While the level of overall health spending is by no means the only determinant of life expectancy, our analysis suggests it has been an important factor in the last 20 years.
Last week, the UK Government announced that many NHS workers would receive a pay rise. It follows years of little, if any, growth in public sector pay which has been a fundamental part of so called austerity policies. These policies, which were intended to reduce the budget deficit by a combination of tax increases and spending cuts, have defined UK and European economic policy for almost a decade. While we recently blogged about the challenges of staffing the health service, and have written about the links between austerity policies and health outcomes for a major report in 2017, this article asks a more fundamental question – does more health spending equate to longer lives. To answer this, we explore two issues at a cross national level using the most up to date OECD data:
- The evolution of health spending and life expectancy across developed countries.
- The statistical relationship between health spending and life expectancy in those countries.
What’s happened to health spending and life expectancy?
The year 2000 is as good a place to start as any. Between the turn of the Millennium and 2015, the broad trend was a gradual increase in health spending and a gradual increase in life expectancy. But in latter years, most countries experienced a reversal or slowdown in the rate of improvement in life expectancy. For some, the reversal had been preceded by a slowdown (or fall in some cases) in health expenditure per head (see chart). A highly relevant question is therefore whether by curtailing health spending, countries are inadvertently reducing their capacity to deliver longer lives.
Source: OECD and author’s calculations
A bird’s eye view of health spending and life expectancy
Between 2000 and 2015, those countries that spent the most tended to have the highest lifespans. There are exceptions to this rule – notably the United States which spends more than anyone else but where life expectancy remains towards the lower end relative to other high income countries. Similarly, Japan has the highest life expectancy despite its spending being closer to the OECD average. The final thing to note is that health spending may suffer from diminishing returns – that is for every additional pound or dollar spent, there is a relatively smaller gain in life expectancy (hence the curved dotted line on the graph below).
Source: OECD and author’s calculations
Notes: Graph shows the relationship between average per head health spending (2000-2015) and average life expectancy (2000-2015).
So far so good. But while the above graph suggests that the overall level of spending matters, it doesn’t tell us anything about changes to health spending over time. In other words, to what extent do changes to health spending lead to increases/decreases in life expectancy. For this, we look at the statistical relationship between the change in health spending and the change in life expectancy over the same period. As the below chart shows, there is a weak positive relationship between spending increases and life expectancy increases.
But what about other determinants of life expectancy?
So far, the data suggests there might be a relationship between health expenditure and life expectancy – both in terms of the level of expenditure and (to a lesser degree) the rate of change. But there are likely to be a multitude of factors determining life expectancy other than health spending. Moreover, the level (and rate of change) in health expenditure might not reflect an active policy choice to boost the population’s health, but rather a function of a country’s age profile since older populations are likely to experience greater demand for healthcare.
In order to be a little bit more scientific, we create a panel dataset of OECD countries ranging 2000-2015 which gives us nearly 500 observations. From this, we construct a regression model whereby life expectancy is a function of health spending, the old age dependency ratio (OADR) and individual country specific characteristics. The model is a good fit for the data – explaining over 90% of the variation in life expectancy experienced across the OECD during this time period. More importantly for our purposes, health spending remains positively correlated with life expectancy (and statistically significant) even after controlling for these other factors (see table).
Regression output: summary
Final thoughts: While the level of overall health spending is by no means the only determinant of life expectancy, our analysis suggests it has been an important factor in the last 20 years. A statistical link between austerity policies and health outcomes might be impossible to prove in the short term, but in the long term prolonged increases/decreases in health spending will affect the longevity of populations.
Assistant Director of Research and Policy