I am currently undergoing research on the relationship between economic growth and inequality using a dynamic panel model. My measure of economic growth is difference in logs of real GDP per capita between period t and t-1.
When I am specifying my regressions, should I express the dependent variable as (lnGDPpc_t - lnGDPpc_(t-1)) or should I first generate a growth variable, defined as the difference in logs, and then include this generated growth variable as my dependent variable?
i.e. should i use (lnGDPpc_t - lnGDPpc_(t-1)) or growth_t as my dependent variable?
When I am specifying my regressions, should I express the dependent variable as (lnGDPpc_t - lnGDPpc_(t-1)) or should I first generate a growth variable, defined as the difference in logs, and then include this generated growth variable as my dependent variable?
i.e. should i use (lnGDPpc_t - lnGDPpc_(t-1)) or growth_t as my dependent variable?
