Sector Income Accounting and Analysis for Latin American and Caribbean Economies--More Appropriate Equations

Abstract

AR E not Latin American and Caribbean economies, as a broad group, sufficiently different in their structure and dynamics to justify, or, in fact, logically require recognition of this fact by national income statisticians and analysts? And if the way in which these economies work is, indeed, sufficiently distinctive, what are the most important, relevant characteristics that should influence national income research, statistics and analysis? What are the principal innovations that are suggested by the peculiarities? This paper offers answers to these questions. I shall point out briefly the main features of the structure of these economies, indicate the principal autonomous stimuli experienced by them, suggest a simple system of sector accounts that would (i) bring together meaningful information on the most relevant magnitudes and (2) reveal relative changes in these from one period to another, and finally, discuss how the information might be used, suggesting also some of the equations that might be derived to guide the formulation of policy as well as for theoretical study. It is not my argument, however, that the aggregative measures which have become familiar are without use for the study of Latin American economies. For certain purposes, summary national account systems of the form recommended by the United Nations Statistical Office are undoubtedly needed. But, as will be explained, the structure of each of the Latin American economies is such that important aggregate-relating parameters which are fairly stable in advanced countries, or subject to change within relatively narrow limits, change greatly in the short run, even in the economies we discuss. The general argument here, in other words, is that much of the information that is needed for shortand medium-term forward looking analysis in Latin American economies is lost in the conventional national accounts. I assume that it is not necessary for me to demonstrate or prove in these pages the now familiar propositions that the equations of an accounting system are simply the formal economic identities written out in a particular way, and that the same estimates of the values of transactions in relation to one another would be less intelligible if they were not presented in an accounting form.