What a difference a few months can make in politics. A poll released at the end of April finds 53 percent of Virginia voters oppose expanding Medicaid to 400,000 additional low-income Virginians. That’s in serious contrast with February, when 56 percent supported the expansion. While we can all agree that low-income Virginians deserve quality healthcare, the question remains: Does Medicaid expansion provide the answer?

In fact, Medicaid expansion would be politically and fiscally unwise. Advocates argue that it wouldn’t cost Virginia any money and that not expanding would cost $2 billion in foregone money from the federal government, but these claims do not withstand scrutiny.

It is true that the federal government will cover 100 percent of the cost at first (though the state would still have to foot the bill for administrative costs, which represent an average of 5.5 percent on top of total — federal and state — benefit costs). But after three years, the federal government will reduce its subsidy to 90 percent of costs — assuming it keeps its commitment over time in the face of its own financial problems.

In addition, an aspect often omitted from these cost estimates is the fact that among the new people enrolling for Medicaid under the new push, many were previously eligible but simply never enrolled. For these people, the old and much less generous formula applies: Virginia foots 50 percent of the bill on top of administrative costs.

As for the $2 billion of federal funds, I have bad news for the state politicians: There are no free lunches. When economists study the question of the pros and cons of federal funding in state budgets, they find that it all comes with serious constraints as to how the money is spent, which reduces the states’ ability to solve its own problems. According to a May 2013 Congressional Budget Office report, this is particularly true for categorical formula grants, such as Medicaid.

My colleague at the Mercatus Center, Matthew Mitchell, wrote about how in 2010 Arizona was denied by the federal government the ability to discontinue non-emergency taxi service for Medicaid patients after the state showed how cost-ineffective the program was. However, around that same time the state’s Medicaid program was so cash-strapped that it stopped paying for certain transplant procedures. In other words, the federal government insisted on cab rides for Medicaid patients but was OK with depriving them of life-saving procedures. Thankfully, the state has since restored the funding.

Federal money comes with costly strings attached that are never accounted for in the original cost estimates. Take the No Child Left Behind Act, for instance. The cost imposed on the states turned out to be gigantic: countless hours spent by state employees to obtain federal funding and millions of dollars in compliance burdens for local schools.

The same is true of Medicaid. Administrative costs are already a burden for states and they will only increase after the expansion. Medicaid also suffers from a massive overpayment problem (payments go to people who aren’t eligible or some eligible people get more than they should), and that will get worse.

Finally, the incentive structure in Medicaid leads to massive opportunity costs for other state programs since states allocate more and more of their budgets to Medicaid at the expense of other programs, such as cash transfers and transportation.

But Virginia should also think about what will happen if it expands Medicaid and the federal government doesn’t deliver on some part of its funding — as may well be the case if the federal government’s current spending trajectory continues — while the requirements remain permanent.

Using data from 50 states over a 13-year period, a 2010 paper by economists Russell Sobel and George Crowley found that every dollar in temporary grants from the federal government to state and local governments caused the latter to increase their own future taxes by between 33 and 42 cents.

What should states be doing instead? They should think about reforming their current Medicaid programs so they can better serve the truly poor in their states.

For one thing, Medicaid is actually a bad deal for the recipients themselves. Study after study, including the academically controlled experiment known as the Oregon study, has shown that the program often provides second-class care. Poor access and poor health outcomes are often the fate that awaits Medicaid beneficiaries — including greater reliance on emergency rooms and higher mortality rates.

Virginians should ask themselves if this expansion is worth the future cost to taxpayers and if it is fair to throw more low-income Americans into costly, substandard health care. As the polls make clear, Virginians are coming to understand that the answer is not Medicaid expansion, but reform.

Veronique de Rugy is a senior research fellow at the Mercatus Center at George Mason University in Fairfax. Her email is vderugy@mercatus.gmu.edu. This column first appeared in The Richmond Times-Dispatch.