Administrative records are collected and stored by different public and private entities to carry out and monitor their own activities. Ministries of education collect school enrollment and academic achievement records; police departments store records of criminal activity; emigration services record movements of people in and out of the country; tax authorities collect information about earnings and employment contained in income tax reports provided by firms and individuals; and civil registries record marriages, live births, and deaths. All of this information, however, is typically stored in individual databases at each agency, missing an opportunity to interrelate them to yield a rich tool for policy advice. For example, combining school placement records with examinations taken at the end of the school year could inform objective calculations of academic school value added. Similarly, linking school-level data with employment and earnings could inform the effectiveness of different schools in terms of labor market outcomes. This individual-level linking is frequently precluded by the absence of unique individual identification numbers registered in the different databases and, more importantly, by the compartmentalization culture usually observed across different agencies.
Against this background, several Caribbean countries have started to break these silos with palpable results. For example, with the technical assistance of the Inter-American Development Bank, the Ministry of Education of Trinidad and Tobago has linked academic assessments taken at primary and secondary schools at the individual level from year 1995 onward. In addition, criminal records from the police service as well as marriages and live births from the civil registry have also been linked. Because unique identification numbers were not recorded in the databases, the individual-level linking was conducted using names, dates of birth, and gender. To guarantee confidentiality, the linking followed strict protocols in the relevant agencies and, after the process was completed, all personal information was removed. The linking process is described in detail in this Technical Note. Using the matched data, several policy-relevant analyses have been conducted looking at the effect of peers on academic achievement; the effect of single-sex schooling on academic achievement, criminal activity, and teen pregnancy; and the effectiveness of school management systems. Similarly, Barbados has already linked primary and secondary schools’ examination records since 1987 onwards within a collaborative engagement between the Inter-American Development Bank and the Ministry of Education, Science, Technology, and Innovation. In Jamaica, administrative records of applicants to the PATH program (the conditional cash transfer program of the country) were linked to academic examinations administered by the Ministry of Education. This linked dataset has allowed an analysis of whether PATH participation increases academic achievement, educational aspirations, and secondary school placements.
The previous efforts demonstrate that interagency communication and harmonization of databases can put great value on dormant information toward guiding evidence-based policy decisions. This collaborative effort between the Inter-American Development Bank and Caribbean countries is ongoing, as many other data sets could also be linked. Specifically, valuable databases containing labor market outcomes such as employment and earnings (collected by social security agencies) and emigration records should be the next step in this effort.
To learn more about how we’re bringing Caribbean data to your fingertips, check out our Data Portal: caribbeaneconomics.org.
Get more special features in your inbox: Subscribe to our newsletter for alerts and updates.
Do you have a strong opinion on this article or on the Caribbean economy? We want to hear from you! Tell us what you think by commenting below, mentioning us on twitter @IDB_Caribbean, or contribute your own op-ed piece to firstname.lastname@example.org.