European Commission logo
NEMESIS

New Econometric Model of Evaluation by Sectoral Interdependency and Supply

EconomyEnergyEnvironmentClimatemacro-sectoralforecastsquantitative forward-looking scenarioseconomic policiesstructural policies

overview

EconomyEnergyEnvironmentClimatemacro-sectoralforecastsquantitative forward-looking scenarioseconomic policiesstructural policies

main purpose

Macro-econometric model designed to assess socio-economic impacts of research and innovation policies.

summary

NEMESIS has been developed by a European consortium coordinated by the ERASME team, a research laboratory common to Panthéon-Sorbonne University and École Centrale de Paris (France). The other contributors to the model are the National Technique University of Athens (ICCS-NTUA), the Federal Planning Bureau (Belgium) and later the UNU-MERIT of Maastricht University and many others contributors. The model of which construction started in the early 2000’s is, since, continuously improved in order to suit to the evolution of theory and to the needs of users.

NEMESIS, is a system of detailed macro-sectorial Economic model (30 sectors) for every European country, that allows (1) to make quantitative transcriptions of storyline scenarios for the building of short- medium- to long-term forecasts, and (2) to assess the socio-economic and environmental impacts of a wide spectrum of policies.  For its running, the model requires as inputs exogenous assumptions on drivers such as demography, fossil fuel and other raw material prices, world trade, and the policy implementations of every European country and of Europe.

The mechanisms of the model are based on behavioural equations for four categories of ‘representative’ agents: firms (one by sector), households, governments and the rest of world. These equations are parametrized by econometrics or calibrations based on econometric literature. The resulting dynamic of the model is truly macro-sectorial in that it combines sectorial and inter-sectorial evolutions to macro-economic feedbacks. This dynamic is time-recursive (adaptative expectations) and shares similarities, on the short-medium term, with the Neo-Keynesian synthesis, while the long- term dynamic of the model is based on the new theory of endogenous growth where the investments in R&D, in ICT and in  other intangible such as training and software, play a key role together with knowledge externalities. The model includes also a detailed Energy and Environment module that retroact on its dynamic.

Along with forecast scenarios, NEMESIS allows for a wide spectrum of policy assessments including fiscal, budgetary, labour market, research and innovation, energy, climate and environment policies. The results are presented in terms of economic activity, external competitiveness, employment per categories, inequalities, GHG emissions, public finances, available, whenever it makes sense, at sectoral, macro, national and European levels.

model type

  • Other

ownership

Third-party ownership (commercial companies, Member States, other organisations, …)
SEURECO sarl, France

licence

Licence type
Non-Free Software licence

homepage

http://www.erasme-team.eu/imgs/FCK/file/publications/NEMESIS_Reference_Manuel_Draft_version.pdf

details on model structure and approach

We present here successively (1) the general characteristics of the model, (2) it representation of innovation and long term growth, (3) its Energy-Environment module and (4) its main applications.

1) The general characteristics of NEMESIS

A fully macro-sectorial model, with a recursive dynamic

NEMESIS is a fully macro-sectorial model, and not a simple split of macro results in 30 economic activities. The model results are influenced by three dynamics: sectoral from the interaction of supply and demand by product, intersectoral from the transactions and the knowledge externalities between production sectors, and macro-economic from households’ savings and consumption decisions, and the action of government.

In the model, the behaviour of economic agents is based on the determination of optimal, or desired values. The adaptative expectations, the existence of price rigidities and of adjustment costs, introduce delays for reaching these optimal values. The expectations structure and the reaction delays can be modified to take into account the influence of important external shocks, for instance on the price of fossil fuels. The delays towards the optimum values are estimated either by error correction mechanisms or by a partial adjustment process. For all these reasons, the adjustment costs and the price rigidities, the dynamic of the model is time-recursive.

The behavioural equations of the model are either estimated by econometrics or, calibrated whenever the econometric literature can furnish robust and well proven values for the parameters. It is notably the case for the endogenous growth module of NEMESIS.

The long-term horizon the economic track obeys to the new endogenous growth theories: it is feasible to endogenize the long-term rate of growth by adapted research and innovation policies. Moreover, a wide scope of policies, such as fiscal, energy and environment policies, can influence the innovation rate and then change the long-term rate of growth.

The behaviour of the firms

On the supply side, each sector is modelled with a representative firm that takes its decisions regarding the production and the use of factors, given its expectations on market expansion and input prices. The production is represented with nested constant elasticity of substitution (CES) functions combining five production factors: low-skilled labour, high-skilled labour, capital, energy and intermediate consumption. In addition, a particular feature of this model is that the firms integrate innovation in their decisions, allowing the improvement of their productivity and of their products, so that the technical progress is endogenous in the model. Innovation is the consequence of investments in three types of as-sets: R&D, ICT and Other Intangibles (including software and training). The specification of the innovation process in the model allows to account for a large range of innovative activities and for the enabling feature of ICT that are considered as GPT. Furthermore, while R&D are mainly the matter of industrial sectors, the other innovative assets allow a more appropriate endogenization of innovation in services sectors. The pricing scheme for enterprises is defined by a monopolistic competition and the margin rates are assumed constant but different between sectors.

The behaviour of households

In an econometric model such as NEMESIS, the maximization of households’ utility is implicit and the welfare function of households is not specified and computed as such. The main determinant of household’s aggregate final consumption is the disposable income which comes from wages, from the capital revenues, and from taxes and redistribution. Other determinants are the structure of the population and the influence of ageing in the long-term, and the unemployment rate in the short-term. The households’ total final consumption is econometrically is disaggregated in 27 consumption functions. The allocation system runs through several steps by assuming separability between different groups of consumption categories. At first, the consumer decides the respective shares that will be allocated to non-durable and durable goods. Then, within each group, the consumption is split again between subgroups, etc. We can add to this presentation that, for specific purposes this allocation scheme can be adapted to traduce, for instance, the modification of households’ preferences towards a more ecological behaviour induced by a clearer conscience of ecological difficulties.

Labour Markets

While the labour demand is determined by the optimisation of the production function, the supply is based on population projections by skill and age. The interaction of labour demand and supply is specified in the model by an ‘augmented Phillips curve’. The augmented Phillips Curves determine the wages by linking rate of growth of wages to unemployment level, labour productivity (in order to assure long term stability of profits/total wages ratio) and consumption prices. We know the importance of this specification for the long-term properties of the model, and it is usual for policy assessments to make sensitivity analysis on the parameters of this specification, especially on the unemployment coefficient that traduces the flexibility of the labour market. Not that wage equations are different for low-skilled and high-skilled labour that belong to two different labour market. The evolution of wages is also different in the different economic sectors, depending in the growth rate of labour productivity in the different economic activities.

External trade

The external trade by country in NEMESIS takes two directions: intra-EU trade and trade with the rest of the world. The exports are driven by two main effects in the model: (i) an income effect which accounts for the economic activity in the other regions and (ii) a price effect which accounts for the relative competitiveness of the considered country with respect to other EU-countries and the rest of the world. Regarding imports, the drivers are similar: the income effect is captured by internal demand, and the price effect by the ratio between the imports price and the price of domestic producers.

Government

The behaviour of government is described by the accounting framework of the model. The model includes the data and assumptions on fiscal policies (indirect and direct taxes with detailed fiscality on energy and environment, social security contributions, etc.), government expenditures (defence, health, education, infrastructures, others expenditures), investments and subsidies. The model provides indicators on government expenditures, incomes and surplus and debt.

 

2) Endogenous innovation and long-term rate of growth: a fully endogenous growth approach

In NEMESIS producers can determine the Innovation they intend to blend with the new production techniques. It is the result of the investment in three innovation factors: R&D, ICT, and other intangibles investments than R&D (Software and professional training, and of knowledge spillovers and technological transfers between economic sectors and countries. As R&D investments concern mainly the industrial sectors, the investments in ICT and in other intangible categories of intangible describe more appropriately the innovation in the service sectors. The innovation takes two forms: process and product innovations. The first increases the productivity and then lowers the production cost, while the second increases the quality of product. In both cases innovation has a direct positive impact on the demand for products.

Analytically, the innovation mechanisms of the model are based on the new growth theories that stress the central role played by intertemporal knowledge externalities for the long-term growth rate of the economy. Two ideas must be retained:  (1) the importance of Knowledge externalities of which rate of growth influence positively the innovation capability of the firms in the different economic sectors, and (2) the possibility to increase durably the rate of growth of the economy by an adequate policy that increases the investment in innovation factors, like in the second generation of ‘fully endogenous growth models’.

However, the possibility of increasing durably the rate of growth of the economy could be limited in time if the supply of the skills necessary for innovation activity is too limited, inducing inflationary pressures that will counterbalance the positive effect on demand of process and product innovations. Therefore, the model will describe in the long-term, after a policy that increases the innovation rate of the economy, a new equilibrium unemployment rate compatible with the rise of labour productivity, and a long-term price growth compatible with the trade balance equilibrium. Innovation policy does not influence any more the growth of employment, while its influence on the growth rate of the economy is equal to the growth rate of labour productivity induced by the policy. This result for employment is conditioned by parameterization of the wage curves of the model, and notably of the Phillips effect: depending of the value of the Phillips effect, the long-term equilibrium rate of unemployment would be different, but without influence on the long-term rate of the economy.

The most important point is that with the fully endogenous growth approach followed in NEMESIS, innovations policies will theoretically succeed to influence the long-term growth rate of the economy, whereas it is not the case for the semi-endogenous growth approach followed with other models, and were the long-term rate of growth is invariant the innovation policies that are introduces by the governments.

 

3) The energy and the environment

Another important specificity of the NEMESIS model, when compared with the other macroeconomic models, is that it includes a “plug-in” module detailing the energy demand and the GHG emissions.

The energy-climate module calculates the energy balances in physical unit for each country. On the demand side and for each economic activity, the aggregated energy demand is split into fourteen different energy products: fossil fuels (coal, gas and oil), electricity, heat, renewable energy (solid biomass, liquid biomass, biogas, renewables municipal wastes, heat pump, geothermal, solar thermal) and others (non-renewable municipal wastes and industrial wastes). In each economic activity, firms’ choice between energy product depends on the aggregated energy demand and the relative prices and it models with embodied CES production functions.

The power generation sector is also detailed by technology and it includes fifteen different production technologies: coal with/without CCS, gas with/without CCS, oil, solid biomass with/without CSS, nuclear, geothermal, hydro, wind off/onshore, PV, CSP, and tidal. The characterisation of the power generation technologies (capital cost, O&M costs, capacity factor, timelife, etc.) allows the calculation of average production costs by technology that is used to choose between technology through diffusion curves.

The model covers all Kyoto gases: CO2, CH4, N2O, HFCs, PFCs, SF6 and NF3. The GHG emissions from fossil fuel combustion is directly related to the energy demand by product and calculated with emission factors. Other GHG emissions depend on specific sector economic activity, on autonomous efficiency and on specific abatement curves, only emissions from LULUCF are exogeneous.

This energy-climate module allows the assessment of a large set of climate mitigation options for EU: EU-ETS market, cap-and-trade policy, carbon tax, etc.

In the field of climate, environment and energy policies, we must emphasize here that the model has been used for numerous study and national and European research projects, alone using its energy-climate module (Capros et al., 2014; France Stratégie, 2019 ([1], see below)), but also linked to other more detailed partial/sector equilibrium models. It is the case at national level for socio-economic impact assessments of detailed energy policy with the POLES model (DGEC, 2015 ([2], see below)), or European studies with GREEN-X, on employment impact of renewable energy sources deployment (Duscha et al., 2016) and European research projects: SET-Nav (www.set-nav.eu), with ENERTILE, ASTRA and INVERT-EE or Paris-Reinforce (www.paris-reinforce.eu) with FORESCAST and the JRC-EU-TIMES model (on-going). A former version of the model has also been linked with a model of the European agriculture, CAPRI, a detailed land-cover model (Dyna-CLUE) and a model of forest and forestry management (Efiscen), in a European research project called SENSOR (Jansson et al., 2008).

 

4) NEMESIS utilisations

A model simulation starts with the exogenous variables: policy variables, demography, prices for raw materials (including Energy), Rest of World demand, and computes the endogenous variables: competitiveness, GDP and growth, employment by categories, wages  and prices, external balance, public finances at Sectoral and Macro Economic level, whenever it makes sense , and this for every Country or the whole Europe. A more detailed list of inputs and outputs is given later as well as a description of the NEMESIS uses. The different properties of the model allow a wide spectrum of utilizations for forecasts and the analysis of policy options.

The model was recently, and is currently used at the European level and national level, for many purposes and by different organizations:

  • By think-Tanks like the Commissariat au Plan, France-Stratégie, la Fabrique de l’Industrie and the IFRAP in France, the CEPS, BRUEGEL and ESPAS in Brussels;
  • By public administrations, different D.G. of European Commission, the O.C.D.E., the French Government with the Ministry of economy, the Ministry of labour and the Ministry of environment, and the French SENAT;
  • By private administrations, as Chambre de Commerce et d’Industrie de Paris in France.

For the reference or prospective scenarios, 4 types of exercises have been implemented with the model:

  1. Ex post scenarios, describing quantitatively the past, that allow, for instance, to make an ex-post assessment of policy measures that were formerly implemented;
  2. Short to medium term forecasts up to 10-15 years;
  3. Medium term  prospects, up to 10-15 years;
  4. Long term forward looking scenarios up 2050.

Long term “Futurology” based on hypothesis or unforeseen evolutions leading to structural modifications of mechanisms of the model. These are breakthrough scenarios describing extreme situations where the usual mechanisms of the model can no longer explain the real behaviour of agents. For instance, the COVID crisis did change drastically the behaviours in terms of consumption and production, as did previously the 2008 Financial. NEMESIS was used for this type of variant in the Flagship European research project (ref). Future evolutions can also be impacted by appearance of breakthrough technologies as it is investigated currently in the CAD study for the European Commission on the consequences of the development of Connected Automated Driving .

For the analysis of policy options, 2 types of exercises have generally considered:

  1. Variant works aim at calculate the consequences on model results of the modification of exogenous variables or mechanisms of the model, comparing the new results to a baseline. These modifications can concern for example assumptions on fiscal or budgetary variables;
  2. Structural policies, like policies for employment, research an innovation, energy, environment and climate change.

We must add to this presentation, of which references are detailed later, that the model has been used coupled with other models or exchanging results with other models, by iteration in order to better describe the Agriculture and Land-Use, the Energy technologies and the Transport. One of the advantages of the structure of the model is its malleability, that allows coupling with other models or to make exchange of data by iterations and this in a detailed coherent accounting framework.

 

[1] France Stratégie. (2019). The Value for Climate Action - A shadow price of carbon for evaluation of investments and public policies. Report by the Commission chaired by Alain Quinet

[2] MEDDE-DGEC, « Scénario prospectifs Energie-Climat-Air pour la France à l’horizon 2035 – Rapport final - Synthèse des résultats », September 2015

model inputs

The parameters used to parametrise the behavioural equations of the model are either econometrically estimated, or their values are retrieved from the econometric literature. The econometric estimations are used for exports and imports functions, household’s consumption, savings and investment choices, and firms’ pricing and production behaviour, including the choice of ordinary production factors (physical capital, high-skilled and low-skilled labour, energy and other intermediate consumptions). The calibrations are used for wages formation and for the decision of firms to invest in process and product innovation, including their investments in innovation inputs (R&D, ICT, software and professional training). 

Scenarios development and policy analysis with NEMESIS require a set of ‘entry’ variables on inputs, to put in a coherent way the assumptions on the main exogenous variables that will condition the track of European economies up to 2060. The main inputs of the model are:

  • Raw material prices:
    • Fossil fuels prices: Oil, gas and coal;
    • Other raw material prices: Metal products and agricultural products.
  • Financial variables:
    • Exchange rates, especially €/$ and €/against other European currencies;
    • Nominal interest rates: Short and long term;
    • Changes in national fiscal and social systems.
  • Demography:
    • Five age cohorts: ([0-14], [15-24], [25-54], [55-64] and [65 and +]);
    • Two skills: ISCEAD 0 to 4 and ISCEAD 5 to 6.
  • Economy:
    • World demand addressed to EU by product category: It is calculated usually from GDP projections for non-EU countries (BRICs, USA, Japan, Rest of OECD and RoW) world I/O tables and bilateral trade flow matrices;
  • Other:
    • Policies: CO2 ceilings, R&D policy, fiscal policies, etc.
    • Expert judgments on specific sectors and variables.

The analysis of NEMESIS results is based on a large number of ‘exit’ variables or outputs, that are calculated endogenously by the model, with more than 200,000 different indicators defined at European, National, Macro-economic and sectoral levels. 

model outputs

The main outputs of the model are:

Economic indicators at sectoral level

  • Production factors:
    • High and Low skilled labour;
    • Energy consumption;
    • Other intermediate consumptions;
    • Investment in physical capital;
    • Investment in equipment and buildings;
    • Investment in ICT;
    • Investment in software and professional training;
    • Investment in R&D.
  • Final demand:
    • Households and Public final consumption;
    • Investment
    • Exports
  • Activity:
    • Production;
    • Value added.
  • Prices:
    • Production;
    • Value-added;
    • Imports;
    • Exports;
    • Final consumption;
    • Production factors;
    • Etc.

Economic indicators at macroeconomic level

  • GDP and GDP counterparts:
    • Final consumption;
    • Public consumption;
    • Investment;
    • External balance
  • Price indexes:
    • GDP;
    • Final Consumption;
    • Investment;
    • Terms of trade.
  • Production factors demand and cost:
    • High and Low skilled labour;
    • Energy consumption;
    • Other intermediate consumptions;
    • Investment in physical capital;
    • Investment in equipment and buildings;
    • Investment in ICT;
    • Investment in software and professional training;
    • Investment in R&D.
  • Labour market and households:
    • Labour force;
    • Employment rate;
    • Unemployment rate;
    • Real disposable income.

Public finances

  • Receipts;
  • Expenditures;
  • Investment;
  • Deficit in % GDP;
  • Debt in % GDP.

Energy and environment

  • Energy consumption by product: Electricity, gas, coal, petrol, bio fuels, etc;
  • CO2 and other GHG emissions: So2, NOX, HFC, PFC et CF6;
  • Carbon and ETS prices.

model spatial-temporal resolution and extent

ParameterDescription
Spatial Extent/Country Coverage
EU Member states 27 and UKIcelandNorwaySwitzerland
Spatial Resolution
National
Temporal Extent
Short-term (from 1 to 5 years)Medium-term (5 to 15 years)Long-term (more than 15 years)
Short- and medium-term forecasting (up to 8 years) at the macro- and sectoral levels; Building long-term baseline scenarios (up to 30 years).
Temporal Resolution
Years