test cycle conversion factor (CF), using Table A Step 2 Convert from original test cycle to target test cycle with the same metric (e.g., convert from CAFÉ mpg to NEDC mpg), using Table A Step 3 Convert from original metric to target metric using Table B. If the conversion is between a fuel economy metric (such as mpg or km/l) and a CO2 metric, use fuel CO2 content factors i Table A: Test Cycle Conversion Source (X) Target (Y) A B Calculate Conversion Factor Calculate Target CAFÉ mpg NEDC mpg -0.1033 1.473 CF = A*ln(X)+B Y=X/CF NEDC mpg CAFÉ mpg 0.0816 0.6243 CF = A*ln(X)+B Y=X/CF JC08 km/l NEDC km/l -0.0841 1.3464 CF = A*ln(X)+B Y=X*CF JC08 km/l CAFÉ km/l -0.2038 1.7618 CF = A*ln(X)+B Y=X*CF Table B: Unit Conversion Original Target gallon liter 3.785 mile km 1.609 lb gram 453.592 Km Miles 0.621 km/l mpg 2.35 Table C: CO2 g/km calculator (Apply after test cycle conversion) Country Ori. Unit Target Unit Petrol Diesel LPG CNG Calculate target US mpg g/km 5497 6315 3806 3862 dividend Japan km/l g/km 2337 2684 1618 1642 dividend China L/100-km g/km 23.4 26.8 16.2 16.4 multiplier CO2 content factor of fuels Page 26 of 28
Petrol Diesel LPG CNG lb/gal 19.5 22.4 13.5 13.7 gram/l 2337 2684 1618 1642 Examples 1. Convert 25 CAFÉ mpg to NEDC g/km Step 1 CF=-0.1033*ln(25)+1.473=1.14 Step 2 NEDC mpg = CAFÉ mpg/cf = 25/1.14=21.9 mpg Step 3 NEDC gco2/km = (CO2 content factor)/nedc mpg For petrol vehicles NEDC gco2/km=$d$23/mpg=5497/21.9=251 gco2/km For diesel vehicle NEDC gco2/km = $E$23/mpg=6315/21.9=288 gco2/km 2. Convert 250 NEDC gco2/km to NEDC l/100-km Step 1 No test cycle conversion is needed Step 2 No test cycle conversion is needed Step 3 NEDC l/100-km = (NEDC gco2/km)/$b$19*100=250/2338.6*100=10.7 l/100-km Page 27 of 28
Endnotes i Automobile fuel economy standards can take many forms, including numeric standards based on vehicle fuel consumption (such as liters of gasoline per hundred kilometers of travel [L/100- km]) or fuel economy (such as kilometers per liter [km/l]) or as miles per gallon [mpg]). Automobile GHG emission standards (expressed as grams per kilometer or grams per mile), even though they are not designed to directly control oil consumption, also affect vehicle fuel consumption. ii Registration data is usually collected by national governments as part of the system of regulation and taxation of vehicle ownership within countries. Official new vehicle registration data typically includes detail about each vehicle, such as its specific make, model, and configuration (e.g., fuel type, transmission type, engine displacement, etc.). iii For example, data available at KBS was found to be too general (though very easily accessible from the agency s Website) as it only contained the general (aggregated) statistics compiled by the Bureau as sent to them by the KRA Registrar of Motor Vehicles Office. This was data on the total numbers of vehicles that are registered per month and/or per year and only broken down to categories such as car, vans, trucks, etc. Page 28 of 28