Wednesday, November 2, 2011

Beale's Law Debunked

The political left and political right are engaged in a knock-down, drag-out competition. Although they seek to sway public approval numbers, the competition is not about public approval per se. They are competing to see who has the defter hand at data manipulation to suit their respective political objectives.

A post today by Linda Beale provides a suitable case study in data manipulation. Beale links to this CivilAmerican report, which presents a series of graphs plotting (i) top marginal income tax rates over the decades, and (ii) income earned by the top 10% of taxpayers and top 1% of taxpayers over roughly the same time period.

According to Beale, the graphs evidence the fact that "income inequality levels inversely track the top tax rate--as the rate increases, income inequality decreases."

Beale's argument has a intuitive appeal, which reflects the beauty of data manipulation. The use of data to support an ideological argument cloaks the argument with a pseudo-scientific legitimacy. Beale's position evokes the law of universal gravitation ("the gravitational force between two objects is proportional to the mass of each, and inversely proportional to the distance between them"). How could we possibly challenge the foundational concept of gravity? And how can we possibly challenge Beale's Law: that income inequality levels inversely track the top tax rate?

But let's take a harder look at the data in the graphs. How does Beale's Law hold up to a skeptical analysis by a political independent? (Spoiler: not so much.)

The first graph from the CivilAmerican report plots the top marginal tax rate applicable to individual taxpayers over time.

The second graph plots the percentage of total income going to the top 10% of taxpayers for a period ending in 2006.

The third graph plots the share of income reported by the top 1% between 1913 and 2007.

I don't have access to the data underlying these graphs, but let's assume that the underlying data and the graphing tools are accurate. We observe that the graphs are not exactly apples to apples -- the first graph plots data for a period ending 2010; the second graph 2006; the third graph 2007. The use of different time periods is somewhat misleading, but bigger problems emerge when examining Beale's Law against the graphical data. So we'll give a pass to the apples to oranges presentation.

The inclusion of data on the top 10% in graph 2 is outright misleading. The first graph and third graph are focused on the top 1% (rates and income). The second graph creates an inference that the top 10% benefit from the same changes in marginal tax rates as the top 1%. But taxpayers in the 90th income percentile are light years away from taxpayers in the top 0.01% and the top 0.001%. The "tippy top" of taxpayers distorts the income numbers. Meanwhile, we don't have data regarding the tax rates paid by the 90th through 99th percentiles. None of this data supports Beale's Law. The graphical presentation is a mess.

Now, on to the graphical data. If your eyes -- like mine -- have seen better days, crack out the microscope.

[1] First Decade of Income Tax (1913-1923)

The top tax rates jumped from less than 10% in 1915 to around 75% in 1917 or 1918 (first graph). That's an increase of more than 750%. The top rates stair-stepped down from roughly 72% in 1920, to 58% in 1922 or 1923, to roughly 25% from 1925 through 1932.

Despite a 750% increase in the top tax rate from 1915 to 1917 or 1918, the top 1% of taxpayers reported a roughly 20% decrease in income from peak (1918-19%) to trough (1922-15%).

We only have data on the top 10% beginning in 1917. Between 1917 and 1923, the top 10% bounced from a low of 38% in 1921, to a high of 43% in 1923. The reported incomes of the top 10% were not apparently impacted by the 750% increase in top marginal rates.

The data on the top 10% doesn't support Beale's Law. However, I question the credibility of all this early data. I'm guessing that compliance/enforcement was selective and that tax shelters were widely available to the top 1%. Those issues make it very difficult to compare data from the early decades of the income tax (generally unreliable) to more current data (generally reliable).

[2] Second Two Decades (1923-1943)

As noted, the top rates stair-stepped down from roughly 72% in 1920, to 58% in 1922 or 1923, to roughly 25% from 1925 through 1932. They jumped back above 60% in 1933, and were close to 90% in 1943.

The reported income of the top 1% plummeted from its peak of 23.9% in 1928 to approximately 15% in 1931 or 1932. Top tax rates were stable, but 1932 and 1933 were the worst years of the Great Depression. Even the richest Americans are not immune from the forces of wealth destruction. (A similar trend emerged in the wake of the 2007 financial crisis.)

From 1933 through 1937, top rates climbed precipitously, but the top 1% saw their reported incomes rebound to approximately 19% of reported totals. Their reported income then began a period of gradual decline, reaching approximately 12% in 1943.

The top 10% of taxpayers reported approximately 45% of total income from 1923 through 1940. Reported income declined sharply around 1941, hitting 34% in 1943.

Lessons from the period? The top 10% were basically insensitive to the top marginal tax rates for most of the period. Beale's Law does not hold if we focus on that group. The top 1% saw income losses while rates were stable, and income gains during a period of increasing rates. In the early 1940s, Team Roosevelt increased the top rate above 90%. The top 1% began to report a lower share of total pre-tax income. However, we don't know how much of that decrease is attributable to weak compliance/enforcement and tax shelter activity by the ultra-wealthy. My guess is that tax sheltering drove much of the data for the next two decades.

[3] Third Two Decades (1943-1963)

The top tax rate was consistently in the 90% range throughout this period.

The top 1% bounced from peak income of approximately 12.5% (1944) to approximately 10% in the late-1950s and early 1960s. These data are somewhat consistent with Beale's Law. However, Beale's Law would not predict a 20% decline in reported income during a period of rate stability. Remember, Beale's Law suggest that income inequality is inverse to rates. During periods of rate stability, we'd expect to see stable levels of income inequality.

The top 10% bounced from peak income of approximately 37% (1946) to approximately 33% in the early 1950s. The reported incomes of the top 10% were very stable, which is consistent with Beale's Law. However, taxpayers in the 90th to 99th percentiles reported income gains, while the top 1% reported income losses over the period. Guess who can allocate more cash to planning and implementation of tax shelters?

[4] Fourth Two Decades (1963-1983)

President Kennedy slashed the top rate from 90% to 70% in 1964. Reagan chopped the top rate to 50% in 1982.

Despite the top rate decrease, the top 1% reported approximately 10% of pre-tax income throughout the period. The trough was 8.9% in 1976. Reported income began to trend upward around the time of the Reagan tax cut.

Similar to the prior period, the reported incomes of the top 10% were very stable. Again, taxpayers in the 90th to 99th percentiles reported income gains, while the top 1% reported income losses.

Lessons from the period? Under Beale's Law, we would expect to see a jump in income inequality after Kennedy cut the top rate from 90% to 70%. Something else was going on, because the top 1% reported declining incomes through the mid-70s. I smell ... tax shelters!

[5] Fifth Two Decades (1983-2003)

From a historical perspective, this period was relatively volatile. The top rates declined from 50% in 1983 to 28% in 1988. Bucking his predecessors, Clinton raised the top rate to 40%. Bush Jr. cut the top rate to 35% in 2004.

The top 1% reported approximately 15% of pre-tax income in the mid-'80s through about 1994. From 1994 through about 1999 or 2000, the group's reported income spiked to approximately 22%. Reported incomes fell when the dot-com bubble popped (2000-2001), and then rebounded to 23.5% in 2007.

The top 10% largely tracked the top 1%. The group reported approximately 36% of total income in 1983, spiked to approximately 47% in 2000, fell for a couple years, and then rebounded to a high of approximately 50% in 2006.

Lessons from the period? Under Beale's Law, we'd expect to see a decline in income inequality after Clinton raised the top rates. In fact, income inequality continued its sharp ascent notwithstanding the tax increase. History repeats itself. The top 1% took a relative hit when the dot-com bubble popped (compare peak years of Great Depression). They rebounded before the Bush tax cuts took effect.

[6] Conclusion

Beale's Law is a pseudo-scientific assertion seemingly grounded in historical data. When we scrub the numbers, however, the theory crashes and burns. During periods that tax rates changed, we don't see "inverse" reactions in reported taxable incomes of the top 1%. During periods of relative stability, we see patterns that aren't consistent with Beale's Law. For much of the history of the income tax, the top 1% (or perhaps the top 0.01%) were able to exploit tax shelters that were unavailable to taxpayers in the 90th to 99th percentile. When loopholes were tightened, their share of reported income rebounded significantly.

Moreover, Beale's Law disregards the macroeconomic currents that cause income inequality to ebb and flow over time. In the late-'90s, for instance, we see increasing inequality despite the Clinton tax increases (no inverse relationship). The evolution of new technologies and increased fluidity of global capital markets have concentrated the rewards to talented innovators and entrepreneurs. Beale's Law, and other pseudo-science on the political left and political right, is guided by political expedience: fitting complex economic issues into a simple box for widespread consumption by the relatively unsophisticated masses.

Although I've spent too much time debunking Beale's Law, her fundamental argument misses the point. Beale anguishes over the fact that income inequality has ebbed and flowed over time. She yearns for the "good old days" when the top 1% reported a smaller percentage of total income (8.9% in 1976). But the key question is whether the top 1% are paying a fair share of taxes necessary to run the government that Americans desire. The top 1% paid approximately 37% of total income taxes in 2009. Is that enough? Too little? How much would satisfy Professor Beale? And what if we increase taxes and don't cause a dent in income inequality? What's the next step from Professor Beale's perspective?

1 comment:

  1. Let's all elect 535 personal shoppers for Christmas. Debunk that and you'll debunk war.