The preparation of national accounts for the Central American republics since 1920 was a necessary prerequisite for the writing of this book. The purpose of this appendix is to explain the methodology underlying these national accounts estimates and to provide for the benefit of scholars working in this field the detailed statistics on which much of the book's analysis has been based.

First, however, some consideration needs to be given to the question of whether such an exercise is justified. Although all scholars are in broad agreement that it is desirable to have statistical series of good quality, it is worth remembering that an unreliable series can do more harm than good. The problem is not only lack of information in certain areas of economic activity, but also published figures in which the authorities themselves place little confidence. For example, the director of the Dirección General de Estadisticas y Censos in Costa Rica complained in 1930 that the statistics on agriculture and commerce were valueless because of ‘the anarchy that reigns in the reports sent to this office by the political authorities.’ The director then published a table on the area cultivated with various agricultural products with a note stating: ‘The object of the above table is to demonstrate the inefficiency of the data collection system in current use, in view of the differences in the figures from one year to the next.’ Similarly, a survey on Guatemalan statistics (published in 1946) had the following to say about agriculture: ‘Each year the Ministry of Agriculture sends questionnaires asking for information with regard to agricultural products, livestock and the number of acres planted and harvested.