Читать книгу Economics of G20 - Группа авторов - Страница 53

Spearman’s rank correlation gives a weaker result than correlation. Thus, this chapter analyses these results further using correlations. Correlations for the same variable (averages) in the pre-crisis and the post-crisis periods for all the countries are calculated. They are found to be significant, re-enforcing the results of the Spearman’s rank correlations.

The correlation between two variables is examined before and after the crisis, to check whether they are significant. Then a t-test is performed at 95% confidence inter val. This is to see if the correlations are statistically significant and whether the relationship among the variables in the precrisis period holds even in the post-crisis period. The results are discussed below.

There is a positive correlation between average GDP growth rate and average GFCF in the pre-crisis and post-crisis periods. However, the correlations are not statistically equal. So, the same relation is not maintained after the crisis. Average export share has a positive correlation with CAB, foreign exchange and total reserves in both the pre-crisis and post-crisis periods. This reinforces the conclusion in the previous section. However, the pre-crisis period correlation is not statistically equal to the correlation in the post-crisis period.

There is significant positive correlation between the average foreign exchange and average total reserves in the pre-crisis and post-crisis periods. The average interest rate and average money stock have a significant negative correlation in the pre-crisis and post-crisis periods. This relationship is maintained in the post-crisis period as well — as interest rate increases, people invest more (or save more, expecting future returns) and thus money stock decreases. This did not change in spite of the crisis.

There is a positive correlation between pre-crisis average foreign exchange, pre-crisis average GFCF and average total reserves with post-crisis average GDP growth rate. While the aforementioned correlations are backed by economic logic, there are other correlations that seem spurious. Not all results are discussed here.