The incongruous result is that workers in some of America’s best-known and most profitable companies are deriving benefits from safety-net programs. According to the Dayton Daily News, 117,890 people working at one of the fifty largest firms in Ohio were receiving food stamps, and 141,182 were on Medicaid in February 2013. The list of employers with large numbers of workers receiving benefits was dominated by restaurant chains (McDonald’s, Burger King, Bob Evans and Wendy’s), large grocery chains (Kroger) and Walmart. Nearly 30 percent of Walmart employees received food stamps, according to the report.

Stagnating incomes at the bottom also make federal programs conditioned on nonwork more attractive. The Social Security Disability Insurance (SSDI) program is a case in point. SSDI was created in 1956 to ensure that Americans who could not work because of serious disability would not become poor. Over the years, the definition of serious disability was expanded so that it now includes general complaints of pain and mental issues such as depression. Vocational factors such as the applicant’s ability to get a job given one’s age, education and work history are also part of the disability determination. The result, combined with the economic difficulties facing low-skilled workers, has been dramatic.

SSDI caseloads have exploded in the last twenty years, and especially in the last decade. In 1989, about three million Americans received SSDI. Today, it’s almost ten million, and about eleven million if you include children and spouses of disabled workers. No one who looks at the program believes the large spike in caseloads is because America’s workplaces have become less safe. Instead, most observers contend that it is because the expanded definitions of disability have intersected with declining opportunities for good jobs for low-skilled, older workers to give rise to a massive temptation for those people.

Consider the case of a man in his fifties whose factory closes. SSDI offers him a steady check of about $1,000 a month. He is also eligible for Medicare if he has been disabled for two years, and the long backlog to process SSDI claims means most applicants can expect to receive Medicare immediately upon enrollment. So, if this worker says, “I can work,” he has to struggle to find a job that may not pay more after tax than the $1,000 a month and may not offer health insurance. If he says, “I can’t,” he gets a low but secure payment for life. And he can qualify based on very subjective, difficult-to-disprove criteria that take his lack of job opportunities into account.

Given this, it should be no surprise that SSDI applications soared in the last decade, even prior to the Great Recession. While SSDI applications have risen and fallen based on the unemployment rate for some time, annual applications remained below two million until 2001. They rose by almost 50 percent between 2001 and 2004, and remained at historically record levels until the Great Recession. They rose another 33 percent between 2008 and 2010 to hit nearly three million. Nearly 1.75 million people have been added on net to the SSDI rolls since 2008. Many observers have noted that labor-force-participation rates are at thirty-year lows; the opportunity to do as well or better on SSDI than in the economy surely has had an important impact on that.

The cost for the American taxpayer for this is huge and rising. SSDI benefits alone were $140 billion in 2013, or a bit less than 1 percent of SSDI. Medicaid payments for disabled people, which mainly include people receiving benefits from another disability program, Supplemental Security Income (SSI), were another $167 billion (at both the state and federal levels) in fiscal year 2012. Medicare spent yet another $68 billion on disabled, nonelderly recipients in 2012, more than double what was spent in 2001. And SSI, which serves mainly disabled children, the elderly and the blind, spent another $56.2 billion in fiscal year 2014. Together, this means that taxpayers pay at least $441 billion in benefits or services for disabled people, or over 2.5 percent of GDP.

This growing reliance on the safety net is likely to increase dramatically under Obamacare. By chance or design, that measure’s key features are most directly beneficial to working Americans with below-median incomes. The Medicaid expansion allows states to extend coverage for all adults in households making less than 138 percent of the poverty level, or $32,918 for a family of four. Households making above that level can get health insurance through state or federal exchanges. If these households choose the cheapest, “bronze” plan, they would receive family health insurance without having to pay any premiums. This is big money for these families. According to the Henry J. Kaiser Family Foundation’s online subsidy calculator, a family of four in Marietta, Ohio, making just above the cutoff for Medicaid eligibility would receive over $7,000 in tax credits for a bronze plan.

This tremendous subsidy is available even for households earning much more. The Kaiser website tells us that this Marietta family would still receive a free bronze plan if it earned up to $53,000, slightly higher than the national median income of $51,000. Even a family of four earning $65,000 would pay only about $1,900 for a bronze plan, receiving over $5,100 in federal subsidies.

Other exchanges in political swing states are not as generous, but in Iowa (Dubuque), Wisconsin (Wausau), Colorado (Pueblo) and Nevada (North Las Vegas), bronze plans are free or cost less than $150 a year for families of four earning up to $39,000 annually.

It is easy to see why the safety net and Obamacare are enormously important for low-skilled Americans making below the median income. Nationally, these people are still outnumbered by those who are making more than $65,000 a year and the elderly. What makes this a particularly difficult political challenge for conservatives to address is the fact that the very states they need to gain to win back the presidency benefit disproportionally from these programs.

THE LIST of battleground states hasn’t changed much in the past twenty years, and is unlikely to change much before 2016. That means the next Republican nominee must pick up the electoral votes needed to win from a small group of states: Florida, Virginia, Ohio, Iowa, Colorado, Wisconsin, New Hampshire and Nevada. Virginia is a special case because of the large number of highly educated Northern Virginia voters attracted to Washington, DC, by the federal government. In each other case, GOP hopes rest on persuading larger numbers of white or Hispanic working-class voters to support the party’s nominee.

Colorado, Florida and Nevada all contain large, Democratic-leaning Hispanic populations that provided President Obama with his victory margins in 2008 and 2012. In 2012, their electorates were each between 14 and 19 percent Hispanic, and the president took between 60 (Florida) and 75 (Colorado) percent of those votes. The Hispanic population in each state continues to grow and in 2016 their share of the vote will be even higher.

Exit polls show that the growth or change in voting preference of the Hispanic community was the sole reason Mitt Romney lost each state. We can see this by comparing the share of the white vote obtained by both George W. Bush, who carried each of these states, and Romney. In 2004, Bush received 57 percent of the white vote in Colorado, 55 percent in Nevada and 57 percent in Florida. Romney received 54 percent in Colorado, 56 percent in Nevada and 61 percent in Florida. However, in each state the share of whites as a percentage of the electorate declined—by 3 percent in Florida, 8 percent in Colorado and a huge 13 percent in Nevada.

That decline was matched by large increases in each state’s Hispanic population. The share of the Hispanic vote in Colorado rose from 8 percent in 2004 to 14 percent in 2012; in Nevada from 10 percent to 19 percent; and in Florida from 15 percent to 17 percent. It’s true that Romney did much worse among Hispanics in these states and nationwide than did Bush, but even if Romney had matched Bush’s showing he would only have picked up Florida. By nearly doubling in size in only eight years, the Hispanic vote in Colorado and Nevada turned what had been reliably Republican states for decades into Democratic-leaning outposts.

Hispanic voters are overwhelmingly non-college-educated and earn below the median income. According to the Census Bureau, only 13 percent of Hispanics had a bachelor’s degree or higher in 2011. Thirty percent had no health insurance, and the Hispanic median income was only a shade over $38,000, significantly below the U.S. national median. Opinion surveys also show that Hispanics are much likelier than whites to support larger government and more services over smaller government and lower taxes. Any conservative effort to win back Colorado and Nevada must take notice of these facts.

A conservative GOP nominee may seek to counter these trends by relying on the swing states of the Northeast and Midwest. These states have few Hispanics and, except for Ohio, have much lower percentages of African Americans and Asians than the rest of the country. They are dominated, however, by blue-collar voters that are not evangelicals. Exit polls from 2008 show that whites without a college degree were a majority of the electorate in Ohio, Wisconsin and Iowa, and 45 percent in New Hampshire. This was much higher than their 39 percent national share. Moreover, blue-collar whites in these states are much likelier to be Catholic or Lutheran than they are nationally. Accordingly, President Obama carried the white working class in Wisconsin, Iowa and New Hampshire in 2008, losing it only in Ohio, where a larger percentage is neither Catholic nor Lutheran.