After the whole data analysis process, the following tentative conclusions can be stated:

1. Both electrification and urbanization are good indicators of economic growth, and they are highly correlated with each other during the


Figure 9.11 Annual percentage change in US GDP and EP indices, 1960-2001


Figure 9.11 Annual percentage change in US GDP and EP indices, 1960-2001

development trajectory, especially for countries that started industrialization late. We chose electrification instead of urbanization for the first proxy mainly because electrification is more directly energy-related and its data quality is better.

2. Per-capita electricity consumption plus some fraction of per-capita oil consumption can be used as a good proxy to simulate economic catch-up by using values relative to that of the US for both the proxy and per-capita GDP.

3. By dividing countries into groups based on criteria such as oil exports, hydro-electricity fraction and geographic latitude, we identified nine non-empty sub-groups, of which four sub-groups accounted for most of the countries (and the GDP). Only two of these sub-groups exhibit significant development vis-à-vis the US. Satisfactory regressions were obtained for these two groups (Table 9.5). Weighting by GDP significantly improves the quality of regressions. We have checked the residuals and found no problem of heteroskedacity.

4. Countries with high petroleum exports did not exhibit catch-up. (This phenomenon has been termed 'the curse of oil'.) Other 'growth stoppers' include ethnic or religious conflict, transition from central planning to market orientation, poor macroeconomic management and corruption.

5. For countries that are actually catching up, the relationship between the GDP gap and the energy proxy gap is not linear, but behaves like a logarithm or square root of the energy proxy. There is not much difference between the logarithm model and square-root model in terms of the fraction of oil added into the proxy, although 0.10 and 0.15 turned out to be a little bit better than higher values.

6. The catch-up countries exhibited very similar growth trajectories, relative to the US, except for four 'young' countries (South Korea, Singapore, Malta and Jordan) whose oil consumption increased relatively faster than electricity consumption and departed from the main trend. It is possible that these countries started from lower levels than others. After dropping these four countries, we obtained very good results for the relationship between the energy consumption gap and GDP gap.

7. The rate of increase in GDP with respect to the energy proxy ('catchup elasticity') decreases as countries approach US levels (that is, the gaps decrease). The differences in 'catch-up elasticity' between the two non-linear models (square root versus natural logarithm) are surprisingly large for countries at early stages of development, suggesting that more research is needed on this question. Here we differ from the 2 percent per annum convergence 'law' that was suggested by a number of economists in the 1980s, and later (for good reason) discarded.

As a final concluding remark, we note that the empirical results obtained in this study tend to support the theoretical basis of prior work on the US economy by Ayres and Warr and discussed in the previous chapters of this book (Ayres and Warr 2002; Ayres et al. 2003; Ayres and Warr 2005; Warr and Ayres 2006). More important, in terms of the need for more credible forecasting and scenario-building tools, we think these results have an immediate application. While the underlying theory - regarding the role of energy and useful work as a driver of growth - is not yet fully tested even for the US, still less widely accepted, we think that the results demonstrated in this chapter are quite sufficient to justify extrapolation in a 'scenario' context, for several decades. In particular, we hope to use our model results to simulate future economic growth for important catch-up countries, such as India and China.

Also, to make the statistical results more statistically persuasive in the future, some additional statistical tests, such as auto-correlation and stationarity, should be run, since the data we use are panel data.

Trash Cash Machine

Trash Cash Machine

How recyclable trash can save the world and bank us huge profits! Get All The Support And Guidance You Need To Be A Success At Recycling! This Book Is One Of The Most Valuable Resources In The World When It Comes To How To Make Profits With Trash!

Get My Free Ebook

Post a comment