Those who cannot remember the past are condemned to repeat it
— George Santanyana
Over the course of the last few years, there has been much propaganda from politicians of all persuasions on what factors are conducive to economic growth and/or the emergence of budget deficits or surpluses. Regardless of the political viewpoint being represented, these claims have usually been supported by essentially no facts, and this absence of facts implies a lack of analysis in a political debate that has displaced by competing ideologies. The breakdown of reason is taking on great importance as the so-called fiscal cliff approaches, and with it, the increasing possibility of a full-blown constitutional crisis. My aim is to identify factors that have been associated with budget surpluses and/or solid economic growth (as measured by annual growth in the gross domestic product (GDP) in fiscal year (FY) 1931 through FY 2012 and to relate them to the emotionally-charged question of taxes.
Before proceeding further, it is important to list several caveats. First and foremost, what follows is an historical analysis, not an economic analysis. For example:
* It does not address the very different nature of federal debt from private debt in creating burden on the economic growth, not to mention the way deficits are accounted for in the National Income and Product Accounts.
* It does not discuss the comparative impact of different types of taxes and different types of government or private spending on the economy.
* It treats inflation-adjusted GDP growth as a simplistic measure of overall economic advancement. In reality, it is also relevant to consider whether GDP growth is broadly shared or narrowly concentrated in one small segment of the population. Similarly, GDP growth treats a billion dollars of activity X as being equivalent to a billion dollars of activity Y, even if X and Y have far different effects on the lives of ordinary people.
* There is no attempt to identify causative relationships between various contributing factors (tax rates, government spending patterns, etc) and political/economic results (budget surplus, strong economic growth). In fact, correlation does not automatically imply causation, but a statistically strong correlation between A and B could plausibly be construed as meaning that A at least contributes to B if there is no credible mechanism for B to cause A and no obvious underlying cause for both A and B at the same time.
Finally, and most importantly, this analysis is not prescriptive. It is simply a description of some basic historical patterns that have evolved in our contemporary political economy.
Keeping these crucial caveats in mind, the following sections present some first cut correlations between deficits, economic growth, employment, and tax rates. The correlations described below suggest that many of the ‘assumptions’ shaping the debate of the so-called fiscal cliff are disconnected from reality. Unless these basic misconceptions are corrected, any resulting grand bargain may well worsen the deficit by impeding employment and economic growth, assuming the inertia of past patterns of correlation continue into the future.
Factors Correlated with Federal Surpluses or Deficits
Unemployment rate: During the 82-year base period, the factor most strongly correlated, by far, with having a federal surplus is the unemployment rate. Out of these 82 years, 41 had an average unemployment rate of 5.6% or higher and 41 had an average unemployment rate of 5.5% or less. To place this in context, the current unemployment rate is 7.9%, and the average rate for 2012 will probably be slightly over 8%. Of the 41 years with low unemployment, three occurred during the World War II spending spree of FY 1943 through FY 1945, when no credible level of revenue could have produced a balanced budget. I have excluded these years from the table below.
There was a federal deficit in 100% of the years with an unemployment rate of 5.6% or higher. The highest unemployment rate during a year with a surplus was 5.5%, during one year of the Eisenhower administration. In 28 non-WWII years with unemployment rates of 4.6% to 5.5%, there were 6 surpluses and 22 deficits. In 10 non-WWII years with unemployment rates of 4.5% or less, there were 6 surpluses and only 4 deficits. Such low unemployment rates are almost always the result of several consecutive years of rapid GDP growth. In other words, a robust economy seems to be an absolute prerequisite for having a federal government surplus.
Overall federal taxes: Assertions about the impact of “high taxes overall” are easily tested, as good information is available about federal taxes as a percentage of GDP, federal deficits/surpluses, and economic growth, is available for every year during the 82-year basis period. The median value for federal taxes (including Social Security, Medicare, individual income taxes, corporate income taxes, etc.) during this period was 17.6% of GDP, with 40 years of 17.5% or less and 42 years of 17.6% or more. By comparison, the values for 2009 through 2012 were all in the 15.1% to 15.8% range. This was the first time during the post-WWII period that taxes were under 16% of GDP for more than two consecutive years, and the first time since the Truman administration that taxes were under 16.2% of GDP at all.
The table below shows a strong correlation between higher than average overall federal taxes and federal surpluses or smaller than average deficits, although the correlation is not as strong as is the case with unemployment rate. In addition, the last budget surplus when federal taxes were under 17.5% of GDP occurred in the Truman administration, more than 60 years ago. At that time Medicare did not exist, and the fraction of the populace eligible for Social Security was far smaller than it is today.
Federal taxes on the very wealthy: Assertions about the impact of “high taxes on the wealthy” are more difficult to test. Information about the maximum tax rate on the wealthy is available for the entire base period, but there are complicating factors:
* The point at which the maximum tax rate goes into effect has varied wildly over the last 80 years. The maximum tax bracket has affected anywhere from several percent of the population down to less than one person per thousand. Currently, the maximum tax rate on ordinary income affects slightly over 1% of the population.
* It is not trivial to combine the maximum tax rate on ordinary income with the maximum tax rate on capital gains in a consistently meaningful manner. There have been several years when the maximum tax rate on capital gains was the same as the maximum tax rate on ordinary income, and years in which the maximum tax rate on ordinary income was up to five times higher than the maximum tax rate on capital gains.
* The tax rate on capital gains has often been constant, without regard for the amount of capital gains, but has often had multiple tax brackets for varying amounts of capital gains.
* For most people, “ordinary income” is the dominant factor. By contrast, the richest 0.1% of the population (over $2.9 million per household, according to the Washington Post (Nov 28, 2012)) mostly derive more income from capital gains than from ordinary income, but the exact ratio of capital gains to ordinary income varies widely within the fortunate few.
* Dividends have sometimes been treated as ordinary income and sometimes as capital gains.
At the expense of being simplistic, I will use the maximum tax rate as the metric for taxes on the wealthy, and I will define the “maximum effective tax rate” as being 0.6*(maximum tax rate for capital gains) + 0.4*(maximum tax rate for ordinary income). This is based on the assumption that the “average rich person” derives 60% of his/her income from capital gains and dividends and 40% from ordinary income.
The average value for “maximum effective tax rate” for the base period is 40.9%, and the median value is 44.8%, compared to a mere 23% today (35% on ordinary income and 15% on capital gains, with dividends treated as capital gains). The current “maximum effective tax rate” is the lowest it has been since 1931. I occasionally see claims that taxes on the wealthy (or overall taxes) were lower in the Reagan-Bush41 era than they are today. The maximum tax rate on ordinary income was in the 28% to 31% range 1988 through 1992, compared to 35% today, but dividends and capital gains were taxed at up to 28% to 28.9% then, yielding a “maximum effective tax rate” of 28% to 29.7%. Moreover, overall federal taxes never dropped below 17.3% during the Reagan era. The real Reagan was much more of a tax-and-spend kind of guy than modern Tea Party adherents realize or acknowledge.
There has been a strong positive correlation between a higher than average “maximum effective tax rate” and budget surpluses or smaller than average deficits, as shown in the table below. Moreover, there has NEVER (within the last 82 years) been a surplus in a year when the “maximum effective tax rate” was under 27.5%, compared to 23% today.
Factors Correlated with GDP Growth
Overall federal taxes: As stated above, the median value for federal taxes during this period was 17.6% of GDP, with 40 years of 17.5% or less and 42 years of 17.6% or more. By comparison, the values for 2009 through 2012 were all in the 15.1% to 15.8% range. This was the first time during the post-WWII period that taxes were under 16% of GDP for more than two consecutive years, and the first time since the middle of the Korean War that taxes were under 16.2% of GDP at all. The table below shows a moderate correlation between lower than average overall federal taxes and stronger than average GDP growth, although there was extremely wide variation within both the “high tax years” and the “low tax years.” For example, three of the five worst years during the 82-year base period were “low tax years.”
Federal taxes on the very wealthy: All of the caveats and complications from the discussion above still apply here. The situation here is radically different than it is with regard to the overall federal tax burden. There is a significant positive correlation between higher than average taxes on the very wealthy and higher than average economic growth, as shown in the table below. That is, high taxes on the wealthy have usually been associated with good economic growth (although I make no claim regarding causation here).
Moreover, the results from the table are NOT sensitive to the manner in which taxes on ordinary income and taxes on capital gains are combined to create the maximum effective tax rate. The same conclusions would result if the table focused only on capital gains, or only on ordinary income. For example, the median value for the maximum tax on capital gains was 25% in 1931-2012. In years when the maximum tax rate on capital gains was 25% or higher, annual GDP growth had mean and median values of 4.2% and 3.7%, respectively. In years when the maximum tax rate on capital gains was under 25%, annual GDP growth had mean and median values of 1.4% and 2.7%, respectively.
The results in the table above are so contradictory to much of the modern political dogma that I decided to delve into the period from 1965 through 2012 in more detail. I divided this era into 6 periods of 4 or more consecutive years, based on keeping years with a similar “maximum effective tax rate” together. I omitted 1981 and 1986-1987, because these years did not fit naturally into any period of 4 or more consecutive years with similar tax rates. The table below shows the average rate of GDP growth during each of the 6 periods. The 3 periods with below-average taxes on the wealthy finished 3rd, 5th, and 6th (last) in GDP, and the period with the highest taxes on the wealthy finished tied for 1st. Clearly, many other factors influenced the rate of economic growth during these 6 periods of time, so the table does not prove that high taxes on the wealthy are good for the economy. Nevertheless, the table strongly indicates that high taxes on the wealthy are fully compatible with rapid economic growth. Hence, the burden of proof should be on people who claim that any increase in taxes on the wealthy would drive the economy into the toilet, not on people who advocate a modestly higher maximum tax bracket.
Corporate taxes: Any such analysis is quite difficult, starting with the choice of metric(s) to be used. The maximum bracket for US corporate income taxes is 35%, for profits exceeding $18.3 million. This is one of the highest rates in the world, and higher than average by US historical standards. On the other hand, the United States has innumerable deductions-exemptions-loopholes for corporate income taxes, so that corporate incomes taxes in recent years have constituted a lower percentage of GDP than the historical average. Moreover, several corporations had profits exceeding $1 billion in 2010, while paying less than 1% of their profits in corporate income taxes.
I investigated two candidate metrics: corporate income taxes as a percentage of GDP and corporate income taxes as a percentage of total US corporate profits. I was unable to find enough data on the second metric, so I settled on corporate income taxes as a percentage of GDP. This metric has at least one significant flaw. Since corporate taxes are based on corporate profits, not total corporate income, these taxes automatically decline during a recession that leads to a major reduction in overall corporate profits and automatically rise greatly during an economic boom. Hence, this metric does not fully capture any “drag effect” that this tax might have on the economy.
The Results: Corporate taxes exceeded 2% of GDP every year from 1946 through 1980, and averaged above 3% of GDP. From 1981 through 2009, corporate income taxes never reached 3% of GDP, seldom exceeded 2% of GDP, and averaged about 1.5% of GDP. The value of about 0.95% in 2009 was the lowest on record (since 1946). GDP growth averaged 3.2% per year in 1946-1980 and 2.8% per year in 1981-2009, which shows that higher levels of corporate taxes are compatible with good economic growth. Modestly higher corporate taxes could, however, reduce the deficit.
Federal spending: From 1931 through 2012, federal spending was at or above 19.5% of GDP in half of the years. This figure has grown over time: it had a median value of 17.6% in FY 1931 through 1971, compared to a median of 21% and a minimum of 18.2% in FY 1972 through 2012. The table below shows that there has been a weak inverse correlation between high federal spending and strong GDP growth, but there has been extremely wide variation within both high-spending years and low-spending years. In fact, the worst year on record was a low-spending year as were three of the worst four years. The only high-spending year among the four worst years was in FY 1946, during the de-mobilization from World War II, and GDP shrinkage during that drastic transition was virtually inevitable.
The situation with regard to federal spending is far more complex than the table can express. Federal spending is not some uniform blob that expands and contracts – the composition of federal spending today is far different from what is was 50 to 80 years ago, and there is no reason to think that expenditures on military procurement, infrastructure upgrades, research on science and technology, education, Social Security, Medicare, Medicaid, and unemployment compensation would all have a similar effect per trillion dollars spent. Sorting out the impact of separate types of spending on the economy would be a major undertaking.
A final relevant point is that some types of federal expenditures – Medicaid, food stamps, unemployment compensation – rise automatically during a recession and contract automatically during an economic boom. Moreover, expenditures on Medicare and Social Security rise automatically when the fraction of the populace over the age of 65 grows. It would be necessary to exclude these types of expenditures from the table to get a more meaningful idea on whether discretionary federal spending is good for the economy.
Conclusions
Several of the tables above conceal extremely wide variation within groups. In such a case, a moderate difference between two groups is of little relevance, due to the large range of variation in the data. Historical data does, however, support two conclusions that strongly contradict Republican orthodoxy and the views of so-called fiscal hawks. These are:
* The key to having a manageable deficit is to have a robust economy. Thus, discussions within the government should focus primarily on reducing unemployment by generating broadly shared GDP growth, without generating excessively inflationary pressure when the economy is near to full employment (which is not the case today). Therefore, measures aimed directly at deficit reduction should be focused on items that clearly pose little risk to employment and/or economic growth.
* The only policy change that could achieve measurable deficit reduction with almost no risk to economic growth is higher taxes on the very wealthy and on corporate profits. High taxes on the very wealthy have been associated with both strong GDP growth and budget surpluses in the past. A credible initial step to keep from falling off the fiscal cliff might consist of modest tax increases for individuals/families making between about $250,000 per year and $1,000,000 per year, and large tax increases for individuals or families making over $1,000,000 per year. Reducing or eliminating the gap between the maximum tax bracket for ordinary income and the maximum tax brackets for dividends and capital gains could also generate significant additional revenue with little risk to the economy.
On the other hand, these historical patterns imply a grand bargain based on cuts in Social Security, Medicare, Medicaid, infrastructure upgrades, education, and research on S&T, while maintaining today’s absurdly low taxes on the wealthy, is likely to push the economy back into full-scale recession, undermine US economic competitiveness well into the 2020s, cause enormous suffering for a significant fraction of the populace, and actually increase the deficit.
Newk Mindshaftgap is the a nom de plume of a man with long experience in the Federal Government and he has three degrees in physics.