Beware of Biased Rate Forecasts!

At almost every year-end since the end of the Great Recession, economists predicted that the benchmark ten-year Treasury rate would rise over the next 12 months. Seven out of nine times in which the outcome is known at this point, they overestimated the next year-end Treasury rate. One exception involved the 2013 Taper Tantrum, when a surprise shift in Federal Reserve policy produced by far the period’s biggest year-over-year yield rise. (See details in the table below.)

Was the forecasters’ 78% overshoot rate merely a chance outcome? Almost certainly not. We can say with 90% statistical confidence that the economists surveyed by Bloomberg systematically overestimated the next year-end rate.

Forecast Error: 10-Year Treasury Rate (%)
Year Median Forecast Actual Difference
2010 3.84 3.30 0.54
2011 3.61 1.88 1.73
2012 2.25 1.76 0.49
2013 2.08 3.03 -0.95
2014 3.06 2.17 0.89
2015 2.39 2.27 0.12
2016 2.39 2.45 -0.06
2017 2.63 2.41 0.22
2018 2.96 2.69 0.27
Sources: Bloomberg, Federal Reserve Bank of Philadelphia

Overshooting the rate forecast 78% of the time is almost certainly not a chance outcome.

A fairly benign explanation is that economists relied on an outdated model of inflation, a key determinant of interest rates. They likely assumed that the Fed’s quantitative easing program (massive bond purchases) in response to the Great Recession would stoke inflation by greatly expanding the money supply. A key reason why that did not happen is that the velocity of money—how frequently each dollar turns over within a stated period—fell off a cliff. The velocity dropoff is not fully understood, but several explanations have been offered. One is that suppressing long-term interest rates produced less economic activity, hence fewer transactions, than existing models suggested it would. That shortfall, in turn, may be related to an unexpectedly sharp rise in the personal savings rate.

Then there is the less benign explanation. Many of the surveyed economists work for financial organizations that depend on consumers investing in common stocks. (Equities, more than bonds or preferreds, are where the commissions and fees are for most brokers and money managers.) Investors tend to buy stocks when they feel bullish about the economy. Financial organizations therefore have an incentive to issue upbeat forecasts of economic growth, perhaps even more upbeat than the facts truly justify. If a firm’s year-ahead forecast includes a healthy rise in Gross Domestic Product, internal consistency requires a prediction that the ten-year Treasury yield will be higher 12 months out, as a function of increased demand for credit.

This is by no means to suggest that all economic forecasting is hopelessly corrupt. Many forecasters display great integrity. But one economist friend of mine was fired from two major investment banks for (correctly) forecasting a recession. I saw another chief economist effectively retired by the salesforce, which preferred his assistant’s more bullish outlook. In a private conversation, an institutional equity salesman praised his firm’s chief economist for his unfailingly rosy forecasts, evidently unaware that prostitution is illegal in most U.S. jurisdictions, including New York.

The post-Financial Crisis evidence suggests that habitually overhauling your portfolio in response to bulletins from economic forecasters is unwise. In particular, income investors should think twice before turning defensive just because financial firm warn yet again that interest rates are headed sharply higher. That view may not be a perfectly impartial assessment of the available facts.