An insistent question of our time is, how much government debt is too much. Is there some debt level that becomes crushing as opposed to merely costly? The controversy over research by economists Carmen Reinhart and Kenneth Rogoff shows how explosive the issue is. * * *
One group of economists and policymakers argues that annual deficits must be cut because they’re creating — or have already created — dangerous debt levels. Another group contends that large deficits are needed to propel stronger recoveries and reduce huge unemployment. It’s “austerity” versus “stimulus.” If debt exceeding 90 percent of GDP is hazardous, then the case for austerity seems stronger. (Already many countries exceed or are approaching the 90 percent mark.) If not, deficit spending remains a possible temporary spur. Which is it? Although the newly discovered errors in Reinhart and Rogoff’s 2010 paper (“Growth in a Time of Debt”) are embarrassing, they do not alter one of its main conclusions: High debt and low economic growth often go together. * * * Whether debt causes low growth or merely reflects economic weakness is undetermined. * * *
What’s sobering about this brawl is that it settles nothing. With some exceptions, most advanced countries, including the United States, seem caught in a similar trap. Their debt/GDP ratios are high and rising, so it’s hard to embrace massive deficit-financed stimulus programs. But austerity programs of spending cuts and tax increases may dampen growth and raise debt/GDP ratios. There is no obvious exit from this dilemma except a burst of spontaneous growth, which is conspicuous by its absence. (Emphasis added) — Robert J. Samuelson 
This is as good a summary as you will find anywhere of the basic features of the Reinhart/Rogoff debate, as it has played out in the neoclassical arena. What I especially like about Samuelson’s summary is his tone of skepticism, and expression of dissatisfaction with the analyses of both sides. The previous two posts in this series have thoroughly reviewed the issues as presented by both neoclassical Keynesians and “supply-side” ideologues, and answered some of the key questions raised in Robinson’s summary of the debate:
1. There is no case for austerity, ever. The “austerity doctrine” is a fictional ideology which, together with the companion “trickle-down” ideology, serves the purpose of arguing for minimal taxation of top incomes;
2. The apparent Hobson’s choice between austerity and stimulus masquerades as a “dilemma” and will continue to do so until the austerity choice is understood to be ideological fantasy: Real-world data emerge from real-world economies, which is why the corrected Reinhart/Rogoff study fails to support their debt threshold theory, and indeed actually refutes the austerity doctrine, of which “debt threshold” is a variation;
3. As Robinson correctly observes, the Reinhart/Rogoff data do show a real correlation between higher public debt and lower growth, and that creates a problem for the neoclassical Keynesians: Running up even more debt cannot, as they argue, provide effective stimulus for recovery because, if higher debt does not directly cause growth to decline (the Reinhart/Rogoff hypothesis), high debt and low growth are jointly caused by other factors. Not recognizing this second perspective is an error in neoclassical Keynesian thinking that results from ignoring the implications of wealth and income concentration;
4. Robinson seems to clearly appreciate that there is no such thing as a “burst of spontaneous growth.” That’s not only anti-Keynesian, it’s virtually anti-Newtonian. Contraction never magically turns into growth — not when taxes on the wealthy are reduced (trickle-down), or when government is reduced instead of raising their taxes back up (austerity) ;
5. There is a real exit from this dilemma — taxation of wealth and top incomes. That this has been obvious to almost no one in recent years will likely go down as one of the most remarkable facts of economic history;
The Reinhart/Rogoff study and the intense debate over it stem from economists asking the wrong questions: Instead of Robinson’s formulation (“How much government debt is too much? Is there some debt level that becomes crushing as opposed to merely costly?”) every economist should be asking herself or himself these questions:
How much income and wealth inequality is too much? Are there levels of income and wealth concentration that become crushing on the bottom 99% as opposed to merely costly?
I started asking those questions more than two years ago, and can now see substantial glimmerings of the answers. There is much work to be done in this area. This post will show that the Reinhart/Rogoff data actually support what I call “redistributional macroeconomics,” and demonstrate that the “dilemma” Robinson has identified is merely a “Catch 22” generated by outmoded economic thinking. Once the true scope of the inequality problem becomes clearer, the answer that continues to elude attention — taxation — becomes obvious.
I appreciate that many readers will find it incredible that economics, a respectable social science, would not have solved a “dilemma” like this long ago; I know I did. But as the first nine posts in this series have explained, Keynes simply did not get it quite right; his failure to account for income and wealth redistribution in his dynamic full employment model was a fatal oversight. Thus, the Reinhart/Rogoff debate has taken place in the “neoclassical arena,” with neither side taking the dynamic consequences of redistribution into account.
Inequality and Growth
Average Annual U.S. Real Income Growth
1947 to 1979 1979 to 2010
Lowest fifth 2.5% -0.4%
2nd fifth 2.2% 0.1%
Mid fifth 2.4% 0.3%
4th fifth 2.4% 0.6%
Top 5th 2.2% 1.2%
The 1947-1979 period was when the U.S. economy grew and prospered, under a system of progressive taxation (the top income tax rate was 91% until 1963, then reduced to 70%) with over 2% growth shared by all five income quintiles. The middle class flourished, and the America of economic opportunity some of us can still remember came into its own. This period began one year after 1946, the year of the deepest collapse and decline in the post-WW II recovery, and the year that led Reinhart and Rogoff erroneously to speculate that, above a “threshold” 90% PD/GDP ratio, debt causes growth to decline.
Before the recent “debt crisis” growth had already declined substantially, for three decades, as a consequence of the “Reagan Revolution.” Deregulation enabled higher profits for large corporations, and much higher incomes for their CEOs and senior management. Perhaps more significantly, however, they were allowed to keep much, much more of these growing profits and incomes: Taxes on top incomes were drastically reduced, and declining tax revenues from corporations soon followed. The marginal tax rate on top incomes was lowered all the way, at one point, to 28%, and corporate taxes declined even more as major corporations began to “locate” in foreign taxing jurisdictions to avoid paying domestic taxes. This materially reduced the federal government’s revenue, and created the need for continuous federal deficits and the accumulation of an enormous amount debt.
What is especially noteworthy, as discussed in earlier posts in this series, is the enormity of the income collapse that accompanied this less progressive taxation See post #5 (here). As shown on the table, GDP (income) growth was cut almost in half for the entire top quintile, and reduced growth was increasingly allocated all the way down to the bottom quintile, which actually experienced negative growth over the entire period 1979-2010.
We now know that scarcely anyone below the top 1% avoided reduced growth over this second period. In their 2011 study,  economists Piketty, Saez, and Stantcheva showed that top 1% real income (per adult) tripled from 1980-2008:
More recently, Saez has informed us  that, as of 2010, 121% of all growth is taking place within the top 1%, which means that for the last few years there has been no growth in the U.S. economy except high within the top 1%.
The Implications for Public Debt
As suggested earlier, although public debt does not directly cause growth to decline, growing public debt and declining growth are causally interrelated with the growth of income and wealth inequality. High public debt in the United States and reduced growth over the last three decades have been causally interrelated via this mechanism:
(1) Overall growth has been curtailed by the growth of income (and wealth) inequality;
(2) The constraint of taxation on the growth of income and wealth inequality has been reduced by a marked decline in the taxation of top incomes (and wealth);
(3) These tax reductions on top incomes (and wealth) have been financed by government borrowing and an ever-increasing level of accumulated debt;
(4) Thus, America’s high public debt is a joint consequence, along with higher inequality, of the reduced taxation of top incomes and wealth.
In the Reinhart/Rogoff debate, the argument about causality centered around whether lower income growth causes higher debt or whether higher debt causes lower growth. The first explanation is basically a truism (any given year, less income means less tax revenue), and the second explanation is both theoretically baseless and statistically disproved. The relationship between debt and growth is explained, however, by their interrelationship with the factors of taxation and inequality.
Here is a graph  of the U.S. federal debt as a percentage of GDP (PD/GDP) from W.W. II to 2010:
The trend is very similar to the trend in the top 1%, 0.1% and 0.01% income shares discussed earlier in this series of posts. Both income inequality and the PD/GDP ratio bottomed out in the 1970s, but have increased substantially ever since. The huge outstanding balance of government debt now totals more than $16.7 trillion, so once again it has reached the level of GDP.
While the top 1% income share (which rose from 8.9% in 1976 to 23.5% in 2007 — see post #2, here) and the PD/GDP ratio have both risen substantially, they are not directly correlated. While income inequality skyrocketed during the Clinton Administration, for example, the PD/GDP ratio nonetheless declined. As is well known, the Clinton years were notable for relatively strong growth, including investment in a vast fiber optic network and the “dot.com” revolution:
1. Because of the strong economy and an increase in the taxation of top incomes (the top rate was raised from 30% to 40%), the Clinton administration was able to maintain a balanced budget, except for the interest on the debt compiled in previous administrations. Consequently, the PD/GDP ratio declined, as shown;
2. However, because the tax structure remained too regressive to control growing inequality, the top 1% income as shown on the previous graph increased sharply, an immediate reflection of the higher corporate profits in those years.
Were it not for the technology boom in the Clinton years, the PD/GDP ratio on its longer trend-line might well have grown today to the 120% level experienced at the end of WW II. The important questions today are why the nation’s debt has grown so much, and why that growth is now accelerating. Answering these questions requires an understanding of the relationship between income and wealth.
Income and Wealth
The process of inequality growth is intimately tied to the growing concentration of wealth. When income becomes more concentrated in the top 1%, and especially the top 0.01%, wealth is increasingly transferred to the top, effectively removing most of that money from active circulation as income and leaving the bottom 99%’s economy with in an increasingly smaller functional money supply. This leads to more income inequality, in a vicious cycle of growing wealth concentration and income inequality. See post #4 (here).
Edward N. Wolff, America’s premiere chronicler of wealth distribution, has observed the high correlation between the upward redistribution of income and the increasing concentration of wealth:
[T]he level of wealth concentration was at a postwar high in 1998. The time series on income inequality indicates exactly the same result for the concentration of household income. Moreover, the run-up in wealth inequality that characterized the 1980s had a twentieth century precedent only during the 1920s. A similar finding can be reported for income inequality.” 
When income inequality grew over the last 30 years, so did wealth inequality. The Economic Policy Institute (EPI) has computed  that 40.2% of the wealth growth from 1983-2009 went to the top 1%. The wealth transferred to the top 1% measure in the trillions of dollars, moreover, and the top 1% share of wealth fluctuates in a way similar to, and reflective of, the fluctuations of the top 1% share of income. For example, based on the Census Bureau’s record of household and non-profit net worth, and Edward Wolff’s and Linda Levine’s wealth distribution statistics over the period 1982-2010,  I have estimated that the top 1%’s share of total net worth grew from 1982 to 2006 by approximately $14.3 trillion, an amount that declined after the Crash of 2008 to approximately $10.7 trillion in 2010.
Clearly, these are very large gains. To put these numbers in perspective, one percent of the U.S. population is about 3.1 million people, so a gain of $10 trillion in net worth by the top 1% works out to a per capita gain of about $3.2 million for the top 1%; and assuming that wealth is distributed within the top 1% roughly in the same logarithmic manner as income, the top 0.01% takes about one-fifth of such a gain, or about $2 trillion. That amounts to a $64.5 million per capita wealth gain, on average, for all of the 31,000 people in that group. 
I created this graph to show the nominal growth of the national debt and GDP over the period 1982-2012 together with the growth of top 1% net worth. The top 1% data points were determined using the Wolfe/Levine wealth concentration percentages through 2010 and the Census Bureau’s data for total net worth.  This graph of the top 1% share of net worth over this time may be a first; I have not seen another. It is quite revealing, especially considered together with the growth of total income (GDP) and public debt (PD):
It confirms Wolff’s observation that changes in the distribution of wealth and income are highly correlated. At it’s peak in 2006, top 1% net worth had grown by about $20 trillion since 1982. Net worth is constantly fluctuating as asset values change; an initial asset value originally recorded when an asset is produced will change with each transfer as its market value changes; there have been big swings in home and equity securities values, for example. Big swings in the market value of such assets show up far more markedly at the top of the income ladder, where valuable assets are more highly concentrated. Hence, the growth top 1% income and wealth shares will not move in exactly the same way. However, when the top 1% share of income is growing significantly, to the extent the top 1%’s growing wealth is not taxed back down (either out of income in the first instance or out of wealth itself) there is an enormous accumulation of net worth. The distribution of income and wealth within the top 1% is so exponential that roughly $4 trillion of that 1982-2006 gain went to the top 0.01%.
As we know, total income (GDP) growth since 1979 has declined significantly. GDP includes, importantly, top 1% income as well as bottom 99% income. It is extremely significant, therefore, that top 1% net worth has grown so much faster than total GDP, reflecting a very high level of inequality growth. In the 1982-1983 period, top 1% net worth (total wealth) reached the level of GDP (total annual income) for the first time since W.W. II, and it has remained higher ever since. According to Wolff’s data, the concentration of net worth in the top 1% rose from 19.9% in 1976 to 33.8% in 1983, a level it has remained above ever since. (I will prepare a graph for my next post extending these data lines back closer to WW II.)
A deeper understanding of the relationships between income and wealth is fundamental to distributional macroeconomics: Tangible wealth can only be created out of work, which is compensated for in the first instance by income. Income that does not compensate for the creation of real value, as explained in earlier posts in this series, is “economic rent.” This rent is used to purchase valuable assets (e.g., mansions. real estate, corporate shares, expensive yachts, valuable art, etc.) that already exist. The process therefore largely amounts to a transfer of asset ownership from the bottom 99% to the top 1%.
The data for GDP includes all of the top 1% income reported to the I.R.S., including the economic rent. Keynes’s assumption that income and production are roughly equivalent was wrong; the GDP income data disguise a hugely slower growth of aggregate net worth, indeed a decline in bottom 99% net worth, as real assets are moved to the top 1% share as suggested by this graph.
The graph only hints at the level of income inequality and associated concentration of wealth that already existed from W.W. II through the 1970s. It is no coincidence that the national debt started to grow in the Reagan Administration. In the 1970s, a progressive taxation system, the main feature of which was a top income tax rate of 70%, kept inequality from growing, but during the Reagan Administration the taxes at the top were slashed, as discussed above, simultaneously creating more wealth at the top and federal deficits as the federal government was forced to borrow the funds to replace those lost to the tax cuts. Ever since then, the increasing debt has continuously financed all of the tax reductions for top incomes, dollar-for-dollar, that resulted in the rapidly increasing top 1% wealth we see on this graph. Effectively, the federal debt has financed enrichment of the top 1%.
Now it is possible to restate the relationship between public debt and growth another way, perhaps more precisely:
(1) Growth is extremely repressed by wealth transfers to the top;
(2) Public debt since the 1970s has financed wealth transfers to the top;
(3) The public debt is the direct consequence insufficiently progressive taxation;
(4) Insufficiently progressive taxation has therefore caused declining growth.
Growth of the PD/GDP Ratio
Once it is understood that the public debt has essentially financed the enrichment of the wealthiest Americans, it becomes clear that the increase of the PD/GDP ratio to over 90% today is also attributable to that enrichment. That fact is inherently reflected in the Reinhart/Rogoff data. Here again is the chart of the U.S. results for growth provided by Reinhart and Rogoff in their initial workpaper on their “Growth in a Time of Debt” (GITD) study (here):
The Reinhart/Rogoff chart shows that over the 200 or so years covered in their database, growth has been slightly lower on average when the national debt was 30-60% of GDP than when it was below 30%, and lower still when PD/GDP was between 60-90%. It also shows that when the PD/GDP ratios were higher than 30%, the average (mean) of aggregate income growth was lower than the median aggregate income growth; this confirms that a higher level of income inequality is embedded in the GDP data for these periods of higher PD/GDP ratios. And, of course, the income growth (GDP) data at any point of time would reflect the degree of income inequality present at that time. Notably, these relationships are present in the data for all periods, both when debt was growing and when it was declining, and the Reinhart/Rogoff data cover a 200-year period.
For these reasons, that these factors are reflected so strongly in a regression of contemporaneous observations of PD/GDP and growth, as opposed to a time-series study, is really not surprising, despite the many factors that affect growth. But it is telling: The Reinhart/Rogoff corrected study strongly confirms the “redistributional macroeconomics” developed in this series of posts and summarized here. The Reinhart/Rogoff corrected study confirm the relationship between growing income inequality and declining growth.
Some avenues for further study, in addition to full-bore analyses confirming the basic relationships I have described, come to mind:
1. The Reinhart/Rogoff database of GDP (100% of income) growth for the entire 2oth Century could, and should, be regressed with the Piketty/Saez databases of top 0.01%, top 0.1%, top 1% and bottom 99% income shares. Their own study shows that a modest correlation between higher debt and growth is found between PD/GDP ratios and total income growth (including the top 1%), which confirms that inequality dampens aggregate growth, not just bottom 99% growth (a fact already apparent from time series data). A correlation of total growth with PD/Top 1% income, or for that matter with PD/Bottom 99% income, should yield some remarkable results, and help identify and explain “tipping points” for income inequality;
2. A factor that contributes to the inequality growth cycle, and becomes more significant as public debt grows, is the interest on the debt. Currently, the interest is about $200 billion/year, and growing. Some of the interest payments are external, to other countries, and whether such payments cause declining growth likely depends on the balance of trade. Whether internal payments cause declining growth, however, depends on their distribution. It used to be said that the national debt is money “we owe ourselves,” but debt interest is not distributionally neutral: The inexorable growth of debt interest on rising public debt is reflected in growing inequality, as the wealthy increasingly “earn” this debt interest from the not-so-wealthy. A study of how this compounding factor affects growth would be very helpful, I suspect, as would a thorough economic rent study.
Here is how the Reinhart/Rogoff debate should be resolved, taking income and wealth redistribution into account:
1. The Reinhart/Rogoff study shows that growth declines somewhat as PD/GDP ratios increase. However, this declining growth has not been the result of financial crises or government borrowing; more accurately it can be said to have caused these crises, and created the need for government borrowing. The declining growth occurs because of redistribution of income and wealth;
2. The national debt problem grows and compounds over time with the non-linear growth of income and wealth inequality. Higher income and wealth inequality both reduces growth and directly increases the level of debt. So, yes, higher debt/GDP ratios are closely related to lower growth, but both are caused by higher concentrations of income and wealth. The remedy is to restore a workable level of inequality;
3. All of the outstanding $16.7 trillion of U.S. debt has effectively financed wealth transfers to the top 1%, through avoided taxation of top 1% income, which makes the inequality problem today far worse than I had previously dared to imagine. The top 1% net worth gain after 1982 peaked at $20 trillion in 2006. These ongoing transfers consistently reduce GDP and raise the level of debt, so the debt/GDP ratio will continue to rise. The implications:
a. The idea that austerity might help at all, never more than a dream and a prayer, is beyond reason;
b. More federal borrowing will aggravate the rise of the debt/GDP ratio and contribute directly to still lower growth and increased income and wealth inequality;
c. In the interest of debt management, fiscal responsibility, and economic growth, it is essential to institute a sufficiently progressive program of federal taxation without delay.
The Reinhart/Rogoff debate has been very useful in developing principles for a new distributional macroeconomics. Its greatest value, for me, is its demonstration of the enormity of the causal link between income and wealth redistribution and prosperity, putting real meat on the bones of my argument that the one essential remedy for the current decline is the one nobody is talking about, namely, using progressive taxation to counter the growth of inequality. All other economic issues are trivial by comparison.
In my next post, I plan to provide an extension of my graph to include much of the period of prosperity since W.W. II. There is more to discuss, and much more urgent work for economists to do. As this post suggests, Joseph Stiglitz’s admonition that inequality growth has gone too far is, to say the least, accurate. It is time to start working out how close to the end of the road we have traveled.
JMH – 7/8/2013
 “The Reinharet/Rogoff Brawl,” by Robert J. Samuelson, The Washington Post, April 24, 2013 (here).
 In post #8 of this series, we pointed to EPI’s observation that for the United States the apparent debt limit of 90% was unreliable because it relied on only a few years of data in the WW II period; notably the PD/GDP ratio was highest in 1946 (over 120%), a year when the economy declined about 12%.
 “Optimal Taxation of Top Labor Incomes: A Tale of Three Elasticities,” by Thomas Piketty, Emmanuel Saez and Stephanie Santcheva, DP No. 8675, CEPR, November, 2011 (here).
 Edward Wolff, Top Heavy: The Increasing Inequality of Wealth in America and What Can Be Done About It, The New Press, NY, 2002, p. 39.
 “Huge disparity in share of total wealth gain since 1983,” by Lawrence Mishel, Economic Policy Institute, September 15, 2011 (here).
 “Income, Expenditures, Poverty, & Wealth,” The 2012 Statistical Abstract, U.S. Census Bureau (here); Edward Wolff, Top Heavy, supra, pp. 82-83; updated in “Recent Trends in Household Wealth in the United States: Rising Debt and the Middle-Class Squeeze—an Update to 2007,” by Edward N. Wolff, Levy Economics Institute of Bard College, Working Paper No. 589, March 2010, Table 2, p. 44 (here); updated for 2007 and 2010 in “An Analysis of the Distribution of Wealth Across Households, 1989-2010,” by Linda Levine, Congressional Research Service (CRS), Table 2, p. 4, July 17, 2012 (here).
 Reports of the Piketty/Saez income distribution data show the top 1% with roughly 20% of income in the last few years, the top 0.1% with 12%, and the top 0.01% with 6%. Here is a recent graphical video representation of this all but unimaginable wealth concentration graphically (here). Although it helps picture the degree of wealth disparity, this video mentions but fails to address the implications for wealth concentration of the increase in income inequality over the last three decades; nor does it point out that the concentration of wealth was extremely high before the 30-year period began.
 The national debt and GDP data are from “National Debt by Year Compared to GDP, Recessions and Other Major Events,” by Kimberly Amadeo, About.com, April 24, 2013 (here); Top 1% net worth figures are computed from the sources listed in fn. 8.