The Economy
Transitions in the bond market—when we see increased NAV volatility or changes in the level of amortization expense due to significant upward or downward moves in rates—are always a good time to review how we manage the primary risk to which we are exposed, interest rate risk. In general, this risk evidences itself when changes in the shape and level of the yield curve affect total return through changes in spread, prepayment speeds, asset duration and absolute price levels. We manage this risk through maintaining a portfolio of short duration assets which by definition will experience less volatility than longer duration assets. Further, our portfolios blend the performance of floating-rate, adjustable-rate and fixed-rate assets, in which some assets are expected to perform better in a rising rate environment via increased spread income and lower NAV volatility, and some assets are expected to perform better in a falling interest rate environment via capital gains. We call this mix of asset types our "barbell strategy."

A critical point to remember when thinking about the mechanics of how these two "sides" of the barbell complement each other is that they do not happen simultaneously. Rather, they occur over time. Thus, as interest rates fell in the first three months of 2004, our portfolio performance benefited as somewhat longer duration ARMs and fixed-rate securities experienced unrealized mark-to-market gains. These gains will in turn offset the spread compression that we are experiencing as prepayment speeds quicken and amortization expense picks up. In the current rising rate environment, the longer duration securities in the portfolio experience unrealized mark-to- market losses. (Relative to other fixed income strategies, our portfolio mark-to-market losses should be mitigated by the assets’ short duration and negative convexity, i.e., the fact that mortgage-backed securities increase in relative value as rates rise because the homeowner’s option to refinance is less attractive.) In turn, these unrealized losses should be offset in future months by improved spread income from declining amortization expense as prepayments slow and from coupons resetting on floaters and ARMs. Moreover, since mortgage-backed securities are amortizing assets, we are constantly reinvesting into the higher interest rate environment. In other words, this month’s NAV decline is next quarter’s yield pickup. The lesson here is that the benefit of the barbell strategy is realized over the long term.

Our track record demonstrates this point. We have been executing this strategy since 1994—in the US Dollar Floating Rate Fund and other funds—through a wide range of interest rate environments. In 1999, the last significant cycle of tightening by the Federal Reserve, the Fed Funds rate began the year at 4.75% and ended the year at 5.5%, and the 10-year began the year at 4.65% and ended the year at 6.44%. During the year the US Dollar Floating Rate Fund experienced NAV decline from $2.08 to $1.95 but still generated strong income, paying out an average annualized monthly yield of 9.7%. The result was a total return for 1999 of 3.26%. (As points of reference, in 1999 the Lehman Brothers U.S. Aggregate Index returned -0.82% and the five-year Treasury returned -2.84%.) The markets recovered, and the total return for 2000 was 10.88%, even as the curve moved back to inversion in March 2000 (and stayed that way until April 2001).

To repeat: Our strategy is like other fixed income strategies in that we will see NAV movement in a volatile bond market. For us, however, we attempt to mitigate this volatility by investing in a short duration portfolio of assets. Most importantly, over the long term and through all types of bond markets the dominant contributor to total return is net interest income.

This review is germane because we are indeed in a volatile bond market. In March, the bond market was rallying. During the month, the 10-year Treasury bottom-ticked at a yield of 3.681%, average mortgage rates hit a low of 5.35% and the MBA Refinancing Index nearly touched the 5000 level. To their credit, investors began asking about our expectations for performance in an environment of falling rates and rising prepayments. April, however, was a crueler month. The unexpectedly strong nonfarm payroll report that came out on April 2 (308,000 new jobs created in March, the biggest monthly gain in four years and almost three times consensus estimates) sparked a sizeable sell-off and caused a 180 degree turn in the market discussion. The rise in rates—the 2-year and 10-year Treasuries rose from 1.576% to 2.319% and from 3.837% to 4.507%, respectively—has the market digesting a change in stance by the Federal Reserve. Investors now ask about our expectations for performance in an environment of rising rates and falling prepayments.

On Tuesday May 4, the FOMC announced that it was holding the Fed Funds rate at 1% but changed the language regarding its stance on monetary policy. In the release, the Fed announced that risks to the goal of price stability "have moved into balance", a change from the March statement declaration that risks were tilted toward disinflation. In addition, the Fed indicated that it was prepared to remove its accommodative policy "at a pace that is likely to be measured." The market will parse for meaning in that one phrase, but the bottom line is that with the next signs of strength in economic data, particularly employment and inflation, the Fed is likely to raise the Funds rate.

Over the past two years, the Fed has been preoccupied with staving off deflation and the unwelcome sluggishness in resource utilization, both human and industrial. In public speeches and testimony, research papers and FOMC announcements, the message has been consistent: The Fed will do whatever it takes to prevent the type of deflationary spiral that has swamped Japan for the past decade. "[W]hen inflation is already low and the fundamentals of the economy suddenly deteriorate," Fed Governor Ben Bernanke said in his now infamous speech from November 2002, "the central bank should act more preemptively and more aggressively than usual in cutting rates." Bernanke went on to suggest that the Fed could crank up the dollar printing press to add inflationary stimulus. Until recently, the effects of accommodative monetary and fiscal policy weren’t really showing up in the economic data. As we have discussed in past Commentaries, inflation had been virtually absent from the conventional economic measures of pricing power, such as CPI, PPI and the personal consumption expenditure price index, and that there was little traction in the employment and capacity utilization numbers. [Arguably, what we have instead witnessed has been the unintended consequences of accommodative policy: A roaring housing market, the great cash-out refi wave of 2002-2003, a US debt bubble, a global commodity boom and America’s anomalously huge third quarter 2003 GDP number. But that is a story for another day.]

In the last month, however, with the accumulation of positive economic data, the Fed has changed its tune. At his semiannual testimony before Congress last week, Chairman Greenspan said that the disinflationary trends had ended, and that "the federal funds rate must rise at some point to prevent pressures on price inflation from eventually emerging." Prominent investors, including Berkshire Hathaway’s Warren Buffett and Pimco’s Bill Gross, have started saying the Fed should have already raised rates. We believe it may still be too early to declare that the trends have definitely changed, as productivity and inflation in the US continue to be abetted by the benefits of free trade with low-cost countries. The strong employment number in April and May (288,000 new jobs, with an upward revision to April’s number), however, and the recent strength in manufacturing and industrial production have been provocative to the Fed. Capacity utilization, while still historically weak, is solidly off its lows of mid-2003. And while inflation is still not a problem based on the empirical evidence apparently viewed by the Fed, at some point, we believe, higher oil (and milk!) prices should have an effect along the chain of production and pricing. These observable points make it reasonable to conclude that deflation may no longer be a problem. In turn, this means that the tennis-loving Chairman Greenspan, who has been hitting nothing but accommodative backhands since January 2001, has indicated that he is getting ready to switch to his tightening forehand. Ultimately what will drive the Fed’s view on what constitutes a "measured" pace of interest rate hikes will be the performance of a number of data points: continuing the trends in resource utilization and pricing, continued solid performance in consumer sales, industrial production and other signs of improving economic conditions. Specifically, now that this morning’s May report is out, the Fed will have one more employment report to consider before its June 30 meeting, and two additional reports before its August 10 meeting. Other influences that the Fed will likely monitor will be the strength of the dollar, the actions of policymakers in China, commodity prices—particularly oil—and the performance of the overall bond market.

The Mortgage Market
With the rise in yield of the 10-year Treasury bond by 67 basis points during the month of April, the mortgage prepayment landscape has changed dramatically. Freddie Mac’s Primary Mortgage Market survey rate rose to 6.01% as of April 29, 2004, a rise of 61 bp from the end of March rate of 5.40%. Further, the MBA Refinancing Index, which measures all applications to refinance existing mortgage loans, went from a seven month high of 4988 in the middle of March to a recent low of 2403. Consequently, most of Wall Street has lowered their estimates for prepayments for the second half of the quarter. According to Bear Stearns’ analyst Dale Westhoff, "less than 30% of the borrower universe is refinanceable." Thus, if mortgage rates continue to climb, the market expects further downward moves in refinancing activity and a decrease in amortization in the coming months.

Despite improving prepayment fundamentals, MBS spreads have widened to treasuries in recent weeks as investors digested the new macroeconomic developments in the marketplace. As these macro themes develop, we continue to monitor how the expected slowdown in refinancing activity will affect the composition of the mortgage market. As refinancing slows, the composition of the mortgage universe may change from one where most securities are backed by loans that have been refinanced to loans that are from purchases. A change to a purchase dominated mortgage market can have varying effects on the prepayment speeds for mortgage backed securities. We will assess this phenomenon going forward so that we may react appropriately. Supporting the theme of a purchase dominated mortgage universe is the current robust housing market. The MBA Purchase Index, a measure of all mortgage applications for a single family home, increased 6.8% during the last week of April, and the National Association of Realtors (NAR) gauge of existing home sales rose 5.7% for the month of March. Further, on a national level, housing remains affordable. The NAR Housing Affordability Index currently stands at 144.3, implying that a family earning the median family income has 144.3% of the income necessary to qualify to buy a home at the median price given today’s mortgage rates. As rates rise one would expect housing affordability to drop somewhat, but the current level of affordability is well above historical averages. This should continue to support housing purchases in the coming months.

The Markets
In April, the bond market had a sizeable sell-off. Yields on the short end of the curve rose more on a percentage basis, but the yield curve stayed steep. Stocks were down slightly, as investors were apprehensive in the face of higher interest rates. The dollar strengthened, gold fell and oil reached even higher levels.