Federal Reserve Chairman Bernanke is likely to get some critical questions from members of Congress when he testifies on Capitol Hill this week. One of the easiest for him to address is whether quantitative easing (QE) poses a risk to the Fed’s balance sheet. Bernanke may be worried about the political optics of potential future Fed losses, but he knows very well that any such losses would be more than offset by gains to the Treasury and that QE unambiguously reduces our national debt burden.
The potential for future losses at the Fed is explored in a recent Fed study (Carpenter et al. 2013 [pdf]) and a study by staff at the International Monetary Fund (IMF 2013) [pdf], which show that Fed profits are likely to decline below normal for a few years in the future, especially if the Fed is forced to raise interest rates sharply to fight a surge in inflation. The IMF’s “tail risk” scenario is the most pessimistic of the scenarios considered in these two studies. In this unlikely development, short-term interest rates would rise 600 basis points and long-term rates would rise nearly 400 basis points abruptly at an unspecified future date. According to the IMF, in this scenario the market value of Fed assets would decline by about 4.25 percent of GDP.1 The Fed does not mark its assets to market value, and in any case, as the monopoly provider of legal tender, the Fed does not need to have a positive net worth.2 But, depending on how long interest rates remain at these levels, the Fed might not have any profits to remit to the Treasury for a long time. By itself, this would increase the federal budget deficit and gradually raise our national debt over time.
But future losses to the Fed from this hypothetical event are only part of the picture. To gauge the overall effect of QE on the US national debt burden, one reinforcing factor and four offsetting factors must be considered. The reinforcing factor is that any rise in interest rates above the pre-QE level would raise Treasury’s borrowing cost temporarily. The offsetting factors are that (1) prior to any rise in interest rates, the Fed would have earned extraordinarily high profits from QE for several years; (2) prior to any rise in interest rates, the Treasury would have been able to borrow at extraordinarily low rates for long terms; (3) higher economic growth and higher inflation created by QE would raise tax revenues; and (4) higher inflation erodes the real value of the debt.
In late 2009, I examined all these factors and calculated the effect of a large expansion of QE (to about where the Fed will be this fall) on America’s future debt burden. In a scenario roughly similar to the IMF’s tail risk scenario, I found that additional QE would permanently reduce the ratio of federal debt to GDP by 1 percentage point.3 In a scenario more consistent with current financial market expectations, I found that additional QE would permanently reduce the debt ratio by nearly 3 percentage points.
These estimates included all the potential costs of QE but not all of the potential benefits, because they did not include the benefits of the first round of QE that was implemented in 2009. In addition, they assumed that the period of low rates would last only one year in the tail risk scenario and two years in the baseline scenario, but in fact they have lasted more than four years already and will surely last at least another year or two.
If inflation remains around 2 percent and interest rates gradually return to normal, as implied by current yields on nominal and indexed bonds, the Fed will make lower than normal profits for several years in the future. However, the present discounted value of below-normal Fed profits after 2015 is smaller than the actual and projected above-normal Fed profits in 2009–15, implying that QE is slightly profitable on net from the narrow perspective of the Fed.4 Under the same assumptions, the Treasury Department will have saved at least 2 percent of GDP in interest expenses on long-term bonds issued in 2009–15 without any offsetting higher interest expenses in subsequent years.5
In addition, to the extent that QE has boosted economic activity and/or inflation, this has increased tax revenues. Fed Vice Chairman Janet Yellen (2013) estimates that $500 billion of long-term asset purchases by the Fed would reduce the debt ratio by 1.5 percentage points in the long run, mainly owing to higher tax revenues. Scaling this up to a projected level of $4 trillion in long-term assets held by the Fed by early 2014, QE would reduce the US debt burden by 12 percent of GDP. Under the tail risk scenario, the debt ratio would be reduced by about 2 percentage points less than it would under the assumptions used by Yellen.6
In my view, the recent good news about the federal deficit owes in no small measure to the Fed’s QE policy. Concerns about potential future losses on the Fed’s balance sheet are misinformed and misguided.
1. This estimate assumes total purchases of long-term assets in 2013 of about $1 trillion.
2. The central bank of Chile, for example, had a negative net worth for many years after incurring large losses on its foreign exchange reserves. The negative net worth of its central bank did not prevent Chile from being the star economic performer of Latin America over the past two decades.
3. In the tail risk scenario, I assumed that the seven-year Treasury yield jumped to 5 percent permanently, the federal funds rate jumped to 6 percent before declining gradually back to 4 percent over two years, and the inflation rate moved up only slightly to 2 percent. Higher assumed interest rates increase the debt ratio whereas higher assumed inflation rates reduce the debt ratio.
4. This assessment is based on figures 3 and 12 of Carpenter et al.
5. According to the Treasury Department, net issuance of Treasury securities with maturities of at least two years totaled more than $5 trillion in 2009 through 2012. At least another $1 trillion are expected to be issued in 2013 and 2014. Assuming an average maturity on these securities of seven years and a reduction in yield from QE of 1 percentage point implies a present discounted value of net interest savings of 2.7 percent of GDP.
6. This equals the difference between tail risk and equilibrium short-term interest rates cumulated over a two-year period and multiplied by federal debt held by the public minus Fed holdings plus Fed interest-bearing liabilities. This overstates the cost to the extent that most long-term debt would not roll over during the two-year period and long-term rates would not rise by as much as short rates, but the higher costs on long-term debt that is rolled over in this period would last more than two years.