Medicare beneficiaries are by definition elderly, disabled, or patients with end-stage renal disease. Private insurance beneficiaries may include a small percentage of people in those categories, but they consist primarily of people are who under age 65 and not disabled.
Naturally, Medicare beneficiaries need, on average, more health care services than those who are privately insured. Yet the bulk of administrative costs are incurred on a fixed program-level or a per-beneficiary basis. Expressing administrative costs as a percentage of total costs makes Medicare's administrative costs appear lower not because Medicare is necessarily more efficient but merely because its administrative costs are spread over a larger base of actual health care costs. When administrative costs are compared on a per-person basis, the picture changes. In 2005, Medicare's administrative costs were $509 per primary beneficiary, compared to private-sector administrative costs of $453.
I don’t know all the methodological details of Book's analysis (and I'm skeptical of any study produced by Heritage), but his essential point should extend well beyond the health care debate. The question of how to compare unique data sets is often very complicated, even in the absence of politics. Basic considerations -- like whether to measure in terms of absolute numbers or proportions -- are often controversial. We should be particularly wary of unqualified statistical comparisons presented in the context of an emotionally charged policy debate.
They’re often very misleading.