Export Performance and Growth:

There have been two stages  in cross-country  regression analysis of  the relation between openness and  growth. In the first stage (surveyed by Edwards (1993) growth was  regressed against variables  representing  export performance. Then, in the nineties, the arialysis moved  towards examining the relation between trade policy and growth rather than between trade performance and growth.

Studies, which tried  to  estimate a relationship between export performance and growth  based on moss section samples  of a  number of countries, measured export performance by indicators such as the share of exports in the GDP, or the change in  the export share or the growth of exports.  Initially, economists found  a  significant positive  relation between growth in  GDP and  export performance (Balassa (1973) and Michaely (1977)).  But  this approach was criticised  on several grounds. Balassa's study  analysed a  small sample  of eleven countries,  including Korea,  Taiwan, India and Chile, which  may  bias  the  results.

Furthermore since the GDP  includes exports, faster export growth would lead to faster GIX  growth. The more appropriate  test to regress pwth of non-export GDP oa export performance often showed no effect. Goncalves and  Richtering  (1986) found  a  significant  Spearman  rank correlation coefficient  between  the growth  rate of GDP, the export growth  rate, and the change in  &are  of exports in  GDP,  but not  between  export performance  and the growth  rate of  non-export GDP. Taylor (1986) also  found 'that differences in  trade policy  orientation had  little to  do  with  how successfully countries responded  to  the  external shocks of  the  seventies.'

Batchelor, Major and Morgan (1980) in a detailed analysis found that  the exports of manufactured goods played a significant role in  the development  of only small semi-industrialised countries where the  small size of the market was an obvious constmint.

In a  remarkable article Sheahy found a  significant relation between growth and almost all other majar categories in the national accounts,  whether  these aggregates were froxn the demand side or hm  the production side. In many cases the impact of  these  other national  income aggregates  was  more important than that of exports. Our  interpretation of Sheahy's result, and thus of all cross  country regressions, is  that a regression of growth and any national income aggregate measures the backward and forward linkages  of  that aggregate. The larger these linkages  are, the  more important is  the interrelation; the more is the sector  an enclave  type, the less significant is  the linkage.