How long does it usually take equities to recover from a bear market?
I don’t just mean how long does it take for a bear market to end. Bears can be officially over in a matter of months.
But how long does it take for us to recover our losses? To get back in the black? In real, inflation-adjusted terms.
Sadly, that’s a much longer slog…
Investing returns sidebar – All returns quoted are inflation-adjusted total returns (including dividends). Fees are not included. The bear recovery column shows you when the stock market fully restored its losses in real terms. Total duration measures the period from the start of the bear market until recovery.
World equities: bear market recovery times 1970-2025 (GBP returns)
| Bear start | Bear trough | Bear real recovery | Fall (%) | Total duration |
| Dec 1969 | Jun 1970 | Jun 1972 | -22 | 2 years, 5 months |
| Dec 1972 | Sep 1974 | Dec 1984 | -52 | 12 years |
| Sep 1976 | Apr 1980 | Mar 1983 | -39 | 6 years, 6 months |
| Aug 1987 | Nov 1987 | Jul 1989 | -30 | 1 year, 11 months |
| Dec 1989 | Sep 1990 | Aug 1993 | -39 | 3 years, 8 months |
| Aug 2000 | Jan 2003 | May 2014 | -51 | 13 years, 9 months |
| Oct 2007 | Feb 2009 | Feb 2013 | -36 | 5 years, 3 months |
Data from MSCI. November 2025. Note: MSCI World monthly returns begin in 1970. The December 1969 bear market actually began before that – see the UK and US bear market recovery tables below.
To summarise:
- Average bear market loss: -38%
- Average bear market recovery time: 6 years, 6 months
- Shortest bear: 1 year, 11 months
- Longest bear: 13 years, 9 months
The real-return figures I’m sharing here are much worse than the nominal ones you’ll see from sources that ignore inflation.
Unfortunately though, the cost of living is real as we’ve seen only too recently.
Inflation-adjusted returns are the ones that put food on the table. So let’s not obscure reality with nominal figures.
That aside, I’m always shocked by the potential depth and severity of really big bear markets.
If you weren’t invested during the Global Financial Crisis (GFC) then you haven’t even experienced an average bear market shock yet.
God knows how awful many of us would feel if the market were to fall by 50%.
So far that’s happened twice in my lifetime. But happily not my investing lifetime.
Smarter than the average bear
Many people seem to believe that they can always ride out a bear because the market will bounce back in a few years.
As the table shows, that could prove a serious miscalculation if you’re gliding towards retirement with a portfolio stuffed full of equities like a jumbo jet carrying too much fuel.
Remember the recovery periods above only get you back where you started.
It’s also worth pondering on that fact that, as I say, since the GFC we’ve enjoyed an exceptionally benign bear-free patch.
Long may that continue, eh?
(Gulp! Should you suddenly feel a desire to dig deeper, I recently refurbished our article on defensive asset allocation.)
UK equities: bear market recovery times 1900-2025 (GBP returns)
Okay, we can’t access World equities data before 1970. So for a longer term picture, let’s turn to the UK and US record of bear attacks:
| Bear start | Bear trough | Bear real recovery | Fall (%) | Total duration |
| Jun 1914 | Dec 1920 | Feb 1923 | -52 | 8 years, 8 months |
| Jan 1929 | Jun 1932 | Feb 1934 | -37 | 5 years, 1 month |
| Jan 1937 | Jul 1940 | Mar 1945 | -40 | 8 years, 2 months |
| Jun 1951 | Jun 1952 | Nov 1953 | -28 | 2 years, 5 months |
| Jun 1957 | Feb 1958 | Aug 1958 | -21 | 1 year, 2 months |
| Apr 1961 | Jun 1962 | Aug 1963 | -25 | 2 years, 3 months |
| Jan 1969 | May 1970 | Jan 1972 | -35 | 3 years |
| Apr 1972 | Dec 1974 | Jan 1984 | -75 | 11 years, 9 months |
| Jan 1976 | Oct 1976 | Aug 1977 | -32 | 1 year, 7 months |
| Sep 1987 | Nov 1987 | Apr 1992 | -34 | 4 years, 7 months |
| Aug 2000 | Jan 2003 | Feb 2006 | -45 | 5 years, 6 months |
| Oct 2007 | Feb 2009 | Mar 2013 | -43 | 5 years, 5 months |
| Dec 2019 | Mar 2020 | Aug 2021 | -25 | 1 year, 8 months |
Data from Before the cult of equity: the British stock market, 1829–1929, (Campbell G, Grossman R, Turner JD, (2021), European Review of Economic History. 25. 10.1093/ereh/heab003.), A Century of UK Economic Trends, and FTSE Russell. November 2025.
Some highlights:
- Average bear market loss: -38%
- Average bear market recovery time: 4 years, 9 months
- Shortest bear: 1 year, 2 months
- Longest bear: 11 years, 9 months
Surprisingly, inking in the period wracked by World Wars and the Great Depression does not make the UK’s bear market recovery stats look any worse than the World index.
That said, my eye is always caught by the UK’s -75% 1972-1974 crash.
Reflecting on that period also reminds me we’ve endured periods of social discontent that makes today’s disharmony look like a primary school nativity play.
Bear country
In some ways, these tables underplay the potential threats to our portfolios.
For one, our tables don’t include the near-bear markets: losses of 15% or more that pockmark the inter-bear periods.
Sub-bear shocks can still be enough to shake someone whose portfolio has galloped ahead in the good times. A few years of worth of wonderful gains can quickly move us from a place where we had little to lose to suddenly having a lot on the line.
In that situation, we may have imperceptibely become less risk tolerant than we thought.
Secondly, sometimes only a few months separates one bear market recovery from the next mauling.
For example there is only a three month respite between the January 1972 recovery and the April 1972 market mutilation. So I personally view that period as one long 15-year bear market rampage. (Perhaps it would be with fees included.)
Similarly, Y2K’s Dotcom Bust and the GFC really amount to a lost decade for UK investors.
Finally, the last of my ‘glass half empty’ / ‘the glass is smashed all over the floor’ points is that the UK stock market has performed pretty well historically.
Yet it’s plausible to imagine a nastier, parallel universe where all equities were ripped up by a Bearzilla disaster on the scale of the Japanese stock market crash.
Incidentally, the December 1989 to September 1990 bear market (in the World equities table) is largely caused by the bursting of the Japanese asset bubble.
US equities: bear market recovery times 1900-2025 (USD returns)
For completion’s sake, here’s the bear market recovery record of the world’s most successful stock market:
| Bear start | Bear trough | Bear real recovery | Fall (%) | Total duration |
| Jun 1901 | Oct 1903 | Dec 1904 | -25 | 3 years, 6 months |
| Jan 1906 | Nov 1907 | Jan 1909 | -35 | 3 years |
| Jun 1911 | Dec 1914 | Oct 1915 | -20 | 4 years, 4 months |
| Nov 1916 | Dec 1920 | Aug 1924 | -47 | 7 years, 9 months |
| Sep 1929 | Jun 1932 | Nov 1936 | -77 | 7 years, 2 months |
| Feb 1937 | Apr 1942 | Apr 1945 | -48 | 8 years, 2 months |
| Oct 1939 | Apr 1942 | Jun 1944 | -38 | 4 years, 7 months |
| April 1946 | Feb 1948 | Oct 1950 | -35 | 4 years, 6 months |
| Dec 1961 | Jun 1962 | May 1963 | -22 | 1 year, 5 months |
| Dec 1968 | Jun 1970 | Nov 1972 | -32 | 3 years, 10 months |
| Jan 1973 | Sep 1974 | Jan 1985 | -49 | 12 years |
| Nov 1980 | Jul 1982 | Dec 1982 | -23 | 2 years, 1 months |
| Aug 1987 | Dec 1987 | Aug 1989 | -27 | 2 years |
| Aug 2000 | Feb 2003 | May 2013 | -45 | 12 years, 9 months |
| Oct 2007 | Mar 2009 | Mar 2013 | -50 | 5 years, 5 months |
| Nov 2021 | Oct 2022 | Mar 2024 | -25 | 2 years, 4 months |
Data from Robert Shiller. October 2025.
- Average bear market loss: -37%
- Average bear market recovery time: 5 years, 4 months
- Shortest bear: 1 year, 5 months
- Longest bear: 12 years, 9 months
Again, you could choose to label the benighted sequence from the Great Depression to World War 2 as one giant bear lasting from September 1929 until April 1945.
Which would have meant over 15 years until you broke even. And then you got a whole 12 months off before the 35% plunge commencing April 1946.
What a time to be alive.
Essentially then, US stocks have suffered three lost decades in 125 years.
Yes, the US – the land of the permabulls!
This might seem like scaremongering. But if an investing lifetime lasts 50 to 60 years (accumulation and decumulation phases combined) then many of us are likely to live through the sharp end of at least one such stagnant period.
Investing in the real world
So far we’ve considered raw market data. But in reality, the bear market recovery time we experience will be further drawn out by investment costs.
And on a brighter note, we can improve our results by pound-cost averaging through the downturn, and by diversifying into defensive assets – such as government bonds – ahead of time.
The chart below shows how a larger allocation to high-quality government bonds sped up the recovery from the coronavirus crash versus a pure equities portfolio:
Source: JP Morgan: Guide to the Markets. 31 May 2022. Page 63.
The All-Weather portfolio is another asset allocation approach that can dramatically reduce the severity of a bear market.
Yes, you’ll probably pay for this cushioning in the form of lower long-term returns. (Though that’s never a certainty).
But experiencing shallower swoons makes it easier to stay the course. And it’s far harder to come back from a bear market if you panic sell after a deep plunge, lock in your losses, and then miss the rebound.
So take the right steps to protect your portfolio ahead of time. It’s usually too late once a bear market runs wild.
Take it steady,
The Accumulator
I notice that the higher the inflation rate the longer the recovery takes and the less effective the equity/bond split becomes.
I have been increasing my exposure to TIPS.
Thank you @TA, interesting to see the different timescales and compare UK vs. US. Although totally agree pound cost averaging in any downturn and diversification can help. However with regards to diversification, I do think its worth noting (just as you noted for Covid Crash how bonds helped for that period), that so far for 2022’s “Bear” they have not helped much, YTD numbers from Morningstar:
VWRL -3.9%
IGLT -11.7%
RL Global Index Linked Fund -7.4%
Its weird to see Lifestrategy 100 doing better -3.1% than Lifestrategy 20 -8.9%.
I also think just looking at this this year to date, some of the Alternatives commonly used as diversifiers are holding up pretty well. My preferred Renewable Energy IT has helped (up 11.6%), as has my main physical Property IT (11.8%), Infrastructure IT is also up just (2%). Not all my diversification has helped though PE (-15.2%). I don’t personally hold any hedge funds, but a mainstream one available to private investors, which @ZXSpectrum has mention previously is up 18.6%.
Great article thanks, oh heck, last year’s just one more year full time working to fire, is looking like best add another year to that or coast in a part time gig.
Damn you inflation thought I could just pretend you would always stay below 5%
Can’t enjoy the bull years without riding out the bear 🙂
Close the hatches and prep for the storm!!!!!
It does not have to bad news that market takes longer to received. It can be the good news for investors depending on where they are on their journey. If you DCA from your salary and still got plenty time to retire then it is a good news it takes longer for the market/your wealth to recover. You will buy cheaper, therefore you will increase the odds for positive returns in a future.
@Peter it’s only good news in the short term though. Every investor who “gains” while in accumulation mode has to save more to hedge the risk of the bears that might occur in decumulation mode. Given that the point of FIRE (and investing more generally) is to get into decumulation mode, the bears just mean more of your life working.
But you don’t have much choice in the matter.
@miner2049r, yep OMY threatening to become TMY – where T could be Two or Three!!!! 🙁
Whettam, my green energy infrastructure trusts have kept me positive YTD, along with holding no bonds over 5 years, no tech, no US equity unless they happen to pay reasonable dividends.
JLEN Environmental Assets Group has been the standout performer for me, up 26% YTD. So much so I have sold down the bulk of the gain to redeploy, as it had grown to be my largest single holding after a dull 2021.
Anyway, my positioning coming into this year was heavily influenced by the CAPE ratio by country, which was covered here the other day 🙂 , and looking back at how long bear markets can run, even for the golden boy index the S&P500.
This article is an excellent reminder that the current bear market still likely has a way to run, and adjusting for inflation there is a lot of lot of ground to make up.
As such, and as I don’t have decades to accumulate, I demand dividends/income as risk compensation now more than ever.
I always find content like this fascinating. Two things leap out to me. Firstly this reinforces that the liability matching approach (covered in your Pension Split Series) best matches my personality to deal with any pre-pension age gap. Falls and recovery times like these just aren’t worth messing with in my opinion (obviously depending on someone’s personal situation).
Secondly, it’s how important the purchasing power of money is, and the impact this has on our potential FIRE date. It’s easy to get complacent with this while inflation plods along at less than 1.5% year after year. But if energy, fuel and food costs have just added £2-£3k to my annual expenses then at a 3%WR that’s an extra £66k-£100k I need to save to cover these core expenses. Ouch! Factor in the market falls too and it’s very easy for FI to suddenly feel a long way off again.
It’s easy to see why people give up or feel this FIRE lark is not achievable for a “normal” person, so never bother starting in the first place.
Guys, I went back and read @TA’s post, the ‘Coronavirus as told by Monevator commentators’! link! It’s fascinating on so many levels.
Please go back and add ripostes to what you said, felt and wrote at the time. So many of you left incredible insights. There are too many to draw attention to, but @TI – just amazing! You were a 10/10 coach, @zx measure in USD, anything else is a lie, but honestly too many smart, prescient comments to mention individually.
@TA it was such a great concept for an article and v valuable record!!
@SemiPassive, ‘I demand income/dividends now – yes!
I left a comment a couple of weeks ago about my response to the best market. I adjusted my annual portfolio from 4% loss to a 7% gain last FY. Obviously, I track reality, but demanding a basic return for risk is a form of defiance. And if I can’t achieve this ridiculous low bar, I would act on this information and make better investment choices.
I did this over C19 and my portfolio the following year returned 20+%, which returned my ongoing returns to < 7%+.
I only do this for the retirement/superannuation component of my portfolio. And yes, of course I pay attention and adjust for currency and inflation.
The markets are meant to work for me not vice versa. So far, they have.
@London a long time ago — Cheers, it was indeed a great idea of @TA’s… I was skeptical at the time as it felt like we’d been overwhelmed in the proceeding weeks, but in retrospect it was both a useful summary of shifting sentiments and, as you say, is now a really interesting historical document.
Incidentally if anyone looks back and sees they wrote something that they wish they hadn’t, I say go easy in yourself. It was a mad time to live through, let alone invest through, and none of us got everything right. Just getting the basics right was job well done! 🙂
There are bear markets and then there are bear markets, secular bear markets that end the long term frenzied up cycle and cyclical bear markets that are pauses in the long term up cycle.
Everyone will have a different opinion on which one is which but it interesting that the longer bear markets (10, 7, 9, 1, 2 on the first chart) have occurred during secular bear markets; 1929-1947, 1965-1982, 2000-2017. The rest have been cyclical bear markets. The average is roughly two years, but a pause for two years is very different to a severe drop, and will definitely feel so economically for most people (rising unemployment etc).
So where are we now? I believe that this is a pause in a secular bull market, a bit like 1987 to 1990, although not accompanied by a slumping housing market, that would mean 10+ years of bull market ahead of us until the secular blow off top.
Some charts to illustrate the above are here:
https://twitter.com/17_6YrStockCyc/status/1396116805974962176?cxt=HHwWgMC50dG1gOAmAAAA
@ Whettam – agreed that conventional bonds haven’t helped this time around and were never going to when inflation got out of control.
What’s become apparent from quite a few voices in various comments threads is that many people expected bonds to automatically work when they quite often don’t. They’re an imperfect hedge to say the least.
And the losses inflicted by long-dated inflation-linked bonds as real interest rates rose are a real shocker for many.
I don’t know if it’s because of the oversimplification of investing education, or it’s indicative of insufficient research, or simply because bonds are so damn counterintuitive.
Short index-linked bonds look OK, but one of the few YTD bright spots in my portfolio is a FTSE 100 ETF clocking in at 3.65%. About time it pulled its weight!
Still, after inflation everything looks awful so I have to cast my eyes at the 10-year annualised returns for solace 🙂
@ London a long time ago – cheers! It was a labour of love putting it together. Especially in the teeth of TI’s outright scepticism. There’s nothing like hearing he’s a 10/10 to soften his ‘tude 😉
On bear markets – it’s pretty amazing how many articles there are out there that gloss over how long recovery can take. I guess it doesn’t make sense for a financial services provider to scare the horses.
Don’t know why, but I’m surprised at the lack of comments with this article, TA must of nailed it especially around how long the actual recoveries can take, for sure I’ve bookmarked this one for guidence for when the bear shit hits the fan 🙂 now, later, next year…?
Is a further element that people project an annual return in their financial plans that is also missing so return to the original £100k is only part of the story ?
I seem to have inadvertently used an Asset Allocation close to the All Weather portfolio for many years albeit without Commodities or Gold -both of which I know very little about
Started Retirement with a 30/70 Asset Allocation-currently 35/65
Retired 23 years so seen some serious ups and downs
Never traded or tinkered except to make a once yearly withdrawal -rebalancing at the same time
No doubt could have made more monies but this AA suited my investing personality and more importantly has done the job-so far!
xxd09
@TA:
Re: “Again, you could choose to label the benighted sequence from the Great Depression to World War 2 as one giant bear lasting from September 1929 until April 1945.
Which would have meant …”
This is what ERN seems to have done in his 2019 post, see: https://earlyretirementnow.com/2019/10/30/who-is-afraid-of-a-bear-market/
Could you provide some explanation as to your bear start and bear end criteria, as I think the latter seems to include something like curtailing the current bear duration if a new bear then starts, ie I think your bears cannot overlap.
Assuming I might be onto something then it may also be worth reconsidering this sentence (and any others like it): “Remember the recovery periods above only get you back where you started.”
Thanks.
I remember 2007/8 when suddenly banks blew up and it was a very real possibility that the whole financial system could break down. It put me off investing for years and led to the opposite mistake, not taking enough risk with stocks when expected returns were good.
Scott Cederburg’s work has changed how I think about this (the interviews on Rational Reminder are good entertainment for long journeys). Risk is a different thing for long-term investors. We need to square stock market risk, inflation risk and longevity risk. Over the long term, stocks become less risky because their returns tend to revert to the mean. The opposite for nominal bonds, their risk increases. Correlations between countries go down; the correlation between stocks and nominal bonds goes up. Diversification between countries becomes more valuable over the long term while nominal bonds become a liability.
The limitation with Scott Cederburg’s extensive historical dataset is that it only includes stocks, nominal bonds and cash. With that limitation, 100% internationally diversified stock portfolios came out as *the least* risky in his simulations. But other diversifiers may well do a better job (I think that linkers, annuities, gold and commodities do). As always we are left to make our own judgements.
Good work, TA.
For me this measure is the wrong way round, however. My question in a dip is “how far has this set me back?”. I.e. how recently was I last at this value.
It seems a bit unfair to compare to the absolute peak of all time prior to the drop. Insane melt-up factored in … is it any surprise it takes a long time to recover? Just as I can’t time the bottom of the market, I’d be unlucky to nail the top too.
So if we fell 50% (vwrl to 61) that would be Apr 2020 or Jan 2019 if you want to ignore the covid blip. 5 or 6 years rather than the 12.
Similarly in 2000 the trough in 2002 (ftse100) only took you back to 95.
I am as defensive as I can stand right now, looking for a reasonable yield and swerving unreasonable growth. I am also trading like I did in 1999, just before taking some shocking losses so there’s that…
Drawdowns and volatility are the price of admission, and bull markets are like playing a video game in rookie mode.
Without the crashes, there’d be no ‘excess’ returns over cash.
And getting cut in half (or worse) is how you earn your way in markets.
These excellent historical reviews always throw up intriguing ‘anomalies’.
The worst real term GBP denominated bear market lasting longer than any for the UK market alone – explain that one! (Global Diversification, Nil: Unwanted Outcomes, One).
A World War wiping out 60 or 70 million people being less of drag on returns than the fall out from either some grumpy petrostates hiking the cost of a single commodity in the seventies, or people getting a bit carried away with a piece of new tech (and over their skis on valuations) in the nineties.
To me the seventies stand out a bit.
In 1999 it was really obvious things were crackers.
Everyone knew that stock price rises like those in 1994-99 were going to be wholly unsustainable, and that there was going to be a heck of hangover.
Granted, the length of the bear market after 2000 and the fake top in 2007 were a bit of surprise, but the direction of travel was not.
But in 1969?
I don’t think it was obvious then that this was a bad time to get into stocks.
And in the 1910s, 1930s and 1940s it was hardly surprising that equities got beat up, given what was going on in the World.
Admittedly, the extent of the sheer collapse from 1929-32 must have taken even skeptics of the roaring twenties aback at the time.
But I don’t think that there was such a prevalent sense in, say, 1969/70, that the long boom up to the end of the sixties would end so badly in immediately subsequent years; or at least not in terms of there having been a vibe (as with the excesses of both the twenties and the nineties) that prices would have to unwind themselves, possibly dramatically and soon.
The new 4% rule…
Buy index linked gilts on 1.22% real yield and draw an index linked 4% that lasts 30 years.
You can get well over 1.22% real now…
Okay I’m not really advocating this but the numbers don’t lie, it doesn’t have to be stocks versus nominal bonds , there are other options out there.
@Hariseldon – strikes me as a perfectly rational plan if you have enough investable cash, subject to risk of (“hard” and/or “soft”) default by the government. Thank you for pointing out that there is often insufficient focus on index linked securities.
@mathmo – would you want to adjust for inflation with that mode of thinking? Another, potentially psychologically helpful, approach could be to adjust your networth calculation so it isn’t current value but some function that is less than that based on historic average draw downs? I.e. you’ve already mentally accounted for some or all of the pain of the next crash. Otherwise you’re always sort of wildly optimistic about your position
@rhino I definitely would, but I’d also want to reinvest dividends so I’m guessing the answer is similar.
@Sparschwein (#17)
McQuarrie’s work (McQuarrie, E. F. (2023). Stocks for the Long Run? Sometimes Yes, Sometimes No. Financial Analysts Journal, 80(1), 12–28 – there’s a version at SSRN too with the title “Where Siegel Went Awry: Outdated Sources & Incomplete Data”) on extending stock return data back to the 18th century forms a useful counterpoint to the work of Cederburg et al. since the abstract states “The new historical record shows that over multi-decade periods, sometimes stocks outperformed bonds, sometimes bonds outperformed stocks and sometimes they performed about the same. New international data confirm this pattern. Asset returns in the US in the 20th century do not generalize. Regimes of asset outperformance come and go; sometimes there is an equity premium, sometimes not.”
@Hariseldon (#20)
Currently a 4.4% payout rate on a 30 year index linked gilt collapsing ladder.
Alternatively:
Single life RPI annuity at 65yo (so roughly a 30 year horizon) 5.2%
Joint life RPI annuity at 65yo (100% beneficiary, same age) 4.6%
So, there are plenty of choices for providing income flooring in retirement (beyond the state pension and index linked DB pensions).
Of course, the so-called ‘4% rule’ only applies historically to the US (more like 3.0 to 3.5% for the UK) and, in the future, doesn’t guarantee that the portfolio will survive for 30 years.
Something I don’t understand here – there’s no mention of rebalancing. Are you taking a snapshot of a portfolio and seeing how long it takes to recover, or running it forward with continuous or periodic rebalancing?
@DH (#19)
Re: late 1960s
A flavour of contemporaneous opinion of the general economic position can be found in the various reports at the BoE (e.g., https://www.bankofengland.co.uk/-/media/boe/files/annual-report/1969/boe-1969.pdf)
The introduction to the 1969 report linked above mentions sterling devaluation, problems with the $, gold, and the balance of payments (remember when that used to be important?). In other words, there was plenty going on.
@Al Cam – Bear market is 20% down from previous market high. Trough is lowest point before recovery. Recovery is first month that exceeds previous market high. All in real terms.
However, I’ve made a couple of exceptions. For example, in real terms the US Dotcom trough was actually the bottom of the GFC. (That’s a -52% loss relative to the market high on the eve of the Dotcom Bust.)
But in nominal terms the Dotcom Bust losses were recovered before the GFC began. When that occurred depends on whether you’re looking at the total return index or the price return index.
So I’ve separated out those two events in the US bear trough column, though by my preferred definition they should strictly both trough out in March 2009.
In the US, the GFC is a bear within a bear – the market didn’t recover to its previous real-terms high before the Credit Crunch hit. But it seemed daft to me to ignore the GFC.
The second exception like that occurs in the MSCI World table. The Sep 1974 trough could be inputted as April 1980 (a -56% slide). But I’ve distinguished between the oil crisis and the later turbulence in the 1970s.
In the same period in the UK, Dec 1974 is the lowest point. I can easily see the case for ignoring the double dip in the case of the World and UK markets as I don’t think it really lives on in our cultural memory. It’s all just the 1970s – kipper ties, loons and other past mistakes.
In the US, part 1 of the Great Depression (-77%) is recovered in real-terms in Nov 1936. Only for the market to plunge again from Feb 1937 (-48%). Then World War 2 rocks up! The market hasn’t reached a new high before WW2 but again, it seems weird not to separate it out.
@DH – yes, the 1970s stands out as a dreadful period not informed by a World War. From my perspective, it’s why inflation protection is essential.
@Sputan – This article is written from the perspective of stock market drawdowns. The final section nods to the fact that you can alter the impact via portfolio management. One of our problems as Monevator writers is knowing where to draw the line 🙂
@TA:
Thanks for the additional details.
In my view the “rules are the rules” and the index you follow should be stated (IIRC ERN did both total returns and index alone in at least one of his Tables) and, there should be no exceptions. That bears overlap (partially or completely) is just one of those curious things that happen from time to time. The problem I see with not sticking strictly to the “rules” is that you can under-play things.
The key message (I assume) is that it can take a very long time to recover from some bears; according to ERN: on average >ten years for the five worst US bears in the last century! That some of the bears overlap is a blessing otherwise you might conclude that you have more than a 50% chance of being in a bear over the last century!
IMO this stat should also mean very different things to you if you are 35 or 75.
@Trevor Eaves (#1): is there any way to hold individual TIPS in the UK or are we limited in practice to, say, ETFs?
@Al Cam – I’m not sure what you mean by the “rules are the rules”. What rules are you referring to?
As you can see I’ve clearly provided links to the data used. Anyone can replicate what I’ve done or ask me questions about the results.
The issue here is that you have to decide whether to present bear markets in price return terms, total return or real total return.
ERN neatly sidesteps the problem I ran into by presenting the bear market trough in nominal terms only.
He then shows months to recovery in nominal and inflation-adjusted flavours.
In other words, nominal peak-to-troughs “underplay” the severity of most crises, but we both get to the same place re: inflation-adjusted recovery.
That’s because the recovery duration is the length of time it takes the index to recover its pre-bear value in real terms. It doesn’t depend on the trough.
If I had daily historical data instead of monthly then we’d probably see that the troughs were slightly worse on an intra-day basis.
Given more time I’d probably add in near-bears, say, anything over a 15% loss. That would add a handful of extra drawdown events. And I’d add in Japan for a taste of a market that’s been less correlated to the UK and US.
You’d see many overlapping bears during their 30-year post-1989 uber-bear market. It’d make sense to include them too, otherwise you’d miss, for example, the -57% hit inflicted by the GFC.
@Alan S – thanks for pointing this out. Astonishing how shaky the foundations are even for the most common investing gospel. Can we get to the bottom of this?
In my understanding, Cederburg’s latest paper (Anarkulova et al. 2025, SSRN #4590406) goes beyond McQuarrie’s points. Their data is from 39 developed countries and goes back to 1890. They model what matters most, whole portfolio outcomes (not individual asset classes) for the probability to run out of money in retirement. They say “Two methodological choices—preserving time-series dependencies via block bootstrap* and including international stocks—appear to explain the disconnect” [between their results and conventional wisdom on bond allocation].
So, in the hypothetical limited case with only stocks, nominal bonds and cash, the international all-stock portfolio is best and least risky over the long term. That doesn’t mean 100% stocks is best in the real world. It just shows that stocks are less risky than thought especially considering inflation and longevity, and that nominal bonds and cash are poor choices. Scott Cederburg himself recommended an index-linked annuity for most people.
*) The statistics is beyond me. I think it means that the property of stock returns to revert to the mean, and of bond returns to do the opposite, is preserved.
Paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4590406
Interview: https://rationalreminder.ca/podcast/350
The Economist summed it up in a useful graph here: https://www.economist.com/finance-and-economics/2025/11/11/old-folk-are-seized-by-stockmarket-mania?giftId=NWExZjlhNTctNzFhOC00ZDhkLTkyZjMtNzMxMTMwZWJkNGI2&utm_campaign=gifted_article
@Alan S – thanks for pointing this out. Astonishing how shaky the foundations are even for the most common investing gospel. Can we get to the bottom of this?
In my understanding, Cederburg’s latest paper (Anarkulova et al. 2025, SSRN #4590406) goes beyond McQuarrie’s points. Their data is from 39 developed countries and goes back to 1890. They model what matters most, whole portfolio outcomes (not individual asset classes) for the probability to run out of money in retirement. They say “Two methodological choices—preserving time-series dependencies via block bootstrap* and including international stocks—appear to explain the disconnect” [between their results and conventional wisdom on bond allocation].
So, in the hypothetical limited case with only stocks, nominal bonds and cash, the international all-stock portfolio is best and least risky over the long term. That doesn’t mean 100% stocks is best in the real world. It just shows that stocks are less risky than thought especially considering inflation and longevity, and that nominal bonds and cash are poor choices. Scott Cederburg himself recommended an index-linked annuity for most people.
*) The statistics is beyond me. I think it means that the property of stock returns to revert to the mean, and of bond returns to do the opposite, is preserved.
[posting again without links, not to bother the moderators]
@Sparschwein – From a quick eyeball of the countries listed and the start dates, Cederburg’s paper is inevitably going to skew towards 100% equities.
Essentially post-war inflation smashes bond investors in the 20th Century in many of those countries.
But the paper mostly doesn’t include countries / start dates where stock investors were wiped out by Communist takeovers.
I need to write a post about this but my background reading has convinced me that mean reversion is an optical illusion and not a force of financial nature.
In short, I think that the case for 100% stocks is overstated when built on 20th Century international stock market survivorship bias.
FWIW, I agree with the index-linked annuity conclusion while diversification remains the only free lunch.
@TA:
By “rules are rules” I mean no exceptions.
1929 your table says 7 years two months to recover; ERNs says 188 months ie 15 years 8 months – these seem very different!
@Alan S #26: very interesting. I must reread Sandbrook’s White Heat 1964-70. One can get a rose tinted view of the latter sixties.
I had been thinking more internationally, i.e. in terms of how poorly a (hypothetical) diversified global investor did from the end of sixties through to the early eighties, even, in some ways, compared to such an investor in the nineteen forties.
Of course the fifties had tremendous stock returns in the US, as AWOCS/Carlson recently covered:
https://awealthofcommonsense.com/2025/11/the-greatest-bull-market-no-one-every-talks-about/
And the later sixties obviously had the pressures on the Bretton Woods system and Guns and Butter issue of funding both the Great Society and the second Indochina war.
But, even so, the seventies are slightly surprising to me in terms of just how bad it got for international investors.
At peak in 1968/69 Vietnam consumed 3% of US GDP (and defence as a whole 9%). In 1944/45 the war effort consumed over 38% of US GDP (even at the peak of the Korean War, in 1952/53, the burden was as much as nearly 14%, of US GDP for defence, much higher than fifteen years later).
It seems that in the seventies perhaps it wasn’t one really big disaster, like in WW2, but a cumulation of much smaller but many mini crises
@TA #34: “doesn’t include countries / start dates where stock investors were wiped out by Communist takeovers”: IIRC DMS did look at this from 1900 and concluded that where this had occurred (the Russian Empire transitioning to Soviet Russia in 1917/18, the Chinese Republic to the PRC in 1949/50) it reduced the real £GBP return of a global cap weight investor from 5.4% p.a. to 5% p.a. (across 125 years)
@ Al Cam – ERN’s 1929 real return recovery ends over 15 years later in 1945 i.e. the same point at which my 1937 bear market recovers. If you merge the Great Depression into one event then that’s what you get. You already pointed out to me that’s what ERN did.
Finally, my data comes from Shiller. His numbers are average monthly returns. ERN’s conclusions will be slightly different if he’s using month-end returns or different inflation data. For example, a slight variation in the numbers would wipe out the few months of real return recovery Nov 1936-Feb 1937 that occur using Shiller’s data. I mentioned in the piece that it could be worth viewing such edge cases as one big bear.
@DH – That’s interesting, cheers. I’ve got a vague memory of a paper – can’t remember if it was DMS – that included a number of Eastern European markets that were previously excluded. Actually, as I’m writing this, I think the author(s) were specifically pushing back against DMS’ rosier numbers. I can’t remember what difference it made but I remember wincing a little. I must try and dig up that paper. I’ve noticed in the years since that DMS have continually updated the number of markets they include in their data.