In an earlier post I showed a chart that attempted to convey the limited progressivity of the American tax system when not only federal taxes but also state and local ones are taken into account. Here are two additional charts. They’re based on my calculations from data for 2004 in a Tax Foundation report (tables 3 and 4 and figure 1) by Andrew Chamberlain and Gerald Prante.
These charts show effective tax rates for each of the five quintiles of households. The effective tax rate is calculated as taxes paid divided by income. For instance, to get the effective rate for the bottom quintile, I divide the average amount paid in taxes by households in that quintile by the average income of those households.
It turns out that whether taxes are progressive depends on how income is defined.
As the first chart shows, if income is measured as market income — income from employment, investments, and a few other sources, but not including government transfers — the tax system is essentially flat. The effective tax rate is approximately 30% for households throughout the income distribution. This may hide some progressivity, since the effective rate may be higher in the top portion of the top quintile, but we can’t be sure because the Tax Foundation data don’t separate out the top 1% or 0.1% of households.
Adding government transfers (as the Congressional Budget Office does in its calculations of federal tax progressivity) increases the average income in each quintile, but much more for the bottom than for the middle or top. This reduces the effective tax rate more in the lower part of the distribution than the upper, resulting in a progressive structure.
What should we conclude? I think the first chart here better reflects the impact of the U.S. tax system. It does very little to alter the market distribution of income. Redistribution is achieved mainly by government transfers rather than by taxes. We aren’t unusual in this respect, though; it’s the case in most if not all rich countries.
Who needs a flat tax? We’ve already got one!
Lane, Thanks for rerunning the data. I like having both sets of numbers here.
I do have a question about the first chart — it just seems very funny to me that the most representative depiction of the relative tax burden of Americans would have the 2nd poorest quintile paying more than the lowest quintile.
I suppose what chart is most telling depends on what question we are trying to answer:
Do we care about the tax burden/outtake of the individual relative to his financial standing/intake (e.g., all incomes, including transfers)?
Or, do we care about the tax burden of the individual relative to his intake from wages and investment income particularly?
I understand that Lane prefers the the latter because it preserves the purity of the tax rate assessment, but this rings a little hollow for me. The fact that some income is guaranteed and tax free (gov’t transfers) doesn’t seem enough to merit exclusion from the denominator. To me, this appears to be systematically excluding a type of income that deliberately augments a group’s income without forcing them to pay taxes on it.
So, once again, do we care about the relative tax burden on dollar EARNED or dollar ACCUMULATED?
Put specifically, if Family A makes $10, gets $5 dollars in government transfers, and pays $2 in taxes, should we say that his effective rate is lower or higher than Family B, which makes $12, gets $0 in transfers, and pays $2.40 in taxes.
They pay the same 20% effective tax rate given Lane’s preferred model. In the second model, the first family pays 13%.
The second characterization seems more fair to me, but I’m open to persuasion.
This is common practice in fiscal analysis. Willem Buiter, for example, expresses the solvency constraint of a jurisdiction subject to the “no-Ponzi game” condition as: PV of exhaustive spending ≤ {(assets – debts) + PV (taxes – transfers)}, where PV denotes the present value of a cash flow.
Well, the top tax rate does a little to redistribute. The share of household income held by the top 1% before taxes is ca. 20%, and 15% after taxes. But yes, most redistribution occurs through the welfare state. Which is why I’m hope Obama comes through in his promises to increase the EITC.
Fred,
Figuring for cash flow seems to be fundamentally different than judging the progressiveness of tax rates, but maybe I’m wrong.
I’ll go back to my final example, would that mean you think the most revealing analysis should show Familes A and B to pay the same?
1) It matters not how we get progressivity, taxes on income or with transfers included, we have progressivity.
2) These gross tax rates leave an important economic incentive aspect of taxes to still be known: the marginal tax rate. SocSec taxes of 12.4% stop at $106,800 so income earners past this point get a reduction from this but then the income tax rate goes up. It means that the lower income people are not properly incentivized to work as too much of their income is taxed away. Social Security taxes should be progressive not regressive because there is no LOCK BOX.
Norman,
“Social Security taxes should be progressive not regressive because there is no LOCK BOX.”
I couldn’t agree more. I hope Obama follows through on his promise to apply to payroll tax to incomes in excess of $150k. That would at least be a step in the right direction.
Yes, the definition of “income” in the CBO report includes a great many things that aren’t normally considered income — such as employer contributions to healthcare plans and both employee/er contributions and retiree withdrawals from tax-deferred savings plans and “imputed taxes” that is, taxes paid by others that would fall on individuals if they didn’t fall on others (the simplest example is employer FICA contributions). This definition is actually substantially broader than typical measures of “total compensation.”
Publius: Why does this seem fair analysis when considering the imputed value of Medicare and cash value of employer contributions to healthcare but not, for instance, the value of natural resources harvested from public lands by the licensed or the reduction in property losses by the propertied from natural disasters due to public services or the reduction in losses from bank failure by the monied due to gov’t intervention? I understand that the analysis is difficult, but there is a presumption here that only those receiving transfer payments are receiving value from the spending of their government. This is demonstrably untrue.
At least in this chart Lane sort of reveals how he produces his numbers [his prior posting was blatantly wrong and intellectually dishonest]. However, his reasoning here is still very flawed and he leaves out a few critical points.
1: Excluding transfers to describe effective tax rates is absolutely nonsensical. Consider the retired person that collects a few thousand each month from social security with little to nothing in the way of market income. If this person spends most of his income (social security/transfer) and pays sales taxes on this or to pay property taxes, Lane’s so-called method would have us believe that his so-called effective tax rate would around 100%. Clearly this is utter nonsense. Likewise, it doesn’t make any sense to include one person’s health insurance [ private health insurance = market income ] and exclude another persons [ public health insurance = transfer] — both directly contribute to their overall well being and allow them to direct their other (e.g., cash) income towards other pursuits like food and entertainment.
You can NOT reasonably tally sales and property taxes without also accounting for a large share of transfers that gives them the money to spend this in the first place.
If we want to measure the ability of each household to pay taxes, his method fails.
If we want to measure what relative burden each quintiles actually bears on net terms, this measure also fails.
2: He totally fails to mention the fact that poorer households are a lot bigger than the richer ones in terms of the number of individuals in each (Tax Foundation does their quintiles by HOUSEHOLDS). In fact, the poorest is almost TWICE as large on average as the richest. The fact that many poorer households have more mouths to feed and shelter is going to drive up their state and local taxes significantly. It’s silly to ignore the effect behavior has on this stuff–especially for a so-called sociologist.
3: His closing statement that taxes do “very little to alter the market distribution of income” also fails reason and mathematical analysis. If the so-called rich weren’t paying such highly progressive income taxes most of these transfers wouldn’t even be possible. Likewise, if the poorer people weren’t huge beneficiaries of “net” transfers, they wouldn’t have the money in the first place to pay the few taxes they actually do pay.
I am working on a project about the rise of a hidden regressive tax state in the U.S. beginning (their expansion) in the 1980s. These would be fees at the DMV, toll collection, phone surcharges, cable, utilities, all used to compensate for the loss of revenue from high income taxpayers.
I am from Netherland and do not know much about taxation methods in US. But your blog is providing plenty of information, analysis and insights about the taxation progress. Also you got nice interactive vistiors who add value to your blog post with their thoughtful comments.