Two Working Papers on the macroeconomic impact of network industry reform have been published. The first is a detailed report in Dutch, the second a summary paper in English. The analysis builds on a simulation by FPB’s macroeconometric model HERMES, and a simulation by IMF’s general equilibrium model GEM. Although both simulations were based on the same exogenous input, they gave significantly different outcomes. This sheds light on the applicability of different modelling approaches to an issue at hand.
First of all, the impact upon micro-economic efficiency was determined. This analysis was limited to allocative and productive efficiency, and to five industries: electricity, gas, railways, postal services and telecommunications.
Allocative efficiency involves the elimination of economic rent. That rent benefited the former monopolistic producers, but also their employees if they were able to receive a ‘wage-premium’, lifting their pay above market level. With operating surplus used as a proxy, economic rent seemed to be very large for electricity, gas and telecommunications. Moreover, and in particular for electricity and gas, there indeed seemed to be industry specific wage-premiums. Productive efficiency involves a more productive use of labour and capital. Because of the labour intensity of network industries the analysis was limited to labour productivity. Reform might lead to a loss of between 24,000 and 36,000 jobs in the network industries. Most of this loss would take place in railways, postal services and telecommunications. The improvements in allocative and productive efficiency would result in falling prices for the network industries’ services. These would range between a few percent for natural gas to about 30% for certain telecommunications services.
Using these efficiency improvements as simulation input in a demand-driven macroeconometric framework (HERMES), the mid-term outcomes are disappointing. Because of the fall in profits and employment levels, consumption and investment would in the short run stay up to 0.4% behind the growth that would have been realised without the reform. The fall in prices, however, would have a positive impact upon competitiveness. In the mid term, exports might become up to 0.3% higher than they would have been without reform. The impact upon GDP would be between 0.1% and 0.2%. The loss of jobs would not be fully compensated for by jobs created elsewhere in the economy. The net impact upon employment would be about -10,000.
Using the efficiency improvements as an input in a structural long-term framework (GEM), more optimistic outcomes are obtained. Because of the fall in prices, real income would increase and demand with it. This would lead to more demand for labour and capital and thus have a positive impact upon consumption and investment. The job losses in the network industries would lead to a higher number of job seekers, and hence more excess supply in the labour market. In the long run the market mechanism would then lead to a new equilibrium in the labour market without additional unemployment. In the steady state, GDP would be between 1.1% and 2.6% higher than it would have been without reform, and job growth between 34,000 and 77,000.
Both models thus give strongly different outcomes although they start from the same input. Most striking is the different modelling of the impact upon consumption and investment. Given the thorough modelling of the behavioural and institutional relationships that characterise the Belgian economy, the performance predicted by HERMES seems more realistic, at least in the short and mid term. The loss of profitability and employment initially lays a burden on the economy, which needs time to be compensated for. However, the simulation by GEM shows that market forces may, in the end, lead to new equilibriums with higher output and employment. In behavioural terms, this implies that investment may be attracted when demand-driven opportunities arise, and unemployed people might accept new jobs at less favourable working conditions. In institutional terms, this implies that flexible functioning markets smooth the adjustment process after an initial shock. An analysis using two different modelling approaches thus seems useful in understanding the economic forces empowered by network industry reform and other structural changes. The approaches are complementary rather than conflicting.