When Congress created Health Savings Accounts in 2003, the underlying theory was elegant: give consumers financial skin in the game, and they will shop for healthcare the way they shop for anything else. Prices would become transparent. Unnecessary care would decline. The system would grow more efficient.

Two decades later, HSAs paired with high-deductible health plans cover more than 30 million Americans. The data is now rich enough to ask a sharper question. Did consumers actually behave the way policymakers predicted, and if so, at what cost?

The answer is more complicated than either side of the original debate anticipated. HSAs did change behavior, but not always in the ways economists modeled. They shifted spending patterns, reshaped wealth accumulation, and quietly redistributed tax benefits across income brackets. Understanding these effects matters because HSAs are no longer a niche product. They are increasingly the default architecture of American private insurance.

HSA enrollment has grown steadily, but the growth has been uneven across the population. Higher-income households and older workers are significantly more likely to open accounts and contribute the maximum. Younger and lower-wage workers often enroll in HDHPs because their employers offer no other option, but they contribute little or nothing to the paired account.

This pattern reveals something important about how HSAs function in practice. For a household with disposable income, an HSA operates as a triple-tax-advantaged investment vehicle: contributions are pre-tax, growth is untaxed, and withdrawals for qualified medical expenses are also untaxed. Some financial advisors now recommend maxing out an HSA before contributing to a Roth IRA.

For households living paycheck to paycheck, the calculation looks entirely different. The high deductible arrives whether or not the account is funded, and setting aside money for future medical costs competes directly with rent and groceries. The result is a bifurcation: affluent enrollees accumulate substantial balances, while lower-income enrollees face the exposure of a high deductible without the cushion the account was meant to provide.

Aggregate balance data confirms the divide. Average account balances have grown substantially, but median balances tell a different story, and a significant share of accounts hold less than $500. The policy created two very different products depending on who was using it.

Takeaway

A single financial instrument can function as a wealth-building tool for some and a source of financial strain for others. The mechanism is identical; the lived experience is not.

The theory behind HDHP-HSA pairing assumed consumers would become more discerning purchasers of care. Faced with real prices, they would skip low-value services while continuing to seek necessary ones. The evidence tells a more troubling story.

Multiple studies, including analyses of employers who switched their entire workforce to HDHPs, show that enrollees do reduce spending, often by 15 percent or more in the first year. But the reductions are not selective. Consumers cut back on preventive screenings, medication adherence, and chronic disease management alongside genuinely discretionary care. Patients with diabetes reduce insulin refills. Women delay mammograms. Emergency visits for conditions that could have been managed earlier tend to rise.

Part of the problem is informational. Healthcare prices remain opaque even to motivated shoppers, and few patients can distinguish a necessary MRI from an unnecessary one at the moment of decision. Part is behavioral. When cost imposes friction, people postpone decisions across the board rather than triaging carefully.

The Affordable Care Act's requirement that preventive services be covered without cost-sharing was designed to address exactly this failure mode, but the carve-out is narrow and poorly understood. Many enrollees still assume anything before the deductible costs them money, and act accordingly.

Takeaway

Cost-sharing is a blunt instrument. It reduces spending, but it cannot distinguish between the care people should skip and the care they will regret skipping.

HSAs are subsidized through the tax code, which means their benefits are worth more to households in higher tax brackets. A dollar contributed to an HSA saves 37 cents for a top-bracket household but only 12 cents for a household in the 12 percent bracket. This structural feature has predictable distributional consequences.

Congressional Budget Office and Joint Committee on Taxation analyses estimate that the majority of HSA tax expenditures accrue to households in the top income quintile. The pattern intensifies over time as balances compound. Because unused HSA funds roll over indefinitely and can be invested, long-tenured account holders effectively build a shadow retirement account funded by public subsidy.

This is not incidental to the policy design. It is a direct consequence of using tax exclusions rather than refundable credits or direct subsidies. The same architecture appears in the employer-sponsored insurance exclusion, the mortgage interest deduction, and the treatment of capital gains. HSAs are simply one more example of a subsidy delivered through the tax code, which almost inevitably tilts toward those with the largest tax liability.

The equity question is not whether HSAs help anyone. Clearly they help many families manage healthcare costs. The question is whether public dollars spent this way could produce more health benefit per dollar if deployed differently, particularly for the households currently exposed to high deductibles without the means to fund an account.

Takeaway

How a subsidy is delivered matters as much as its stated purpose. Tax-based benefits redistribute upward almost by default, regardless of the underlying policy goal.

Health Savings Accounts did what economists predicted in one narrow sense: they made consumers more price-sensitive. But price sensitivity is not the same as informed decision-making, and the reductions in care have not been confined to waste.

The broader lesson is that consumer-directed healthcare works best for people with resources, information, and predictable health needs. It works least well for those with chronic conditions, tight budgets, or complex medical situations. That is a significant portion of the population.

As HSAs continue expanding, the policy question is no longer whether they change behavior. It is whether that behavioral change, and its distribution across the income spectrum, aligns with what we want a health system to accomplish.