Most Recent Stories

Be Careful Relying on Historical Market Returns

If you read just about any document published by a Wall Street firm you’ll inevitably run across some form of this statement:

“Past performance is not indicative of future returns”

We all seem to implicitly know that the future will not necessarily look like the past.  But there is, arguably, no approach more often utilized than the analysis of past returns leading to an “empirical” conclusion about the future performance of asset classes.  For instance, Eugene Fama’s famous 3/5 Factor Model approach is based almost entirely on historical market data citing the tendencies of certain asset classes to perform in certain ways.  Jeremy Siegel’s work on “Stocks for the Long Run” is based almost entirely on historical market data citing the tendencies of certain asset classes to perform in certain ways.  Robert Shiller’s CAPE is a perspective of future potential returns placing valuations in a historical context.  “Value” approaches of all types rely on using some historical context to gauge how inexpensive or expensive the market is.

The problem with all of these views is that the future never perfectly reflects the past.  And I think there’s a strong argument to be made that today’s environment is more unique than any we’ve seen in the historical data.  Yet we continue to see many investors relying on expected future returns based largely on historical data.

One thing we know, for a fact, is that investors who think the bond market will generate the types of returns that it did in the last 30 years, will be sorely surprised.  With 0% interest rates there is about a 0% chance that the next 30 years in bond returns will mirror anything like the last 30 years when the aggregate bond index returned an astounding 7.5% per year with virtually no negative volatility.  This means that an investor who uses the historical returns of a balanced portfolio is using a framework that looks nothing like what one should really expect.

Some investors like to think that they don’t make projections about the future.  Or worse, they imply that the future will look like the past just because the data says stocks and bonds perform in a certain way over the “long-term”.  But there’s one certainty we know based on the structure of interest rates today – we’ve never been in an environment like this.  And future returns are likely to be lower than most people expect given the same amount of risk taken.  And that means that your use of historical data has to be placed in the proper context or it will likely lead you astray.  Worse, if you’re not trying to look forward in today’s environment you might as well not be looking at all….


  1. Frederick

    Historical returns provide context. Not guarantees. But I agree, the future returns of bonds are likely to leave a lot of investors very surprised about their returns in the future.

  2. Cullen Roche

    It’s pretty interesting. I think the “balanced” asset allocation approach is among the most popular it’s ever been. And yet this allocation is really just being chased by indexers who expect the past to reflect the future….

  3. tealeaves

    Looking back at the history here, the 1940s had similarly low yields (and choppy equity markets) but that was in WW2 economy. I could imagine people living in that era could not imagine the economy and markets would begin to turn up with the start of the industrial revolution and impressive growth rates (and rising rates).

    Another low interest rate era is 1890 to 1915 which had decades of low interest rates with a sideways equity market which quickly moved into manufacturing, automobiles and the roaring 20s (and higher rates).

    Two data point and neither very applicable to our current highly indebted private sector economy though I might convince myself that in the face of despair, human innovation will lead to unexpected growth rates and a more optimistic future. Or more pessimistically, that the US and the world follows Japan for the next decade with even lower rates. Or maybe this ends up somewhere in the middle of those.

  4. Tomas

    I’d like to turn your argument against you. Wasn’t it observed that argument “this time is different” is consistently wrong?

  5. Jonah Thomas

    That argument only works when you expect it to be different in a good way.

    If something bad has consistently happened before, it will mostly keep happening, though the form might change somewhat. “This time it’s different, I’ll get a good result!” will fail.

    If something good has been happening, and you say “It’s going to keep happening, this time won’t be any different” then that is likely to fail too.

  6. Cullen Roche

    I guess the real question is, with 0% interest rates, how is it not “different this time”? Bond returns going forward are 100% guaranteed to resemble nothing like we’ve ever seen in the past. I don’t think this is controversial. It’s a reality. Those relying purely on historical return data to make forward forecasts are not being realistic with themselves….It is indeed different this time. It has to be.

  7. Jonah Thomas

    Tomas, I don’t understand what you are saying. Maybe I’d get it if you spelled it out farther.

  8. Tomas

    Don’t get me wrong, I do agree that current circumstances are unique. However circumstances are not everything and historically argument “it is different this time” ended up being very bad every time.

  9. Cullen Roche

    I don’t know what that means. Are you assuming that stocks and bonds will return their same historical returns just because that’s what they did in the past? I find this rationale, which I see quite often these days when people construct asset allocation plans, to be incredibly naive. We know, for a fact, that bonds will not generate the same historical returns as they did in the past. It is an impossibility given the structure of interest rates at present.

    Now, will stocks make up for the lack of bond returns? They might. But that means the investor has to take more risk to achieve the same return. I think that’s a pretty important thing to consider when constructing your portfolio. Simply pointing to the past and saying “it’s never different this time” is simply not a very intelligent way to go about this process in my opinion….

  10. Tomas

    There is no way to tell who is right, but thinking about the process there are few factors to consider:
    1. Current unique circumstances cannot and will not last forever.
    2. Risk of being wrong on “this time is different” seems higher than being wrong on “it’s never different” (might be my assessment)
    3. I do not advocate keeping only cash, however I do think that when allocating assets, one has to consider scenario that this time is not different (and not with probability of 2%)

  11. Cullen Roche

    I don’t understand how you can refute the point being made. US 10 year bonds have averaged a return of 5.5% over the last 90 years and 8% over the last 35 years. The current interest rate is 2.5%. It is virtually impossible for bonds to generate those kinds of returns in the future if held to maturity without exposing the investor to principal loss and high real, real return loss. This time IS different. We know, for a fact, that this time is different because the structure of interest rates has NEVER looked like this….

  12. Tomas

    I think we misunderstand each other. Once again, I do not deny that current situation is different, but that in most likely scenario it will revert to normal (or average).

    For example, you mention Robert Shiller’s CAPE. If you look at the chart, that indicator is almost never at it’s long term average. But it reverts back.

    This time it is different (for me) = this time it won’t go back to average.
    It is as always = it fluctuates (sometimes a lot and for long term), but finally reverts to average

Comments are closed.