THE METHOD OF SIMULATED MAXIMUM LIKELIHOOD FOR THE ESTIMATON OF DYNAMIC ORDERED PROBIT: AN APPLICATION TO COUNTRY-RISK FOR NON-DEVELOPED COUNTRIES

This paper aims to give a detailed explanation of the econometric methodology necessary to estimate dynamic probit models with ordinal dependent variables. A typology of cases are established which appear when considering different choices of individual heterogeneity along with time correlation. To be able to estimate by maximum likelihood the models which come out of the different alternatives proposed, simulation techniques are used and put into practice by the GHK simulator and, in this way, estimators by simulated maximum likelihood are obtained. Finally, all the models described are used to measure and determine the macroeconomic factors which explain the ratings of country-risk in non-developed countries.