Secular Changes in Interest Rates

I really like the way this chart, by Joseph Ellis, presents this information:


Long-term interest-rate changes and the stock market

click for larger graph

W1303

Source:  Ahead of the Curve

>

Ellis observes:

Because of the inverse relationship between interest rates and stocks’ price-earnings ratios, rising interest rates from 1960 to 1982 contributed to a compound annual appreciation of only 2.9% in the S&P 500 Index.

Conversely, falling interest rates from 1982 to 2003 were a major long-term stimulus to the stock market, helping produce compound annual growth in the S&P 500 of 10.5%. Note that bear markets from 1960 to 1982 were more frequent and longer, whereas from 1982 to 2003 they were less frequent and shorter in duration.

Print Friendly, PDF & Email

What's been said:

Discussions found on the web:
  1. Alex Khenkin commented on Feb 16

    While this observation makes intuitive sense, it’s only one data point, so, strictly speaking, the implied causation may be tenuous at best.
    -Alex.
    Small Investor Chronicles

  2. Tom commented on Feb 16

    Profit margins improved over the same time period, so a couple of things worked together. I’m intrigued by the comment about recessions. If we assume that interest rates have entered a relatively stable period (moving sideways, as opposed to trending up or down), then what happens to recessions?

  3. spencer commented on Feb 16

    Regression results for the past 50 plus years imply that the relationship between the stock market pe and interest rates is roughly one to one — a 100 basis point rise in rates generates a 100 basis point drop in the stock market pe.

    Roughly, from 1960 to 1980 the s&p 500 fell from 17 to 7 and has since rebounded to 17. Because of this from 1960 to 1975 the S&P 500 fell despite a strong increase in earnings and since 1975 the change in the PE has accounted for approximately twice as much of the stock market increase as earnings growth.

  4. B commented on Feb 16

    Indeed Spencer. The correlation is extremely positive. The $64 question is if rates will remain low. I tend to think not based on the data we see daily and historical precedence. Price pressures are everywhere. I am not sure there is anything anyone can do to stop these long cycles. Long periods of stability create instability to steal a line that I love. So within the long wave cycles we have oscillations or shorter cycles of stability/instability yet continue in trend within the long wave cycles. That likely means a volatile reversion of S&P dividend yields and PEs if rates do continue to cycle upward over time.

    People forget even Intel, as an example, had a PE of 9 around sizteen years ago. Uh, a PE of 9 would be quite a haircut from Intel’s continual haircut that now sits at what, 80%?

    I wonder what causes this behavior. I tend to think it is almost entirely psychology or human behavior. Or, as has been coined by Barry and others, behavioral economics. I think that is significantly responsible for commodity price increases similar to the 70s. So while nearly everyone tells you we are running out of oil or copper is in high demand, there is quantitative, measurable evidence commodities are going up for other reasons. If that is so, Greenspan’s control over the economy during the bubble was an illusion. While he might have nuanced the situation a little differently to reduce the fervor somewhat, the end result would have still been the same.

    I think there are reasons to be cautiously bullish for the future but I also think the market risks are well, well above normal.

  5. Joe P commented on Feb 16

    The graph is interesting, and may spark some thinking on long-term interest trends in general. However, a correlation between rates and “compound appreciation” (what is that? does it include dividends?) does not prove causation or even contribution.

    Additionally, the “source” link on the “ahead of the curve” page points to a “page not found” at the FRB site.

  6. JSchreiber commented on Feb 16

    I’ll bet in real terms the discrepancy is even greater.

  7. Earvin Johnson commented on Feb 16

    So basically, if / when the yield on the 10-yr starts ramping up that’s when the we begin to get a pullback in the US market.

  8. Hit The Bid commented on Feb 16

    If you can get the right REPO rate, sounds to me like a great time to Sell 10 year strangles any time Vols creep up.

  9. Elambend commented on Feb 16

    So for someone too young to remember, what did the real estate market do in the early, mid seventies.

  10. algernon commented on Feb 16

    The symmetry is remarkable. Roughly 22 years on each side of the peak.

    It is interesting that few people seem to envision us as currently being in an interest rate trough.

  11. Jay Walker commented on Feb 16

    Definitely, when you see interest rates around 4% for extended periods, you have to know it’s not going to last forever. When commodities start heating up, inflation must so too.

    Having said that, I don’t think you’ll see central bankers make the same mistakes they did in the 70s, where they allowed inflation to get almost out of control, to the point where 18-20% first mortgages were not unusual.

    In terms of real estate, it did very well during those inflationary years (the 70s), until the central bankers tightened the money supply to produce 20% first mortgages. That killed the housing market for a while. In fact, 1979-1981 in the real estate market reminds me of what happened to the NASDAQ in 1999-2000.

    Of course, the next couple of years weren’t that pretty!

  12. B commented on Feb 17

    So was 1990ish when we had the S&L crisis. I’m still convinced the consumer being tapped out is erroneous. The big problem is going to be inflation. And while we all hope we won’t see 18-20 percent rates, remember it took a decade to get there. “Experts” just officially raised their estimates to 5% on Fed Funds. I remember when those same experts were betting the Fed would stop at 3.75%. For the first time yesterday I heard 5.5% muttered.

    Don’t think the Fed does not understand what is at stake. Regardless of what they blather in public about long term inflation in check and other ramblings, I am quite confident in private they have looked at the same charts we are. They may not all agree on the final outcome but…..

    And for those who say wage inflation is in check because of globalization, I tend to agree but what if we are wrong? Remember the malaise of the 70s? Jobs flowing overseas, worry about Germany and Japan crushing us, America becoming an also ran? Yet….regardless of the misery and stats, did we see wage inflation? I know many, many markets are tight. Not the Pizza Hut market but the white collar markets. They are responsible for a disproportionate amount of wages.

    I tend to discount this week’s pricing action in the markets because the options distribution showed a high probability rally was coming. Let’s see how much further it goes before anyone believes this was nothing more than a quick trade. Especially since many of the largest NAS 100 stocks are still puking.

    Btw, find the ten year bond on a log scale and chart it. You can use linear regression to draw the high and low support lines or if you are confident to do so, draw upper and lower support lines. There is a clearly defined down channel from 1980. It’s nearly beautifully perfect. We are at the top of the monthly BBs showing strength in interest rates and we have made a bottom and are currently making higher lows. For the markets to advance significantly, rates will likely have to be stopped at the top of the channel and fall as they did in 1995……….which so many of the “experts” are comparing this year to. In 1995, ten year rates dropped about 40%. And guess how much the market went up that year? 40%. One cannot predict future equity action without intermarket analysis. The 1994-1995 comparative to 2005-2006 is totally inaccurate on nearly every front….except the equity charts look similar. If rates go higher, you can likely fuggedabout a rally in late 06 or 07 that will lead us to significantly higher highs.

    Btw, for those who eschew TA, what do you think program trading uses? A little gift for the very few which will read yesterday’s posts.

  13. miami commented on Feb 17

    His numbers are way, way off:
    SPX Jan 01, 1961- Dec 31, 1982
    Price Appreciation: 4.1% annually
    Dividends reinvested: 9.07% annually.

    His below 3 number is way, way, way off when you compound over 22 years. Terrible math.

    source: Bloomberg markets

    miami

  14. Barry Ritholtz commented on Feb 17

    S&P dividend yield was 9.07% annually? Um, I dont think so . . .

    I assume you mean Yield plus Cap Appreciation — even then, 9% sounds high . . .

  15. miami commented on Feb 17

    Sorry, dividends reinvested was shorthand for ‘with Divs reinvested,’ or Total Return of SPX Index with Dividends reinvested. Nowhere did the word ‘yield’ appear in my post.
    9.065% to be exact, I’m looking at the data and tables right now.

    If it sounds high, that may be because you don’t know the facts going back decades if you haven’t parsed the original data.
    If you are telling me that Bloomberg’s data on the SP500 is wrong [ha, right], give me your number with citation, please. Use Ibbotson, Compustat, take your pick.

    Let us know what you find…

    best,

  16. miami commented on Feb 17

    Okay, made the same typo TWICE so I feel like a complete idiot.
    SPX total return over those 22 years was 8.065%, not 9, so I’ll go stand in the corner but even my typo was much closer to reality than Ellis’ claim. I emailed him as well.

    miami

  17. B commented on Feb 17

    S&P went from 57 to 138 in the time miami quoted. That is 4% annually as is stated per a quick calculation. BUT, if the total return including dividends is 9%, that means dividends contributed 280 S&P points on top of the market’s 4% return. ie, The market’s return was less than 100 points but the dividend was over 3X that.

    Do they say what the average dividend was during that period? Does sound high. But then, we’ve been in a 100 year low S&P dividend trough since 94.

  18. miami commented on Feb 20

    http://www.bos.frb.org/economic/neer/neer1997/neer597b.pdf

    Annual return of SP500 in
    1960’s: 8.0%
    1970s: 5.8%
    1980s: 14.2%

    Notice how all of these numbers are much higher than the phony 2.9% quoted by Ellis.

    Even the 1970s, which stunk for equities, the entire decade is contained in his sample, has double the return he quotes, and that’s the lowest one in the sample. 8.065% is the correct number for the time period I listed. If you tell me a particular quarter to start, I will give you a different return number.

    Daniel Patrick Moynihan: “Everyone is entitled to his own opinion, but not his own facts.

    It’s fine to be bearish. I was fairly bearish last year and only got back into the markets because I luckily[skillfully] picked the market low on April 20 and figured that was a good time to put money to work. This year I am in the ‘soft landing’ camp.

    It’s not fine to woefully understate historical SP500 returns, however, particularly when they are so easily obtained in the age of the Internet.

  19. Barry Ritholtz commented on Feb 20

    I’ll check with Ellis and see if those rates are “real” — i.e, inflation adjusted — or not.

    But when you consider the 10 year was paying between 6-10% risk free — even 5.8% in the 1970s sound pretty pathetic

Posted Under