While relaxing after several hard days of skiing in Telluride, my brother Dick suggested a change to the chart I featured in a post earlier this month. I liked his suggestion, which was to add color to the bars in order to distinguish between periods of rising (red) and falling (green) debt burdens. And I like the chart even more now—though it has been highly controversial, generating more positive and negative comments than almost any post I can remember. Since this is a very important issue, I think it bears repeating. What follows is adapted from the original post, with additions meant to clarify some of the issues that came up in the comments.
The chart above addresses the issue of how much federal debt was accumulated during past presidencies. There are many ways of calculating this, and it is of course difficult to lay the blame on any president for increasing the national debt, since only Congress can spend money. But the president sets the tone of the national debate, and like any chief executive, he must bear the ultimate responsibility for what happens during his watch.
I think there is only one correct way to calculate the burden of our national debt, and that is to compare the debt owed to our national income (GDP). The nominal size of our economy is a critical variable, since it is one thing to owe a million dollars if your income is 10 million, and it is quite another to owe a million dollars if your income is only one million. Using this method, if GDP grows by a nominal 10% and our debt grows by the same amount, then there has been no net increase in the debt burden. Some might argue that the average interest cost of the outstanding debt should be factored in, since it is one thing to owe a trillion dollars when interest rates are 1%, and quite another to owe a trillion dollars if interest rates are 10%. But interest rates tend to track inflation over time, so while higher interest rates increase debt service costs, they don't necessarily increase the overall burden of the debt, since the economy tends to grow by at least the rate of inflation over time. Debt service costs are quite low right now because interest rates are exceptionally low, but this could change dramatically once the Federal Reserve starts raising short-term interest rates, and if inflation starts to pick up. On average, and over time, the effect of inflation tends to cancel out the effect of interest rate costs.
Furthermore, I note that Obama's contribution to our debt burden during his first term is likely to be about 25% of GDP, while the next biggest post-war contribution came during two terms of Bush II (15.3%). If Obama wins a second term and the economy and fiscal policies evolve along the lines currently projected by OMB, Obama could end up adding more to our debt burden that all past presidents combined. I don't think there is any way to escape the conclusion that fiscal policy under Obama has been conducted in reckless and unprecedented (with the exception of WW II) fashion, especially since all that deficit spending has done practically nothing to improve the economy—indeed, as I have been arguing for the past three years, "stimulus" spending has only weakened the economy, since it has consisted mainly of transfer payments which create perverse incentives.
For Federal debt I use the amount of debt held by the public, not total debt (which includes debt the federal government owes to itself, e.g., to social security). If I included all debt, then the current debt burden would be more than 100% of GDP. I think it's reasonable and conservative to use debt held by the public, since debt that the government owes itself is not necessarily an inescapable obligation; social security benefits are not cast in stone, and they are not an individual's property, whereas Treasury debt held by an individual is sacrosanct. Social security obligations can change at the whim of politicians (e.g., by raising the retirement age), and the tweaking of benefit formulas—for example, adjusting future payments by the rate of inflation instead of by the rate of wage inflation (which includes a real and an inflation component) would cause the future liabilities of social security to decline dramatically.
For the amount of debt incurred during each presidency, I have compared federal debt outstanding to nominal GDP at the end of each calendar quarter immediately following a president's assumption of office. Thus, Bush II starts in Mar. '01 and ends in Mar. '09. For one, this allows me to use GDP data which is only available on a quarterly basis. I also think it makes sense to give a new president a month or two to get his feet on the ground, and to credit outgoing presidents with the policies they set in motion before they left. I recognize that a new president may find it difficult to immediately and significantly change the policies he inherits, but he has at least four years to get things right. In this regard, I note that there have been several sizable reversals in the chart above (e.g., Clinton inherited a rising debt burden but managed to reverse that in significant fashion, while Bush II did just the opposite).
In order to project the federal debt burden as of Mar. '13, I assume that the federal deficit in the 15 months prior will be an annualized $1.2 trillion, and that nominal GDP will grow by an annualized 5%. I believe these are conservative assumptions for the purpose of this analysis, since the OMB is projecting a higher deficit and there are many analysts projecting slower GDP growth. A bigger deficit and slower GDP growth would result in a larger increase in the debt burden than shown here.