Note to Readers:  This is not our usual economic commentary.  We use this vehicle to introduce a new index called the CUMB-E Index of Federal Reserve Policy Flexibility.  It attempts to capture the extent to which the Fed can change interest rates as part of its exit strategy from its quantitative easing policy without exhausting its capital.  The measure also provides insights into the feasibility of engaging in asset sales as a means to tighten monetary policy, should it become necessary to fight inflation.  The index can be viewed as a measure of the Fed’s interest rate risk exposure.  The following explanation references four charts that can be accessed via our website by either clicking on the CUMB-E Index box for attachments or just using this link and scrolling through the six pages, (http://www.cumber.com/content/index/duration.pdf).  The following explanation should be viewed as a technical note and, as such, may not appeal to all readers.

In early July, David Kotok and I published an economic commentary titled “What If the Fed Were a Bank?” (July 7, 2009).  The focus was on the Fed’s balance sheet and what might happen – given the recent expansion of its assets into private-sector longer-term mortgages, debt instruments, and Treasuries – if the Fed had to sell those assets quickly in order preemptively to curtail inflation and withdraw the extraordinary amount of liquidity it has injected into the economy.  In particular, we noted that “Over half of its outstanding loans and other asset holdings have a maturity of over one year, and over 32% have a maturity of five years or more.  This means that the durations of the Fed’s assets have increased substantially during the crisis period and are likely to increase even more as its mortgage-related holdings continue to expand.  The increase in duration plus expansion of its mortgage-related assets to potentially $1.25 trillion means that the Fed’s interest-rate and credit-risk exposure has increased substantially.” 

We conjectured that, assuming an average asset duration of five years and an average liability duration of eight years, a calculation using a measure of the duration of capital, an increase in interest rates of one percentage point would reduce the Fed’s equity by 38%.   After publishing that piece, we began to think even more about the Fed’s interest-rate risk problem and how it might affect the feasibility of employing asset sales or increases in interest rates as policy tools.  We decided to create an index of “policy flexibility” based upon measures of the Fed’s asset durations to highlight the kind of information that a truly transparent Federal Reserve should be providing.  By necessity we had to make many simplifying assumptions.  The purpose of this commentary is to introduce the index and to detail as much as possible the assumptions that were made in its construction.  We invite comments and suggestions as to how our methodology might be improved, and we sincerely want to take those suggestions into account going forward.  Of course, our efforts are no substitute for what the Fed might be able to do, because of the information it has on its asset holdings.

We will make the index, which we have named the CUMB-E Index, available weekly on our website in two forms.  One takes the Fed’s balance sheet at face value, while the other recognizes that the Fed has one significantly undervalued asset which, if were it marked to market, would provide a significant cushion to absorb capital losses if it had to increase interest rates or sell assets: that is its holding of gold certificates. 

Perhaps the most important change from our original approach was to rethink how we treat the Fed’s liabilities.  Previously, we attempted to proxy the durations of liabilities in attempting to estimate the duration of the Fed’s capital account.  Upon further reflection, however, we realized that essentially all the Fed’s assets, except for its capital, are carried on the books at their market value, which is the same as their nominal value.  For example, the market value of currency is equal to its book value.  The same is true for member-bank deposits held at the Fed.  Even though the Fed now pays interest on excess reserve deposits, they are payable upon demand and essentially have a zero duration for computational purposes.  (Note: We are interested in any suggestions as to a better assumption that might be made on this point.) 

The real nuances of the computations center on the asset side of the Fed’s balance sheet.  It has loans of various maturities, private-sector assets, and long-term Treasuries, which are now detailed in Table 2 of the Board’s H.4.1 release each week.  The two asset categories that matter the most in terms of approximating the Fed’s interest-rate risk exposure and translating it into a measure of policy flexibility are the large holdings of Treasury securities in the 1 to 5 year, 5 to10 year, and the over-10 year maturity categories, and the very large portfolio of mortgage-backed securities that have maturities over 10 years.  Approximately 87% of its $731 billion of Treasury securities are currently (as of August 12, 2009) over 1 year in maturity, and 100% of its $609 billion in mortgage-backed securities have a maturity of over 10 years.  And this $1.4 trillion of long-term assets is soon to become even larger if the Fed lives up to its pledge of expanding its mortgage-related assets by $641 billion to $1.25 trillion by the end of this year.  Fortunately, we were able to obtain reasonable approximations for the duration of the Fed’s Treasuries and mortgage-backed securities by using the modified durations provided by Barclays Capital on their website. (Note: Again, we invite suggestions as to how to improve the estimates and what alternative data sources might exist.  https://live.barcap.com/BC/barcaplive?menuCode=MENU_FI_WELCOME)

Chart 3 lays out the basic issue.  On a book-value basis, the Fed has historically been highly leveraged when compared with commercial banking organizations.  For the period 1989-2002, its leverage has averaged between 40 and 60 times, but then trended down significantly until the financial crisis hit in the fall of 2008.  Leverage then spiked up to between 80 and 90 to one, reflecting the rapid expansion of excess reserves held on deposit by member banks.  An astute student of the Fed’s balance sheet will note that this leverage ratio is misleading, because the Fed is carrying its holdings of gold, in the form of gold certificates, at the book value of $42.22 per ounce.  This asset has experienced significant capital gains, which if recognized would add to the Fed’s capital account.  Indeed, it would add over $235 billion to the $49.8 billion that it currently lists as total capital.  Chart 4 shows the impact that revaluation of this gold holding would have on the Fed’s capital ratio, compared with not recognizing the unrealized capital gains. 

The question is, how would a change in interest rates affect the Fed’s capital position and hence, its policy flexibility?  Using an approximation as proposed by Kaufman and Bierway (1983), we have created two indices, shown in Charts 1 & 2.  The index essentially provides an estimate of how large of a percentage-point change in interest rates the Fed might be able to engineer without generating sufficient capital losses on its balance sheet to wipe out its equity if its assets were marked to market.  This would happen if either the Fed attempted to sell assets to raise interest rates or if the term structure shifted in a parallel fashion.  Chart 1 suggests that the Fed would only have about a 78-basis-point of policy flexibility to increase interest rates without reducing its capital position to zero.  This is down from the cushion that had built up from 2004 until the fall of 2008.  The Fed’s cushion is now back down to where it was in the late ’90s, during the rapid expansion.    This picture is not so bleak when one recognizes the impact that a revaluation of gold would have, as is shown in Chart 2.  The same basic pattern emerges, but now the Fed could increase interest rates by about 396 basis points before it would be capital-deficient.  Of course, revaluing gold once would seem to require the Fed to change its accounting system and reflect gains and losses due to fluctuations in the price of gold (and other assets) on its asset and liability accounts.  Finally, should the Fed increase its mortgage portfolio to the projected $1.25 trillion level, then the index would drop to .45 without gold and 2.35 if the capital gains in the Fed’s gold stock were recognized.

The point of this entire exercise is to demonstrate that the Fed has substantially less flexibility to exercise an exit strategy, given the present long maturity of its assets, than may have previously been apparent.  The concern is real because of the huge deficits the country faces and the implications that this has for the level of interest rates.  Accounting gimmickry can only paper over this issue for a short while.  Should markets or foreign governments suddenly become concerned about the solvency of the Fed, rationally or not, this would have significant implications for not only the Fed but the entire nation as it attempts to raise trillions of dollars in its efforts to engage in deficit spending to stimulate the economy.