The respected Tax Policy Center (TPC), a non-partisan think tank set up as a joint venture of Brookings and the Urban Institute, released on November 8 its first estimates of the distributional impact of the November 3 Republican tax proposal.  The TPC runs a sophisticated micro-simulation model of the US tax system, using a large but anonymous set of tax returns to determine the impact of proposed changes in tax law.  The model is similar in nature to models run by the US Treasury and in Congress, but with full disclosure of the results and how they arrived at the estimates that they obtained.

As was discussed in an earlier post on this blog (on the Trump plan released in late September), these Republican plans will provide huge cuts in taxes for the very rich, and far more modest cuts for others.  The November 3 plan released by the Republican Chair of the House Ways and Means Committee (the committee responsible for all tax issues) is the first fully detailed plan for which good quantitative estimates can now be made of the impact by income group.  While certain elements in the proposal are changing daily, with the Republican leadership in a rush to push their plan through quickly before there can be extensive analysis and a solid debate, there is enough now available to make reasonable estimates of the impact.  And so far, the changes adopted have been minor, with the basic structure of the original proposal maintained.  It remains a plan where the rich will benefit disproportionately.

The chart at the top of this post shows the estimated average cuts in taxes that would be provided under this plan, categorized by income group:  in quintiles plus for the top 1% and top 0.1%.  The quintile groups are defined by equal shares of the population (20%), with the bottom 20% in those households in the lowest income rank, and so on up to the richest 0.1%.  The breakpoints (in 2027, but in terms of prices of 2017) for household incomes for each group are:

a)  Bottom 20%:  $0 to $28,099 of annual income;

b)  Second 20%:  $28,100 to $54,699;

c)  Middle 20%:  $54,700 to $93,199;

d)  Fourth 20%:  $93,200 to $154,899;

e)  Top 20%:  $154,900 or more;

f)  Top 1%:  $912,100 or more;

g)  Top 0.1%:  $5,088,900 or more.

One would have to be earning over $900,000 a year (in 2017 prices) to be part of the top 1% in 2027, and over $5 million a year to be in the top 0.1%.  These groups are very well off.

As the chart shows, the cuts are overwhelmingly clustered at the top.  Indeed, the cuts are relatively so small for the first 80% of the population that they do not even register at all on the bar lines.  And even if one re-draws the chart to remove the tax cuts the top 0.1% would gain (which stretches the scale), there is not much of a difference:

One can still barely see the cuts that would go to the first 80% of the population.

However, supporters of the Republican plan respond by saying that the rich get most of the cuts because the rich pay most of the taxes.  It is true that the rich do pay more in the current system when expressed in absolute dollar terms.  (It is not always the case when the taxes are expressed as a share of their incomes, but that is a separate issue and one I will not go into here.)  But is it the case that the tax cuts that will go to the rich under the Republican plan are proportionate to their incomes and/or the taxes that they pay?  As we will see in the remainder of this post, they are not.

First, one can calculate from the TPC figures (available for 2027 for the groups analyzed here in spreadsheet form at this link) the share of the tax cuts (in total dollar terms) that will go to each of the income groups identified:

The rich will receive a much higher share, in dollar terms, of the tax cuts than would go to the others.  The richest 20% will receive 73% of the dollar value of the cuts, while the subset within this group that are the top 1% will receive 47% and the top 0.1% will receive 25%.

But how does this compare to their shares of overall income?  After all, if the top 20% receive 73% of the tax cuts, but also account for 73% of overall income, one could say the tax cuts are proportionate to their incomes.  But that is not the case:

The top 20% account “only” for 49% of the nation’s household incomes (as forecast for 2027), but will obtain 73% of the tax cuts (as also forecast for that year).  And one can look at the similar relative amounts for the other income groups.

Perhaps the argument would then be made that the tax cuts should not be compared to the income shares, but to the shares of taxes each group would be paying (before the tax cuts, as under current law).  There is still a degree of progressivity in the tax system, even though the progressivity was reduced sharply in the Reagan and Bush tax cuts.  For the shares of taxes that would be paid (under current law) in 2027, one can calculate:

But even here, the share of the top 20%, say, comes to 69%, which is still less than the share of the tax cuts (73%) they would receive under the Republican plan.  And the shares are far more disproportionate for the top 1% and especially the top 0.1%.

One can look at this systematically, for all groups, by calculating the ratio of the shares of the tax cuts which would go to each group, to the shares of what each group would be paying in taxes (under current law).  By definition, the average ratio will be 1.0 (weighted properly), but the individual ratios will be above or below this.  What we find by dividing the shares of the tax cuts (Chart 3 above) by the shares of taxes to be paid under current law (Chart 5) is:

The ratio of what the bottom 20% of the population will be receiving in the proposed tax cuts, to what they would be paying in taxes under current law, is only 0.4.  They would be receiving, under the Republican plan, a disproportionately low share of the cuts.  And the ratio is even worse for the second 20% (which one might consider as the lower middle class), at just 0.3.  At the other end of the scale, the ratio of what the top 1% will be receiving in the proposed tax cuts (47% of the cuts) to what they would be paying in taxes under current law (31% of the taxes) is very generous, at 1.5.  And the top 0.1% would enjoy a ratio of tax cuts to what would they pay in taxes under current law of an even higher 1.8.

Finally, what do the ratios look like if one compares the share of the tax cuts that would go to each group to the shares of the incomes of those groups?  For the purposes here, I have taken the incomes as they would be after taxes, but only after the taxes as they would be under current law.  This is the base one is starting from, and thus the basis for such a comparison.  One finds:

Here the ratios are even starker.  The ratio of the share of the tax cuts that would go to the bottom 20% (0.4% – see Chart 3), to the share of income of those in that group (5.3% – see Chart 4), is just 0.07.  Put another way, the share of the tax cuts that would go to the poorest 20% of the population is 93% below what it would be had the tax cuts been allocated equally according to income shares.  It is almost as low for the second 20%, at just 0.12 (or 88% below what it would be had the cuts been allocated equally according to income shares).  And the ratio remains below 1.0 for all but the top 20% of the population.

But the tax cuts are generous for those with high incomes.  For the top 1%, the share they would receive in the proposed tax cuts is over 3 times their share of income.  And it is 3.6 times as high for the top 0.1%.

It is nonsense to claim that this would be a “middle-class tax cut”, as the Trump administration and the Republican leadership in Congress have repeatedly asserted.  A dramatically disproportionate share of the cuts will go to the extremely rich, while those at the lower end of the distribution will receive only token amounts.

A recently released report from the president’s Council of Economic Advisers (CEA) claims that cutting the headline corporate income tax rate from the current 35% to 20% would lead to a jump in the real incomes of American households by a minimum of $4,000 a year and possibly by as much as $9,000.  Others have criticized those forecasts for a variety of reasons, and Larry Summers has called the estimates “absurd”.

Indeed, they are absurd.  One way to see this is by looking at the historical evidence.  This is not the first time the US would cut its corporate tax rates.  Did such cuts in the past then lead to a jump in real wages?  As the chart above suggests, the answer is no.  This blog post will discuss that evidence, as well as other issues with the CEA analysis prepared under (and it appears largely by) its new chair Kevin Hassett.  But first some background on the CEA and its new chair, and what this recent incident portends for the Council and its previous reputation for professionalism.

The Council of Economic Advisers has, until now, been a highly respected office in the White House, set up to provide the president with objective and professional economic advice on the key economic issues of the day.  The Council was established in 1946 during the Truman administration, and has had as its chair and its members many illustrious and well-respected economists.  A number later received the Nobel Prize in Economics and similar awards.  While the CEA can be and has been political at times (it is located in the Office of the President, after all), it has had an able staff who were expected to provide professional assessments of the issues as a service to the president.  Many came on leave from academic posts.  As an example of the type of staff they could draw, both Larry Summers and Paul Krugman, then young and rising economists, were on the Council staff in the early 1980s during the Reagan administration.

The current chair is Kevin Hassett.  Trump did not nominate someone to the position until April, and Hassett took up his post (following Senate approval) only on September 13.  Prior to this post, Hassett was perhaps best known for co-authoring (with James K. Glassman) the 1999 book titled “Dow 36,000”, in which he forecast the Dow Jones Industrial Average would reach 36,000 by 2002 and certainly no later than by 2004.  In the event, the Dow never exceeded 11,750 (in January 2000) and dropped to 7,200 in October 2002, as the Bush administration’s first recession took hold.

Hassett has now, as one of his first official acts, released a formal CEA study that claims that if the Trump Tax Plan were enacted, with the headline corporate income tax cut from 35% to 20%, household incomes in the US would rise by a minimum of $4,000 per year, and possibly by as much as $9,000.  Larry Summers has termed it “dishonest, incompetent, and absurd”, and other economists have been similarly scathing.

The study really is pretty bad, and must be an embarrassment to the CEA staff. The report starts (Figure 1) with a chart that shows average real wage growth over the last several years (2013 to 2016) among the 10 OECD member countries with the highest statutory corporate income tax rates, as compared to that for the 10 OECD members with the lowest rates.  Between 2013 and 2016 (but essentially just in 2015) the wage growth was higher by a few percentage points in the set of OECD countries with the lower tax rates.  But the 10 OECD member countries with the lowest corporate tax rates were mostly countries from Central and Eastern Europe (Estonia, Latvia, and so on to Slovenia).  They were starting from lower wage rates than in the richer countries, and benefited as they opened up to globalization and in particular to the EU markets.  It is difficult to see how this simplistic correlation tells us much about what would happen if the US cut its corporate income tax rate.

Hassett then quantifies his estimate of the dollar gains per household by citing a number of obscure articles (several of which were never published in a peer-reviewed journal) to come up with estimates of possible elasticities (explained below) that relate how much household incomes would rise if corporate taxes were cut.  He concludes this review by asserting that an elasticity in the range of -0.16 to -0.33 would be reasonable, in his view.  The -0.16 figure came from a study from 2009 published in the “Federal Reserve Bank of Kansas City Economic Review”.  That is not exactly a prestigious journal.  And the -0.33 figure came from a 2007 paper that was presented at a conference, and does not appear to have ever been published.

An elasticity of -0.16 means that if the corporate tax rate were cut by 1% (not 1 percentage point, but rather by actually 1%, e.g. from 35% to 34.65%), then real wages would rise by 0.16%.  A 10% cut in the corporate tax rates (e.g. 35% to 31.5%) would lead, according to this assumption, to real wages rising by 1.6%.  And a cut in the corporate income tax rate from 35% to 20% (a 43% fall), as proposed in the Trump tax plan, would raise real wages by 6.9% under this assumption.  Hassett then applies this to the wage portion of household incomes to arrive at his calculated gain of $4,000 per household.  And the $9,000 gain is based on assuming an elasticity of -0.33.

There are numerous problems with this analysis, starting with the assumption that correlations are the same as causation.  There is also a question of what correlations are relevant.  The study that came up with the -0.33 elasticity, for example, looked at correlations across a panel of 50 countries.  It is not clear that such correlations would be of much relevance to judging the impact on real wages of a change in the US on corporate tax rates, as real wages across such a range of countries are driven by many factors (including, not least, the level of development).  And the -0.16 elasticity came from a study that examined correlations between real wages and corporate tax rates across the different states of the US.  But labor is mobile across US states, as is capital, plus the range of variation (state to state) in corporate tax rates is relatively modest as state taxes are relatively modest in size.  And indeed, it is not even clear how many companies actually pay the headline corporate income tax rates on the books, as states routinely grant them special tax holidays and other favors in order to try to get them to move to their states.

One would have thought that the most interesting investigation as to whether changes in corporate income taxes would matter in the US to real wages, would have been to see what actually happened in the US when such rates were cut in the past.  The fact that Hassett ignored this obvious question in the new CEA report is telling.  And there have indeed been earlier changes in the corporate tax rate, most notably (in recent decades) in 1987/88, following from the Tax Reform Act of 1986 during the Reagan administration.

The impact (or rather the lack of it) can be seen in the chart at the top of this post.  As had been discussed in earlier posts on this blog, real wages have been stagnant in the US (for the median wage earner) since around 1980.  The chart at the top of this post is an update of one prepared for a post from February 2015 that looked at the proximate causes of stagnant wages over this period, despite growth of real GDP per capita of more than 80% over the period.  While real GDP per capita is now 82% above what it was at the start of 1979, real wages (as measured by real median weekly earnings of full-time workers) are just 5.7% above where they were at the start of 1979.  Furthermore, the current “peak” of 5.7% growth can all be attributed to growth in the period since mid-2014, as the economy finally approached full employment levels in the later years of the Obama administration (having been held back by government spending cuts from 2010), with this carrying over into 2017.

The top corporate tax rate on profits was cut from 46% in the years up through 1986, to 40% in 1987 and then to just 34% in 1988 and thereafter to 1993 (when it was raised to the current level of 35%).  Did the cuts in 1987/88 lead to a sharp jump in real wages?  There is no indication of that at all in the chart.  Indeed, real wages fell by close to 6% between late 1986 and 1990, and then stayed at close to that low level until they started to rise some in the mid to late 1990s.  And there is no indication that the small increase in the corporate tax rate in 1993 to 35% led to wages then declining – indeed, they started to rise a few years later.

Based on this, one might come to the conclusion that a cut in corporate tax rates will lead to a fall (not an increase) in real wages, as seen following the 1987/88 cuts.  And also that a modest rise in the tax rate (such as in 1993) would lead to a gain in real wages a few years later.  But I would not claim this.  Rather, I would say that real wages and corporate tax rates are simply not closely linked to one another.  But for Hassett and others to claim that cuts in corporate taxes will lead to a significant jump in real wages, the exact opposite outcome following the 1987/88 cuts needs to be explained.

The CEA report was badly done, and must be an embarrassment to the professional staff there who certainly know better.  And as Larry Summers remarked in his blog post:  “Considering all this, if a Ph.D student submitted the CEA analysis as a term paper in public finance, I would be hard pressed to give it a passing grade.”