Financing universal health coverage--effects of alternative tax structures on public health systems: cross-national modelling in 89 low-income and middle-income countries

Abstract

Background: How to finance progress towards universal health coverage in low-income and middle-income countries is a subject of intense debate. We investigated how alternative tax systems affect the breadth, depth, and height of health system coverage.

Methods: We used cross-national longitudinal fixed effects models to assess the relationships between total and different types of tax revenue, health system coverage, and associated child and maternal health outcomes in 89 low-income and middle-income countries from 1995-2011.

Findings: Tax revenue was a major statistical determinant of progress towards universal health coverage. Each US$100 per capita per year of additional tax revenues corresponded to a yearly increase in government health spending of $9.86 (95% CI 3.92-15.8), adjusted for GDP per capita. This association was strong for taxes on capital gains, profits, and income ($16.7, 9.16 to 24.3), but not for consumption taxes on goods and services (-$4.37, -12.9 to 4.11). In countries with low tax revenues (<$1000 per capita per year), an additional $100 tax revenue per year substantially increased the proportion of births with a skilled attendant present by 6.74 percentage points (95% CI 0.87-12.6) and the extent of financial coverage by 11.4 percentage points (5.51-17.2). Consumption taxes, a more regressive form of taxation that might reduce the ability of the poor to afford essential goods, were associated with increased rates of post-neonatal mortality, infant mortality, and under-5 mortality rates. We did not detect these adverse associations with taxes on capital gains, profits, and income, which tend to be more progressive.

Interpretation: Increasing domestic tax revenues is integral to achieving universal health coverage, particularly in countries with low tax bases. Pro-poor taxes on profits and capital gains seem to support expanding health coverage without the adverse associations with health outcomes observed for higher consumption taxes. Progressive tax policies within a pro-poor framework might accelerate progress toward achieving major international health goals.

Funding: Commission of the European Communities (FP7-DEMETRIQ), the European Union's HRES grants, and the Wellcome Trust.

Figure 1

Tax revenues and health spending in low-income and middle-income countries, 2009 Source: World Bank Indicators. Excludes Botswana, Hungary, Seychelles, and St Vincent and Grenadines. Some countries are not labelled for sake of clarity. Tax revenue and public health spending is adjusted for inflation and purchasing power. PPP=purchasing power parity.

Figure 2

Tax revenues, GDP, and the breadth and depth of coverage in low-tax revenue countries Source: WHO, World Bank Indicators and International Labour Organization. All models estimated using ordinary least squares. Low tax revenue countries are those where revenues are less than $1000 per capita. Tax revenue and GDP are adjusted for purchasing power parity and inflation, per capita. GDP=gross domestic product. *Proportion of pregnancies. †Proportion of the population not exposed to severe financial costs. For full models see the appendix.

Figure 3

After correcting for health spending, the association of alternative tax regimes with child survival, 89 low-income and middle-income countries, 1995–2011 Source: World Bank Indicators and IHME. SEs are adjusted for repeated observations. All models correct for country-specific differences and time trends. The natural log of the dependent variable is used in these models. All models adjust for total public health spending and other tax revenue. *Adjusted for purchasing power parity and inflation, per capita. †Deaths per 1000 livebirths per year (before age of 1 month). ‡Deaths per 1000 livebirths per year (before the age of 1 year). §Deaths per 1000 livebirths per year (after the age of 1 year and under the age of 5 years); ¶Deaths per 1000 livebirths (under the age of 5 years). For full models see appendix.