In an article published earlier this week by Kiplinger, financial advisor (and fellow CFP®) Mark Cortazzo suggested that due to our current low-interest-rate environment, it is now 10 times riskier to retire than it was a mere decade ago. I immediately found his conclusion to be a particularly ironic one, since a decade ago turns out (given the benefit of hindsight) to have been a particularly terrible and “unsafe” time to retire—the stock market plunged precipitously in 2008, leading many retirees to delay retirement, and some unlucky folks to have to return to work as their retirement account balances dwindled. In fact, the experience was so damaging for some prospective retirees’ portfolios that my firm now uses the 2008 “bear market” scenario as a standard part of our portfolio stress-testing procedures.
Sure, some lucky folks may have had their portfolios sufficiently “recession-proofed” (i.e., bond-heavy) that the plunge in stocks did not affect them, but most anecdotal evidence suggests that few retirees were so fortunate. Even setting that issue aside, however, I had to wonder whether Cortazzo’s argument held water, stock market crash or none. First, let’s let him state his case:
You need to add significantly more risk to your investment portfolio in order to pursue the same rates of return as 10 years ago, according to a report from Callan Associates, one of the largest pension consulting firms in the world…
According to Callan’s data, a pension fund targeted to achieve a 7.5% return in 2005 could have more than 50% of the portfolio in bonds, reducing the risk and resulting in a standard deviation of 8.9%. To pursue the same 7.5% rate of return today, the pension fund needs to have a much smaller percentage in bonds (12%) with the balance in equities, raising the standard deviation to 17.2%.
Using that significantly higher total-portfolio standard deviation, Cortazzo then took it upon himself to run a Monte Carlo simulation (a fairly standard risk-assessment tool for advisors) to determine the impact on an individual investor’s retirement picture.
For the scenario of the individual retiree in 2005, we used a 7.5% expected rate of return, 5% withdrawal rate and adjusted for 3% inflation. The result was a 2.4% fail rate. This means only one in 40 people would have run out of money during a normal retirement beginning in 2005. To apply this to someone starting to withdraw today, we ran a simulation with the same withdrawal rate and inflation factors as before but, this time, the result was a fail rate of almost 25%.
Right away, I couldn’t help but notice a couple of fairly glaring flaws in the analysis, as presented. Here’s what you should look out for, and why you should be at least a little bit suspicious of sensationalist headlines (and conclusions) like the one presented above.
Nominal versus real interest rates
The first problem I have with Cortazzo’s analysis is that the current low-interest-rate environment is presumed to persist indefinitely, even though inflation is simultaneously assumed to run at a level well above the Federal Reserve’s stated target (an assumed 3%, versus a Fed target of 2%). To build that assumption in as a basic statistic in a Monte Carlo simulation is probably an overly conservative (and potentially unrealistic) decision.
To understand why requires a quick discussion of “nominal” versus “real” interest rates. The nominal interest rate is simply the rate paid on current bonds (recently close to 0% for short-term bonds, up to 3% for long-term Treasuries or 4-5% for investment-grade corporates). The real interest rate, though, subtracts inflation from the nominal rate to arrive at an inflation-adjusted rate—in other words, what’s left of your investment return after inflation has taken its toll.
The distinction is important because inflation and interest rates tend to move in accordance with each other—as inflation rises, so do nominal interest rates, since investors require a greater return to compensate them for inflation’s bite. To assume that one will remain low while the other remains high is at odds with macroeconomic realities.
Historically, inflation has indeed run close to the 3% figure that Cortazzo used in his analysis, and it was a fairly appropriate figure to be using in 2005, as well.
But the 3.39% inflation rate that prevailed in 2005 is not today’s reality—the annual average inflation rate since 2009 has been just 1.48%, less than half of the historical average. As a result, the Federal Reserve (and other market forces) have kept nominal rates at a much lower level, as investors have surely noticed. Real interest rates, however, have been at least somewhat stable (though still noisy), despite persistently low nominal rates.
While the real (inflation-adjusted) interest rate is certainly lower today than it was 10 years ago, the magnitude of the difference is not nearly as dramatic as Cortazzo’s Monte Carlo analysis is told to assume. A current real interest rate of 0.5% is only about 1-1.5% lower than that of 2005; Cortazzo’s assumptions, though, would likely assume a persistently negative real interest rate, an assumption that would doom just about any long-term financial forecast.
Any Monte Carlo simulation is only as good as its inputs, and running a simulation that is based around an assumption of steady negative real interest rates is simply unfair, and unsupported by any historical data.
What is the maturity of your “bonds”?
Furthermore, Cortazzo’s analysis seems to make a fairly common error—it paints “bonds” as one, static asset class, rather than a multi-faceted, dynamic one. One of the benefits of bonds versus stocks is that while stocks are a perpetual asset without a maturity (with stocks, there is no requirement or guarantee of any repayment of capital, at any point in the future), bonds do have a definite term, and invested assets have a pre-determined time horizon. What’s missing from Cortazzo’s analysis (and possibly from the Callan analysis that he cites; I don’t know because he doesn’t cite Callan’s specific assumptions) is any statement about what type of bonds he’s assuming are owned.
In his article, Cortazzo writes that “surprisingly, most financial software uses historical rates of return that create assumptions for fixed income that are almost mathematically impossible to achieve in the current interest rate environment”. That’s fair criticism, but should we assume, then, that the “current interest rate environment” is likely to persist for the entirety of an investor’s retirement years? That the investor is locked in to the low interest rates, and owns investments with a long enough average maturity that a rising interest rate environment will provide no net benefit to the investor (from reinvestment of interest payments or reinvestment of maturing principal)?
Just about any prudent advisor will have encouraged clients to shorten the average duration of their fixed income portfolios over the past decade, so as not to “lock in” these historically low nominal interest rates. Yes, that shortening of duration comes with a short-term cost (since short-term rates are generally lower than longer-term rates), but it also helps keep investors nimble, so that (hopefully) the current low-interest-rate environment will not, at least by definition, become a permanent portfolio phenomenon.
But Cortazzo makes no statements either way about the maturity of the bond portfolio. It’s simply assumed (in a somewhat hand-waving manner) that rates are low, that they’ll remain low, and that investors will never have an opportunity to reinvest at higher rates, even as inflation returns to historical norms. They’ll therefore have no choice but to buy stocks, own stocks, and never sell any of those stocks in order to shift back into bonds. That assumption seems unlikely to me, and it should seem unlikely to you, as well.
In the fixed income world, it’s not so much “if you own bonds” that matters, it’s “which bonds you own”. Look at the performance (total returns) of three separate types of bonds from the last 5 years alone:
To talk about the “current interest rate environment” without acknowledging the significant differences in performance among various types of bonds is simplistic, if not outright invalid.
In 3 of the last 5 years, the spread between the “worst-performing” and the “best-performing” bond sector has been greater than 10%, which is a whopping difference. Does it really matter what the nominal interest rate is on a Treasury bond, when the total-return disparity between various types of bonds dwarfs that nominal rate?
In the last two months of 2016 alone, US Treasury bonds (of a 7-10 year maturity, as represented by iShares’ IEF fund) logged a total return of -4.3%, while corporate bonds of a similar maturity (as represented by Vanguard’s VCIT fund) logged a -2.4% return. That 1.9% difference is material, particularly if compounded over a long time horizon, as a Monte Carlo analysis assumes.
In other words, there isn’t just one “interest rate environment” at any given point in time—in fact, there are several, and all are relevant. Yes, the current macroeconomic situation probably calls for more diligence and attention to “what you own” than may have been the case a decade ago, if only because risks increase as nominal rates decline.
But we’ll leave discussions about bond convexity and its implications for your duration strategy for another day. For now, I think it’s sufficient to note that one’s retirement strategy should account for two very important dynamics: 1) that real interest rates, not nominal interest rates, are the important driver of long-term investment success; and 2) that there is not one asset class called “bonds”, that there are in fact several, and that using them wisely can help an investor adjust and respond to just about any macroeconomic reality.
Unfortunately, both the analysis and the sensationalist headline in the Kiplinger article omit those pieces of information, and an unhelpful conclusion is the inevitable result. When analyzing long-term risks to your portfolio, don’t fret about outcomes that haven’t been seen in economic history (like persistently negative real interest rates). Spend your time instead thinking about how and why to move in and out of the various types of fixed income investments, and your portfolio will almost certainly thank you.