Does Education Pay in Bolivia?

By: Lykke E. Andersen*

Returns to education in Bolivia have been dropping steadily over the last 15 years, to the point that some researchers have argued that education no longer pays[i]. Nevertheless, a record number of young people are in school or university. Are they all wasting their time?

In this blog we will explore what has happened to the returns to education in Bolivia over the last 15 years, using standard household surveys (as well as the students in my Microeconometrics II course at UPB).

After running several hundred Mincer earnings regressions of many different specifications, it is beyond dispute that returns to education have dropped dramatically between 1999 and 2014 (see Figure 1). In 1999, one additional year of education implied about 11% higher hourly earnings, but this benefit of additional education had dropped to just 4.3% by 2014.

This is really shocking, and if the trend continues, education would indeed be worthless when the kids in primary school now finish their education.

Figure 1: Average returns to education in Bolivia, 1999-2014

Source: Author’s estimation (with the help of my Microeconometrics II class of the diploma course “Quantitative Methods for Economic Analysis” at Universidad Privada Boliviana during June 2016).

While very low, the returns to education are not yet zero. Not at the national level, and not for any significant sub-group that I could identify. In 2014, the group with the lowest returns to education were workers in the transport sector, followed by workers in the hotel and restaurant business. The ones with the highest returns to education were rural workers and workers in the department of Potosí. Everybody else earned just about average returns to education (see Figure 2). Even the highest returns in 2014, however, are very low by historical and international standards.

Figure 2: Average returns to education in Bolivia, by sub-group, 2014

Source: Author’s estimation based on Mincer earnings regressions using the logarithm of hourly labor earnings as the dependent variable. By taking logarithms, people with zero earnings get eliminated from the regression.

Figure 3 shows the changes in the returns to education in more detail, by years of education. In 1999 returns to education were approximately linear, with each extra year of education adding about 11% to hourly labor earnings[ii]. By 2014, however, returns had decreased substantially at all levels, except at the very lowest levels of education.  Indeed, it now seems that all the years of primary school and secondary school add very little to labor earnings per hour. You would need 16 years of education to earn significantly more than a person with 1 year of education.

By 2014, the three years after high school (13, 14 and 15 years of education) earned only a fifths of the returns that they did in 1999. The benefits of a complete university education (bachelor, master or Ph.D.) in 2014 was only a third of what it was in 1999 (see Figure 3).

Figure 3: Returns to education in Bolivia, by education level, 1999 and 2014

Source: Author’s estimation based on Mincer earnings regressions using the logarithm of hourly labor earnings as the dependent variable, and including dummies for each year of education.

I hate to admit it, but Werner Hernani is disturbingly close to being right. While technically, the returns to education are still significantly above zero, about two thirds of the 20-25 year olds in Bolivia currently have less than 13 years of education, and with that, they barely earn more than people with one year of education.

This is a huge problem. While the changes in the returns to education have brought about a significant reduction in inequality in Bolivia, no country can be competitive in the global economy with an uneducated labor force, so the current situation is clearly not sustainable.

The reasons for the current dismal situation are multiple and difficult to disentangle. One important factor is changes in the supply and demand of skilled labor. With the impressive increases in education levels in the country, the supply of skilled labor has increased, so it is natural that its price should decrease. At the same time, and due to the construction boom caused by the Dutch Disease caused by the boom in the prices of our main export products, there has been an increase in demand for unskilled labor (construction workers), which has put an upward pressure on the price of non-skilled labor.

A different issue, however, is the quality of education. Right now, it seems that the education that the vast majority of school children in Bolivia receive simply does not help them improve their employment prospects. Bolivia has not participated in any of the international education quality surveys since 2002 (when results were frankly disappointing), but the regression results reported above suggest that our education system does little to improve local prospects, and probably does even less to help them compete internationally.

Conclusion: The situation is much worse than I thought. Returns to education in Bolivia are technically still positive, but alarmingly close to zero. If this situation is maintained over a prolonged period of time, it will severely harm the long run growth prospects of the country.


[i] Specifically, Werner Hernani from Fundación ARU, with whom I have made a bet that returns to education are still positive, while he claims they are 0. This post is my official argument, which is to be judged by Professor Miguel Urquiola of Columbia University — Bolivia’s foremost expert on education — and compared with Werner’s reply, to determine who is going to organize a barbecue party in October.

[ii] Linear returns to education means that the percentage increase in earnings is the same for each additional year of education. However, the same percentage increase implies exponential increases in the absolute level of earnings, which is why the curve with linear returns to education looks exponential.

* Lykke E. Andersen is a Senior Researcher at INESAD. She greatly appreciates feedback on this post either in the comments below or directly to:

The views expressed in this article are those of the authors and do not necessarily reflect the views of Fundación INESAD.


Check Also

Impressions from the 8th Bolivian Conference on Development Economics

By: Lykke E. Andersen* Despite my initial hesitations about going ahead with the 8th Bolivian …


  1. Would you like to spare a moment for Jesus Christ?

  2. La verdad es que no me sorprende este hallazgo, pues veo que se halla plenamente en línea con el modelo aplicado en Bolivia relacionado con el extractivismo, el rentismo y el prebendalismo, muy acentuados en los últimos años. Cada uno de estos ámbitos constituye un campo de acción en sí mismo y tienen la común característica de no requerir en buena medida de niveles significativos de educación.

    Así como no se requiere de un gran nivel de educación para extraer y transportar minerales del socavón al puerto (pasando por un ingenio) o gas y petróleo del pozo al ducto, tampoco se requiere un gran nivel de educación para obtener ingresos incluso significativos en actividades rentistas como la venta de ropa de usada, el contrabando, la producción de coca, la piratería, el comercio minorista, etc. Algo similar puede decirse de la mayor parte de comportamientos prebendales y clientelares que aprovechan un vinculo con el poder político para obtener ciertas ventajas extraordinarias que pueden traducirse en grandes ingresos (como se vio en el caso de las cooperativas mineras), sin que ello exija niveles de educación significativos.

    En este marco, es oportuno recordar que el 80% de la economía se ha refugiado en el llamado sector informal, que se caracteriza por no cubrir una serie de costos relacionados con impuestos, legislación laboral, seguridad social, etc., debido a su baja productividad e ingresos, todo lo que también contribuye a explicar la escasa importancia que se le asigna a la educación.

    Dicho en breve: la educación debe estar claramente integrada en el modelo productivo, debe ser requerida y necesitada, como para que el retorno sea creciente o cuando menos constante, asunto que por lo que muestras está muy lejos de ocurrir en Bolivia. Por ejemplo, en el caso de Galicia, España, el rendimiento de la educación para estudios superiores es del 28,2% y de 14,6% para aquellos que poseen estudios anteriores al superior.

    Gracias por ponerle cifras y tendencias a esta situación tan preocupante.

  3. Evelin Mamani Huayta

    I wish we can know the methology used in this analysis for a better consideration. Nevertheless, I have two coments. The first one, this issue might be a problem in many other countries, developed and developing ones since in the first group (developed countries) when analizing by cohorts the lower quality institutions does not assure the deserved returns to education in the future, while high-ranked institutions more than assure it.

    The second one which was mentioned by Lykke E. Andersen, and is related to my fisrt observation, is the quality of education. The 2015 MDGs pushed to countries like Bolivia to invest on programs to have every child enrolled in primary education, achieving this objective, however, it did not take into account the quality of education the countries were giving, Bolivia’s education problem might be bigger than other developing countries due to the lack of comparative measures of education’s quality.

    Finally, I think the core of this problem is on quality of education and the blindness of the policy makers that are launching high-cost programs on education (Bolivia’s investment on education as percentage of GDP is the highest in LA) without first doing a serious research.

  4. Ernesto Bascopé

    Plantearía otra causa que, quizás, merece análisis: el enorme crecimiento del trabajo en el sector público. Los trabajos que se han creado en el Estado central (y en menor medida en municipios y gobernaciones) pagan salarios que no necesariamente guardan relación con el nivel de formación de los trabajadores o su productividad. Además, los salarios más altos del sector se encuentran limitados por factores políticos y los salarios más bajos superan con mucho a lo que una persona con poca o sin formación esperaría ganar en otros sectores. ¿qué tanto contribuye esto a entender la correlación entre años de formación e ingreso?

    • Lykke E. Andersen

      Estimado Ernesto: Muy buen punto. Definitivamente merece investigación. Como un primer paso he calculado los retornos a la educación en el sector público y privado por separado. Resulta que los retornos son significativamente más altos en el sector público (4.1% versus 3.0% en el sector privado), así que el mecanismo que tu mencionas no parece ser la principal causa de las caídas en los retornos a la educación en Bolivia. Es aún peor en el sector privado.

      • Cesar Santa Gadea

        Personally, I am not surprised. Time and time again I was confronted with a particular brutal reality that almost destroy any motivation left on me. You can have a fine education, but if your working environment lacks o a multigeneration knowledge base from which you can build your skill base, then you are doom to a low productivity career path. Being in a workplace where all the organisational structure is dominated by people that lack the knowledge base and the working skills to carry their job responsibilities, generate a poisoning mix.
        Usually, under this scenario, the hiring decision are politically driven and expose a string of chauvinism. I believe this is the reason that explains the fact that you can travel all around the developing world and find several technicians with PHDs from fine academic institutions that truncated their professional development because of lack of practical and state or the art job experience.


Get every new post delivered to your Inbox

Join other followers: