Tim Carney has a terrific column on the legislative history that led up to the prescription drug kerfuffle I read about last week.  It makes something clear that I found hard to explain during the back and forth with commenters (as I said to Tim last week, this is the most semantically complicated, and yet least actually complicated, issue I've tried to explain in a long while). 

Say you have a company that in 2002 was providing drug benefits to its retirees at a hypothetical cost of $2500* per retiree.  Each $2500 you pay reduces the taxes you owe by about a third of that amount, so the actual cost of the subsidy to the corporation is about $1700.

In 2003, the Congress passes a Medicare prescription drug benefit.  Worried that corporations would drop their benefits and stick the government with the tab, they offer a subsidy for firms that continue their retiree drug benefits.  That subsidy, 28% of the total benefit cost (up to a cap), costs the taxpayer on average much less than providing the drug benefits, so it's a good deal for the taxpayer.

So now the cost to the company for these benefits is $1700-700 or about $1000.  The cost to the taxpayer is $2500-1000, or about $1500--still slightly less than Medicare pays for the average Part D beneficiary, but for much more generous benefits.  One can argue about the economic distortion, but it's not your traditional "corporate giveaway".  If it's a giveaway to anyone, it's to corporate retirees.

However, had the Congress structured it the way that is now mandated by the new reform, the company would minimize its tax bill by about $600 instead of $800.  The cost of the drug benefits to the company would be $1200; the cost to the taxpayer, about $1300.

So this really is an instance of giving the subsidy, and taxing part of it back, because the company was already getting the standard tax subsidy for its retirement benefit.  On the one hand, you're giving them a new subsidy; on the other hand, you're taking away part of the old subsidy.

Now, that doesn't tell you whether this was the right thing to do; I don't know how many companies will drop their now-more-expensive retiree benefits as a result, which is the relevant question.  But I feel like a lot of people were struggling with the notion that this was the equivalent of taxing the subsidy, so I thought it worth highlighting.

*This is an arbitrary number mostly chosen because the math is easy.  Please do not attempt to engage me in a discussion of whether or not this corresponds with real-world averages.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.