Things Change (or Dealing with my Obsession with Interest Rates)
On a recent drive home from Palm Springs my wife and I were trying to resort to the usual incentives and rational discussion to get our two kids to stop fighting and to allow a little civility to enter into their relationship. We even suggested that they try to fall asleep. My daughter, ever the honest and outspoken one, immediately retorted, "I'm not tired!" Eventually cooler heads prevailed, or more precisely, our bribes worked, and the tension in the air dissipated to a calm peace. Sure enough, within thirty minutes both of our kids were sound asleep, despite my daughter's protests. Later in the evening, after we were settled in our house, I asked my daughter what happened in the car. After all, she was not tired and then, before you know it, she was in a deep sleep. She didn't have an answer for me and being ever so concerned about using every moment as a learning experience that will somehow help her get into a fine university (she's 8), I suggested a response should she give me in the future: "Things change."
As I have mentioned before, just because something hasn't happened, doesn't mean the odds of it happening have gone down, in fact they probably have gone up. Unfortunately there are over 225,000 people in Asia who have paid the ultimate price for us to learn the hard lesson of small probability events taking place after a 9.0 earthquake and resulting tsunami in the Indian Ocean.
It is hard to make money following conventional wisdom. One year ago, the consensus was for the 10-year Treasury security to yield 5.10% by the end of 2004. So where did it end the year? 4.25%, 17% lower than anticipated. What is the forecast for next year? According to a Wall Street Journal (1/3/05) poll of 56 economists it is 5.10% again with only 4 of 56 believing it will end 2005 at 4.50% or lower. To add more fuel to the consensus fire, the Bond Market Association's annual poll also resulted in a 5.10% prediction for the end of 2005. The participants probably socialize together in the same New York City clubs.
I have a real aversion to consensus forecasts, particularly when it comes to interest rates. Nevertheless, we all need to place our bets and this requires some view of the future. My view is somewhat shaped by what most people think will happen and betting on the opposite. The ultimate purpose of this article is to let you know our latest thinking regarding interest rates and how this will impact our properties in 2005 because an increasing percentage of our debt is variable rate. For what it's worth, our 2004 budgets anticipated that the Federal Reserve would raise interest rates four times in 2005 for a total of a 1.00% increase by the end of the year. While we were off in that the Fed raised rates five times for a total increase of 1.25%, it started its rate increase cycle later than projected such that our debt service came in below budget. So what is ahead in 2005?
According to the Wall Street Journal poll, the consensus estimate is for the Federal Funds rate to increase from 2.25% to 3.60% by the end of 2005. So what do I think? I will start by giving you some very logical reasons for this to take place based on commercial paper rates, a good proxy for Fed-influenced, short-term interest rates. They are as follows:
- Alan Greenspan has been uncomfortable with the "emergency" rates of our post-September 11 world. In August 2001 commercial paper averaged 3.54% so I think he would feel more comfortable if they were closer to this level.
- In this cycle commercial paper rates peaked at 6.53% (June 2000) and reached a low of 0.99% (Jan.-March 2004). The average of these two is 3.76%. One occurred at the height of the internet/dot com/telecom boom and the other while the economy was quite weak. In the last three years, over $3 trillion of additional debt has been created in the U.S. economy, which I believe makes it more difficult to hit the previous mid-point.
- Median inflation as reported by the Cleveland Fed is running at a rate of approximately 2.33% over the last twelve months. For over thirty years the average premium of short-term rates over the median inflation rate has been approximately 1.33%, bringing the target to 3.66%.
- Between 2/1/62 and 8/31/01 the 10-year Treasury security has yielded on average 128% more than the 3-month Treasury bill. The 10-year Treasury as of this writing is 4.20%, which implies a more normal T-Bill yield of 3.28%, a good approximation for commercial paper rates as well.
If you were to do a simple average of these four data points, then it would come out to 3.56%, dangerously close to the consensus of 3.60%. Because the logic is so elegant and the results so in synch with the economists' view of the world, and that it's probably more prudent to be conservative regarding our interest rate assumptions, we are budgeting our variable rate loans based on five interest rate hikes such that the Federal Funds rate is 3.50% by the end of the year. To quote my favorite accountant, William Shakespeare:
"When clouds appear, wise men put on their cloaks;
When great leaves fall, the winter is at hand;
When the sun sets, who doth not look for night?
Untimely storms make men expect a dearth."
So what do I think will really happen? Let's turn to one of my favorite market commentators, William Shakespeare:
"Oft expectation fails and most oft there where
most it promises; and oft it hits where hope is
coldest, and despair most fits."
As embarrassing as it may be to admit, I am fascinated by interest rates. I obsess about them like some people do over fine wine or art. In fact, I'm going through a mini personal crisis right now because I can't seem to locate my copy of the classic A History of Interest Rates. I fancy myself as a quasi-interest rate historian so I thought I would go back in time in order to put this projected interest rate increase into perspective by looking back at 157 years of interest rate history tracked by the National Bureau of Economic Research (NBER). This same organization is also responsible for determining when recessions begin and end. In case you were wondering, since 1857 there have been 31 official recessions as determined by the NBER.
Embedded in yield statistics is so much about hopes and fears of investors and societies and their responses to critical events shaping economics, politics, culture, business, peace and war, and international relations. On its website, the NBER has monthly yield information for commercial paper placed in New York from 1857 through 1971. The Federal Reserve has the more recent data since 1971. Although the two data sets differ slightly, they are still very useful for how I will present my information.
This period of time includes the Civil War, World War I, World War II, Korea, the Cuban missile crisis, Vietnam, 9/11, ten banking crises prior to 1934 and one in the early 1990s, two depressions, Wall Street crashes, a bailout by J.P. Morgan, bailouts by the Federal Reserve, unemployment as low as 1% and as high as nearly 26% (this change occurred within a four-year period), deflation, inflation, gold standard, Bretton Woods, flexible exchange rates, Monica Lewinsky, Marilyn Monroe, Presidential assassinations, resignations, and impeachments. And the list goes on and on. Yet the money kept getting priced. That is what is so fascinating. Investors assimilated all of these critical events into how they priced money. So let's take a ride back over 150+ years of history and see what we can learn and apply to today. I want to warn you now that you will be completely inundated with numbers from this point forward. If you are overwhelmed by a lot of numbers, then I suggest you stop reading at this point.
Let's first take a look at business cycles because interest rate movements are highly correlated with economic activity. During the first half of an economic expansion interest rates fall 67% of the time (56% since World War II) and increase 83% of the time (89% post WW-II) during the second half.
As we transition to recessions rates have risen 80% of the time (70% post WW-II) during the first half and have fallen 97% of the time (100% post WW-II) during the second half of the recession. Thus, recoveries are aided by dropping interest rates and recessions are typically induced by higher rates. We are currently in the second half of our recovery (my guess) and are experiencing higher interest rates that will probably continue until it helps bring about a recession in 2006 or 2007.
Since 1857 recessions have lasted from 8 months (August 1918-March 1919) to 66 months (October 1873-March 1879). Between June 1857 and December 1945, the U.S. was in a recession approximately 44% of the time, or every 27 months with the average duration being 22 months. Since 1946 the recession frequency has dropped to 16%, or every 57 months with an average duration of 11 months. Our current expansion has lasted 37 months. If recent history is any guide, then we have approximately two years before our next recession.
It is important to put this recent interest rate increase cycle into perspective. Commercial paper yields reached bottom between January and June 2004 at a yield of 0.99% to 1.00% (using monthly averages). Since this time, commercial paper rates have increased to 2.22% as of December 2004. I went back and looked at the percentage change in interest rates as compared to the same month one year earlier. I did this for every month between January 1858 and December 2004. Between December 2003 and December 2004, the commercial paper rate increased by 116% (1.03% to 2.22%). To put the magnitude of this increase in perspective, only 20 months out of 1,686 experienced annual increases of 100% or more, equivalent to 1.2% of the time. Even when I isolated those months in which the percentage increase from the previous year was positive, the frequency was 2.2% (20/889). The following table highlights some of the key data points regarding these rare tightening cycles.
The table shows that interest rates peaked between 2 and 24 months after they had risen by 100% or more as compared to the previous twelve months. The average peak occurred ten months later, taking rates approximately 38% higher. This implies a peak rate in this cycle of approximately 3% with a range of 2.40% to 3.60% (there's that number again) using the extremes. Looking ahead eighteen months resulted in an average interest rate approximately 6% lower, or 2.88% assuming an implied future interest rate of 3.06%, although the range is quite wide (-47% to 24%). Using these extremes would result in an average over the next 18 months ranging from approximately 1.15% to 2.75% (rounded) as compared to the 2.22% December rate.
When the rate of change is compressed to six months, 100% increases become even more rare (12 out of 1,634 months or 0.2% of the time). One of these occurrences took place in November 2004. During the seven other periods when this occurred, five of them were during recessions (1857, 1860, 1893, 1896, and 1932), one in 1958 in which a recession ensued 16 months later after rates went up by 47% (2.95% if this happened today), and in 1980 at the very peak of interest rates (18.95%) with the exception of one month in 1857 (24%). A recession ensued seven months later and interest rates eventually dropped dramatically.
Because the consensus forecast calls for interest rates to reach 3.60% by year-end 2005, this is an 18-month time frame from when interest rates bottomed in June 2004 at 1.00%. To convert this to a percentage increase results in rates rising by 260%. The question I am interested in answering is how does this compare to all of the other 18-month increases going back to 1858? To be blunt, unprecedented!
The largest percentage increase in history occurred between July 1958 and January 1960 in which rates went from 1.50% to 4.91%, representing a rise of 227%. If this were to happen during this cycle then rates would peak in July 2005 at 3.27%. With commercial paper rates at 2.25% and four Federal Reserve meetings between now and then, it is conceivable that rates could rise by 1% if the Fed raises rates 0.25% at every meeting.
Since 1858, there have been only two months in which rates have increased by more than 200% as compared to 18 months previous and 34 months with 100% increases or more. To be fair, the recent decline in interest rates has been unprecedented as well as they have dropped by over 70% during an 18-month period. This has happened only 12 months in history with seven of them occurring in the recent cycle. The other five took place in 1857 (twice) and rates went up 254% from their low over 31 months, 1893 and rates went up 202% from their low over 16 months, and 1933 (twice) and rates went up 44% over 41 months. Even using the two extreme increases in our analysis would produce future peak interest rates of 3.02% and 3.54%.
Despite having access to over 150 years worth of data, I think it might be most instructive to compare interest rates in this economic recovery to the last one that took place between April 1991 and February 2001.
The first chart on page 5 shows commercial paper rates for 80 months of the previous economic expansion as compared to the first 37 months of the current one. It's quite impressive how strikingly similar the two paths are (.66 correlation) and how steep the recent increase in interest rates has been. If recent history is any guide, then interest rates do not have much more to rise in this cycle.
Finally for those skeptics out there who cannot fathom interest rates staying at very low levels, then let's look at Japan. Some financial analysts believe that the United States is following Japan's trajectory towards deflation with about a ten-year lag. This chart compares Japanese interest rates between 1983 and October 2004 and U.S. commercial paper rates between May 1921 and February 1943, a period of time that covered the roaring 20s, the Great Depression, and part of World War II. As a point of reference, commercial paper rates reached 1% in December 1946 and 2% in March 1951.
So what do I think will happen? I would be very surprised to see rates go much beyond 2.75% because this would represent one of the most significant increases in interest rates in history. Even at 2.50% it's a large increase on a historical basis. Nevertheless, we are budgeting for short-term rates to go up to 3.50%. I also believe that there is a not insignificant probability (I sound like Alan Greenspan) that we will have a scenario in which our unprecedented debt creation will be reversed through a deflationary contraction and result in low interest rates for a long period of time. Thus, our variable rate debt has protection if interest rates go higher because of our interest rate caps and the ability to capture any "benefit" from a deflation scenario.
Finally, I am uncomfortable being too close to the consensus regarding my projection for short-term interest rates. Here is where I differ from the consensus by a wide margin. The most accurate predictor of a recession is an inverted yield curve in which short-term rates are higher than long-term ones. The United Kingdom is already facing this as its 3-month interest rates are 4.72% and 10-year rates are 4.55%. I foresee the United States having an inverted yield curve by the end of the year if the Fed continues down its rapid tightening path. How will this present itself? Short-term interest rates going to 3.50% and the 10-year Treasury security's yield dropping to less than 3.50%, far below the consensus estimate of 5.10%. So there's my call. Then again, things change.