Joined at the hip: The Macroeconomy and the Stock Market

This is a topic I wanted to cover for quite a while and I think this is the perfect time for it: I got a few posts lined up already dealing with the intersection of macroeconomics and (personal) finance. Of course, I can already hear one objection:

“Oh, come on Big ERN, did you see the volatility in the stock market last week? My portfolio went down by 10+% since the January 26 peak and you want to talk about macroeconomics now? The first week of February 2018 proves that the macroeconomy doesn’t matter for stocks!”

And my response: Yes, this is actually the perfect time to talk about macroeconomics! See, the main reason I was not overly concerned about the market volatility earlier this month is that the economy seems to be running just fine and there don’t seem to be strong signs of any impending recession. Of course, the stock market has some wild swings, sometimes during a recession, sometimes outside of a recession. But all the really bad disasters, the bear markets that sunk retirement dreams in my Safe Withdrawal Rate historical simulations, they all occurred during recessions. And not just any minor garden-variety recession but the big ones! In other words, this is how I insist macroeconomics matters for the stock market:

What do we make of a 10% drop in the stock market? It’s a buying opportunity during an economic expansion! But during a recession, the market might potentially get a lot worse before it gets better!

And in today’s post, I like to provide some empirical evidence for my view…

Normally, people look at the historical returns post 1926 (see Ben Carlson’s post for example) rather than going back all the way to 1871. I figure the idea is that if we go too far back beyond 1926 the economy and the monetary policy response would become less comparable to today’s economy. So let’s plot the cumulative equity return during the last 92 years, starting in January 1926. As always, I use real CPI-adjusted S&P500 Total Returns (incl. dividends) and this is all plotted on a log-scale to make comparisons easier. I also include the U.S. recessions shaded in orange. See chart below:

Cumulative real S&P500 return. Major drawdown always occur around recessions! NBER recessions are shaded (Source: NBER).

Notice how there were even some (brief) recessions where the stock market advanced. There were also some wild swings during economic expansions (think 1987). But nevertheless, all of the really deep and extended stock market meltdowns occurred during recessions. Why does this matter? When I saw the stock market swings last week I wondered what regime we’re going to be in; is this the precursor to a recession or just some market volatility that just simply happens every few years? The economy seems to be doing just fine (how I reached that conclusion is a topic for a future blog post) and I concluded it’s probably not the time yet to run for the hills and sell equities. Quite the opposite, last week, I funded our IRAs for the 2018 tax year and I bought more FUSVX!

Another chart I always like to look at is the “drawdown chart,” i.e., plot how far the market fell below its previous all-time high. Why does a drawdown chart matter? Think Sequence of Return Risk! I have written extensively on the topic (see Part 14 and Part 15 of the SWR series) and also did an entire podcast episode on ChooseFI about the topic. Deep drawdowns mean that retirees are taking money out of the portfolio at discounted prices. That’s bad for the long-term survival prospects of the portfolio! The exact same way dollar-cost-averaging sometimes helps the saver, the retiree loses from an extended market downturn early during the withdraw phase! In any case, you obviously get some drawdowns during economic expansions. But the drawdowns are only moderate, around 10-20% (30% once during 1987), and they are very short-lived. But the really traumatic, deep and extended drawdowns always occur during recessions. And of course, sometimes the recovery and the path to the new all-time high will take years and extend for years into the next expansions.

Cumulative real S&P500 return. Major drawdown always occur around recessions! NBER recessions are shaded (Source: NBER).

Not only that, some recessions were so bad that the new all-time high wasn’t even reached until after the succeeding recession:

  • The peak of 1929 before the Great Depression was not reached until 1945, i.e., after the 1937 recession.
  • The pre-1973 high wasn’t reached until after the 1980 and 1982 recessions.
  • The pre-2001 high wasn’t reached until after the 2008/9 recession. (and again, that’s because I plot real/CPI-adjusted returns. In nominal terms the market had recovered in 2007!)

And do we all notice what these three episodes have in common? The first two are the classic examples of when the 4% Rule failed (both over a 30-year horizon and most definitely for a 60-year horizon). And the third episode? Well, the jury is still out on that one but I’m not too hopeful, see a case study I did a year ago.

I’m not too optimistic about the 2000 retirement cohort!

And finally, here’s another chart to drive home my point: Here’s the cumulative return during the roughly 75 years of expansions (black line) and the roughly 17 years of recessions. Think of this as plotting the cumulative return chart once by splicing together the white shaded regions (i.e., expansions) in the chart above and once by splicing together recession portions only:

Cumulative returns during expansions and recessions. Notice a pattern here? How can anyone claim that the macroeconomy doesn’t matter for stock returns?

It’s a different chart, a different way to present the same data but the same story: The macroeconomy matters a lot for stock returns. +11.44% average real return during expansions and -11.32% during recessions. But I’m the first person to concede that 15.61% annual volatility during expansions and monthly returns ranging from -23% to +42% can still look pretty wild. But again, during expansions, we have a 62%+ share of positive returns and significantly more attractive returns than during recessions!

How is this analysis different from Ben Carson’s calculations?

You might have seen the post about stock market performance around recessions on Ben Carlson’s site A Wealth of Common Sense. He makes it sounds as though recessions don’t really matter so much for the stock market. Of course, I use the exact same data but only present it differently. (OK, for full disclosure, I use real inflation-adjusted returns, which I find more comparable across time, while he uses the less informative nominal returns.) And I think my way of presenting the data makes a pretty compelling case for a strong link between macroeconomics and stock returns. For example, he (correctly) notes that there have been a few recessions where the stock market advanced. I don’t disagree with that but I just think that this it is a red herring and it distracts from the more meaningful hard numbers I provide.

But, but, but … it’s difficult to declare recessions in real-time!

Brett Arends, a MarketWatch columnist, had this pretty cringe-worthy quote in a post in 2012 making fun of how long it takes economists to actually declare the NBER recession dates:

“Economists are almost always slow to spot key trouble signs. […] Remember the Great Recession that began in December 2007? The economists at the National Bureau of Economic Research, who are basically the official scorekeepers of recessions, didn’t discover the recession until December 2008 – a year late, and only a few months before the episode (officially) ended.”

The NBER business cycle dating committee is so late in declaring the macroeconomic turning points because they want to be 100% sure when they call the recession start and end. Even though everyone already knows the recession has started/ended if there is still some residual uncertainty about what exact month it was the NBER will not make their official statement. Take the following analogy: the U.S. Presidential election. We all normally know the election result on the evening of the election (or the early morning hours of the following day) but it takes another formal vote in the electoral college, and then until January 20 of the next year until the new president is actually sworn in. The NBER waits until the recession is “sworn in” while everyone else already knows with 99.9% certainty that the economic turning point has occurred. For example, economists definitely monitor the economy in real time and there is actually a whole “industry” in economics trying to estimate NBER recession probabilities in real-time before the actual announcement is made. It’s actually something I would occasionally do in my day job!


Chill out, everybody! The big drop last week wasn’t so bad! Volatility will probably stay with us for at least a few more weeks. Maybe months. But to me, it looks less like the beginning of another Global Financial Crisis and more like another false alarm like 2011 (U.S. downgrade), 2013 (Taper tantrum), 2015 (Chinese devaluation), Jan/Feb 2016 (Fed rate hike jitters) and June 2016 (Brexit). And I probably forgot a few other events that were paraded as the end of the world over the years.

Of course, I haven’t really told you why I think the economy is way too strong for me to be concerned about a recession right now. That will be the topic for an upcoming post. Blogging 101: always put in a cliff-hanger at the end of the post to make the readers come back. And while we’re talking about cliff-hangers, I also got a fun new joint venture with a fellow blogger, Actuary on FIRE, about the impact of inflation on early retirement! That will be a multi-part series and AoF will post part 1 on his blog next week. So, stay tuned for that one as well!

We hope you liked today’s post! Please leave your comments and suggestions below!

38 thoughts on “Joined at the hip: The Macroeconomy and the Stock Market

  1. This correction definitely seemed like an overreaction to rate increases and inflation fears than anything meaningful. Intuitively I thought that it would be hard to imagine much further of a draw down when most economic indicators are positive but I had no idea about what the data actually indicated in these situations. It’s nice to see that a big draw down in an economic environment like this is unlikely.

  2. Not to forget the pleasures of 2010 (Greece bailout), 2011 (Debt crisis, EU) and 2011 (Debt ceiling,EU), all in the range of ~10-19% drops in equities. All within two years of the 2008/09 debacle.

    I have been amazed also with the chatter and advice of pundits who tell us what we should do with our portfolios in this recent period of (ahem…) turmoil. Here are the three key moves that we made with our portfolio recently:

    1. Nadda
    2. Zilch
    3. Nothing.

  3. A word of caution. The one rule governing correlation based indicators is they eventually fail (think about the mountain of PhD economists and Nobel Laureates at LTCM who presided over its failure based on correlation). The other rule determined by the math of compound returns is one failure destroys a lifetime of compound growth. Stated another way, I fully agree with your premise that draw downs of statistical relevance tend to correlate with recessions. Absolutely correct! However, implementation timing is tricky (as you elude to) so that the horse isn’t out of the barn before it’s too late, and there must be a fail-safe for the inevitable time that correlation fails (because it will).

    1. Thanks, Todd! That’s why I’ll never have any hope that any indicator will ever be able to time you out of the stock market in October 2007 and back into the market in March 2009. Macro mostly tells you when not to panic. For example, the much-hyped 10M momentum signal would have given you at least two false signals (2011 and 2015). And the much-hyped momentum signal will also miss the turning points by a few months (by definition!). So supplementing the completely price-based signal from momentum with some additional information from macro will be better.
      Also, just to be sure I like to stress what I mean by a success in market timing: If the index falls from an all-time high of 100 down to 50 at the trough, I would call selling equities at 80 and buying back at 60 a huge success. In other words, I’m not too worried if I suffer the initial 20% downdraft in the market.
      About correlation: The correlation between the business cycle and the stock market is the one I’m not worried about ever going away. Other correlations, such as Stock/Bond etc. are a lot less stable. Completely agree with that and finance is littered with victims of unstable correlation patterns. 🙂

  4. You not only left us with one cliffhanger, you left us with TWO! Excited to hear you’re working with Actuary On FIRE, inflation’s one of my bigger concerns. Great post showing market cycles and economic connection. You’re pretty good at this blogging stuff, just quit doing that double cliffhanger thing at the end, you’re killing us here…..

  5. I have to admit the recent stock market event made me a bit nervous since I’m preparing for my early retirement exit. It’s hard to see my taxable portfolio drop quite a bit since I’ve got 100% in equity. But I stayed calm and did nothing.

    Looking forward to the post on inflation!

  6. The 10% drop is not bad at all. The market shot up so much in January. It can’t keep up that pace. The economy is still doing very well so I think 2018 will be a good year for equity. We didn’t make any drastic changes. I just contributed to my i401k as usual.

  7. Thanks for this timely piece. I was hoping you would weigh in on what was likely on the minds of so many of your readers. Much appreciated!

  8. ERN – Thank you for the continued sharing of interesting, empirically driven research! Your work in the FIRE community is differentiated and very interesting.

    Given that NBER is ex-post, perhaps you could opine on real-time (or near real-time) economic indicators that can help investors to gauge forward recession probabilities. When combined with valuation and basic trend indicators, this might be marginally helpful (I fully acknowledge the poor track record of short-dated market predictions).

  9. ERN – Great technical post as always. Appreciate the data and detail you put into the posts.

    You may recall, I’m kinda a real estate guy, (385K into a flip that ya liked)… But, I am also big into understanding the general equity markets and always looking for that way to reduce loss and volatility while keeping returns close to market. Such as synthetic indexing using options..

    Sequence risk is the big concern for all of the FIRE crowd without large pensions. (what’s a pension?) With that said and relating to perhaps this article and your previous, do you know of any analysis that is more intelligent about withdrawals in down times. Such as, if the year was down, you pull from your bond/cash portfolio instead of stock portfolio?

    Another alternative that would reduce early sequence risk is to start retirement with a lower equity position, continuing a dollar cost averaging system the first N years of retirement. Obviously, if the market significantly is up the first year of retirement, portfolio will end lower. What would this do to the overall success rates given different portfolio options? Such as, starting with 50/50 equity/cash-equivalents and legging in 10% a year until 75% equity? What about 5% a year? Starting with all cash and 20% a year until you’re 75/25?

    Adding additional intelligence to the calculators would be a worthwhile project. If the kid does not land a CS internship this summer, guess what this he’ll be doing with his old man/xEngineer/xSoftwareGuy?

    1. Withdrawals that depend on equity performance enter the realm of active allocation. Easier said than done. It raises the question of when you rebalance again to the target. One could also do a glidepath, which is more passive. Check the SWR series parts 19-20!

  10. Here is a good read for you, it is based on the giant real returns of everything study and I think you may like the read. Obviously from a bunch of real estate bugs, who can be as bad as gold bugs or individual stock bugs at times. 🙂

    On Wed, Feb 14, 2018 at 9:01 AM, Early Retirement Now wrote:

    > posted: “This is a topic I wanted to cover for > quite a while and I think this is the perfect time for it: I got a few > posts lined up already dealing with the intersection of macroeconomics and > (personal) finance. Of course, I can already hear one objection: “Oh, ” >

    1. Yes! I liked the paper! Real estate has been a pretty good and stable investment. Not only higher returns butalso less sequence risk due to the significant rental yield –
      much higher than the dividend yield!

  11. I disagree with the suggestion that current economic indicators do not indicate a very high recession risk.
    The indicators themselves are a month or a quarter behind. E.g. if consumer spending, jobs, sentiment, etc. was plummeting at this very moment, you would not hear about it for weeks/months. By that time, the recession would already be here, stocks would have already declined, and we would probably be halfway through the bottom already.
    More importantly, all these things revert to the mean. It is more likely that unemployment could go from 4% to 6% than 2%. (Think about the massive societal transformation required to hit 2%!). A period of historically great performance occurred in the 90’s and mid-2000’s followed by largely unforeseen recessions, with few leading clues from the statistics. One definition of recession is that the economy grows until it must shrink.
    Now, some claim you can watch the Dow Transports, but I’m not sure that works so well in an economy more driven by information than by trucks. We’ll see!

    1. As I said, there is a lag between the recession start and some indicators to show the signs. But the 2008/9 recession started in January 2008. Even with a few months delay you would have had plenty of time to get out of equities before things fell apart in September 2008. By the summer, most macro indicators were already flashing red.

  12. The recent drop was easy to not stress out about – it was a technical correction of forced selling of equities by institutional investors that got clobbered by the short vol unwind and deleverage.

  13. Hello ERN, detecting recessions is an interesting topic. Here are a few references:
    1) Philosophical Economics:
    2) Investing for a Living: they’ve removed some of the articles on this topic when they started their subscription product, but here’s a taste:
    3) Elliot Middleton’s paper “Animal Spirits” :
    4) site name says it all.

    1. vesc, those are great links, especially the Philosophical Economics one. Thanks for sharing.

  14. I believe there are some volatility “amplifiers” that exist now compared to the past. The robots, levered ETF’s, the fact that 25% of the market is indexing, fake news, 24/7 talking heads, etc. Since interest, bond yields, and stocks have been 90% off the mean due to the Fed I think that adds to the vol as well since the Fed seems resolved to be more hands off now. My hope is a more normal relationship between stocks bonds and interest re-establishes with some reversion to the mean. That will mean a more normalized diversity, and people haven’t yet levered up to toxic levels of debt. My impression is we narrowly missed deflation.

    Can’t wait for the inflation series. I’m sure the government can’t wait for some inflation either. Debt? what debt.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.