…because I’m an economist and a mom–that’s why!

Refutation by Redefinition: Feldstein’s Redo of TPC’s Analysis of the Romney Tax Plan

August 30th, 2012 . by economistmom

Earlier this week, Martin Feldstein, a Romney campaign economic adviser and Harvard professor, published this op-ed in the Wall Street Journal, critiquing the Tax Policy Center’s viral (thanks to President Obama) analysis of the implied distributional effects of Mitt Romney’s self-proclaimed-revenue-neutral tax reform plan.  If you recall, the Tax Policy Center’s analysis showed that it was mathematically impossible to cut marginal tax rates as much as Romney proposes, not increase capital income taxes, and broaden the tax base in a revenue neutral way, without the reform resulting in a shift of tax burdens away from the richest households and towards other households (the “non-rich” you might say)–in other words, a “regressive” distributional effect.

Feldstein decided to do the calculation for himself, looking into which tax expenditures he himself could find to reduce/broaden the tax base that would reverse the conclusion that the Romney plan would cut taxes for the rich and raise them on everyone else (…remember, this is relative to Obama’s tax proposals, not relative to current law).  He reports his discovery, which he characterizes as not just a critique of the TPC analysis, but an outright refutation (emphasis added):

The key question raised by the Romney plan’s critics is whether this revenue loss can be offset by broadening the tax base of high-income individuals. It is impossible to calculate the exact effects of the future reforms since Gov. Romney hasn’t specified what he would do. But refuting the Tax Policy Center’s assertions doesn’t require that. It only requires knowing if enough revenue could be raised from high-income taxpayers to cover the $186 billion cost.

The IRS data show that taxpayers with adjusted gross incomes over $100,000 (the top 21% of all taxpayers) made itemized deductions totaling $636 billion in 2009. Those high-income taxpayers paid marginal tax rates of 25% to 35%, with most $200,000-plus earners paying marginal rates of 33% or 35%.

And what do we get when we apply a 30% marginal tax rate to the $636 billion in itemized deductions? Extra revenue of $191 billion—more than enough to offset the revenue losses from the individual income tax cuts proposed by Gov. Romney.

In other words, Feldstein refutes that the Romney plan would raise taxes on the non-rich by redefining the non-rich.  Obama, the TPC, and I’ll bet Romney himself, don’t consider households in the $100,000 to $200,000 range the “rich.”  We know President Obama has always made the dividing line between the “rich” and the “middle class” somewhere in the $200K to $250K range.  Households in the $100K to $200K range are squarely within Obama’s definition of the middle class households who would never be subjected to any increase in tax burdens under Obama tax policy.  (By the way, those households also happen to be the households that tax policymakers often talk about as unfairly bearing the bulk of the burden of the alternative minimum tax, in contrast to the truly “rich”–say, millionaires–who are typically not on the AMT because their marginal tax rate puts their ordinary income tax burden above their broader-based AMT burden.)

So as the Tax Policy Center counter-responded today:

Writing in Wednesday’s Wall Street Journal, Romney economic adviser Martin Feldstein attempts to contradict our finding. Instead, his analysis actually confirms our central result. Under the stated assumptions in Feldstein’s article, taxpayers with income between $100,000 and $200,000 would pay an average of at least $2,000 more. (Feldstein uses a different income measure than we do – see technical note at end.)

Taxes would rise on families earning between $100,000 and $200,000 in Feldstein’s analysis because he considers a tax reform that would completely eliminate itemized deductions for taxpayers with income above $100,000. In 2009, taxpayers earning between $100,000 and $200,000 claimed more than half of these itemized deductions. Eliminating itemized deductions would raise more in taxes from people in this group than they would save from the rate reductions and other specified features of Governor Romney’s plan.

Gee, let’s repeat that Feldstein version/reinterpretation of the Romney plan (emphasis added):

a tax reform that would completely eliminate itemized deductions for taxpayers with income above $100,000. In 2009, taxpayers earning between $100,000 and $200,000 claimed more than half of these itemized deductions. Eliminating itemized deductions would raise more in taxes from people in this group than they would save from the rate reductions and other specified features of Governor Romney’s plan.

One has to wonder:  did the Romney campaign really want Feldstein to “refute” the TPC analysis of the Romney tax plan this way–in effect spelling out that it’s “just” the $100K to $200K households that might get socked with the burden of paying for the net tax cuts for the above $200K households?

I don’t get it.  But that’s probably why I’m not cut out to ever advise a political campaign. (I think I would have said “keep this quiet.”)

There are other, less-fundamental problems about Feldstein’s analysis including his use of 2009 tax year data (an unusually low-revenue year) which you can read more about in the same TPC blog post.

18 Responses to “Refutation by Redefinition: Feldstein’s Redo of TPC’s Analysis of the Romney Tax Plan”

  1. comment number 1 by: Patrick R. Sullivan

    ‘Households in the $100K to $200K range are squarely within Obama’s definition of the middle class households who would never be subjected to any increase in tax burdens under Obama tax policy. ‘

    Isn’t it about INDIVIDUALS with incomes over $100k, and households of $200K or higher?

    But, anyway, the latter claim has already shown to be untrue with the revelation (via the Supreme Court) that the ACA was in fact a tax, contrary to Obama’s specific denials.

  2. comment number 2 by: Brooks / Gordon

    Feldstein responds to critics at

    I’d like to see response from TPC and others re: the reasonableness (1) of Feldstein’s revised set of assumptions (and data and related calculations) and (2) of his justification of $100,000 as the cut-off point for “middle class”, and thus (3) of his conclusion.

    I’ll be checking TPC’s site for a response to Feldstein’s response, but if anyone has links to other such responses, please share.

  3. comment number 3 by: Brooks / Gordon


    I just noticed your comment over at

    That thread may be dead, so I’ll ask you my question here (this thread doesn’t have much activity either, but at least is not cluttered).

    From a quick look at your comment, it seems to me that your criticism of TPC’s methodology is an invalid criticism.

    Taking your example of a taxpayer, prior to change in tax policy:
    $1,000 income
    $200 deductions
    $800 taxable income after deductions
    35% tax rate
    So, $280 tax

    Continuing with your example, after elimination of deductions and change from 35% to 28% rate:
    $1,000 income
    $0 deductions
    $1,000 taxable income after deductions
    28% tax rate
    So, $280 tax

    It so happens, result in this illustration is no change in revenue.

    But let’s consider how we we get there.

    A proper approach would be to consider the revenue loss from the tax rate reduction:
    $1,000 income
    $200 deductions
    $800 taxable income after deductions
    28% tax rate
    So, $224 tax

    That’s a loss of revenue of $56.

    Then consider the increase in revenue from elimination of the deductions, using the 28% rate, so $200 X 28% = $56.

    What I assume TPC is pointing out (apparently validly) is that (apparently) Feldstein instead calculated the revenue gain from eliminating deductions along the lines of $200 X 35% = $70. That would imply a net revenue gain of $14 from the combined tax rate reduction and elimination of deductions, which, unless I’m missing something, would be incorrect, per your own illustration.

    Am I missing something, or does the above imply that your criticism of TPC was invalid?

    (As a note, of course the above is all static analysis, not considering behavioral (dynamic) effects, but that’s beside the point.)

  4. comment number 4 by: Vivian Darkbloom


    My issue with the TPC was not its critique of Feldstein’s math—I was questioning their own math on their own analysis. If you read my first comment carefully, you would have noticed that I gave the same example you did as to how to arrive at the same result. The question is whether this is what the TPC did in their own calculations. Their past emphasis on the importance of the order in which tax cuts and deduction cuts occur led me to believe it might not have been. Of course, we can’t check TPC’s math directly because they have not disclosed anything but their conclusions. I challenged the TPC to set out how *they* did the math for *their* numbers, but in typical fashion they don’t provide details. It was a little exercise in what is good for the goose is good for the gander—trying to smoke out those details. It strikes me as blatantly unfair for the TPC to be criticizing Romney’s failure to disclose his details and then for it to refuse to disclose the details in their critique of his plan. It struck, and still strikes me as unfair that Feldstein should clearly put on the table his sources and numbers and how he got to his result (even if flawed) and for the TPC to not disclose theirs. It is not only unfair to Romney and Feldstein, it is unfair to private citizens like you and me. I’m sure the TPC does not like me being the L’ enfant terrible, but to quote the late Studs Terkel: “I’d like to be remembered as someone who caused trouble where trouble was due”.

    Second, I’ve long questioned a number of TPC’s assumptions on the Romney “plan”, which I believe are unrealistic. Their conclusion is only as valid as their assumptions. Some of their assumptions are, frankly, not realistic and not at all fair to Romney. As I see it, the TPC is hiding behind their model and a lot of data implicit in that model that we will never know about. It’s kind of like using a model to show the effect of fiscal stimulus. The result is only as good as the model. And, it’s bad enough that you think the results are valid because your own model says so, but it’s even worse when you don’t even reveal what the equivalent of the “multiplier” is.

    The TPC has come out with a revised set of numbers based on changing a couple of assumptions, but not all the ones they need to. That has brought the Romney plan much closer to break even. And yet, several very unrealistic assumptions remain. One big assumption they made and have yet to correct was that the Romney plan would entail *both* eliminating the estate tax and maintaining the existing step-up in basis for income tax purposes to the heirs of those assets. When Bush proposed (and succeeded for one year) to eliminate the estate tax, no step-up was given to heirs with the exception of an exempt amount which would be trivial for someone like Warren Buffett). Correcting this, would entail, I think, another adjustment of $56 billion (but again we don’t know the effect of that assumption in the TPC’s analysis because they have not revealed the detailed numbers or even their sources). We know the source of Feldstein’s data—the 2009 IRS SOI. So, his numbers are based on the most recent data of actual returns filed. Can you tell me the source of the TPC’s data? What were *their* sources and assumptions regarding income and deductions that went into their model and conclusions? How do we not know this is not GIGO?

    Interestingly, I noted that Feldstein, in his most recent post at Mankiw, also raised the possibility of taxing capital gains to heirs at death (this is a somewhat different thing than not taxing those accrued gains at death but denying the heirs a step up). In the long run, it may not make much difference whether one 1) taxes accrued gains at death; or 2) foregoes taxing at death but taxes those accrued gains in the hands of the heirs by not stepping up basis because it is largely a timing issue. However, in the context of a 10 year budget window, it could make a swing of as much as $112 billion between the TPC’s numbers and Feldstein’s most recent suggestion.

    Finally, the TPC and most other folks are assuming a top statutory rate of 28 percent to invalidate his plan. But, when I go to the Romney website and read what his proposal said, it says a tax rate cut of 20 percent across the board. But, Romney doesn’t say which baseline that cut would occur from. If it is a current law baseline, it would be a cut from 39.6 percent, not 35 percent, for example.

    The TPC’s report suggests that they gave Romney all the benefit of the doubt. My closer examination has revealed that while that that is the headline story, in reality it is simply not true.

  5. comment number 5 by: Brooks / Gordon


    Perhaps I’m missing something in what you said, but I just re-read your comment over there, and it seems you are making the opposite contention vs. what I’m saying. You seem to be saying that, to calculate the revenue gain from the elimination of tax deductions, we should apply the old tax rate (prior to tax cut) rather than the new tax rate, meaning you think Feldstein’s approach was valid and TPC’s is not (and thus, that you think TPC’s criticism of Feldstein on that point is invalid). Your whole comment seems to be about that assertion, and it seems that you were referring to TPC’s point about using the new tax rate when you said “…their math strikes me as wrong…” and “That makes no sense to me…”

    Are you saying (now) that you agree that the new tax rate is the one that should apply, that Feldstein applied the old one, which was an error, and that TPC’s criticism on that point is valid?

  6. comment number 6 by: Vivian Darkbloom


    There are two ways to get to the same result. One is to take the old tax rate times times taxable income or the new rate times AGI. My iniitial confusion with the TPC was that they seem to think it makes a difference in which order a computation is made when there is both a tax rate cut and elimination of expenditures. They say, for example, that cutting tax rates makes it harder to eliminate tax expenditures rather than eliminating tax expenditures makes it easier to cut tax rates. This makes no sense to me in a contemporaneous policy change. This then led me question what the TPC was doing with their own calculations and I challenged them to show the math on that. That’s what I meant when I wrote this, which you apparently have not read:

    “Of course, another way to express the same result as the second example would be to take for the revenue loss due to the rate reduction 28% times the taxpayer’s historical *taxable income* rather than *gross income* Thus, (35%-28% ) x $800 = $56, the same as the gain in revenue by taking .28 times $200. But, that does not seem consistent with the TPC’s prior statements on this and the comment above. It makes as much sense to me to say that “eliminating tax expenditures makes it easier to pay for the reduction of tax rates” as it does “reducing tax rates makes it harder for eliminating tax expenidtures to pay for them” and yet the former is the only thing the TPC has been emphasizing now for several months now.”

    In other words, we’ve seen Feldstein’s math; we have not seen what the TPC did in *their* original analysis.

  7. comment number 7 by: Brooks / Gordon


    I still don’t see what fault you see in TPC’s point re: which tax rate to apply, and what fault you see in their criticism of Feldstein on that point. I’m not talking about some phrasing TPC had in this analysis or some prior analysis (nor, for that matter, am I addressing your claim that TPC should show their math). I’m talking simply about TPC’s point that Feldstein seemed to be using the old tax rate to calculate the revenue gained from eliminating tax deductions, and that it was incorrect for Feldstein to do so.

    Are you (now) agreeing with TPC on this point or not?

  8. comment number 8 by: Vivian Darkbloom


    I did not agree or disagree with the TPC on their take on Feldstein’s math. I disagreed with them on not revealing their own with respect to their own work and demonstrated my reasons to suspect it. I think Feldstein did make a mistake on his calculation. Notice that he was not simply working with an assumed highest statutory rate. He was working with an assumed (average) marginal rate of 30 percent times the amount of historical tax expenditures. Based on what he said about historical marginal rates in the prior paragraph one is tempted to think he was simply averaging those out. Assuming that historical average would be correct for those cohorts, I think he missed the step of subtracting 20 percent from the result of $191 billion to be consistent with the way he accounted for the revenue in the first part of his analysis (since he was just calculating the total revenue loss rather than for a particular cohort, taking that shortcut seems appropriate).

    Feldstein seems to have made a mistake on his initial calculation; however, with a few minor adjustments to the assumptions (not merely the math), his basic premise sounds correct.

    I would like to see the TPC reveal more of their methodology and underlying numbers so we can check theirs.

  9. comment number 9 by: Brooks / Gordon


    I still don’t get your argument.

    Seems to me that TPC pointed out an error that Feldstein made.

    I don’t know why you then seized on that point — which you don’t dispute, although I’m still not clear if you agree with it or not — as evidence of a need for TPC to be more open about its methodology. Nor do I see why you apparently saw their point as illogical, as not making sense, and as mathematically (analytically) incorrect.

    Did you simply make an analytical mistake yourself (thinking that the old tax rate should be applied to calculate the revenue gain from eliminating the deductions, as Feldstein apparently did), and upon realizing it (after others pointed it out to you), decide to shift to some broad, unrelated criticisms of TPC?

    Or is there some real reason why you are focusing on what seems to be a rather simple, obvious, valid correction by TPC to argue that TPC should be more open about their methodology/calculations (and should phrase their points more politically neutrally), and some related point that your illustration actually illustrates? (And I don’t think the issue here is the order of calculations — it’s simply that Feldstein apparently did it incorrectly, and TPC pointed that out.)

  10. comment number 10 by: Vivian Darkbloom


    I’ve already given you the answers those questions–see above.

    Now, if you want t point out where I made a math error, please do so. You’ve got all the information you need. If you want to prove TPC’s math in their original analysis, please do that as well. I suspect you do not have all the information you need. That was *my* point and don’t try to evade it. And, stop twisting what I say—I did *not* say the TPC’s analysis was “mathematically incorrect”—I said they would need to disclose the math on their original analysis to make that judgement and that based on their various narratives I had reason to suspect it. They’ve critiqued Feldstein’s, but that have still have not fully disclosed theirs.

    By the way, you are not the only one entitled to ask questions here. My queries above to you are still unanswered.

  11. comment number 11 by: Brooks / Gordon


    First of all, if I’ve missed questions you’ve asked me, please tell me what your questions were.

    As for your comments, I don’t think I’m “twisting what [you said.” It does indeed seem that, in your TaxVox comment, you were asserting (although with argumentation that didn’t really support the assertion) that it at least appeared that TPC was wrong, their math incorrect, and their approach illogical and nonsensical (just to paraphrase your wording a bit). Anyone can go to the link and judge for himself. It seems everyone who responded their had the same impression as I did (and still do) re: what you were asserting. Oh heck, I’ll just paste your assertion, which you followed with an illustration and argument that didn’t really support this assertion. (Bolding mine)

    “He assumes that each dollar of itemized deductions lost by households with income above $100,000 would generate 30 cents in revenue. However, the Romney plan has a maximum tax rate of only 28 percent and most households with income above $100,000 would face an even lower rate on some or all of the additional income from eliminating deductions.”

    I’ve been following this line of thinking (I’m not going to call it logic) of the TPC for some time. An earlier paper by the TPC indicated that reducing the marginal tax rate would make it harder to make up the loss of income by eliminating exemptions. That makes no sense to me if the two things (reduction of rates and elimination of tax expenditures) happen, as proposed, in the same package. In other words, the revenue gained from eliminating tax expenditures should be calculated against a current baseline and not a *new* baseline that will be created after tax rates are lowered.

    What the TPC seems to be doing in its revenue estimates is a two stage analysis that creates a new baseline for calculating the benefits of eliminating deductions. In stage 1, to account for revenue *lost* they appear to reduce revenue by taking current revenue (before tax exenditures) against current marginal rates and then reducing the anticipated revenue per that differential (e.g. current revenue times 35 percent minus current revenue times 28 percent). Then, in stage two, to account for the revenue *gained* by eliminating tax expenditures, they take the *new lower* marginal rate (28 percent in my example) times historical tax expenditures. If this is what they are doing, and the comment above and elsewhere suggests that is the case, their math strikes me as wrong. And, it appears to be what they have been doing

    That sure seems to me to be an assertion by you that TPC was apparently nonsensically applying the new tax rate to calculate the revenue gain from eliminating tax deductions.

    As for evasion, you still seem to quite deliberately avoid saying whether or not you agree with TPC that Feldstein was wrong to apply the old tax rate to the revenue gain from eliminating tax deductions. Do you agree with TPC on that point or not?

    My sense is that you realize at this point that you were wrong in asserting an apparent error on TPC’s part.

    If not, perhaps you could explain how your illustration in your TaxVox comment shows some likely error on TPC’s part, or even why you see in this point by TPC (this correction of Feldstein) a reason to “suspect” TPC’s approach?

  12. comment number 12 by: Brooks / Gordon


    I realize now that what you probably were saying is that you think TPC calculated the revenue loss by applying the change in tax rate to income prior to any reduction from tax deductions (as if tax deductions had already been eliminated), and then counting the revenue gain from eliminating the tax deductions. Is that it? If so, why would you suspect that? What do you see in TPC’s writing that is cause for such a suspicion?

  13. comment number 13 by: Vivian Darkbloom


    Realizing that “now” comes rather late in the game. As to why I questioned it and simply asked that they demonstrate *their* math, has already been discussed ad nauseum. I’m not about to repeat that again. Read again and maybe *now* you will get it, or maybe you never will, but frankly I don’t think you want to do anything but play your usual game of gotcha, on your terms, of course. Why should I have time for that? If you want to know what the questions were, read my comment again and search for the sentences that end with this: ?

  14. comment number 14 by: Brooks / Gordon


    I don’t have some “usual game of gotcha”. It may seem that way to people who are uncomfortable being asked to explain/defend an argument that turns out not to hold water, and who become evasive in the face of such a request.

    Yes, I should have realized sooner that your suspicion — for some reason — was that TPC (for some reason) calculated the revenue loss and the revenue gain, respectively, in such an odd way. That was my error, and I apologize. That said, you seem to be seizing on my oversight as some sort of validation of your argument, and to pretend that your ad hominem is somehow a substitute for having some (reasonable) reason to respond to TPC’s correction of Feldstein with such a suspicion. It seems you set up a nonsensical straw man, claimed — without any basis — that TPC seems to have done it that way, and then called it nonsensical. Now I ask you why you would think they did it that way, and you won’t provide a reason. I just scanned your exchanges over at TaxVox, and I don’t see there your offering any reason for thinking that’s what TPC seems to be doing.

    So we’re left with (at best) TPC pointing out that Feldstein made an error (given his own methodology) by using the old tax rate to calculate the revenue gain from eliminating deductions, and your responding by fabricating some supposed error you claim TPC seems to be making, but offer no reason for such a suspicion.

    And you still won’t say if you think TPC’s criticism of Feldstein was valid (which, by the way, isn’t affected by whether or not TPC made the error you [for whatever reason] suspect they made in their own analysis). Want to say now if you think TPC’s criticism of Feldstein was valid?

    Oh, and as for your questions, seems your only questions were essentially (at one point within one long comment) asking if I knew TPC’s methodology on this point. No, I haven’t checked, so I don’t know for a fact that they didn’t make the error you allege. I also haven’t checked to make sure they didn’t add 2+2 and get 5, but I wouldn’t just say they seem to have made such an error without some good reason to suspect it, let alone respond to TPC’s correction of Feldstein with such a charge.

  15. comment number 15 by: AMTbuff

    Here’s a different critique of the TPC analysis, that it ignores the economic growth that is the objective of the Romney plan:

    My critique of that critique: If you are counting on additional economic growth to make your plan revenue neutral, you can’t also count that money as helping to close the fiscal gap. I don’t know whether Romney did that, but politicians love to double count savings.

  16. comment number 16 by: Vivian Darkbloom

    “If you are counting on additional economic growth to make your plan revenue neutral, you can’t also count that money as helping to close the fiscal gap. I don’t know whether Romney did that, but politicians love to double count savings.”

    Fair point. If I read Rosen’s table correctly, he estimates based on 2012 law that the plan would not only be “revenue neutral”, but revenue positive. To the extent it is positive, it is fair to count that money as closing the fiscal gap.

    This also demonstrates the limited usefulness of analyses that focus only on whether one pays more or less tax, pre- and post-tax policy change.

    If you had the choice between a) Earning $100 and paying $28 in tax; or b) Earning $110 and paying $$30 in tax, which would you choose, all else the same? Shouldn’t you (and the government) be better off? And, if I’m better off, I doubt I am going to get up in arms if someone else’s relative improvement is a couple of percentage points more (or less) than mine.

    As an economic matter, Romney’s plan makes sense. The whole debate about distribution of tax benefits and costs is not about economics—it is about politics. And, when relatively minor discrepancies are noted and those discrepancies do not account for margin of error, it is about using rhetoric to pursue political ends.

    All this back and forth on Romney’s plan really ignores the margin of error issue: despite powerful number crunching models, revenue estimates are not scientifically accurate. If Romney’s plan “fails” one such model by, say, $50 billion, I’d say this is well within the realm of statistical error. Even more so when that estimate was based on a number of unwarranted assumptions to begin with.

  17. comment number 17 by: Dave

    Look, it’s time to be adult about this matter of taxation. The country is in recession/stagcession AND mired in 16 trillion dollars of debt.

    This notion that we should cut taxes on some by heaping them on ’someone else’ is ridiculous. Unavoidable fact: other than those in dire poverty, EVERYONE in this country is going to have to pay up. Soaking the rich (only) is tantamount to handing them a first-class plane ticket to Singapore. French Pres. Hoilland has undertaken to soak the rich and his wealthiest citizen, Bernard Arnault, owner of Louis Vuitton Moet Hennessy, is now a Belgian citizen. If voters want to lower the deficit by soaking someone, … the should volunteer themselves, not someone else. Anyone remember kids waiting in a line giving their friends ‘backsies’? Same problem.

  18. comment number 18 by: Terry Sawyer

    If some one needs to be updated with newest technologies
    afterward he must be pay a visit this web site and be up to date all the time.