There are few issues more politically contentious than tax reform. Any government effort to increase domestic revenues risks stirring up opposition. In the years leading up to the COVID-19 pandemic, high fiscal deficits and already high and rising debt levels made improving domestic revenue mobilization (DRM) a priority. It has become even more urgent with the drop in tax revenue due to the onset of the pandemic and growing demands for spending on everything from health sector capacity building to climate change mitigation.

According to most recent World Bank’s Global Economic Prospects report, “substantial revenue gains could be generated in low-income countries (LICs) by measures such as broadening tax bases, strengthening tax administration, and enhancing efficiencies.” However, revenue collection lags government needs in developing and emerging economies because of narrow tax bases, large informal sectors, and the often-difficult political economy of tax reform.  The challenges are even greater in LICs which face capacity constraints on tax administration.

So how can the World Bank most effectively support countries in managing the difficult terrain of tax reform?

To help answer this question, the latest Evaluation Insight Note (EIN) reviewed the results of closed World Bank projects and operations in support of domestic revenue mobilization (DRM) in five countries, and one sub-national state - Croatia, Guatemala, Liberia, Pakistan, Panama, and the state of Rio Janeiro in Brazil - to draw from existing evidence new insights on what worked well and less well, and under what conditions.

This EIN aims to contribute to the Bank’s goal of supporting DRM in client countries, which has become an increasingly important part of the World Bank Group’s global and country-level policy reform agendas.

The EIN has four main insights:

Political economy and timing

Across all the interventions reviewed, political economy constraints were the most frequent challenges to the sustained achievement of DRM reforms. Some interventions aimed to leverage changes in the country context that appeared to open a political window for reform. The World Bank launched a DRM project in Croatia to support the country’s bid to join the European Union, and a new government in Pakistan led to the approval of a 2015 Development Policy Credit (DPC) to support tax reforms. Yet the problem with relying on changes to the context, is that they can change back again, removing the conditions that supported reforms.

Pakistan was (at least initially) the most successful of the interventions assessed, and the budget support operation was the World Bank’s first policy-based loan to Pakistan in more than a decade. But underlying conditions eventually changed, and tax reforms stalled in the face of opposition from interest groups. The World Bank moved fast in Croatia, but limited time for stakeholder consultation and project preparation led to problems. Subsequent frequent changes in government also meant a lack of steady ownership of the reform process. While a crisis or a sudden change may create an opportunity to launch reforms, they are not enough to sustain momentum over the long term.

Design and combination of interventions

Reforms in Development Policy Operations (DPOs) often benefitted from being paired with investment projects or technical support. A stand-alone DPO in Rio de Janeiro, for example, was hampered by the weaknesses and lack of capacity in the tax administration of the state. Pairing it with technical assistance, perhaps as part of an investment project anchored on an institutional partnership with state implementing agencies, might have helped.

Support for Pakistan, on the other hand, had a combination of interventions that included programmatic budget support paired with adaptable, comprehensive technical assistance.  In Guatemala, the effectiveness of the tax reforms depended on the ability of the tax administration to apply and enforce the new provisions. While program documents envisaged donor support to strengthen tax administration as mitigating measures, IEG’s Project Performance Assessment Report for this project found little evidence of targeted technical assistance, and the reforms struggled against political economy constraints.

Partnerships and results

While coordination with development partners did support results in some of the cases reviewed, it showed little impact in others. This suggests that, on its own, partners coalescing in relation to an agenda may not be enough to deliver results.

In Pakistan, the World Bank’s development policy operation’s prior actions were more effective when they were coordinated with conditionality in an IMF program and with technical assistance from the UK’s then Department for International Development. While in Panama, there was supportive collaboration with the Inter-American Development Bank and IMF, but it did not prove sufficiently catalytic for lasting impact.

Results Frameworks

A well-articulated and designed results framework is critical to guide implementation, track the progress on DRM, and supporting learning. The EIN finds that the DRM-relevant results indicators sometimes had limited links to the specific reforms or prior actions supported. Indicators related to tax were often too high level, limiting their use in monitoring the impact of the specific reforms supported by the operation or project.

For example, a project In Liberia to support fiscal policy management included revenue mobilization and administration as a subcomponent. Results indicators for the associated objective focused on multiyear fiscal planning, expenditure policy, and budgeting, but did not track progress with respect to revenue mobilization and administration.

The findings of this EIN also informed the IEG’s recently released evaluation of World Bank Support for Domestic Revenue Mobilization.