Tracing convergence across countries has vast literature. To put it simply, a negative relationship between growth and initial income refers to Beta convergence, and declining variance across countries refers to Sigma convergence. Some argue that Sigma convergence is the 'real test' of convergence and the other is not. Can someone give updated knowledge on this? Moreover, does the convergence theory holds true only for neo-classical approach or for other models too?
1. Both beta and sigma convergence are somewhat unreliable: beta crucially depends upon the period when it is measured, therefore might turn out unstable; sigma might mask individual churning movements.
2. Beta convergence is grounded in neo-classical theory, but the negative relation, if any, can be found and be compatible with other theories, too
1. Both beta and sigma convergence are somewhat unreliable: beta crucially depends upon the period when it is measured, therefore might turn out unstable; sigma might mask individual churning movements.
2. Beta convergence is grounded in neo-classical theory, but the negative relation, if any, can be found and be compatible with other theories, too
Elisabetta,
Thanks. Yes, while some people use initial income level, others like hotelling go for final period. But if both are unreliable, which i could use for checking convergence ?
Thanks in advance.....
Definitely the initial income level. The link with neoclassical theory is that initial income level must be a good predictor of subsequent growth rates. By using their final income level, one incurs also in the critique of unfairly choosing the countries ex post. Regressions, in this way (with final rather than initial income)usually yield better results, which could be the reason why they are employed ;-).
Interesting, and thanks again. Should there be any rationale for choosing the initial year ? Instead of choosing, say, the first year of a decade or some subjective years, could one go for some other methods to decide which should be an initial year ?
well, the initial year makes the "long run" longer. you need to evaluate growth in the long run. different years, like the last one - I have already answered about it - cannot be predictors of the future. Solow model predicts that in the long run the steady state will be shared, therefore all countries will converge to the same equlibrium. the implication is that the futher away from the steady state you are (the poorer you are) the quicker you will reach the steady state (the higher will your growth rate be). you, of course is impersonal to mean a country. this is why the negative relation is hypothesized and verified when the collection of countries is such that the other hypotheses (e.g. sharing the same technology, absence of disturbances) can be verified.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Writing