Socioeconomic Downscaled Projections

Country-level GDP Downscaled Projections

Along with population growth, economic growth rates are a second, exogenous, assumption incorporated within the four IPCC SRES scenario families. As explained in the SRES report (see especially sections 4.2.2 and 4.3) economic growth rates were assumed to be "very high" for the A1 family, "medium" for the A2 family, "high" for the B1 family and "medium" for the B2 family. Quantitatively these assumptions translated into World GDP for 2100 of approximately 525-550 trillion US1990$ (market exchange rates)/year for the A1 family, 243 trillion US1990$/year for the A2 family, 328 trillion US1990$/year for the B1 family and 235 trillion US1990$/year for the B2 family. The corresponding per capita GDP growth rates depend on the corresponding regional population data used in the SRES report.

At the present time, we have downscaled the SRES GDP projections for individual countries out to 2100, using the same regional growth rate method applied to the population data. In other words, SRES regional GDP growth rates were calculated from the marker model regional data and applied uniformly to each country that fell within the SRES-defined regions.

A key difference between the application of this procedure to GDP and population is that uniform GDP growth rates were applied starting in the base year of 1990. With population, uniform growth rates were applied only after 2050. (Prior to that UN 1996 Revision population data were available to simulate growth rates changes over time.) Therefore our GDP downscaling introduces inaccurate national GDP growth rates in the near-term, when compared to actual near-term data for countries, because national GDP growth rates are obviously not uniform within regions.

Base Year Issues:

The 1990 base year GDP data were downloaded from a national accounts database available from the UN Statistics Division. The data were accessed most recently in May 2002 at: http://unstats.un.org

From this database we originally selected the series titled: “GDP at market prices, US$, current prices (for 1990) (UN estimates).” The UN definitions for market and current prices are given in the notes below. Also given is the source for the “UN estimates.” However, for reasons that are not fully clear, the GDP data from the UN for Eastern Europe and the former Soviet Union (which together comprise the REF SRES region) are significantly too high compared to the SRES REF estimate.

To assess this discrepancy further we downloaded from the same UNSTATS database a second GDP series list entitled: “GDP at market prices, current US$ (for 1990) (World Bank estimates)”. This data derives from the World Bank’s Development Indicator Reports. When summing this data, we find it shows a much closer agreement with SRES for the REF countries. However, the WB country list is shorter than the UN’s list. As an interim solution, in the interests of developing as global a database as possible, we have decided to use the World Bank estimates for as many countries as they provide, and especially for the REF countries. For missing countries in other regions we use the UN estimates.

One initial finding of the downscaling exercise is that the regional growth rate methodology is unacceptable for some countries with high initial incomes that also happen to lie within very high SRES GDP growth rate regions. This occurred especially for the following countries: (1) Singapore, (2) Hong Kong, (3) French Polynesia, (4) New Caledonia, (5) Brunei Darussalam, (6) Renuion, (7) Republic of Korea, (8) Gabon, (9) Mauritius. Unacceptably high per capita incomes in 2100 occur. We would prefer to not list such GDP data for these countries however excluding them from the spreadsheets introduces large regional discrepancies with SRES. (Other countries might have to be excluded for similar, if not as extreme, reasons.) As an interim measure, we have retained these anomalously high-income countries in our spreadsheets but they are color-coded blue to indicate to users that they should be treated as artifacts. This will have to be resolved with further work.

GDP Comparisons:

A1B: Spreadsheet 1 presents results for the SRES A1B scenario. We use the regional economic growth rates from the Asian Pacific Integrated Model (AIM). This model was the 'marker model' (see SRES report for the definition of marker models) for the A1 scenario in general. The 8 regions in this model are: United States, Western OECD, Pacific OECD, Eastern Europe and the former Soviet Union, Centrally planned Asia, South & South East Asia, Middle East, Africa, Latin America.

As can be seen from the spreadsheet, in the base year 1990, the country level data we have downloaded from the UN and World Bank sources shows some regional differences from the estimates used by the AIM modeling team in the SRES report. The discrepancy is greatest for the REF region at 7.24%. (Note the linearity of the downscaling procedure for this region, which happens to be a single model region in the AIM model. As explained in the previous notes, the 7.24% base year difference is exactly maintained throughout the projection period.) The other regions show smaller differences. Unlike REF, these differences are not constant over the projection period because these regions comprise more than one AIM model region and the changing weights of the model regions affect the overall SRES reporting region differences. Generally, the agreement shown is characteristic of the data available at this time and we deem it acceptable for this initial generation of the database.

For further details on the regional projections from this model contact: Tsuneyuki Morita and Kejun Jiang, National Institute for Environmental Studies (NIES), Tsukuba, Japan

A2: Spreadsheet 2 presents our country-level GDP downscaled projections for the A2 scenario using regional economic growth rates from the Atmospheric Stabilization Framework (ASF) model from ICF Consulting in the USA. This model was the 'marker' model for the A2 scenario. The 9 regions in this model are: Africa, Centrally Planned Asia, Eastern Europe and newly independent states, Middle East, OECD-East, OECD-West, South East Asia and Oceania, USA. For the list of specific countries included in each region see page 339 of the SRES report.

As seen in the spreadsheet, there are significant differences between the summed base year GDP values for the REF, OECD90 and ALM regions from the A2 marker model as compared to the country data available currently from the UN and World Bank sources. Most of these discrepancies, however, can be explained by the fact that the A2 marker model had significantly different regional estimates for 1990 GDP for REF, OECD90 and ALM, when compared to other marker models in the SRES report. For example, the ASF model estimate for A2’s 1990 REF GDP is ~13% lower that the B2 1990 REF GDP used in the B2 marker run. Similarly, the ASF model estimate for A2’s 1990 OECD GDP is ~6% lower than the B2 1990 OECD GDP from the B2 marker. Finally, the ASF model estimate for A2 1990 ALM GDP is ~26% higher than the B2 1990 ALM GDP from the MESSAGE marker. These differences, combined with the original smaller differences between our summed country list and the MESSAGE (B2) marker sums, explain the overall discrepancies seen for A2, and the linear downscaling procedure simply preserves these differences over the projection period. More importantly, these differences imply that a single base year country GDP list cannot be made consistent with all the SRES marker models.

The only remedy for our database would be to develop a second base year GDP country list that is more consistent with the ASF model assumptions. However, we felt that presenting model-specific base year country lists would be confusing and difficult to justify for users, and have decided to simply present the current numbers as they stand.

For further details on the regional projections from this model contact: Alexei Sankovski and William Pepper, ICF Consulting, Washington, DC, USA

B1: Spreadsheet 3 presents for the B1 scenario using regional economic growth rates from the Integrated Model to Assess the Greenhouse Effect (IMAGE) from RIVM in the Netherlands. This model was the ‘marker’ model for the B1 scenario. The regions in this model are: Canada, USA, Latin America, Africa, OECD Europe, Eastern Europe, former Soviet Union, Middle East, India, China, East Asia South, Oceania and Japan. For the list of specific countries included in each region see page 340 of the SRES report.

As seen in the accompanying second tab of this spreadsheet, the regional sums for the data differ significantly in the REF region. This initial discrepancy is essentially maintained throughout the downscaling period. The main cause for this discrepancy is similar to the discrepancies explained for the scenario A2 above – the marker model for B1 had a large difference in the 1990 GDP estimate for REF compared to the 1990 REF GDP estimate for the marker models for the other scenarios. Specifically, the B1 marker 1990 REF GDP is ~10% lower than the B2 marker 1990 REF GDP. The remainder of the discrepancy for the B1 REF GDP relates to the smaller base year GDP differences between our country list and the SRES B2 marker regional sums.

Once again this shows that a single country-level base year GDP list cannot be consistent with all the base year marker model regional GDP estimates. Rather than try to develop multiple, marker-specific, base year country GDP lists, we simply present the numbers as they stand.

B2: Spreadsheet 4 presents results for the SRES B2 scenario. We used the regional economic growth rates from the IIASA MESSAGE model. This model was the so-called 'marker' model for the B2 scenario in general. The 11 regions in this model are the same as those listed for the B2 population data: North America, Western Europe, Pacific OECD, Central and Eastern Europe, Newly independent states of the former Soviet Union, Centrally planned Asia and China, South Asia, Other Pacific Asia, Middle East and North Africa, Latin America and the Caribbean, Sub-Saharan Africa. For the list of specific countries included within each region see pages 332-333 of the SRES report.

The spreadsheet shows that the regionally summed GDP values compare fairly well with the regional sums in the SRES book report. The base year differences are a maximum of about 7% for the REF region and lower for the other regions. These base year differences are largely preserved throughout the projection period with small fluctuations due to changing B2 model regional weights.

For further details on the regional projections from this model contact: Nebojsa Nakicenovic, Arnulf Grübler, R Alexander Roehrl, and Keywan Riahi, International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria

NOTES on UNSTATS Database Terminology & Definitions:

“market prices” [code 284] The actual price agreed upon by the transactors. In the absence of market transactions, valuation is made according to costs incurred (non-market services produced by government) or by reference to market prices for analogous goods or services (services of owner-occupied dwellings). (42, para. 2.68) Reference: United Nations, Commission of the European Communities, International Monetary Fund, Organisation for Economic Cooperation and Development and World Bank (United Nations and others, 1994). System of National Accounts 1993 (SNA 1993), Series F, No. 2, Rev. 4 (United Nations publication Sales No. E.94.XVII.4). para. 2.68.

“current prices” [code 238] A fundamental principle underlying the measurement of gross value added, and hence GDP, is that output and intermediate consumption must be valued at the prices current at the time the production takes place. This implies that goods withdrawn from inventories by producers must be valued at the prices prevailing at the times the goods are withdrawn and consumption of fixed capital in the System is calculated on the basis of the estimated opportunity costs of using the assets at the time they are used, as distinct from the prices at which the assets were acquired. (42, 1.62). Reference: United Nations, Commission of the European Communities, International Monetary Fund, Organisation for Economic Cooperation and Development and World Bank.