You, like me, have probably heard for the last 40 years that tax cuts create jobs. When I went to find this in actual data I found a different relationship. Tax policy affects the labor market with long lead times. So not only will unemployment not get much better for 6 years it will probably be above 9% sometime in the next two years and wages will continue to fall behind inflation. There will be another recession, probably starting next year and stocks will be hammered.
First let's look at the data and then I will offer some possible explanations why reality is different from what I and perhaps you have been conditioned to believe. There are two employment related charts, each with several graphs. The first one is about unemployment and the second about workers' share of the economy.
The first graph of the first chart shows how the capital gains tax rate influences unemployment six years later. It might be more accurate to say this is a statistical correlation that I believe represents an actual influence. The top left graph in the chart below is a scatter plot where each point represents the average annual unemployment rate for one year and the capital gains tax rate 6 years earlier. Statistically the correlation is stronger with the capital gains tax rate leading 6 years, than the other dozen lead times I tried.
The blue curvilinear fit through the scatter plot suggests a capital gains tax rate of 25% minimizes unemployment. The high unemployment associated with the 1981-82 recession corresponds with a tax rate that is too high at 39.9%. The high unemployment rate of the last four years matches a capital gains rate that is too low at 15%. The 15% capital gains rate appears to be consistent with a 9% unemployment rate. This data is also shown in a time-series format to the right of the scatter plot.
While the current measure of unemployment only goes back to 1948, the capital gains tax rate also has implications for what happened to employment in The Great Depression. The capital gains rate was cut from 73% to 12.5% in 1922 and remained at 12.5% until it was raised to 31.5% in 1934. If I take the mathematical formula for the curvilinear fit above and plug in a 12.5% capital gains tax rate it estimates a 12.5% unemployment. Using the six year lead time, the 12.5% capital gains rate corresponds with a measure of unemployment that was available during the Depression averaging about 13.9% from 1928 through 1939. Of course, it hit a high of 25% in 1933.
Employment also benefits from having a high top tax bracket, in part because it lowers the average tax rate. Currently the top bracket for a married filing jointly tax payer is $388,350. So a couple with ordinary income above that would pay a 35% tax rate on the portion of income above the bracket and lower tax rates on the portion below the bracket. If the top tax bracket was instead $1 million, it would lower the average tax rate of high earners even though their top tax rate might be the same.
In trying to understand the relationship between the top tax bracket and unemployment I experimented with different ways of looking at the top bracket and different lead times. The best way I found, or at least the one with the strongest correlation, was to measure the bracket as multiples of per-capita GDP with a three year lead time. The current top bracket of $388,350 is 8 times last year's per-capita GDP. The three years with the lowest unemployment in the early 1950s corresponds to the $400,000 bracket being over 200 times per-capita GDP.
This relationship is shown in the chart above by the two graphs in the middle. The best fit line in orange slopes downward in the scatter plot, which means lower brackets correspond to higher unemployment. In the time-series plot on the right this is shown with the inverted orange scale that has zero at the top.
You may be wondering why I don't show the influence of the top marginal tax rate. I did look at it and found an inverse relationship where a higher top rate corresponded with lower unemployment, but when I tried to put it in a model with the top bracket the top rate became statistically insignificant.
The influence of the capital gains rate and the top bracket are combined into a model shown in the red line of the bottom graph. The implication is that unemployment will average 9% for the next three years. The 9% rate is estimated to continue six years if the top bracket is not raised dramatically.
Obviously, other factors also influence unemployment, mainly the business cycle which pushes unemployment above the model during recessions and below during expansions.
The future of tax rates and brackets are up in the air. As it currently stands the capital gains rate will go from 15% this year to 20% in 2013 plus the 3.8% healthcare surtax will bring the rate to 23.8%. This is pretty close to optimal and implies unemployment would drop to 5% in 2019. Raising the top bracket could bring help sooner. House Speaker John Boehner proposed a top bracket of $1 million; this would be about 21 times per-capita GDP and a step in the right direction, but still modest historically.
There are two measures of employment compensation. The one shown in the chart below approximates take home pay for workers. Its share of GDP has dropped nine of the last twelve years. The 3rd quarter of 2012 had an all time low share of 43.5% of GDP. Prior to 1975 it was consistently above 50%. The other measure not shown includes fringe benefits and payroll taxes. It recently dropped to a 56 year low share of GDP at 55%.
Labor's share of the pie goes up and down with the top marginal tax rate. This positive relationship shows in the green upward sloping best fit line in the scatter plot below. A seven year lead time has the strongest correlation. The times series graph to the right of the scatter plot shows labor's actual share of GDP in black and the estimate based on the top rate in green.
The difference between the actual share in black less the estimated share in green is the residual compensation shown in grey in the middle two charts. The residual is the variation in labor's share that is not explained by the top rate. This scatter plot looks at the residual compensation compared to the five year moving average of the capital gains tax rate with a 7 year lead time. The best fit curvilinear line shown in blue implies a capital gains tax rate of 27.4% gives labor a larger share of GDP than a higher or lower capital gains rate.
The model shown in red on the bottom graph combines the influence of the two tax rates. The model suggests that in the next two years compensation will drop to 42.5% and then stay there another 5 years. So the squeeze on labor and the middle class will likely intensify in the next two years.
Perception vs. reality
The difference between the reality represented in the data above and what we have been told the last 40 years about low tax rates creating jobs may be a classic wolf in sheep's clothing story. Those who benefit from tax policy that shifts rewards away from labor toward capital convince much of the working middle class it is in their best interest. The middle class embraces what looks like a cuddly sheep and a few years later finds their take home pay is missing an arm.
We often hear the meme, "Low tax rates leave more money in the hands of business owners to create jobs." Low marginal tax rates have certainly left more money in the hands of business owners, but if the policy had been effective at creating jobs it would mean more money was paid to workers to be productive. The fact is a low top marginal tax rate means business owners take more money as personal income and labor gets a smaller share of the pie.
To my understanding, prosperity benefits from high marginal tax rates with high enough tax brackets that average tax rates remain low.
Part of the reason for this relationship is that low marginal tax rates increase the after tax cost of labor. In the 1950s and early 1960s when the top tax rate was 91%, the after tax cost for a wealthy business owner to hire a dollar of labor was nine cents. With the top rate at 35% the after tax cost for a dollar of labor is 65 cents, or if the owner can get money out of the business at the dividend or capital gains rate, a dollar of labor costs 85 cents after tax.
The lower the marginal tax rate is the bigger the after tax profit a business owner gets from squeezing the pay of labor. Performance based pay for CEOs leverages the incentive from low tax rates to squeeze worker pay even more.
Owners are reluctant to raise wages. Workers are eager for their wages to be raised, but quite resistant to reductions. It takes a long time to squeeze compensation as demonstrated in the 7 year lead times above. In the case of the capital gains rate there is also a multi-year compounding effect so that the full impact of the 2003 cut to capital gains won't be felt on labor's share of GDP until about 2014.
Contemplating what is a fair tax rate or a fair distribution of income is a pointless exercise in subjectivity where almost everyone thinks it would be fair for them to have a lower tax rate or a bigger share. A more valuable question is what distribution of incentive maximizes prosperity.
The last 12 years should make it obvious the current distribution is sub-optimal. The stock market after inflation and dividends is 5% lower than in 2000, or 24% lower if you don't count dividends. Private sector jobs have grown at an annualized rate of 0.02% vs. a long run average of 2%. Labor force participation has plunged. A shift in distribution that had more people working and fewer engaged in financial speculation would likely increase prosperity and the long run return in the stock market.
Intensifying pressure on labor will almost certainly bring a recession, which I think will start in 2013, if it hasn't started already. In the last year, corporate earnings dropped and CEOs intent on protecting profits and their bonuses will aggressively cut costs, which necessarily includes squeezing workers in hours, numbers and wage rates.
In the last two and a half years the annual growth rate in real earnings of the S&P 500 (shown below) has fallen from an all time high of 776% to -2.2%. This record volatility probably has a lot further to go on the downside. Usually when this growth turns negative a recession is near.
Unemployment at 9% or above with plunging earnings and stock prices are probably baked into the cake. I expect helpful policy to emerge in 2013 where marginal tax rates move toward levels that will give a more balanced distribution of incentive and tax brackets will be high enough to keep the average tax rate where it is still worthwhile for the wealthy to run businesses. There is light at the end, but lag times mean the tunnel is long.
Disclosure: I am short DIA, SPY. There is no guarantee analysis of historical data and trends enable accurate forecasts. The data presented is from sources believed to be reliable, but its accuracy cannot be guaranteed. Past performance does not indicate future results. This is not a recommendation to buy or sell specific securities. This is not an offer to manage money.