CBPP Critique of State and Local Burden Methodology Is Baseless and Hypocritical
CBPP Critique of State and Local Burden Methodology Is Baseless and Hypocritical
The Center on Budget and Policy Priorities (CBPP) has once again released a short report criticizing the Tax Foundation’s state-local tax burden report, despite the methodological improvements made in this year’s study. All the criticisms are baseless, and some of them are hypocritical given that the so-called faults they cite are replete in the Center’s own reports.
We analyze them in turn:
(1) CBPP: When the Tax Foundation revises its methodology, the resulting tax burden estimates contradict previously published estimates.
TF: New data and new data sources are constantly being published by governments and private sources, and economists that do not revise their methodologies to take them into account are either lazy or ignorant. Mr. Johnson (the CBPP report’s author) doesn’t point out where he believes the new methodology is wrong. For him, because it’s different from last year’s, it can’t be trusted. Pure silliness.
CBPP itself will demonstrate how silly that is in about three weeks. That’s when the Census Bureau will publish its report on how many Americans lack health insurance. The Census’s estimate will probably be between 45 and 50 million, and the Center will hype this report as it does every year. But just last year, the Census suddenly revised their previously published estimates of the uninsured for years 1995 – 2003 (as well as other recent years) due to technical issues.
By Mr. Johnson’s silly argument, the upcoming Census numbers should not be trusted due to previous revisions. The same would go for methodological changes that the Bureau of Economic Analysis (BEA) makes to GDP, methodological changes that the Joint Committee on Taxation, Treasury, and CBO have made to its scoring procedures, and so forth. Anyone who has worked in Washington with tax burdens, revenue scoring, tax models, etc. understands this. But Mr. Johnson tries to scare readers by implying that methodological revisions are a Tax Foundation trick, when they’re actually a staple of every economic series of this type.
(2) CBPP: The Tax Foundation’s methodological change is just another example of the Tax Foundation’s previous year-to-year revisions.
TF: That’s false, and Johnson has written enough about this report to know it. There are two, clearly distinct types of revisions that change economic studies like this one: (1) new and revised data from government and private sources that are part of data series’ already used and (2) revised methods of calculation. Prior to this year’s report, the changes made in the Tax Foundation’s state-local tax burden series were almost entirely due to changes in the underlying data, such as more up-to-date revenue figures from Census and income amounts from BEA.
The changes to this year’s paper are methodological; that is, we used an expanded definition of income (which actually makes tax burdens smaller), and we allocated more types of taxes back to the states where the taxpayers live. In all cases, we identify our tax burden estimates as just that – estimates. Using BEA terminology, the FY 2008 numbers are “advanced estimates,” and the FY 2007 numbers are “preliminary estimates.”
In light of this CBPP animus toward recent estimates, it’s also funny to see on their website a report showing the size of FY 2009 budget gaps. So at the same time the Center is putting out projections for FY 2009 budget gaps (expenses minus costs), it is criticizing the Tax Foundation for making estimates of the previous fiscal year’s revenues.
(3) CBPP: The Tax Foundation will be making future adjustments like the ones in this report, so we know that the current estimates must be unreliable.
TF: In an echo of his previous critique, Mr. Johnson says the current estimates should not be relied upon because the Tax Foundation will probably be making future methodological adjustments. The fact of the matter is that the Tax Foundation is merely using the most up-to-date data sources that are available. If more states release information on the amount of income taxes paid by non-residents, we will take that into account and possibly make methodological changes. If new empirical evidence on the incidence of various taxes comes out, we will take that into account. If BEA provides any more data or takes away some data, we will make adjustments. If we didn’t, ours would not continue to be the most widely quoted private estimates of state-local tax burdens.
Let’s look at another example of the CBPP’s use of estimates. Just as the Tax Foundation uses BEA estimates in the state-local burden study, the CBPP commonly uses estimates from the Urban-Brookings Tax Policy Center (TPC) to highlight distributional effects of various federal tax policies. The TPC estimates are based in large part on data provided by the IRS from its Public Use File, which is republished each year yet with a rather significant time lag (most recent available is 2004). Currently, the TPC Model is based on the 2004 IRS Public Use File. Yet the Tax Policy Center is making projections starting with the 2004 base year data out to years as late as 2018, and the Center is frequently citing those figures.
When TPC updates its model with a newer IRS Public Use File (maybe the 2005 PUF when it is released), the figures they have published will all be revised, for past and future years. In fact, TPC has done this already, updating its model on numerous occasions due to improvements in available data, technology, and economic literature. The TPC model also uses data from the 2005 Current Population Survey, which is lagged like all data sources, and the Survey of Consumer Finances, the American Housing Survey and Medical Expenditures Panel Survey. This does not mean that the TPC figures are useless or untrustworthy (they are from it), but it shows that the Center is hypocritical when criticizing the Tax Foundation for merely using newly available data sources that may make its work better.
(4) CBPP: The burden estimates of the Federation of Tax Administrators (collections divided by income) are more commonly accepted than the Tax Foundation’s measure.
TF: It may be the case that the FTA figures have been used more frequently throughout history, but that still doesn’t mean they are proper measures of tax burdens. CBPP fails to delve into the pertinent question: (1) What an average effective tax rate really means (taxes paid divided by income) and (2) that the legal incidence of a tax is not the same as the economic incidence of a tax.
To show how ridiculous using taxes divided by income as a measure of tax burdens is, look at Nevada. The FTA finds Nevada has a heavy tax burden because total collections are high compared to the state’s total income. But those collections include the large amount of hotel and gambling taxes paid by tourists in Las Vegas. Of course, those hotel taxes are not paid out of Nevadans’ personal income; they’re paid out of the incomes of tourists who live all over the world.
The Tax Foundation is measuring Nevadans’ ability to pay their taxes by counting all the income they earn and all the taxes they pay out of that income. Mr. Johnson is making a mockery of the ability-to-pay theme that is at the core of the average effective tax rate concept by trumpeting the FTA’s method of simply dividing collections by income.
We invite CBPP to produce its own report detailing the state and local tax burden in each state. But if they choose to continue citing the FTA measures of tax burdens, they should not be taken seriously by anyone who understands what an average effective tax rate is.
Mr. Johnson also questions the Tax Foundation’s projection of capital gains because 2006 is the most recent year for reported state-by-state data on capital gains realizations. He also says that it is “unclear how the Tax Foundation generated these data.” If Mr. Johnson had read the methodology report (page 27), he would have realized that the study used actual data on dividend income by state up through three months of fiscal year 2008 to project the capital gains income. (The two are highly correlated.)
This is another hypocritical charge, because CBPP has repeatedly used projections of capital gains realizations and assumptions to argue how the Bush tax cuts are too costly or that they favor high-income earners over others. How reliable are those estimates and projections that are often significantly revised?
(5) CBPP: The Tax Foundation’s methodology has never been formally published or subject to outside review.
TF: While it has not been published in an academic journal, a 30-page detailed methodology is available on our web site. It details all the assumptions made and data sources used in the report. Why doesn’t he specifically question the assumptions made or the data sources used in the report? That would require debating specific issues of economic incidence and data reliability, which Mr. Johnson appears not to want to do. Instead he just throws mud, saying our report “blurs the actual data with extensive adjustments made by the Tax Foundation that are supposed to reflect the extent to which one state’s residents pay taxes to other states.” Could he be any vaguer?
(6) CBPP: The Tax Foundation’s figures are wrong, and therefore should not be relied upon.
TF: Finally, to close the report, Mr. Johnson says flat out that the Tax Foundation figures are likely to turn out wrong and should not be trusted. Are the estimates 100 percent accurate? Have the numbers been handed down from God? No. They are estimates of state and local tax burdens, a point that is clearly stated time and time again in the report. That’s what all tax and income data are for current or recent years, whether it’s a Tax Foundation report or a CBPP report.