I didn't very much like the Wall Street Journal op-ed by George Schulz and his colleagues at the Hoover Institution this Monday, but I did appreciate that it discussed the problem of poverty traps. This is the phenomenon of poor and near-poor people facing very high marginal tax rates due to the fact that they lose government benefits as they earn more income.

This is a real problem that can keep people from climbing out of poverty, and it's also a reason to be wary about a solution that conservatives often propose to reduce the cost of entitlements: means tests.

Means tests save money by reducing the generosity of public benefits for wealthier people. They are often proposed as a common-sense idea: Why should the government send Social Security checks to people who are already rich? But means tests worsen poverty traps by penalizing people for having more earned income or assets.

Take proposals to increase means-testing in Social Security. For example, some conservatives have proposed offering more generous benefit increases to poor people than to wealthy people. This would save money, but it would also be an effective tax-rate increase -- earning more money would have a negative implication for your Social Security benefits.

Currently, the payroll tax that funds Social Security is one of the least-economically damaging taxes, partly because workers know that added taxes they pay now are partly offset by added benefits in retirement. More means-testing would weaken that link, making the payroll tax into more of an ordinary tax on labor and discouraging work.

Similarly, Medicare already has a modest level of means-testing: Premiums for Medicare Part B, which covers outpatient care, are linked to income. Proposals to contain Medicare cost growth often contemplate further premium increases for higher-income seniors. This would save money and might be a good idea, but the cost savings must be weighed against the fact that this would be an effective tax on retirement saving.

These trade-offs can't be totally avoided. Universal entitlements, with no means-testing, are drastically more expensive than programs targeted only at the needy. But as Schulz and his co-authors note, people with low to moderate incomes can already face marginal tax rates over 50 percent, which makes me wary of adding more means tests that would further raise those tax rates.

If we hope to reduce poverty traps, there are roughly two available strategies. One is to actually reduce the level of means-testing in entitlement programs, spreading the phaseout of benefits over wider ranges of income so that effective tax rates are lower.

Roughly, this is what the Patient Protection and Affordable Care Act does, by offering subsidies for health insurance on a sliding scale. Today, if you are poor and you start to earn more income, you may suddenly fall off a benefit cliff where you lose access to Medicaid. Under the act, the loss of benefits will be much more gradual, which ideally should reduce incentives that keep poor people from working. The downside of this approach is that it is expensive -- and so while you reduce the implicit tax imposed by means-testing, you must replace it with an explicit tax to fund the expanded benefits.

The other option, advanced by Schulz and his colleagues, is to reduce the basic level of benefits, so the phasedown starts from a smaller benefit value. This approach saves taxpayers money and makes poverty traps less severe, but with the major downside of reducing the standard of living of the neediest people. The key problem with the Schulz op-ed is its failure to acknowledge this trade-off.

The happiest solution to the poverty-trap problem would be to achieve an overall reduction in health-care costs. Medical services in the U.S. cost drastically more than in other advanced countries without apparently higher quality. Bringing those costs into line with our peer countries would make it possible to reduce the cost of benefits like Medicare and Medicaid, which would reduce overall fiscal pressures and shrink the poverty traps created by means tests of these programs.

Unfortunately, the market-based reforms advanced by Schulz et al. are unlikely to produce such savings on their own. (The main reason health care is so expensive in the U.S. is that we impose fewer price controls than other countries.) But consumer-directed approaches like health-savings accounts that conservatives favor are not necessarily incompatible with top-down approaches to cost control like the ones favored by Democrats. They can be combined together to achieve more cost control than either approach would separately. A holistic approach to health-care cost control, with conservatives behaving like partners in the search for savings rather than concern trolls, could provide the path we need out of the poverty-trap problem.

(Josh Barro is lead writer for the Ticker. Follow him on Twitter.)