Using Big Data for Anti-Poverty Programs to Protect Latvia’s Poor

April 15, 2014


As Latvia was recovering from the 2009 global financial and economic crisis, the World Bank partnered with the Government over 2012-13 to investigate tax, benefit, and training policies to combat long-term unemployment and draw people into the workforce. The study used detailed administrative data tracking individuals’ labor market status and use of benefits and employment/training programs over time to provide policy advice on reform of the tax and benefit system and the development of labor market programs. The study’s recommendations were utilized by the Government to inform future policy reforms.


Latvia was hard hit by the global financial crisis, as combined GDP declined by nearly 25 percent and unemployment tripled from 6 to 20 percent between 2007 and 2009. As the economy gradually recovered, unemployment remained elevated at 15 percent by end-2011. The lower skilled suffered disproportionately from the deterioration in the labor market. Of particular concern were the socially vulnerable, long-term unemployed, and those who are not active in the labor market. Getting more people back to work was critical to reversing the deepening of poverty and social polarization that has occurred since 2007. Nearly two-thirds of Latvia’s poor lived in households with low work intensity.

The Government required answers to questions about benefit dependency, how the tax and benefit system could be changed to offer adequate protection and promote employment, the direction job placement and formal training programs should take to bring people back to work, and how the Government can use its wealth of administrative data to inform evidence-based policy making in all of these areas.


Latvia’s Ministry of Welfare engaged the World Bank to evaluate the country’s employment and social protection situation coming out of the crisis, noting especially the unemployed and inactive. A series of background notes were produced using detailed administrative household survey and spending data in collaboration with Government technical experts.

The study demonstrated how to use big data to inform a difficult policy debate on welfare dependency, tax benefit policy for lower-income groups, and employment programs. To investigate these issues,  a large panel database was created linking data on individuals in the social security, social assistance, employment services, population registry, and health insurance databases. The data for 91 monthly waves from January 2006 to July 2012 covered 43 percent of the total population and allowed the Bank to investigate—with a strong evidence base—benefits dependency issues and the impact of employment programs.       

By involving a large number of stakeholders in the study from inception to completion and making all the analysis available online and at public conferences, the project was seen as objective and without a political agenda.


The main findings of the study (completed in May 2013) are:

  • While the recovery has begun in Latvia, labor market demand has yet to rebound; unemployment is still high and labor market participation is lower than before the crisis.
  • Those with unstable or no work are from various groups; the 50+ population, youths with low education, and mothers with household responsibilities constitute a large share.
  • There is no evidence of large-scale dependence on benefits. Benefit programs are not generous, have low coverage, and generally act as a stop-gap for many and not a permanent income source. The tax and benefit system could be modified to be more generous for low-income households. In particular, a more gradual phase-out of benefits is recommended.
  • Spending on means-tested benefit programs is low.
  • Labor market programs appear to improve employment rates, though variations in outcomes exist between types of programs.

The study provided strong evidence of a lack of benefit dependency in Latvia, which was widely reported in the media and used by the Government to argue for the need for anti-poverty programs—a new development.


Bank Group Contribution

Latvia graduated from the World Bank in 2007, but re-accessed financing on an exceptional basis following the 2009 crisis. As part of this support, the World Bank provided a €200 million programmatic loan to protect vulnerable groups with emergency safety net support during the economic contraction. The study focused on permanent policies to protect low-income and vulnerable groups as the economy recovered.

The activity was structured as a reimbursable advisory service (RAS). The Government of Latvia provided most of the funds through its European Social Fund allocation (lats 200,000), while the World Bank provided upfront funds to set up the project (US$30,000).


The study involved the social protection, fiscal policy, education and training, and health sectors. Although the study was requested by the Ministry of Welfare and the State Employment Agency, the World Bank worked closely with a steering committee made up of technical staff from the Ministries of Finance, Health, Economy, and Education, and the Central Statistical Bureau.

The European Commission (EC) and International Monetary Fund (IMF) were key external partners. The IMF was involved in the dialogue with the Government and provided a fiscal economist for the team. The EC provided funding for the study and is using the analysis as background for dialogue and planning for the next European Social Fund programming round (2014–19).

Moving Forward

The World Bank is currently concluding a new RAS agreement with Latvia for a follow-on study of promoting active aging. Latvia’s population is aging quickly and the study found significant early labor market exit or low-quality employment among the older population. The World Bank will work with the Government on policy options for taxes and benefits, health policy, training and employment programs, and labor market legislation to keep older people working in rewarding jobs.

Other European Union countries and the EC have expressed interest in replicating the Latvia study. Sharing the methodology and approach began with the other Baltic countries—Estonia and Lithuania—during country visits in March 2014.