For a financial advisor focused on bonds, there is no more vital issue than where interest rates are going. Yet, making reasonably accurate predictions is where uncertainty sets in. The problem is no mere quibble, for income targets with defined risk have to be achieved. This question about interest rates can make or break a retirement plan.
The yield curve shows interest rates on trend-setting government bonds prices of every term along the curve. There are yield curves for every class of corporate bond, too. All link time and return. To the extent that future rates are baked into present conditions, the yield curve is accurate. But its predictive power is questionable. When conditions change, so must the curve.
A study by the Federal Reserve Bank of New York entitled Discounting the Long Run, which was released on Aug. 31, concludes: “Looking solely at market rates to project future rates may generate severely biased forecasts.”
The study compared current term premiums with future rates and found coefficients of correlation of 0.019 and 0.165. A 0.0 correlation means events have no relation whatsoever. A 1.0 correlation means they are in lockstep. The bank’s research found that betting on terms is a poor way to make a living.
The yield curve prevailing in the late 1970s and early 1980s, when bonds paid double-digit interest rates, did not anticipate the dramatic moves by the U.S. Federal Reserve Board and the Bank of Canada to break the back of inflation in 1983. Today’s very low rates do not anticipate a return to historical norms and, indeed, predict very low, recession-level interest rates out to 30 years.
You can make theoretical predictions based on the yield curve in the context of other indicators. The problem is not the target, which can be calculated, but the lag in hitting it.
“You can make a five-year call based on nominal [gross domestic product (GDP)] growth with inflation included,” says Jack Ablin, executive vice president and chief investment officer with BMO Harris Private Bank in Chicago. “That is where rates should be. Given that nominal U.S. GDP growth is 3.5% and that U.S. inflation was 0.2% as of the end of August, the five-year bond should pay 3.7%. It is currently at 1.52%, which implies that bond prices should fall to raise the yield. This is known, but it has not happened in spite of several years of prediction that it must.”
Part of the problem is that the yield curve’s short end is not driven by the market, but is administered by central banks. Therein lies an unusual problem, because low rates develop a constituency of their own. As Ablin points out, very low yields increase the demand for bonds by institutions and investors even though the bonds, which pay little, are less desirable with every downtick of interest.
“If you have a $5-million portfolio and a $100,000 bond- income target, then – with 10% interest rates that prevailed 30 years ago – you would have needed only a fifth of it – [that is], $1 million – in bonds,” Ablin says. “Today, with the 10-year Treasury bond rate at about 2.0%, you need all of the portfolio in bonds to get that $100,000.”
It is a question of what is and what ought to be. As Chris Kresic, senior partner and head of fixed-income with Jarislowsky Fraser Ltd. in Toronto, notes, “We have emergency-level bond interest rates, but there no longer is any emergency.”
You can add that the virtually zero inflationary expectations in the present yield curve ignore history and the likelihood that central banks will raise rates, if only to control asset mispricing. Examples of mispricing include: stock prices higher than they should be; stock buybacks financed by cheap money; and debt-financed acquisitions that would not pay if interest rates were higher.
Today’s yield curve is rooted in the rates set by quantitative easing and easy central-bank monetary policy. “Predictive inaccuracy exists because as duration rises as you go further on the curve, there is more risk and more rate and time uncertainty,” says Charles Marleau, president and senior portfolio manager with Palos Management Inc. in Montreal. “Most people who buy longer bonds don’t hold them to maturity. They will instead take a string of shorter bonds and give up the term premium to reduce risk. Over time, interest rates have to rise. If you are risk-averse, you want to stay with shorter bonds with terms of five years or even 10 years. That’s as far out as I would go.”
In the end, interest rate uncertainty embedded in the yield curve justifies the term premium and drives duration risk. Going from a five-year U.S. treasury at 1.52% to 10 years at 2.19% or 30 years at 2.97% – or a Government of Canada bond of five years at 0.77%, 10 years at 1.47% or 30 years at 2.24% – the yield staircase rises by 67 basis points (bps) and by 78 bps for treasuries and by 70 and 77 basis points for Canadas. That’s the norm. It’s not a lot of bonus for taking on the certainty that the yield curve is going to rise one day.
As Kresic says, “Interest rate prediction is as much art as science.”
Thus, the dilemma: take a low yield with high duration risk or beat the government yield curve by going with investment-grade corporate yield (such as a 10-year AltaGas Ltd. bond with 3.84% coupon due Jan 15, 2025, and priced at $1.02 to yield 3.55% to maturity, a 208-bps pickup over Canadas of the same term).
“Grabbing yield,” Kresic concludes, “is easier than trying to predict it.”
© 2015 Investment Executive. All rights reserved.