126.96.36.199 THE ROLE OF R&D EXPENDITURE AND INVESTMENT The technical progress indicators are used as a measure of innovation and product quality and estimate the non-price competitiveness of an industry. Ideally, E3ME could incorporate measures of innovatory activity in each EU member state, relative to the same activity in its main competitors, at a detailed industrially disaggregated level on an annual basis, covering the period 1970-2002, but limited data restrict the choice of inputs. The decision as to which data to use as a representation for innovation comes down to a choice between two alternatives: patents; and research and development activity. In each case, the available information has to be mapped onto the industrial classifications used in the E3ME model. This requires a very detailed examination and comparison of the systems of classification used in each case, sometimes involving comparisons across countries.
Patenting activity represents one measure of innovatory activity. It is well established that different industries have different propensities to patent. A potentially valuable data set exists in the form of the series collected in the United States by the Department of Commerce, Office of Technology Assessment and Forecasts (OTAF). These data indicate the level of patenting activity by industry conducted by most major states within the US. Given the latter's key role in innovation and the world economy generally, these data provide a potentially very useful measure of relative innovatory activity in different countries. However, the data suffer from some limitations as far as the present exercise is concerned: the industrial classification used is a US one and, by their nature, patents tend to focus attention on manufacturing industries rather than the service sectors. It should be noted that there is ongoing work at the OECD to compile an industry database of patents, which may be a valuable input to E3ME in the future.
188.8.131.52 R&D EXPENDITURES
An alternative measure of innovation is R&D expenditure (and related employment). In contrast to the patents indicator, this is a measure of input rather than output from innovatory activity. The OECD publishes a series of data on R&D activity by industries for major economies, known officially as the ANBERD database. These, in principle, enable relative measures of member states' performance to be constructed. In practice, the OECD data are based on irregular surveys conducted within each individual country. There are therefore large numbers of missing observations. A considerable amount of interpolation and adjustment is therefore necessary to convert these data into a usable form for time-series analysis. They also suffer from similar problems of matching industrial classifications and time-scale coverage as the other series already discussed. These data have been extended to 2003 and are the basic measure of innovatory activity used in the estimation of the equations of E3ME.
The ANBERD database covers only business enterprises, that is, not the public sector, which means that extra work was required before public sector R&D, including defence spending, could be identified. For this type of disaggregation, which was also required in E3ME30, a separate OECD survey based on the Frascati Manual was used, which distinguishes military R&D expenditure for each member state. In addition, the IEA has published annually for the last few years a detailed analysis of OECD states' spending on energy-related R&D, including energy-conservation R&D. This provides detail of large-scale R&D programmes in EU member states funded by national governments.
Unfortunately the data on privately funded energy-saving R&D are partial and incomplete, so it is not possible to present data or results for total energy-saving R&D.
Detailed data for investment demand (Gross Fixed Capital Formation) in European countries, on both a constant and current price base, are published by Eurostat and are available as part of the OECD's STAN database. E3ME disaggregates investment into the 42 industry sectors used throughout the model, and gaps in the published data were filled using the previous version of the E3ME database (version 3.0). The units are standardized to EUR (current prices) and 2000-valued EUR (constant prices) for all the E3ME countries.
Investment is a component of GDP, but expenditure on investment and innovation mainly enters E3ME's equations indirectly, by its use in formulating a measure of technical progress. The approach to constructing the measure of technological progress in E3ME is adapted from that of Lee et al. (1990). It adopts a direct measure of technological progress (Tt) by using cumulative gross investment, but this is altered using data on R&D expenditure, thus forming a quality-adjusted measure of investment. The equation for Tt is written as
where dt(r1) satisfies the following recursive formula dt(r1) = r1dt_i(r1) + (1 - r1)log(GIt + r2RDt) (7.2)
where GIt is the level of gross investment; RDt is constant price research and development expenditure; r1 is a measure of the impact of past quality adjusted investment on the current state of technical advance, while r2 is a measure of the weight attached to the level of R&D expenditure.
To initialize the recursive process for dt, the assumption is made that in the pre-data period the process generating log(GIt) is characterized by a random walk. Under this assumption, the first value of dt can be written as d0 = log(GI) (7.3)
where the right-hand side represents the average of gross investment over the first five-year sample period. The values of r1 and r2 were set at 0.3 and 1.0 respectively, while noting that more sophisticated procedures could have been adopted, such as a grid-search method based on log-likelihood values. The series dt(r1) is then calculated by working the recursive procedure forward given the initial value, d0.
In E3ME41, there are two technical progress indicators, one which measures technical progress related to ICT (Information and Communications Technology) investment in the new economy, and one which is related to all other investment. The construction of the two indicators is similar, with investment split up into ICT and non-ICT related investment, and r2 set to 0 in the non-ICT investment measure (i.e. all R&D expenditure influences the ICT measure).
The two sets of technical progress indicators appear together in the equations outlined below, and separate long- and short-term parameters are estimated for each one. Due to a lack of data, a single set of indicators is maintained for the EU's newer member states.
184.108.40.206 USE OF THE TECHNOLOGICAL PROGRESS INDICATOR MEASURES
The variables used to represent technological progress enter a variety of equations in E3ME version 4.1, including:
The technological progress variables are included as part of the implicit production function that lies behind the factor demand equations in E3ME. The effect on employment demand is deemed ambiguous, as this greatly depends on whether the type of technical progress is labour-saving or labour-augmenting. The extra activity of R&D itself, however, is likely to be more labour-intensive than average production in most industries.
The presence of technical progress in determining average hours worked originates in the determination of the optimal number of hours worked as part of the representative firm's cost minimization process. A negative sign is imposed on the coefficients for technological progress, based on the assumption that an increase in investment R&D will improve the efficiency of the capital stock, thus requiring fewer average hours worked for a given number of employees.
• industrial prices
A positive effect was imposed on the technological progress variable to cope with the quality effect that increased investment/R&D is expected to have, that is, the role of product innovation. The effect of process innovation (which would be expected to lower prices) is taken account of by a measure of unit costs, which is a separate variable in the equation.
• export and import volumes
The technical progress indicators are included to help capture the role of innovations in trade performance. This variable could be measured relative to that of competitors, but this has not been implemented in E3ME4.1. (However, if it was only relative technical progress that improved performance and if all countries experienced such progress simultaneously, then there would be no effect on economic growth; this does not appear to be very plausible as a characterization of modern industrial and service economies). The anticipated effects are a positive elasticity for export volumes and a negative elasticity for import volumes.
• export and import prices
The measures of technical progress are included to cope with the quality effect on prices caused by increased levels of investment and R&D, and progress is assumed a priori to have a positive effect on export prices and a negative effect for import prices.
• energy demand
The energy demand equations include both gross investment and R&D spending directly as explanatory variables. These terms are intended to capture the effects of new ways of decreasing energy demand (energy-saving technical progress) and the elimination of older inefficient technologies. This will also take into account the introduction of new energy-saving techniques and methods of energy conservation, and hence is expected to be negative. In particular, technical progress in the industries producing machinery, electronics, and electrical equipment is expected to reduce aggregate energy demand, and technical progress in the motor vehicles industry is expected to reduce the demand for oil, as transport equipment becomes more efficient and alternative energy sources are adopted.
7.2.4 The effects of GHG and energy taxation 220.127.116.11 INTRODUCTION AND ASSUMPTIONS
One of the purposes of E3ME is to provide a consistent and coherent treatment of fiscal policy in relation to greenhouse gas emissions. Figure 7.3 shows how tax rates affect prices and wage rates in the model; the mechanism is the same for taxes on other emissions and energy.
There are inevitably certain simplifying assumptions required in this kind of modelling.
18.104.22.168 FIRST ASSUMPTION
The first assumption is that the effects of the tax in the model are derived entirely through the impact of the tax on fuel prices, and through any use of the subsequent revenues from the tax in reducing other taxes. Other effects are not modelled. For example, if the introduction of such a tax caused the electricity industry to scrap coal-burning plant in advance of what might be expected from the relative price change induced by the tax, this effect would have to be imposed on the model results. The one exception to this rule is the announcement effect of the UK climate change levy (CCL) (see Section 7.6).
All the energy and emission taxes are converted into a consistent set of units (C/toe). These taxes are then added to the costs of the fuels. Tax revenues can be calculated from fuel use; the revenues are reduced according to the fall in use, but rise according to price inflation and any escalator in the tax rates. In the baseline case, effective tax rates are calculated by dividing fuel use by raised revenues; this includes any exemptions and non-payments, which the raw data for tax rates on its own does not.
The second assumption is that imports and domestic production of fuels are taxed according to the energy content of the fuels, but that exports are exempt from the tax coverage. The treatment is assumed to correspond to that presently adopted by the authorities for excise duties imposed on hydrocarbon oils. It is assumed that industries and importers pay the tax, and that it is then passed on in the form of higher fuel prices paid by fuel users. A further assumption is that industrial fuel users may pass on the extra costs implied by the tax in the form of higher prices for goods and services. The increase in final price is a result of the direct and indirect energy content of each commodity distinguished in the model. If the revenues are used to reduce employer tax rates, then industrial employment costs will fall and these reductions in costs are also assumed to be passed on through the industrial system.
22.214.171.124 ETR EFFECTS ON THE E3 SYSTEM
In considering the competitive response of different sectors and companies, there are two important questions to be answered:
• Are the prices of the product set in the world markets or by the producer or in the local markets?
• How flexible is the process of production in responding to an increase in costs?
If the price is fixed in the world market, then no increase in costs arising from an increase in energy taxes can be passed on to final product prices. If the process of production is also fixed (e.g. because the product requires long-lived capital stock), then it might be very expensive to change the technology or move the plant; so all extra costs must be paid out of profits. If the industry or the company is not profitable, then the extra costs could lead to plant closing. However, this is an extreme outcome and most industries and companies have the ability to pass extra costs on to their customers and to change their production process to reduce emissions and avoid some of the increase in taxes.
For these reasons, changes in manufacturing export and import volumes do not give enough information about the effects on competitiveness. The effects on unit costs can be compared to those on export prices to see which sectors have their profit margins squeezed by being forced to accept world prices for their products, while at the same time being unable to avoid increases in their unit costs (see Barker, 1998).
The net effect on industrial and import prices will eventually feed through to consumer prices and will affect relative consumption of goods and services, depending on the carbon/energy content and on their price elasticities. The higher consumer prices will then lead to higher wage claims.
Figure 7.4 shows the effects of these price and wage rate changes on fuel use, CO2 emissions, and industrial employment. The changes in relative fuel prices as a result of the tax will change fuel use, depending on substitution elasticities. The fuel price increases will be passed on as more general increases in prices, which will cause substitution in consumers' expenditure, in exports, and between imports and domestic production. These changes will feed back to fuel use. CO2 emissions are derived directly from the use of different fuels. If employment costs are reduced when tax revenues are recycled, then industrial employment will be stimulated directly, with a further indirect effect as labour-intensive goods and services gain in relative price competitiveness.
7.2.5 The effects of the various revenue recycling mechanisms
The COMETR scenarios assume that the ETRs in each country are revenue neutral in each year; in this way, the results presented in this chapter
describe a shift in the tax burden to energy use from the more general economy, rather than an increase in the overall tax burden. Of the six countries examined, a variety of methods were used to recycle tax revenues and in some cases a combination of methods was used. These methods are outlined in the list below. For several countries, no explicit revenue recycling mechanisms were put in place, either because the tax reforms were part of such a wide package that it was impossible to determine what the alternative tax arrangements would be, or simply because no provisions were put in place at the time. In these countries, it is assumed that revenue neutrality is achieved through a shift in income taxes. This was judged to be the most non-controversial way of maintaining revenue neutrality, as the ETR revenues are very small when compared to the overall level of income tax revenues.
The revenue recycling methods considered in E3ME were changes in:
• employers' social security contributions;
• employees' social security contributions;
• investment in energy-saving technology.
The first four methods of revenue recycling relate to the labour market. By changing income tax, the government is directly affecting (nominal) disposable incomes. For example, a reduction in income tax will increase disposable income, all other things being equal. The immediate effects are likely to be a boost to household spending, particularly on luxury goods. Through multiplier effects, there will be further increases in average incomes, as domestic firms require more inputs to meet the extra demand, including labour. In the longer term, a 1 per cent increase in household income will lead to an equivalent 1 per cent increase in overall consumer spending, in line with conventional economic theory. Higher employment and lower unemployment may cause some increases in wages. The case of social security payments is interesting. The effects of changing employees' contributions are almost identical to the effects of changing income tax rates. This makes intuitive sense as the payments by employees are the same, even if the treatment of the tax by government is different (the government sector is largely exogenous in E3ME). The main reason for including this separately from income tax is to simplify the processing in Germany, where there were equal reductions in employers' and employees' contributions.
The initial effect of reducing employers' social security contributions is to lower the cost of labour to firms, and hence should lead to a direct increase in employment. The effects of this will be twofold: first, there will be an increase in average household incomes from higher employment rates; there could also be a slight increase in average wages if unemployment falls. Both effects would lead to an increase in average household incomes and, possibly, a short-term increase in household consumption. The overall inflationary impact will be dependent on the relative strengths of these two effects (e.g. labour-intensive firms will have the largest initial fall in costs and wages will increase faster if the economy is close to full employment), but overall the effects are similar, an increase in average household incomes driving forward consumer spending.
The effects of changing benefit rates are slightly different. If the government increases benefit rates, this increases the disincentive to work. The magnitude of this effect varies across the countries of Europe but, overall, we would expect to see a decrease in employment and labour market participation. In pure economic terms, increasing distortions in the labour market is not seen as a way of increasing productivity and economic output. This must be balanced against equity issues.
By increasing investment in energy-saving products, a government is hoping to achieve a decrease in energy consumption over and above the reduction due to the price effect of higher taxes alone. There will be some other positive effects, however, as investment tends to improve the overall quality and desirability of an industry's output (i.e. increased non-price competitiveness). Previous research has found these effects to be particularly strong in international trade in the long term. More immediately, there will be a boost to industries producing capital goods, such as construction and motor vehicles.
7.2.6 A brief description of E3ME's labour market
E3ME includes a detailed treatment of the labour market with stochastic equations for employment (as a head count), average wages, hours worked, and labour market participation. This plays an important role in the scenarios, particularly in cases where tax revenues are recycled through the labour market. Unemployment is calculated as the difference between employment and the active labour force and is a key explanatory factor in determining wages and household consumption. Unlike many equilibrium models, E3ME does not assume full employment, even in the long run.
Employment is modelled as a total headcount number for each industry and country as a function of industry output, wages, hours worked, technological progress, and energy prices. Industry output is assumed to have a positive effect on employment, while the effect of higher wages and longer working hours is assumed to be negative. The effects of technical progress are ambiguous, as investment may create or replace labour; this will vary between sectors.
Hours worked is a simple equation, where average hours worked by industry and country is a function of 'normal hours worked' (expected hours worked based on patterns in other industries and countries) and technological progress. It is assumed the effects of technical progress gradually reduce average hours worked over time as processes become more efficient. The resulting estimate of hours worked is an explanatory variable in the employment equation (see above). Hours worked is defined as an average across all workers in an industry, so incorporates the effects of higher levels of part-time employment in certain countries and industries.
In E3ME wages are determined by a complex union bargaining system that includes both worker-productivity effects and prices and wage rates in the wider economy. Other important factors include unemployment, tax rates, and cyclical effects. Generally, it is assumed that higher prices and productivity will push up wage rates, but rising unemployment will reduce wages. A single average wage is estimated for each country and sector. The estimates of average wages are a key input to both the employment equations and the price equations in E3ME. In the absence of growing output, rising wages will increase overall unit costs and industry prices. These prices may get passed on to other industries (through the input-output relationships), building up inflationary pressure.
126.96.36.199 THE LABOUR MARKET PARTICIPATION EQUATIONS Labour market participation is estimated as a rate between 0 and 1 for male and female working-age population. Labour market participation is a function of output, wages, unemployment, and benefit rates. Participation is assumed to be higher when output and wages are growing, but falls when unemployment is high, or benefits create a disincentive to work. In addition, there is a measure of economic structure and the relative size of the service sector of the economy; this has been found to be important in determining female participation rates. The participation rates determine the stock of employment available (by multiplying by working-age population, which is exogenous). This is an important factor in determining unemployment, which in turn feeds into wages and back to labour market participation.
Was this article helpful?