GINFORS-E

The GINFORS-E (global inter-industry forecasting system – energy) model is a bilateral world trade model based on OECD data, which consistently and coherently models exports and imports of 25 goods groups for 64 countries and one ‘rest of the world' region. It incorporates a macro-model, consisting of exports and imports, other core components of final demand (private and public sector consumption and investment), markets for goods and the labour market, for each country. The models are also divided into 36 goods categories in accordance with the latest OECD internationally harmonised input-output (IO) tables. For every country OECD bilateral trade data on industry level is linked to the IO tables. For EU countries, UK and Norway transport is further distinguished in land, water, air plus warehousing and support activities based on the WIOD World Input-Output Database. These variables are reported for 40 industries.

GINFORS-E can be used to analyse the macroeconomic effects of a variety of price changes and policies in individual countries. It flexibly models trade structures, labour markets, energy intensities and energy source structures, taking into account price dependencies and the situation in specific countries. The use of intermediate inputs, domestic and imported, labour demand and foreign trade are modelled price dependent. Changes in prices due to tax adjustments will be accounted for. The parameters used in the model equations are econometrically estimated (OLS) on the basis of time-series data.

Production prices of industries are driven by unit costs. If prices of electricity in the steel industry increase, producer prices will increase according to their electricity price share. Higher producer prices will influence global competitiveness of the respective industry and other downstream production (e.g. in the automotive industry).

Important behavioural parameters of the model are estimated econometrically, and different specifications of the functions are tested against each other, which gives the model an empirical validation. An additional confirmation of the model structure as a whole is given by the convergence property of the solution which has to be fulfilled on a yearly basis. The econometric estimations build on times series from OECD, UN, IMF and IEA from 1990 to 2000 onwards.

Each national model is linked to an energy model, which determines energy conversion, energy generation and final demand for energy for 19 energy sources disaggregated by economic sector. The model considers technological trends and price dependencies.

The model is solved simultaneously year after year. Almost all model variables are endogenously determined via identity or behavioural equations. Behavioural variables are econometrically estimated as far as possible. Only a few variables, such as population development and international energy prices, are exogenously specified based on international projections or kept constant such as tax rates.

Data for 64 countries plus one region for rest of world include:

• Macroeconomic data as GDP and components (consumption, investment, exports, imports), in constant and current prices plus deflators
• Bilateral trade by 33 product groups
• Population, employment, unemployment, wages
• Input-Output tables (https://www.oecd.org/sti/ind/input-outputtables.htm)
• Sector data for 36 industries: output in constant and current prices, value added, employment, and final demand
• Energy balances
• CO2 emissions by sector and fuel, other GHG emissions
• Energy prices by user and fuel, including tax rates (VAT, energy)
• Carbon prices

Due to the modelling approach, all input variables determined ex-ante by the model can also be output variables. The most important of these are macroeconomic indicators on national level, as well as corresponding sector variables, which are calculated for all countries considered for all years up to 2050. The most important among them are:

• GDP and its components (household consumption, government consumption, investment, exports, imports)
• Employment, production, value added and prices on sector level
• International trade flows by product group, origin and destination
• Energy demand by sector and fuel, energy prices
• CO2 emissions by sector and fuel

The model is flexible to reflect, for example, different uses of CO2 price revenues to reduce labour costs, increase (specific) government spending, or reduce government debt. Various other policy measures can also be mapped quite easily.

Although GINFORS-E can be used for forecasting, the model is mainly used for evaluating the impacts of policy scenarios, changes in assumptions such as international energy prices or another change to model variables. The model can be enlarged to include more detail on interesting datasets, currently e.g. on the bioeconomy and energy technology goods.

The analysis is mainly forward looking (ex-ante), but can also inform implementation or evaluate previous developments ex-post.