 |
 |
Henry's response makes a series of common judgments that are highly suspect. The reasons they are suspect were provided within the string, I believe, but so it goes. By making this entirely a matter of theoretical assumptions, rather than engaging empirical reality, he allows himself to criticize people who disagree with him as biased. In typical Henry fashion, he asserts they do not have their "heads screwed on straight." I'd suggest we're trying to keep people from getting screwed by his thinking. I'll insert the issues within his text, and readers can judge for themselves. Best, Joe On Tue, Jan 12, 2010 at 5:47 PM, Adam Oliver <[log in to unmask]> wrote: > > > > ------------------------------ > > *From:* Henry Aaron [mailto:[log in to unmask]] > *Sent:* 12 January 2010 22:10 > *To:* [log in to unmask] > *Cc:* Oliver,AJ > *Subject:* > > > > Adam Oliver kindly shared the exchange regarding the excise tax debate here > in the United States. Although not a member of the US/UK group, I am going > to take the liberty of posing a question to U.S. opponents of the proposal > to rein in the income tax exclusion of health insurance with an excise tax. > > > > Let’s imagine a world in which everyone had the same health insurance plans > that they have today but that the value of such insurance is included in > income subject to the personal income tax. In this imaginary world, the > value of health insurance is not subject to payroll tax. Employers can (as > now) deduct the cost of such insurance as a legitimate business expense in > computing business income. > > > > To start with, I hope that we can all agree that the value of health > insurance is a valuable form of compensation and that if one is imposing > income tax on earnings, the value of such insurance is properly taxable. > > All Agreed. But not the assumption Henry is making here: that employers > are providing a fixed amount of insurance value to each employee, and that > this can be divided up among employees into equal amounts (save for > adjustments for family size) and counted as if each employee received the > same fixed amount of extra compensation in the form of health benefits. > This is an extremely questionable assumption. First of all, we know it's not true in practice because no employer determines compensation based on family size. No employer is going to say, "well, Joe is single but Allen is married so we'll give greater total compensation to Allen." What's really going on is, the employer has decided to provide medical benefits, and the chips fall where they may in distribution among employees. The employer is treating medical benefits differently than it is treating wages. It is doing that partly because it could not treat benefits the same as wages because of the family size problem, but also because the employer is prohibited by law from treating benefits the same way as wages. An employer that gives one employee twice the wages of another cannot give one employee twice the health benefits of another. So the basic setup of Henry's thought-experiment is wrong. There is no reason to think that employers are consciously providing equal compensation to each employee. In order to assume that there is a tax benefit that can be calculated per employee, and that it is based on equal costs for all employees, you have to make unreasonable assumptions about what employers are really doing. And this matters because, as Tim and I keep saying, it is quite possible that employers are providing more cash value in health benefits to two groups of employees -- those with lower incomes and those with larger families -- than they would have provided in cash if they only paid wages. If we are right -- and we're surely right about family size and I think the only question about income levels is the size of the effect -- then there is a transfer WITHIN EMPLOYERS to start with. > A group of left-leaning progressives comes along and proposes two > alternative plans to modify this arrangement. You are asked to choose > between them. > > > > Plan 1: every person who heretofore has had to pay income tax on the value > of his/her insurance will henceforth be excused from that tax. The savings, > of course, will be equal to the value of insurance, which will be excluded > from income under the new arrangements, multiplied by each filer’s marginal > income tax rate. The cost of this plan will run about $150 billion a year. > The tax cut will be highly regressive because the value of the exclusion > rises with tax rates (which rise with income) and the value of insurance > (which also rises with income, on the average). > > Here is assumption 2: that the way to think about the tax is in terms of > marginal rates, rather than the value of the tax savings relative to the > level of income. The point of the Commonwealth study is that the value of > the insurance is a much greater share of income for low-income workers than > for high-income workers. Therefore the actual effect of the tax break on > taxes paid as a proportion of income is somewhat favorable to lower-income > workers compared to higher-income workers. Essentially, the higher share of > income cancels out the lower rates. Now, Henry may have taken care of this > by increasing the rate effect, which he does by excluding payroll tax from > his setup. But the bills in question do not excluse the payroll tax. > Bottom line: No, the tax cut is not regressive, by standard uses of the term > "regressive." You can define it the way Henry does here, but it's by no > means obvious. To put this another way, a bigger new tax expenditure on > some people than on others does not necessarily make the overall tax code > less progressive. Does this mean I even like the tax expenditure? No. But > it means the equity argument against the tax expenditure is much less > obvious than lots of economists think. > > Plan 2: the tax will be repealed on the first $21,000 of insurance for > families and $8,000 for individuals, but retained to the extent that health > insurance costs more than these amounts. The cost of plan 2 is a small > fraction of plan 1 (I don’t have the numbers at hand). > > > > Plan 1 is clearly less progressive in its distributional effects than is > plan 2, although the pattern is not neat or precise. Plan 2 is more > progressive because the average generosity of insurance tends to rise with > income as, of course, do marginal tax rates. To be sure, some costly plans > are used by lower income people, but the overall pattern is clear. > Yes, the overall pattern is quite clear: most of the variation in costs of insurance is due to variations in cost of care, not due to variations in generosity of benefits. I am sure there is some general trend for more extensive benefits for higher-income workers. And that must be especially true over the general population, simply because many more lower-income workers don't have insurance now, or are under-insured in the small-group or individual markets. But we are talking here about variation within the employer-sponsored market that does not include the currently uninsured or individually- or small-group insured. So I find it hard to believe -- though I don't have data -- that, within this population, the cost of plans is determined more by the average income of workers within a given employer's pool, than by the risk-profile and location of those workers. If Henry has data that shows income variations within the groups of those currently insured by employers with 50 or more employees are a more important source of cost, I'd like to see it. But I haven't found any insurance experts who've suggested that's what is going on here. > > > I believe that anyone interested in the well-being of low- and moderate > income households who has his/her head screwed on straight would flee plan 1 > and regard plan 2 as clearly superior. This is especially the case when one > recognizes that either other taxes would have to be increased or other > spending would have to be cut to cover the revenue reduction of plan 1. > After all, revenues have to approximately equal expenditures over time if > one includes the costs of debt service. > Note what Henry is doing here: he's simply ignoring the question of why the most expensive plans are most expensive. He is also ignoring the simple solution that Tim and I and others have been proposing, which addresses legitimate issues. If (as most of the people who look at it believe) most of the more expensive plans are more expensive because their members have greater needs (either due to health risk or more expensive location), then the excise tax approach targets the pressure to reduce benefits on the more needy groups of enrollees. As a matter of risk pooling, this makes no sense at all. If you really believe that there are excessive benefits out there, simply define what is excessive and only allow the tax exclusion for total benefit packages below that amount. If more generous benefit packages are in fact related to higher incomes, that will address Henry's concern. But it will not catch sicker groups along the way. Note something else Henry is doing here: ignoring completely the fact that the excise tax is designed to apply to more and more people over time. Basically, the threshold increases at the rate of inflation plus 1%. Health care costs to employers have been rising at about inflation plus 3.5%. CBO is assuming they will come down, but not to inflation plus 1%. That is why CBO assumes more and more peoples' plans will become subject to the excise tax, and so it will cause more and more scaling back of plans, so transference of health benefits into wages and taxes, over time. As it is designed, if other cost controls were effective (lord knows for what reason), the excise tax would increase federal revenues (and so pay for reform) less than projected. But if cost increases followed recent trends, the tax would raise MORE money than CBO has projected. So it is an automatic mechanism to respond to cost increases by cutting benefits. Gee, I wonder why that would make anyone nervous... > > > The current arrangements are ones that none on the left would endorse if > they were proposed anew. The excise tax is, admittedly, not the best way to > deal with the problem. The best way, actually, is close to that proposed by > John McCain—get rid of the exclusion entirely and give people cashable > (refundable) tax credits (unlike what McCain proposed, the credits should be > inversely related to income). But the question now is whether the excise > tax helps some or not. > > And we're back to the master assumption, which is that if employers cut > benefits they'd pay the money out in wages, and in equal amounts of extra > wages to even higher-and lower-income workers, except (I guess) that they > would pay extra to the workers with families (who will have to pay much more > for their insurance, and the difference in the size of the tax credits was > not enough to make up for that). > > The reactions of those on the left who oppose the roll-back of the tax > exclusion is, to my mind, a bizarre example of status-quo bias. > And the reaction of the supporters of the excise tax is, to my mind, a totally typical example of letting theoretical perspectives obscure the balance of evidence about likely actual effects. > > > > > *Henry J. Aaron* > > Bruce and Virginia MacLaury Senior Fellow > > The Brookings Institution > > 1775 Massachusetts Avenue N.W. > > Washington, D.C. 20036 > > Phone: (202) 797-6128 FAX: (202) 797-6181 > > > > Please access the attached hyperlink for an important electronic > communications disclaimer: > http://www.lse.ac.uk/collections/planningAndCorporatePolicy/legalandComplianceTeam/legal/disclaimer.htm >