In the first part of Fears of Soaring Rates are Overblown, I reviewed Federal Reserve actions for dealing with the credit crisis since 2008. These include four major actions that have increased the monetary base by roughly $1.9 trillion, of which $1.6 trillion are in marketable securities.
The Fed first provided liquidity to banks and depository institutions in an attempt to keep them solvent. Second, they bought more than $1 trillion in mortgage-backed securities to break a logjam in the frozen mortgage market, as well as stimulating the economy. The last two actions involved purchasing $600 billion in Treasury securities to lower interest rates, and currently Operation Twist is extending maturities in $400 billion of Treasury securities in an effort to flatten the yield curve.
Someday, these Fed actions will be unwound. There is no timetable for an unwinding, although the Feds forecast is for accelerated GDP growth through 2014, and then lower growth in 2015. This may suggest some unwinding in 2014.
The question addressed in this article is not when an unwinding will occur, rather, what effect the unwinding could have on interest rates. Specifically, how much should we expect interest rates to rise when the Fed begins selling its Treasury and mortgage portfolio?
I believe that Newton’s Third Law applies to financial market abnormalities. Newton’s law states that for every action there is always an equal and opposite reaction. When we look at broad market pricing abnormalities, we find many examples of price bubbles deflating to a long-term average and crashes recovering to the same mean.
History is full of examples where fear- and greed-pricing eventually reverse in near equal and opposite price movements. Fear of hyper-inflation in the late 1970s sent gold prices up more than 400 percent only to flame out and fall down to its centuries old inflation-adjusted average price within a few years. Sky-high technology stock valuations during the late 1990s plummeted to normal valuations by the early 2000s. Inflated housing prices in 2006 have recently collapsed down to their inflation-adjusted growth trend line.
Today, the global credit crisis has caused a Treasury bond bubble. Overwhelming demand for Treasury securities from skittish global investors, along with Federal Reserve and foreign Central Bank buying, has lifted prices and pushed yields well-beyond normal. Intermediate- and long-term Treasury yields are at historically low yields, and T-bill yields have actually been negative for a few days since the crisis started.
Table 1 is a list of current interest rates provided by the Federal Reserve. The rates are lower than what “normalized” yields would be if the credit crisis did not occur. My goal in this article is to estimate what those normal rates should be, and then use this information to determine how much risk bond investors face as yields move back to normal levels.
Table 1: Current interest rates as of December 12, 2011
Newton’s Third Law will undoubtedly take hold when investors become less risk-adverse and the Fed starts unwinding its $2.8 trillion balance sheet. Interest rates will go up across the board, although some rates will go up more than others.
T-bill yields will undoubtedly go up the most because they’ve been pushed down the most. The current 0 percent yield puts T-bills at 2.0 percent below inflation. Historically, the yield has averaged 0.5 percent over inflation. The cumulative real gain on T-Bills since 1926 is illustrated in Figure 1.
Figure 1: Long-term Cumulative Real Return for 1-month T-Bills (1926 = 100)
Unprecedented demand from around the world has driven T-bill yields to 0.0 percent (and even negative on occasion). In a non-crisis scenario, T-bills would yield about 2.5 percent, assuming inflation expectations stay around 2.0 percent.
The 2.5 percent jump in T-bill yields will not extend over the entire Treasury yield curve. The 10-year Treasury bond yields are currently at 2.0 percent, which is also the forecasted inflation rate. Historically, 10-year Treasuries yield about 1.7 percent over the inflation rate. This would land the normalized 10-year Treasury at about 3.7 percent, an increase of 1.7 percent.
AAA corporate bonds historically yield about 1.0 percent over the 10-year Treasury. Given a 3.7 percent normalized rate for 10-Year Treasuries, AAA industrial corporate bonds land at a normalized 4.7 percent yield. This is a 2.0 percentage point increase over current rates.
Normalized yields for mortgages, BBB corporate bonds, and 20-year municipal bonds can also be estimated by their historical spreads. Figure 2 illustrates the historical 30-year mortgage rate as a percentage of 10-year Treasury yields, and BBB corporate bond yields as a percentage of AAA yields.
Figure 2: Spread on Mortgages and BBB Corporate Bonds
The rate on a 30-year fixed mortgage has historically averaged about 140 percent of the 10-year Treasury yield. It’s currently at 200 percent of that yield. This spread may narrow as the Fed continues to influence the mortgage market, either symbolically with policy announcements or through their balance sheet. More likely, the spread will move back to a lower range when the Fed begins selling its Treasury holdings.
BBB corporate bond yields are also high relative to AAA bonds. The spread is sitting around 135 percent today versus about 120 percent normally. This spread will likely narrow as AAA corporate rates increase during a Fed unwinding because BBB yields shouldn’t increase as much.
Municipal bonds are a different animal. The yield spread on the 20-year municipal bond average tracked by the Fed is extraordinarily high, relative to 20-year Treasury bonds. This high excess municipal yield is due in part to less demand for tax-free bonds during the crisis and in part to credit risks from enormous unfunded pension liabilities at state and local levels. Figure 3 illustrates the historical spread.
Figure 3: 20-Year Municipal Bond Yields Relative to 20-Year Treasury Yields*
*There is a gap in Federal Reserve data from 1987-1994
The normalized municipal to 20-year Treasury spread is 85 percent. I’ll bump that up to 90 percent to estimate the normalized yield given the liability risks in the municipal marketplace.
We now have enough data to make normalized interest rate estimates. Table 2 shows current interest rates, my estimate of non-crisis normalized rates, and the increase in rates.
Table 2: Estimates of Normalized Interest Rates
Table 2 tells an interesting story. Interest rates will go higher if inflation expectations remain the same, however, some rates will likely go higher than others relative to their current rates.
T-bills will increase the most because they are trailing well below their inflation-adjusted normal yield. The good part about holding T-bills is that you can’t lose nominal dollars. Unfortunately, you can lose money to inflation. Every dollar invested in T-bill will be worth less money as you wait for higher yields.
Bond prices go down when interest rates go up. Since interest rates are very low, the risk of owning intermediate- and long-term bonds is abnormally high. A small move up in rates will cause prices to fall meaningfully. This is called the delta of duration. I’ll skip the math.
10-year Treasury bonds and AAA corporate bonds will lose about 10 percent of their principal value as interest rates rise during the Fed’s unwinding period. This means 10-year Treasuries will experience at least a 3-year period with a 0 percent total return and AAA corporate bonds will probably do the same.
Mortgage rates will also rise, but not as much as Treasury bond rates. This is good news for homebuyers but not good news for mortgage investors. The duration (risk) of mortgage-backed securities extends as interest rates increase because fewer homes are refinanced and pre-payments slow. Accordingly, the price of existing mortgage-backed securities may drop considerably even though mortgage rates have not increased as much (again, I’ll skip the math).
Riskier BBB corporate bonds and 20-year municipal bonds will see smaller price declines because their yields will increase less than Treasury and AAA corporate bond investments. One could argue that the “safest” bonds during a Fed unwinding are those with the most credit risk.
My firm holds a core position in the Vanguard Total Bond Market Index Fund Admiral Shares (VBTLX). This fund currently has a 2.3 percent yield to maturity (YTM) and 5.2 year duration (up from 4.8 percent last year). It’s composed of 43 percent government and agency bonds, 30 percent mortgage-backed securities, and the rest in investment grade corporate and Yankee bonds.
If the Fed unwound its entire balance sheet during 2012, raising interest rates to the levels noted in Table 2, I estimate that the YTM on VBTLX will rise to about 4.2 percent, and cause the fund to lose about 10 percent in net asset value (NAV). It would take about 3 years of interest payments and maturing bond reinvestment for the fund to make up this NAV loss. This assumes interest rates don’t boomerang past 4.2 percent due to renewed inflation fears or other reasons.
Individual bonds and bond mutual funds are going to deliver a punch sometime in the next few years, but it won’t be a deadly blow. I don’t have a good answer for what is coming, except to say that broad diversification helps, and don’t fight the Fed by trying to predict the timing of their balance sheet unwinding. We’re prepared to muddle through a difficult bond market in the same way we’ve been muddling through a difficult stock market since 2007 – diversify, rebalance, and stay the course.
Fed Chairman Ben Bernanke and the FOMC will likely begin unwinding their balance sheet in a slow and controlled manner starting around 2014. The process will reduce the total return of every bond portfolio over the next decade whether an investor holds individual bonds or mutual funds. Think of it as giving back gains that we should have never earned. On the positive side, I don’t believe rates will soar during the unwinding. This should make the reversal somewhat palatable, whenever it comes.