It can be seen that in most sectors also labour-productivity growth is better explained if we
account for country-specific fixed effects. Moreover, the speed of convergence increased in
most sectors as compared to Table 5, with the half-life between 4 years (Transport) and 36
years (Agriculture), and with Machinery as the most important exception (with its half life
increasing to 367 years). Table 7 shows that wages have a statistically significant positive
effect on labour-productivity growth in Agriculture and Textiles, while this is negative in
Transport Equipment. Moreover, similar to energy-productivity growth, the impact of the
other explanatory variables on labour-productivity growth remains mixed, with economies of
scale having again the largest statistically significant effect, being positive in Transport, Non-
Ferrous Metals and Paper, while it is again negative in Textiles. We found openness to have a
statistically significant positive effect in Paper only, while for specialization this is again only
the case in Iron and Steel. Finally, the results do not give any support to the vintage effect,
with the only statistically significant estimates displaying a negative sign.
6. CONCLUSIONS
This paper extends the existing empirical analyses of convergence patterns by providing a
systematic comparison of energy- and labour-productivity convergence at a detailed sectoral
level for 14 OECD countries, covering the period 1970-1997. A σ-convergence analysis
revealed that the development of the cross-country variation in energy- and labour-
productivity performance depends on the level of aggregation, with different patterns of
productivity convergence and divergence across sectors. At the macroeconomic level we
found evidence of energy-productivity divergence, driven by aggregate Manufacturing, as
well as labour-productivity convergence, mainly driven by Services. The Manufacturing
energy-productivity divergence turns out to be caused by the Iron and Steel and the Non-
Ferrous Metals sectors. Moreover, despite a lack of evidence of labour-productivity
convergence at the aggregate Manufacturing level, there is evidence of labour-productivity
convergence in several Manufacturing sub-sectors, with Machinery as the most important
exception in that it shows a clear pattern of divergence (in particular after 1985). A β-
convergence analysis, using a panel-data approach, led tot the conclusion that in most sectors
energy-productivity growth is relatively high in countries with relatively low initial
productivity levels, while in several sectors this is also true for labour productivity. This
result supports the hypothesis that relatively backward countries tend to catch-up to more
advances economies, in particular in terms of energy productivity, possibly because they can
18