The Japanese stock market crash: the bursting of a bubble [Members]
For MAVENS and MOGULS by The Accumulator
on December 10, 2024
The Japanese stock market crash of the early 1990s is the investing equivalent of a scary bedtime story. “What about Japan?” the old hands mutter darkly whenever the youngsters get overly excited about their S&P 500 profits. As well they might, because the Japanese nightmare has an irreducible ‘There but for the grace of God…’ quality about it.
Partly that’s because the bursting of the Japanese bubble was such an extraordinary fall from grace.
This article can be read by selected Monevator members. Please see our membership plans and consider joining! Already a member? Sign in here.
Comments on this entry are closed.
Sequence of returns? Catastrophic failure?? Fine line; scary indeed. Thank you @TA for a fascinating analysis. We may have had a bit of a dart at Rob Dix’s choice of sampling over the weekend, but there’s definitely a whiff of the lottery ingrained in all of our retirement investments!
Super interesting stuff. Especially the “What if you weren’t 100% Japan?” analysis, which was always the question that’s been in my mind in response to “what about Japan”. That analysis has calmed my slight nervousness of most of my wealth sitting in a global tracker and having so much exposure to the US.
> stories of Japanese investors paying record amounts for art and golf clubs.
Diversification in action. Buy land – they’re not making any more of it. Wonder how they fared relative to the stock market?
Out of interest, d’you know what the PE was at the high water mark?
Brilliant piece. Many thanks @TA
Ben Carlson of AWOCs has Japanese stocks and real estate up to 1990 as the GOAT bubble:
https://awealthofcommonsense.com/2016/09/the-greatest-bubble-of-all-time/
https://awealthofcommonsense.com/2021/02/the-defining-trait-of-all-bubbles-the-willful-suspension-of-disbelief/
Whilst his colleague Barry Ritholtz’s choice for the worst % stock exchange index crash is Cyprus from November 1999 (that index only bottomed in 2015, the year after the article below):
https://ritholtz.com/2014/03/the-worst-bear-market-in-history/
@ ermine
Grantham has it as 65x https://www.gmo.com/europe/research-library/waiting-for-the-last-dance_viewpoints/
What a thrilling read! I also loved the what if part, which is mostly missing from Japan stock market articles.
Seeing the volatility there was also a lot of money to be made, if you were lucky! I’d love to talk to a Japanese investor who made money through all those years
@tetromino #5 thanks, and – well, 65! holy crap – irrational exuberance as they said of the dotcom bubble, just twice as much
Excellent (if scary!) article, @TA. I’d be interested to know how quickly (slowly) an investor would have broken even if they had been yen cost averaging from the peak rather than putting in a lump sum?
Windy
@Ermine – I’ve seen quite a few different estimates from P/E over 100 to somewhere in the 70s. I haven’t seen Tetromino’s link before which is different again. Obviously it goes off the charts no matter how you cut it, although P/E is a relative number i.e. best compared against the (higher) Japanese average than the US or UK mean.
The fall in property values was dreadful. As I understand it, many non-real estate Japanese firms speculated on land values and became known as property bets tethered to a stock price.
@pourquoi pas – absolutely. I haven’t come across an account from a Japanese investor who lived through it. There must be some translations out there.
@Windy – yes, it’d be interesting to run through some scenarios. You could imagine a 22-year old with little exposure at the peak, who only starts investing serious amounts at age 35. They’d be OK. Versus a retiree, fully invested at the peak, with only withdrawals to come, they’d be in trouble.
That said, I think Japanese bonds were fine – boosted by falling interest rates and deflation
I’ve read a withdrawal rate analysis that had a 60/40 Japanese investor retiring in 1989 running out of money after 22 years using the 4% rule. They just scraped over the 30-year mark using a 3.3% SWR.
The ‘Japanese investor’ example reminds me how odd it can be to deal with Yen amounts when you’re used to GBP. The 1,000 Yen quoted is only a fiver, so feels a little out of place. I wonder if Japanese sites show returns per 2,000,000 Yen or similar.
Ha, I didn’t check the value of 1,000 Yen cos it was just notional. Probably wasn’t too traumatic to lose that amount 🙂
This is absolutely fascinating, and a good reminder about what CAN happen at a time when a new Memecoin is released daily…
The only thing I’m curious about is Currency. I presume that all these charts are in Yen. It would be fascinating to see them reset to e.g. US$ (or GBP). I suspect that certainly the great results of the last 10-20 years wouldn’t look so good.
The hidden worry for long term investors – especially any looking to use stock markets to try and build up inter generational wealth – is “are we going to go Japanese?”
China already seems to be transitioning to a low to no economic growth / declining working population / worsening ratio of dependents to workers scenario.
In recent weeks the yield on Chinese Government bonds fell for the first time below those of Japanese Government bonds.
A very deflationary signal I fear.
Arguably Europe has already made that transition.
Demographics is destiny, and where Japan and South Korea led the way it seems the rest of the world will follow (very long term global population and dependency ratio projections from the Public Library of Science below):
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0298190
If this comes to pass, then the very long term returns on even diversified global equities could look distinctly like the lost decades for Japanese investors since 1989, albeit perhaps without the very jarring sequence of returns which the Japanese suffered.
@Delta Hedge – it’s a fascinating question. Partially we’ve staved off stagnation by being open to immigration. It remains to be seen how much more room there is for manoeuvre on that score. I suspect more than we might think if the government can demonstrate “control” and make a persuasive economic case.
Also, one consequence of the bubble was that Japanese firms slashed investment. I was slightly discombobulated yesterday by a US economist mentioning that the UK was the 2nd ranking G7 nation for productivity growth post-2019.
I zoomed in on his chart and discovered it was true. Not saying much though.
One question I have is why are we simultaneously worrying about demographic decline and AI stealing all the jobs?
Also note, that the last 20 years do not look like lost decades for Japanese equity investors. Economic stagnation and return stagnation aren’t necessarily linked – valuations matter. In much the same way that investors haven’t made a fortune out of China on its way up.
@Tom – yes, all in yen. The yen has depreciated against GBP since the peak, so would look worse for a lump sum investor in 1989. Maybe not so bad for a drip-feeder though?
“Versus a retiree, fully invested at the peak, with only withdrawals to come, they’d be in trouble.” – except they had presumably invested into the mother of all Bull runs? Assuming they drip fed in over the proceeding 20 years their number should have been huge. If you got a bit greedy/hasty and decided you could retire early because you’d hid your (inflated) number then not so much
@TA #14: “why are we simultaneously worrying about demographic decline and AI stealing all the jobs?”: It is logically disjunctive isn’t it? Humanoid robotics,
AI agents and automation of all sorts would seem to be the only very long term way out of deteriorating dependency ratios as, eventually, we will just run out of migrants to import as all countries complete their own demographic transitions. Japan, of course, back in 1989/90 was the most automated and technologically savvy country of the time, and it didn’t spare them what followed. I’m not sure infinite digital humans and billions of person like robots are on the horizon soon. I’ve just read a piece that, even if LLM scaling curves hold, then we’re still 9 Orders of Magnitude off AGI on compute (i.e. a billion fold increase in FLOPS needed for model training) and 5 OoM (100,000 x) off on the requirements for fresh high quality training data (predicting we hit a wall on the latter in 2026).
> why are we simultaneously worrying about demographic decline and AI stealing all the jobs?
Because the efficiency gains won’t be equitably spread. We are slightly the beneficiaries of this, the top .1% are very much the beneficiaries of that.
Rob Dix was telling us last time that a sense of meaning and value through work is important to lots of people. They’ll be mightily hacked off – their Dads built cars, or dug coal out the ground, they are driving Uber on ZHC or pulling coffee. Half my sarf east London grammar school left school at 16, fixing cars, making stuff in the small firms that had not been purged out of London by money. They were earning while I didn’t for another five years and then I exited into Thatchers first big recession. Ha ha bloody ha.
Their grandsons and granddaughters are Millennials, what the hell does Gramps say to his Millennial grandkids still living at home because they can’t afford a house? The siren song of Farage and Trump will play strong to this. “Remember when you could go to a factory and earn a decent wage and buy a fixer-upper, Grandad? Where did that all go whistling down the wind?
MAGA and MBGA will be a stronger siren song than Yeah but the world in general is richer now and your phone isn’t screwed to the wall. It’s not an unreasonable question to ask but why is our end of the boat sinking to make people we don’t know richer while we end up destitute?.
Who do you think will capture the value from the AI boom, should it be viable? Unlikely to be the 99%
@Sharkey – yes, agreed, we know by now that if you retire after a massive run-up in valuations then you should probably trim back your SWR. Still a terrifying prospect to watch that much wealth disappear. We know in retrospect a 3% SWR would have been okay. But managing that for real? I guess I’d be leaning into my side hustle 🙂
@Delta H – I suspect you’re right that the technology curve will prove steeper than advertised and it’s probably best not to rely on the rise of the robots to save the day. (Never works out in the movies.) Certainly the tech was too immature to compensate for early onset Japanese demographic decline and, as I understand it, they didn’t turn to immigration as a solution.
@Ermine – Fewer workers probably favours labour over capital unless automation can take up the slack. Just happened in the US post-covid though it didn’t seem to make the MAGA half of the country any happier. Still, I wasn’t thinking of who would take the spoils. Demographers are concerned about future worker shortages eroding the tax base / weakening the dependency ratio and social safety nets. But if AI / automation increases productivity then that’s surely less of a worry? That is if, as you and Delta say, the technology fulfils its promise.
I suppose much also depends on which jobs get automated. We don’t have many jobs left in manufacturing to lose. If low paid service jobs get automated away then ideally we learn from the 1980s and support people to reskill rather than leave them to sink or swim. If well-paid service jobs get automated away, then the same but trickier. The fewer pathways to economically desirable jobs, the greater the chance of discontent as you point out.
What does someone say to their kids or grandkids who can’t afford a house? Probably something like, hey kids, let’s pass legislation that makes it easier to build, harder to object, and encourage empty-nesters to free-up family sized homes?
According to the Economist, few companies seriously use AI, and adoption is slow*.
This certainly fits my partner’s experience who is in charge of data strategy at a mid-sized, profitable pharma. It’s actually not obvious to find use cases. You need large amounts of good data (or crap in/crap out), which often requires cleaning up messy data, which no one wants to do. Implementation often involves redesigning the whole business process from scratch. Then there is the usual inertia from political turf wars.
As for MAGA: The average worker had little benefit from the everything-bubble in financial assets, and got hit hard by inflation. People may not understand how central banks screwed them over with negative interest and “QE” (aka money printing and state funding). They do notice that the economic system doesn’t work for them anymore. That’s apparently enough to support politicians who appeal to primitive, nasty instincts.
*) https://archive.is/GhdjA