Subsidizing private health insurance

12/02/2013 12:07:00 PM
Tweetable
They might save money by reducing the size of that pill.
Terence Chai Cheng has a new article making it's way to the pages of the Journal of Health Economics (there is an older ungated version here) using evidence from Australia to show that in countries with both a universal government-funded public health insurance and a private health insurance market, reducing subsidies (ie tax-exemption) for employer-sponsored private plans results in a net reduction in government healthcare spending. Now it may seem obvious that eliminating subsidies reduces government spending, but this is actually non-obvious because we must remember that these countries also have a government-funded universal healthcare system to fall back on--we might have hypothesized that the reduction in subsidies would be offset by increases in the cost of public health insurance.

That is more or less the line that some other papers studying "dual" systems have argued. From Chetty and Saez (2010), for example:
"Taking crowd-out into account lowers the estimate of G(b) by a factor of more than 100. An analyst who ignored crowd-out and applied existing formulas (e.g., Chetty 2006a) would infer that a $1 million expansion in public health insurance programs would generate $280,000 in surplus net of the required tax increase needed to finance the expansion. This analyst would mistakenly conclude that an expansion in the overall level of public health insurance would yield substantial welfare gains. Taking crowd-out into account implies that we are near the optimum in terms of aggregate public health insurance levels, as a $1 million across-the-board expansion would generate only $1,700 in net social surplus."
Now, Chetty and Saez are talking about the US, not Australia, and they are talking about measures social wellbeing, not government expenditures. Still I think their findings add some relevant perspective.

Cheng notes that there are two other countries--the UK and Spain--that have the same dual system with subsidized private insurance alongside universal public health insurance. But in a somewhat older paper, Amy Finklestein reminds us that Quebec was once also a member of the dual health systems club, with both a universal Medicare public health insurance, and a subsidy for employer-sponsored private health insurance. Now, the Canadian and Australian health systems are very different, so it isn't at all surprising that the elasticity of private insurance with respect to the public subsidies is wildly different in the two countries. What is somewhat surprising is that the elasticity was actually much a much bigger magnitude in Canada, where a 1% reduction in subsidies lead to a 0.5% reduction in private insurance coverage, than in Australia where it results in just 0.13% reduction in private insurance coverage. Finklestein does not do the public expenditure analysis that Cheng did (and I'm not sure "public expenditure" is necessarily the right metric to care about), but comparing these elasticities it certainly seems immanently plausible that the Canadian government saved money by reducing subsidies for private coverage.