A Case for a Steep Yield Curve
Introduction: All investors spend their time trying to divine the future direction of interest rates. Equally important, particularly to investors engaged in spread investing, is the alignment of rates, or how various rates relate to one another, such as short-term rates versus long-term rates. This particular alignment of rates is called the yield curve. The yield curve for Treasury bills, notes and bonds can be expressed in tabular form.
| Active Issue | Rate |
| 3-Month | 1.56% |
| 6-month | 1.51 |
| 2-year | 1.69 |
| 5-year | 2.56 |
| 10-year | 3.60 |
| 30-year | 4.67 |
On September 30, 2002, the Treasury yield curve was upwardly sloping, as long-term rates-expressed as the cost for the U.S. Government to borrow for 10 years or 30 years-were higher than short-term rates. Chart 1 below depicts the current Treasury yield curve, and also plots the curve as it stood one-year ago and the end of 1997 and 1999. As is clear from this chart, in the last five years the curve has migrated lower and steeper.
Another way to express the alignment of interest rates is the difference between long rates and short rates. A positive number would represent a positive slope to the yield curve, while a negative number represents a negatively sloped yield curve. Chart 2 shows the relative alignment of short-term and long-term rates since 1962, expressed, in basis points (a basis point is 1/100th of a percentage point), as the difference between the yield on two points on the curve, such as the 10-year Treasury and the 1-year Treasury. The curve is normally positively sloped: The difference between the yield on the 10-year and the yield on the 1 year has been positive 78% of the time since 1962, the average spread over that period is 77 basis points.
It is folly to try and predict the future of any aspect of the financial markets. But the future is all that matters to investors, and therefore we try to hazard our best guesses using the inputs we find useful and then assigning a range of probabilities to potential outcomes. But in today’s volatile markets, where stocks and bonds go through multistandard deviation weeks, “visibility” is limited and every economic statistic is pored over for its predictive qualities, forecasting is difficult. We are in the midst of a seemingly intractable recession, the stock market is going through the long torturous decline of a post-bubble hangover, our currency brand-the dollar is the Coca-Cola of the monetary world-is losing market share, the Federal government’s checking account is now operating with overdraft protection, and we are at war (with the prospect for more to come). As former Federal Reserve governor Laurence Meyer said recently, “We’re in a post-bubble economy….[I]t’s very difficult to figure out the forces that interact with each other.”
Systemic Arguments: With all these caveats in place, we nevertheless will make bold to lay out our case for why the Treasury yield curve will stay positively sloped, if not steeply so. There are systemic reasons that make the case, and there are technical reasons as well. The systemic reasons for a positively sloped yield curve are:
The United States economy is weak, if not still in a recession.
The United States is operating with a budget deficit.
The country is continuing to pile on debt.
Chart 3 sets forth the yield curve since 1962, with shaded areas representing recessions. What is clear from the chart is that the yield curve begins to steepen )i.e., the spread between the one-year and ten-year Treasury widens) at or near the beginning of recessions, stays steep through the recession, and in some case steepens further once the recession has been declared over. The National Bureau of Economic Research, the agency that determines the dates of business cycles in America, declared that our current recession began in March 2001 and has not yet declared it over. Assuming the economy has not recovered (the NBER makes its determinations after the fact), at 19 months the current recession will be the longest since the Kennedy administration. This recession comes on the heels of a 10-year expansion, the longest in NBER chronology. Interestingly, the chart points out that the steepening process begins with the short end of the yield curve falling faster than long rates, reflecting the easing actions by the Fed to stimulate the economy, but inflationary forces putting upward pressure on long rates then takes over once Fed policy takes effect.
Chart 4 plots the yield curve time series against the federal budget deficit. The annual government budget surplus or deficit (blue line), corresponding to the right-hand axis, shows that from the early 1960s to the end of the 1970s, our government operated near breakeven, with only small surpluses or deficits. Interest rate spreads during this period were more closely tied to economic conditions rather than fiscal ones. However, the climate of continuous budget deficits continued and worsened in the 1980s and early 1990s, as the elder Bush administration continued the fiscal policies of the Reagan administration. In September 1981, gross federal debt was 31.50% of GDP, and by December 1993 it had risen to 67.65%. The weight of federal spending and borrowing dominated the bond market at this time and spreads widened and stayed wide. The process of bringing the budget back into balance, which began with the Budget Enforcement Act of 1990, resulted in the yield curve flattening and ultimately inverting as we moved into surplus.
We are currently witnessing fiscal history. The current sharp and unexpected turn in the operating accounts of the government-for the fiscal year just ended, the Congressional Budget Office now projects a budget deficit of $165 billion, a 180 degree change from last year and more than $300 billion below the level it had projected just a year ago-has also ushered back in a steeper yield curve. In his testimony before the House Budget Committee on September 12, Alan Greenspan addressed the risk of operating at deficit. “Restoring fiscal discipline must be a high priority….Returning to a fiscal climate of continuous large deficits would risk returning to an era of high interest rates, low levels of investment and slower growth of productivity. To be sure, at the moment, Treasury rates are at the lowest level in more than forty years, and I can scarcely argue that deficits are pressuring interest rates. And, indeed, our current fiscal situation remains more favorable than it has over much of the past few decades. But history suggests that an abandonment of fiscal discipline will eventually push up interest rates, crowd out capital spending, lower productivity growth, and force harder choices upon us in the future”
An aspect of the financial system that cannot be ignored with regard to its potential impact on the shape of the yield curve is the amount of leverage extant in the US economy. Leverage, or debt, is an important issue because of the potential impact on the borrower. In a growing economy, cash raised through leverage can be used to invest in projects and investments that return an amount in excess of the cost of the borrowing, i.e. a positive net present value. In a slow or falling economy, it becomes more difficult to generate returns on investment that exceed the debt service costs. Similarly, an inflationary economy helps the debtor because prices charged by companies rise while debt service costs generally stay the same; a disinflationary or deflationary economy is the debtor’s enemy. Not only is the debt load a potential drag on the economy, but the need to continue to borrow to fund cash shortfalls (i.e., borrowing to cover operating expenses rather than for capital expenses) offers a technical reason for higher long-term rates.
“The US economy has never been more leveraged than it is today,” writes Morgan Stanley’s money market strategist, William V. Sullivan. Indeed, as Chart 5 shows, the most recent data from the Federal Reserve’s Flow of Funds statement indicates that financial and non-financial debt grew to $29.8 trillion in the second quarter of 2002 (that’s about $100,000 for every man, woman and child in the country).
Charts 6a and 6b show that households and businesses have larger nominal debt loads than the federal government, and that the relative levels of debt have increased over time. In 1965, Federal debt was $262 billion and accounted for 26% of total domestic nonfinancial debt, while household debt was $338 billion, equal to 1.3 times federal debt and 33% of total nonfinancial debt. By the second quarter of 2002, federal debt has grown to $3.5 trillion, and accounts for just 17.5% of total nonfinancial debt. On the other hand, household debt, which includes mortgage debt and consumer debt, has grown to $8 trillion, 2.3 times federal debt and over 40% of total nonfinancial debt.
Not only are the debt loads growing, but their growth is accelerating and their growth relative to national income has grown. Leading in growth rates over the long-term has been the growth in the home mortgage sector, which has a compound annual growth rate from 1965 through the second quarter of 2002 of 9.3%, the highest of any of the non-financial sectors over that period. In the last 18 quarters, mortgage debt has grown at an average annual rate of 9%, thanks in large part to low interest rates, soaring house values and the well-oiled mortgage finance machine.
While the economy has much to be thankful for in the heroic growth rate of household debt in the 1990s, it will nevertheless hand over the growth rate title going forward. We expect federal debt to grow faster than other sectors of domestic nonfinancial debt in the years ahead, particularly as refinancing activity slows down and the consumer slows down. Chart 7 shows the growth trajectory of federal government debt. After annual growth rates of 9.3% in the 1970s and 13.0% in the 1980s, federal debt growth slowed considerably during the 1990s as the budget moved into surplus. From 1995 to 2001, federal debt actually fell at an annual rate of -1.2%. The stunningly swift reversal of the federal budget in the past year has been accompanied by a concomitantly swift and huge reversal in the growth of federal debt outstanding: In the second quarter of 2002 federal debt leapt by 13.3% from the prior comparable quarter. To illustrate the connection between leverage and the budget, Chart 7 plots the growth rate of federal debt with the budget deficit.
The relationship shown in the chart above may seem obvious, but the more troubling aspect of the leverage in the system is its size relative to national income. Surplus or deficit, high growth rate or low growth rate, chart 8 illustrates how every dollar of GDP supports more debt every year. Until the 1980s, the ratio of total domestic debt was less than 1.7:1; in other words, it took $1.70 of debt to create $1 of GDP. Today, generating $1 in GDP requires $2.89 in total debt. Any credit analyst worth his or her salt can tell you that chart 8 is a picture of a deteriorating credit condition. And, very simplistically, an entity with a deteriorating credit condition-a person, a company, an economy-has to pay higher interest rates.
Of course, a country with a currency printing press is not subject to the same credit analysis as a family of four or a C-corp. But the methods by which an indebted nation manages to meet its obligations in times of stress-printing more money, pumping up the monetary aggregates, flooding the banking system with liquidity, devaluing the currency, fomenting inflationary tendencies-are a recipe for higher rates. To bring the growth in debt back to spreads, the sector of total domestic nonfinancial debt that has the greatest correlation with the shape of the yield curve is, not surprisingly given that it controls the reins of fiscal policy, the debt of the federal government. Chart 9 shows that there has always been a loose correlation between the growth in federal debt and the spread between the 10-year and 1-year; and that the correlation has gotten tighter over time. From 1964 to the second quarter of 2002, these two time series have a correlation of 0.41 (perfect negative correlation is -1, no correlation is 0, and perfect positive correlation is +1). From 1983 to date the correlation is 0.51, from 1992 to date is 0.69, from 1998 to date is .73, and from 2000 to date is 0.80. Clearly, as the budget cycle has led to deeper deficits and greater surpluses, the effect of changes in federal borrowing practices has had a greater impact on the yield curve. Given the trend in the correlations, a continued high rate of growth in government borrowing anticipates wide spreads.
Technical Arguments: We have thus far been discussing systemic arguments for a continued positively-sloped yield curve. There are some technical arguments that bear mention as well, beginning with the entity that controls the short end of the yield curve, the Federal Reserve. In setting monetary policy, the Fed has a few tools at its disposal, the most visible (but not always the most powerful) of which is setting a target for the Fed Funds rate. At its latest meeting on September 24, 2002, the Fed voted to keep the Fed Funds rate target at 1.75%, the lowest in two generations, with a bias towards continued accommodation. “Over time,” the Fed announced, “the current accommodative stance of monetary policy, coupled with still robust underlying growth in productivity, should be sufficient to foster an improving business climate. However, considerable uncertainty persists about the extent and timing of the expected pickup in production and employment owing in part to the emergence of heightened geopolitical risks.”
Given the continued sluggishness of the economy, the lack of inflation, the weakness of the equity markets and the loss of paper wealth for so many Americans, the upcoming elections and the uncertainty in the global financial markets, it appears unlikely that the Fed will be raising rates at any point in the near future, at least not with anything other than symbolic significance. The Fed’s hand is stayed by these factors, and as a result it is likely that the very short end of the curve will stay at or near its current levels for some time to come.
The question becomes, then, what happens on the long end. Fed watchers have always debated the long-end ramifications of any Fed move. That said, we believe the systemic reasons outlined above will help to keep long-term rates higher than short-term rates. There is one other technical market factor to consider: On September 16, 2002, Fannie Mae released its monthly summary of activity in its mortgage portfolio, including mortgage commitments, purchases and sales, portfolio growth rate, net interest margin, delinquencies and duration gap. The headline piece of data emanating from this release was that the company’s duration gap had widened to a negative 14 months on August 31, reflecting the recent sharp decline in mortgage rates. Fannie Mae’s duration gap measures the difference, in months, between the durations of the assets and the liabilities in its mortgage portfolio. The target for the duration gap, established by management, is to be within a band of plus or minus 6 months. On October 1, Fannie Mae rushed out to the market the news that its September duration gap had narrowed slightly to 10 months.
It is not unusual for Fannie Mae’s duration gap to be outside of this band; over the past dozen years the duration gap has been outside the target range one-third of the time. Nor is it unusual for the duration gap to turn negative during periods of high refinancing (the Mortgage Bankers Association Refinancing Index hit an all-time high in the fourth week of September). What is noteworthy about the recent disclosure is that with the September data point, it is now three months in a row that the company has had a duration gap outside of its target band. Clearly, the models that Fannie Mae employs to manage interest rate risk did not contemplate current record-low levels of interest rates or the scale of refinancing activity in the market. (Freddie Mac, Fannie Mae’s twin in the mortgage finance market, disclosed that in the same period its duration gap was zero, circumstantial evidence that Fannie Mae’s models might not be working as well as Freddie Mac’s in this regard.)
On a company-specific level, this condition suggests that the current wave of refinancing has resulted in a substantial shortening of the duration of Fannie Mae’s assets relative to is liabilities. Since Fannie Mae generates a significant portion of its revenues through the spread between what it pays on its liabilities and what it earns on its assets, a refi-induced shortening of the duration of its assets will lower their yield and may end up squeezing net interest margins. Going forward, Fannie Mae has different methods it can employ to bring its duration gap to within its target band, all of which ultimately either lengthen the duration of its assets or shorten the duration of its liabilities or both. To shorten duration on its liabilities, Fannie can call its longer duration debt or issue shorter-term liabilities. To extend duration on assets Fannie can use interest rate derivatives or purchase longer-duration assets like other mortgage-backed securities or Treasurys. The market action of Fannie Mae (and not just Fannie Mae, but also any other financial institution or investor that is combating a duration problem) has placed an artificial bid in the market for long-dated Treasurys. Chart 10 shows the relationship between Fannie Mae’s duration gap and the 10-year Treasury.
When we talk about the relationship between Fannie’s duration gap and the 10-year, it is hard to separate what is cause and what is effect. Do swings in the 10-year yield make the duration gap move, or does the move in the duration gap cause Fannie Mae to manipulate the 10-year through its rebalancing operations? We believe that in a normal environment-that is, when Fannie’s duration gap is within its band-the 10-year is the causative agent and Fannie’s duration gap is adjusted. However, as Fannie attempts to bring its duration gap back to within its self-imposed band, it goes into the market and buys or causes others to buy Treasurys, thereby pushing yields down even further. James Bianco of Bianco Research has posited that Fannie Mae bought the equivalent of $60 billion of 10-years in the month of September just to move its duration gap from 14 months to 10 months. The irony in this case is that Fannie Mae is now such a huge player in the credit markets that its struggle to bring its duration gap into more acceptable levels actually exacerbates the problem. We don't know how long this will continue, but it is clear that once Fannie Mae manages its duration gap to within its band, the bid related to its rebalancing effort will lessen and a weight on the market will be lifted. This should prompt long rates to rise relative to short rates, maybe even sharply.
Conclusion: In the last few months, the yield curve has flattened as yields on the 10-year have come down while short-term rates have stayed fairly constant. In May, the spread between the 10-year and the one-year was Fed Funds rate was 286 basis points; by the end of September that spread had narrowed to 205 basis points. The curve is still positively sloped and, on a historical basis, steeply so, but this movement in rates begs the following question regarding this analysis: While this paper is arguing for a continuation of a steep yield curve, what could cause the curve to continue to flatten? The short answer is: Mostly disaster scenarios. First of all, when yields continue to drop as they have, investors are pricing in expectations for continued low or no growth and low inflation, all other things being equal. As long as this expectation obtains-the usual calendar of economic statistical releases and corporate profit reports should be monitored for signs of strength-then the bond market will continue to be strong. Second, the change in the slope of the yield curve has become closely linked to the performance of the equity markets. Tom Sowanick, strategist at Merrill Lynch, recently pointed out that over the last 6 months the correlation between S&P futures and the 10-year to 30-year spread is close to 93%. Third, the downside of the heavy debt load in the US could be continued record defaults, leading to a situation where the price of credit is cut as banks and other lenders compete for the business of creditworthy borrowers. Fourth, if fiscal and monetary policy proved ineffective in jumpstarting the economy, combined with record defaults impairing credit creation in the banking system, were to lead to a Japan-style deflation in the US, then our whole yield curve adjusts downward to resemble Japan’s. While this possibility has been hotly debated over the last several months, most observers believe this to be an unlikely outcome given, among other things, the vastly superior structural soundness and capitalization of our banking system.
With any forecast there are flies in the ointment. To us, the weight of evidence supports the case for a continued steep yield curve.
October 16, 2002
By Jeremy Diamond, Executive Vice President
Annaly Capital Management
This commentary is neither an offer to sell, nor a solicitation of an offer to buy, any securities of Annaly Capital Management, Inc. (“Annaly”) or any other company.
All information contained herein is obtained by Annaly from sources believed by it to be accurate and reliable. However, such information is presented “as is,“ without warranty of any kind, and Annaly, in particular, makes no representation or warranty, express or implied, as to the accuracy, timeliness, or completeness of any such information or with regard to the results to be obtained from its use. While Annaly has attempted to make the information current at the time of its posting on the site, it may well be or become outdated, stale or otherwise subject to a variety of legal qualifications by the time you actually read it.
©2002 by Annaly Capital Management, Inc. All rights reserved. No part of this commentary may be reproduced in any form and/or any medium, without express written permission of Annaly.