Tuesday, August 4, 2020

What the Experts Get Wrong about Inflation

“In the following analysis, we assume future inflation of 2% per year.”  How often have you seen something like this in investment projections or other financial analyses?  This kind of assumption leads to biases that can invalidate a financial analysis.

Even the great Benjamin Graham wasn’t immune.  In the early 1970s, he wrote the following in his book The Intelligent Investor:

“Official government policy has been strongly against large-scale inflation, and there are some reasons to believe that Federal policies will be more effective in the future than in recent years.  We think it would be reasonable for an investor at this point to base his thinking and decisions on a probable (far from certain) rate of future inflation of, say, 3% per annum.”

The following footnote was added by Jason Zweig in a revised edition of Graham’s work:

“This is one of Graham’s rare misjudgments.  In 1973, just two years after President Richard Nixon imposed wage and price controls, inflation hit 8.7%, its highest level since the end of World War II.  The decade from 1973 through 1982 was the most inflationary in modern American history, as the cost of living more than doubled.”

Readers may think I’m calling out Graham for his wrong guess about inflation.  This isn’t my point.  Graham’s mistake was that he guessed at all.  Future inflation is unknown.  Just as we treat future stock returns as a range of possibilities, we should be doing the same thing with inflation. 

We might guess that annual inflation over the next decade is likely between 1% and 3%.  But we can’t say for sure that it won’t shoot up above 5%.  You may judge this to be unlikely, but do you really want your financial security to depend on inflation definitely remaining below 3%?

The biggest effect of assuming future inflation to be at some known level is to make long-term bonds seem safer than they really are.  Once we consider the possibility of rising inflation, 30-year government bonds look a lot riskier.

Because inflation affects how much we can get for our money in annual spending, it’s better to focus on investment returns after subtracting inflation (called “real returns”).  Stock markets look volatile no matter how we view them.  Their nominal returns are volatile.  So are their real returns.  Even if we just treat inflation as a known constant, stock returns are volatile.

However, a 30-year government bond looks very different depending on how you think about inflation.  The bond’s nominal return over the full 30 years looks completely safe.  If we assume inflation stays at some fixed level, the bond still looks completely safe.  But if we correctly assume that future inflation is unknown, the bond suddenly looks a lot riskier.

I have no reason to think Graham would make the mistake of investing everything in 30-year government bonds; he understood the risk of high inflation.  But many people who use spreadsheets and Monte Carlo simulation tools don’t understand the implications of fixed inflation assumptions on their simulation results.  Personally, I avoid all long-term bonds.

When we run financial projections assuming fixed inflation, we make bonds (particularly long-term bonds) look safer than they really are.  We need to get out of the mindset of trying to guess a single value for future inflation and treat it as uncertain.

8 comments:

  1. Good point, Michael. Using that logic in reverse, does that mean that stocks, which is traditionally deemed riskier than fixed income, are safer than they look in a low-interest environment?

    ReplyDelete
    Replies
    1. Samson: I would say that the gap between the risk of stocks and the risk of long-term bonds is smaller than it appears when we take inflation risk into account. However, stocks are still risky, particularly individual stocks. With the prices of both stocks and bonds very high right now, we have lots of risk everywhere. But cash pays very low interest now as well. If I were still in my working years, I would consider now to be a good time to pay down debts (not that I ever thought there was a bad time to pay down debts).

      Delete
    2. True, it's always a good time to pay down debt (guaranteed high ROI). That said, here's my point: Assuming that you don't have any outstanding debt and that you're investing into broad-market assets, that seems to suggest that increasing the equity portion of your portfolio is less risky than it used to be in the past.

      Delete
    3. Hi Samson, I'm afraid I don't agree. It's true that it's not hard to imagine a bad outcome for bonds right now, particularly long-term bonds. However, it's not hard to imagine a bad outcome for stocks as well. I don't trust my ability to judge the relative risks of stocks and bonds, so I don't try. I avoid long-term bonds (but hold short-term fixed income) to reduce a risk rather than to trade it for other risks. I choose my asset allocation based on my life situation (now a fairly young retiree) rather than my outlook on different asset classes.

      Delete
  2. Thank you for alerting me to this risk in long term bonds.
    And thank you for writing for our benefit.

    ReplyDelete
    Replies
    1. Hi Eric,

      You're welcome. Over the years I've benefited greatly from reader comments.

      Delete
  3. About long term bonds, the following are some comments. By taking the difference between inflation indexed bonds and nominal bonds, you can get the market's inflation prediction (the wisdom of crowds). Insurance companies and pension funds tend to have long term nominal obligations. So conventional wisdom is that they drive down the yields of long term nominal bonds. Finally, long term government bonds have some interesting diversifying properties, especially in market crashes. Finally, although I think it very unlikely with fiat currencies, don't exclude deflation. In the Depression, prices went down 25%. So the risk management properties of long term government bonds may alter demand and therefore the yield of them.

    ReplyDelete
    Replies
    1. Anonymous: All true. This makes it possible that long-term bonds will give returns that aren't terrible. Another possible outcome is a period of high inflation that makes them give significant losses. In my own portfolio, I choose to make my fixed-income component safe, which excludes the use of long-term bonds.

      Delete