Illiteracy and the mirage of its figures

On the podcast Café Ejecutivo, Federico Castillo asked me about the country's illiteracy rate. I answered with a number that had stayed with me for years: between 11% and 12%. That figure matched what the 2010 Census had shown and a historical floor that rarely seemed to really move. I left…

On the Café Ejecutivo podcast, Federico Castillo asked me about the country's illiteracy rate. I answered with a number I had carried with me for years: between 11% and 12%. That figure matched what the 2010 Census had shown and a historical floor that rarely seemed to really move. I left the interview with a methodological doubt and decided to dig in. The conclusion is uncomfortable: the Dominican Republic has several official illiteracy rates, all of them current, none reconciled with the others.

The government announced 6.0% for 2024. The figure comes from ENHOGAR and is reproduced in IDEICE studies as evidence of a sustained literacy effort. The Central Bank's Encuesta Nacional Continua de Fuerza de Trabajo reports 5.6% for the same year. UNESCO and the World Bank project 4.5% using the GALP cohort model. The 2022 Census —the most complete and expensive instrument the Dominican State produces— suggests, according to MEPyD's Monitor de la Frontera, a national rate close to 11% when the literacy questions are cross-referenced with the education module. Four official instruments from the same State, in almost identical years, separated by up to six percentage points. The difference amounts to half a million people.

The first methodological break goes unnoticed by the casual reader and is decisive. ENHOGAR changed its protocol in 2019 and 2023 when it joined UNICEF's MICS rounds. ONE itself acknowledges in its internal report that the series are not comparable. The press, however, publishes them chained together as if they showed a downward trend. Until 2024 the rates were calculated using population projections from the 2010 Census, which overestimated the young population and underestimated adults. The denominator was inflated, which artificially lowered the percentage of illiterate adults. That correction is only beginning with the results of the 2022 Census, and the next national census will not arrive until 2032. We have ten years of public policy ahead of us resting on a figure whose consolidation will take at least another three.

There is a demographic component that the official narrative omits with discipline: mortality. Illiteracy is concentrated among people over sixty-five, where the additional probability of not knowing how to read or write reaches almost thirteen percentage points. The COVID-19 pandemic shortened the lives of many older adults with low levels of education, which reduced the national rate without anyone having been taught to read. On top of that comes rural-to-urban migration, which dilutes the rural indicator in the urban average. The Dominican countryside has had illiteracy rates between 12.8% and 13.2% since 2016. The border region —Pedernales, Elías Piña, Independencia— stands between 13% and 16%. The national improvement rests on the movement of people and on selective mortality, rather than on a learning program.

Public spending doesn't match the celebration either. Quisqueya Aprende Contigo and its successor SEPA spent more than 5,300 million pesos between 2021 and 2024. SEPA certified 2,719 people in 2024, an approximate unit cost of RD$1.48 million per head, according to MINERD's own figures. To move the needle in a base population of 7.65 million adults, orders of magnitude different are needed.

Not the whole amount is allocated directly to SEPA. The line item «adult education and literacy» also includes the 610 schools of the adult education subsystem, offering primary education with flexible schedules. No official report publishes the exact breakdown of SEPA spending for 2024. The State reports results without reporting the unit cost that corresponds to them.

This inconsistency matters beyond education statistics. Dominican technology policy rests on a fragile assumption: that we have a literate human capital base at 94% that can climb to the next digital step. If the real figure is closer to 89% and the average quality of learning equals 5.48 years of schooling —according to MINERD's own APEA indicator— then the skills pyramid on which artificial intelligence strategies, chip factories and global services are being planned has a narrower base than the speeches admit. The digital braceros we export are the visible layer of a very unequal educational distribution.

The window until 2032 calls for three pending tasks that cannot wait for the next census:

  • The first is to demand that ONE issue a technical note reconciling the figures from ENHOGAR, ENCFT, ENA-2019 and the 2022 Census, with their methodological breaks made explicit, and updating the population denominators for the entire historical series.
  • The second is to publish the full audit of the literacy programs 2013-2024, verifying actual competencies rather than declared enrollment.
  • The third is to break down the reported reduction into its demographic, migratory and programmatic components, to distinguish what the State did from what happened on its own.

That number I recalled on the podcast had a statistical basis. It was the last robust data point before the statistics began to move at the pace of the political cycle.

The Dominican Republic may have a real illiteracy rate close to 6% or close to 11%. Between one figure and the other there is half a generation of people who cannot read this text.

Digital public policy for the next ten years will be built on whichever figure we decide to believe, with no new census to correct our course until 2032.

Fede, there's the data.