On Forecasting, and the Recent Bond Rout

It is dangerous to make forecasts, especially about the future. 

— Mark Twain

Much of our work is forecasting. We weigh the prospective returns and risks for stocks and bonds. We consider your tolerance and capacity for the next market downturn. And, most importantly, we analyze the likelihood that you will have enough money during your lifetime.

Unfortunately, most of us humans are not naturally built for good forecasting. We identify patterns everywhere, though such explanations are often illusions. We see recent trends as predictive of the future, though the opposite is often the case. And we anchor on initial beliefs, though radical adjustment is often more appropriate.

As far back as I can remember, my grandfather cursed Jerry Taft, the weather forecaster from Chicago’s NBC affiliate. According to Grandpa, it rained all too often when poor old Jerry said “it wouldn’t” rain. Somehow, Grandpa decided that when Jerry said the odds of rain were less than 50%, he meant that it wouldn’t rain — no chance. Of course, Jerry was regularly wrong in the other direction too. Such errors negatively affected Grandpa’s golf game. They were damning offenses that put Jerry’s eternal future in peril.

The recent Presidential election saw similar analysis. It’s become common knowledge that anyone who professionally predicted the election outcome in favor of Secretary Clinton was obviously “wrong”. A competitor of ours said the following: 

Nate Silver, the High Priest of Election Models and founder of the FiveThirtyEight blog, won the distinction of being the least wrong in predicting the outcome. [Silver’s final forecast showed Clinton with a 71% chance of winning the election.] Although he continually warned us that he might be wrong, his clear implication was that Hillary would win. …These are a lot of high-powered minds with deep understandings of statistics and modeling but they were all – to quote The Donald – Wrong!

The single most important aspect of forecasting is a probabilistic mindset. Sometimes, Jerry said there was only a 10% chance of rain, yet it would rain. Many, including Grandpa and our unnamed competitor, would say that Jerry was “Wrong!” on those days. But when Jerry said there was a 10% chance of rain, he didn’t mean 0%. He meant that if tomorrow were to occur 100 times, based on what we know today, he would expect rain on 10 of those tomorrows. Sometimes, it’s not that the forecast is wrong, but simply that the unexpected occurs. 

Our competitor went on to say the following:

But the 2016 election was a one-time event. Trump won 100% of the election and Hillary lost 100% of it. 

That’s the thing about the future — it’s uncertain, and more things can happen than will happen. Yes, tomorrow will only happen one time. Yes, good forecasters must constantly strive for improvements in their always-imperfect models. Yes, good forecasters must constantly work to improve their only-human judgment. And, yes, these realities can be most unsatisfying at times. In fact, it may feel good to blame the forecaster for an undesired outcome, but blame tends to exceed blameworthiness by a rather large factor, and it fails to shed useful light.

The future and certainty will forever share little. Probabilistic thinking, however imperfect it may be, is the only logical framework for analysis of what lies ahead. 

If you really don’t want to get wet, you need to have an umbrella every day that the chance of rain is greater than 0%. Since we find nearly all 0% (and 100%) forecasts to be wildly overconfident, we would recommend an umbrella on 0% days as well. But if you find carrying an umbrella to be a literally too-heavy burden, then you must necessarily accept the risk of getting wet, or maybe even soaked. 

In the case of the bond market, we have chosen to carry a sturdy but refreshingly lightweight umbrella for several years. It’s not that we necessarily forecast that interest rates would go up (and bonds down). We simply believed that there was so little to be gained, and potentially so much to be lost, by trying to eke out a slightly higher yield in longer-maturity or riskier bonds. Because we don’t know anyone who invests in conservative bonds primarily for capital appreciation. For most, bonds are intended to either maintain an already sufficient standard of living or, more commonly, cushion potential losses in stocks. 

In recent months, the market has begun to anticipate rising inflation and, as a result, rising interest rates. Just 90 days ago, the 10-year Treasury note yielded 1.5%. Just two weeks ago (the day before the election), it yielded 1.8% per year. Today, it yields 2.3%. That marks a 53% increase in yield in the last three months, and the biggest two-week jump in 15 years. 

While interest rates have increased by less than 0.8% over the last three months, the global bond market (i.e., the Barclays Global Aggregate Index) has quietly slipped 7%. In contrast, STUDIO’s bond portfolio* has been flat over the same period. Now imagine what might happen to global bond prices if 10-year Treasury bond yields continue increasing toward a more historically normal 6%-ish level. For many, their bonds may cause a pain they are not anticipating.

Again, we encourage ownership of bonds not for their appreciation potential but instead for the security they can provide a stock portfolio. We do not believe bond portfolios are an appropriate place to seek return maximization, nor do we evaluate the success of your bonds in relation to the overall bond market. Instead, we purchase bonds with the aim of providing some level of positive return over the long term, while partially insulating stock portfolios from major downturns. 

We believe the recent steadiness of STUDIO’s bond portfolio was driven by its short-maturity orientation (i.e., nearly all bonds mature in less than three years) and its inflation-protection exposure (through U.S. Treasury Inflation Protected Securities (TIPS)). We anticipate these attributes will dominate your bond portfolio for some time to come, and, while nothing about the future is certain, we believe they will continue to provide the buffer we seek.

Probabilities are not always comforting. We’d rather be sure when it will rain. We’d rather be sure who will win the election. And we’d definitely rather be sure that we won’t run out of money. Since the future doesn’t allow us the sureness we prefer, thinking in probabilities is still the best approach to forecasting and making decisions.

By the way, Jerry Taft is still on the air in Chicago, making useful forecasts every night. I suspect Nate Silver will be around for a while yet too. And may your bonds continue to offer the protection we expect.

We hope everyone had a Happy Thanksgiving.

Short Note: Care for a Netflix Bond?

STUDIO's View of the 2016 Presidential Election