The surprise contraction in fourth quarter GDP may just mean the third quarter's strength came from pulling growth from the fourth. While a recession could have started, it is probably still a few months away. Like other recessions, most economists will fail to warn that it is coming.
In the 30 plus years I have watched, I have never seen a consensus of economists correctly forecast a recession before it started. Economists in positions of authority, particularly at the Federal Reserve and in a President's administration, virtually never publicly forecast a recession, perhaps out of concern it would be a self-fulfilling prophecy. Economists on Wall Street rarely predict recessions, perhaps because it reduces their firm's effectiveness in selling product. Wall Street economists, who predict recessions, even if they are correct, often find themselves unemployed.
Misunderstanding how marginal tax rates affect growth contributed to missed calls of the last two recessions and blinds economists to the one coming. The correct implications of tax policy suggest the economy will have no growth the next two years, while the consensus calls for continued modest or perhaps improving growth. The discrepancy between modest growth and no growth could be the difference between U.S. stock prices continuing to advance and a 50% to 80% decline from present levels.
Economists' Blind Eye
It is hard to recognize reality when you have been trained your whole life to believe something else. There appears to be a near universal belief that low marginal tax rates improve economic growth, while empirical evidence points the other way. Even the economists and policy makers proposing higher marginal tax rates don't claim it will be good for growth, but that raising tax rates will not hurt growth much. The theoretical economic models that ignore tax rates or presume low marginal tax rates benefit growth may have contributed to flawed forecasts of about twice the growth that actually occurred in the last 12 years. Economists expected the Bush tax cuts to improve growth. Yet President Bush's 8 years annualized 2.0% GDP growth, about half of the 3.9% under President Clinton. In mid-2007, none of the so called "blue chip" economists thought there was any chance of recession in the next two years. The great recession then shrunk GDP about 4.7%. At the end of 2010 when the Bush tax cuts were extended, the Congressional Budget Office ("CBO") predicted growth would annualize 4.5% in 2011 and 2012. Instead, we got 2%.
Economists expect modest growth to continue. The Conference Board expects growth of 1.8% this year and 2.4% in 2014, which would annualize 2.1% for the two years. According to money.cnn.com, the consensus of economists forecasts 2.4% growth in 2013. There is recognition that the normal or baseline GDP growth rate has dropped below the 3.3% it annualized since World War II. Looking at the mid-points between the peak rate of growth and the worst rate of growth in the last several business cycles suggests the current base line growth is much lower than realized.
I expect annual GDP growth rate for the current business cycle already peaked at 2.4% in late 2010. Even if growth accelerated higher than any forecast I have heard and hit 3.7%, which was the fastest growth rate of the previous expansion, the mid-point would still paint a dismal picture.
I believe tax policy precipitated the decline in the baseline growth rate. This case may be more understandable if we first clarify the difference between average tax rates, marginal tax rates, average marginal tax rates and the top marginal rate. Then, I'll give a brief review of what the economic literature says and doesn't say about these tax rates. I will then begin building the case on the empirical relationship between top tax rate and growth since 1920. We will broaden the case with the top tax bracket. This will set up looking at growth since 1967 and how it has been impacted by the top rate and capital gains tax rate. This is where I get the forecast of near 0% growth. Finally, I'll offer some conclusions.
What Tax Rate
Let's start with the average income tax rate. You can figure yours by looking at the adjusted gross income on your tax return and seeing what share was paid in tax. In using summary IRS data shown here, the income tax rate for the U.S. has averaged about 13.5% the last 30 years, and hit a low of about 11% in 2009 -- the last year of available data. To get the marginal rate, or the rate on the last dollar of income, you would look to see what tax bracket you were in and the rate for that bracket. In our marginal tax rate system, which also has deductions and exemptions, virtually everybody pays zero percent tax on their first few thousand dollars of income. Then the next few thousand is subject to a 10% tax rate. A few thousand after that is subject to a 15% rate. Currently, there are 6 tax brackets, or potentially 7 with the health care surtax. Beginning this year, the portion of ordinary income above $450,000 for a couple is subject to the top rate of 39.6%, plus some people will pay the health care surtax of 3.8%, for an actual top marginal rate of 43.4%. There are also marginal tax rates for capital gains income that top out at 23.8%, including the surtax.
A person with a great deal of income would have portions of their income taxed at all the rates from 0% to 43.4%. If you took the portion of federal income tax that each individual paid in the highest bracket they were subject to, you would get the average marginal tax rate. Economists estimate the average marginal rate peaked at about 30% in 1981, and has since dropped to about 22% in 2006. The top rate, which applies to income above the top bracket, will be taxed at 43.4% with the surtax, up from 35% last year.
The economic literature has both theoretical and empirical support to show a low average tax rate benefits growth. The literature also shows a low average marginal tax rate increases the hours workers are willing to work. This has strong theoretical support and empirical support up through at least 250% of the average wage. This should mean a low average marginal rate improves growth, but something offsets this positive influence because empirical studies show no robust relationship between growth and the average marginal rate.
Martin Feldstein shows the top marginal rate has a significant relationship with high end income, where cutting the top rate corresponds with rising income for the wealthy. This fact is often used to assert that a low top rate improves growth; however, this assertion has no empirical backing in the literature. There are numerous theoretical studies claiming a low top rate improves growth with no empirical work to back it up. I presume wealthy interests give generous consulting fees to economists who claim low top rates benefit growth, and if the economists could show it with data they would, but since they can't, they weave a logical sounding hypothetical, perhaps with a bit of anecdotal data.
In the last 20 years, the data has become progressively less friendly to the belief that a low marginal tax rate improves growth, and more friendly to the conclusion that it harms growth. The last article I found in the literature studying the empirical effect of any income tax rate on growth was published back in 1993.
Looking at the top tax rate and growth rate on the same chart tells the story. The chart below shows annual GDP growth since 1920 and the top tax rate with a two-year lead time. It also shows two possible relationships: a linear correlation in red and a curvilinear fit in green. On the left, there is a scatter plot, and on the right, the same data is shown in a time-series format. The linear relationship is positive, as shown by the upward sloping red line in the scatter plot. This means a higher top tax rate corresponds with faster growth. A two-year lead time (shown in chart) has the strongest correlation, but every lead time from zero to seven years has a positive relationship with growth. While the slope of the line would turn negative with an 8-year lead time, there is no lead time with a statistically significant negative slope. So the belief that a low top rate benefits growth has no empirical leg to stand on.
The best curvilinear fit of the top rate also has its strongest correlation leading growth two years. This fit has a stronger correlation than the linear fit, and suggests the growth maximizing top rate is 66%.
The bottom left point on the scatter plot represents GDP shrinking 13% in 1932 and the top tax rate of 25% in 1930. The second weakest growth on the chart is the -10.9% growth in 1946, corresponding with a top rate of 94% in 1944. The strongest growth of 18.4% is influenced by an 81.1% top rate. Outside of World War II, the strongest growth in 1923 was influenced by a 73% top rate. The next strongest, 1936, corresponds with a 63% top rate.
The effect of the top rate is more clearly understood in the context of the tax bracket at which it becomes effective. The next four charts look at the rates and brackets.
The top rate has ranged from 7% to 94% and averaged 58.8%. The capital gains rate has ranged from 7% to 77% and averaged 25.7%. The top bracket where the top rate begins to apply (shown below) has ranged from $29,750 to $5 million, or from 1.5 times the prior year's per-capita GDP to 8,688 times. I estimate the $450,000 bracket will be about 9 times last year's per-capita GDP.
High tax brackets correspond with stronger growth rates and lower the average tax rate. For example, if the 43.4% top rate had a bracket of $1 million rather than $450,000, the average tax rate would be lower because a smaller portion of income would be taxed at the top rate. The two charts below show the relationship between what the bracket was one year and GDP two years later. This lead time has the strongest correlation using the last 60 years of data. I examine the top bracket as a multiple of per-capita GDP rather than in actual or inflation adjusted dollars, because the relationship is more robust.
The upward sloping best fit lines suggest growth benefits from higher brackets. As noted above, economists have shown a low average income tax rate encourages stronger growth. What I have found in the U.S. data since 1920 is that if you control for the influence of the top bracket, the influence of the average marginal rate becomes insignificant. So the whole empirical case for lower tax rates boosting growth may instead be a case for higher tax brackets. Curiously, I couldn't find any articles in the literature examining the relationship between the top bracket and growth.
In addition to influencing growth, I have concluded the top bracket also influences what top rate maximizes growth. So while the growth maximizing top rate for the last 90 years of data appears to be 66%, I have found when the top bracket is below 100 times per-capita GDP, the growth maximizing top tax rate is lower.
In 1965, the top bracket was cut from $400,000 to $200,000, and the bracket as a multiple of per-capita GDP plunged below 100 for the first time ever. With the two year lead time, this began affecting growth in 1967. The next chart with several graphs on it shows what I believe to be the influence of the top rate and capital gains tax rate on growth since 1967.
There is no guarantee analysis of historical data -- their trends and correlations 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.
The growth maximizing top rate for the period appears to be 54%, as shown in the green curvilinear fit above. The strongest growth in the period corresponds with Reagan's 50% top rate. This chart does not show a linear fit, like the chart going back to 1920 did, but the linear fit does have a positive correlation with every lead time between zero and thirty years. The string of positive relationships may go further, but I quit checking at 30.
The curvilinear fit of the capital gains rate, shown in blue, suggests that a lead time of four years has the best correlation with growth. Longtime readers know that previously I had estimated a 5-year lead time based on using the last 90 years of data. However, using either the last 30 or 60 years of data suggests the best lead time is 4 years. This is the first time I have presented forecasts based on a 4-year lead time.
The curvilinear fit suggests a capital gains rate of 27.9% maximizes growth. The strongest growth in 1984 corresponds with a 28% capital gains rate. One of my friends believes Reagan's actions made growth strong under Clinton. I partly agree, but probably not for the reason he thinks. I believe Reagan raising the capital gains tax rate back to 28% in 1987 improved prosperity under Clinton. On the other hand, the back to back recessions of 1980 and 1981-82 correspond with the capital gains rate being too high at 39.9%. The great recession and weak growth since corresponds with the rate being too low at 15%.
A regression model based on the two tax rates is shown in red above. The bottom graph in the chart above shows the five-year growth rate and the estimate of the five-year rate based on the model. The model estimates that the baseline GDP growth rate from 2007 through 2014 is 0.1%. Growth has come in a little stronger, the last point on the chart representing the 5 years 2008-2012 annualized growing 0.6%. In the last 35 years, the long-term growth rate has risen and fallen almost like clockwork in response to marginal tax rates.
The model suggests the baseline growth rate will remain 0.1% for the next two years. Since growth has annualized about 2% above this baseline the last 3 years, it might be normal for growth to drop a couple of percent below the baseline the next couple of years.
After a long lag, better times are on the way. The top rate of 43.4% this year implies the baseline growth rate will rise to 1.3% in 2015. The capital gains tax rate of 23.8% with the four-year lag time should begin influencing growth in 2017. If current tax rates continue, the model would estimate GDP growth of 4.3% in 2017.
Implications For Stocks And Earnings
If GDP grows around 2%, earnings of the S&P 500 tend to hold fairly steady and neither grow nor shrink much. If GDP growth accelerates to 2.5%, companies can leverage that modest growth into significant earnings growth. However, if GDP growth drops below 2%, earnings usually begin to decline. If GDP shrinks, earnings plummet. This relationship is shown in the chart below. Over the last year, real "As Reported" earnings of the S&P 500 shrunk 2.2%. In the last three decades, companies have leveraged their earnings to GDP growth. Approximately 2% GDP growth is the fulcrum point around which the leverage multiples earnings grow or decline. If we have two years with zero or negative GDP growth, the leverage will drive S&P earnings down at least 50%. Heck, real earnings fell 51% when annual GDP got down to 0.9% in the 2001 recession. The economic weakness implied by the lagged effect of tax policy would spell disaster for stock prices (SPY), (IVV), (DIA).
If the tax policy correlations with growth shown above represent an actual influence rather than coincidence, there would need to be a logical explanation for the correlation. Here is my attempt at that.
Tax policy is pro-growth if it encourages the wealthy to avoid taxation with deductible or depreciable expenditures that grow the value of a business. The average tax rate on the wealthy and talented must be low enough that running and/or funding a business is worthwhile. The marginal tax rate must be high enough that the wealthy shelter wealth within a business where the growing value of the business, held as unrecognized capital gains, can remain untaxed for years or decades. If marginal tax rates are too low, personal income of the wealthy rises at the expense of lower investment in productive capacity and less compensation of employees (see article). At the low end, marginal income tax rates are primarily a tax on labor. At the high end, marginal rates are mostly a tax on money not plowed into growing a business.
The extraordinary fit of the five-year growth rate with the model's estimate suggests the wealthy's response to tax policy dominates the long-term growth rate. Most other variables may just push or pull growth from one period to another without much long-term influence.
The beginning of the year tax deal to avoid the fiscal cliff was beneficial for growth in three ways. It kept marginal income tax rates low at the bottom of the income spectrum, which is good for the supply of labor. It raised marginal tax rates at the high end, which should shift money away from personal income for the wealthy toward growing value in their businesses and the economy. Third, it raised the top bracket, which will help keep the average tax rate low enough to protect incentive.
Further steps could be taken. If the 50% top rate were to come back, I estimate the growth maximizing bracket that best balances a low average rate with a high marginal rate would be about 35 times per-capita GDP, or around $1.7 million. If we were to get up to the 66% top rate mentioned above, the bracket should be at least $10 million and probably over a hundred million. It should only affect the handful of people with the most ability to plow money into businesses.
The lagged effect of the low top rate and low capital gains rate suggest the base line growth rate is close to 0.1% for the years 2007 through 2014. The last three years grew at about 2.1%, well above this baseline. Growth could revert down to or below the baseline during the next two years.
Mainstream economic forecasts are in a fog that either ignores marginal tax rates or gets their influence backwards. It is like flying a plane in dense fog without instruments, where you can't see the mountain in front of you. The forecasts of continued modest or improving growth will crash, along with the stock market.
As far as I can tell, the crash is a still a few months away, but it could come at any time.
Additional disclosure: There is no guarantee analysis of historical data -- their trends and correlations 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.