At Economics 21, James Capretta has a short essay on how PPACA introduces a serious inequity in the extent to which insurance purchases by similar situated households are subsidized:

Under the new law, workers with incomes below 400 percent of the federal poverty line but above the level for Medicaid eligibility who get their coverage through the new state exchanges can get federally financed “premium subsidies.” These subsidies will be phased down as incomes rise but are quite substantial for persons at about twice the poverty rate or just above that level. At the same time, workers who get their insurance from their place of work also get an indirect federal subsidy because employer-paid premiums are excluded from workers’ taxable income.
The problem is that the subsidies inside the exchanges will far exceed the tax subsidy for employer-paid premiums at the lower end of the wage scale. How big is the gap? CBO quantified it for us in its recent analysis. For a family of four at twice the poverty rate ($50,000 in annual income in 2016), the cost of health insurance would be $11,300 less in the health exchanges than in employment-based coverage, largely because the federal subsidies in the exchanges far exceed the tax benefit for health insurance for families in low marginal tax brackets. For a family of four at three times the poverty rate, the cost of health insurance would be $3,000 less in the exchanges than in employment-based insurance, again largely due to the fact that the government would be paying a larger share of the costs for them in the exchanges.
So, if workers would be better off in the exchanges, why does CBO assume so few of them will migrate out of employer plans? CBO acknowledges that the numbers could be far higher than its 3 to 5 million estimate. As the analysis documents, there are 48 million people in households who today get their insurance through their place of work and have incomes above the level for Medicaid eligibility (138 percent of the federal poverty line) but below 300 percent of the federal poverty line. These are the people who would benefit substantially from getting dumped by the employer plans into the exchanges. Why does CBO assume that won’t happen?
In sum, CBO believes that other preexisting labor laws will make it very difficult for employers to selectively dump their workers into the exchanges. Due to non-discrimination rules, an employer can’t offer insurance just to high-wage workers. That means an employer that stopped offering insurance would have to dump higher-paid workers into the exchanges too. And if higher-wage workers stopped getting employer-paid health insurance, that would reduce the government’s deficit as these workers would get far less in exchange subsidies than they would lose in terms of forgone tax benefits for employer-paid premiums.
But what if CBO is wrong and employers are able to find ways around the existing rules to create a two-pronged approach to coverage, with high-wage workers staying in tax-preferred employer plans and low-wage workers migrating to the exchanges? In effect, that’s what CBO modeled (though with very modest assumptions) in what it called “scenario 4.” In that simulation, CBO assumes that 20 percent of those with income below 250 percent of the federal poverty line but above Medicaid eligibility would migrate from job-based plans to the exchanges. Net federal costs under this scenario would be $36 billion higher than CBO’s base case, even after assuming some higher federal revenue from the assumption of higher levels of taxable compensation. And that’s assuming just 20 percent of these households would find their way into the exchanges. What if the number is 50 percent? Or 75 percent? Then the hit to the budget would be far higher. [Emphasis added]

At the very least, it seems likely that firms will invest considerable effort in finding ways around non-discrimination rules while the federal government will have to be vigilant about monitoring and policing this behavior. Had we equalized the tax treatment of workers received employer-sponsored insurance and those who do not, we might have avoided this problem. Yet that would have been extremely politically difficult, both because of the Obama campaign’s opportunistic attacks on the McCain campaign’s proposal to do exactly that and an understandable fear that Republicans would launch new opportunistic attacks on any Democratic measure that would contribute to the unraveling of existing health insurance arrangements.