Japan's Millionaires Didn't Vanish - They Just Stopped Growing
In 1995, if you were in Japan's wealthiest 3%, the minimum threshold was $1.25 million. In 2023, that threshold is almost unchanged - still $1.25 million. Over the same 28 years, America's top 3% doubled their wealth; Germany's top 3% grew 70%. Japan didn't get poorer. Japan just stopped getting richer. Richard Katz spent four essays answering why. The surprise: Japan's tax code looks like the harshest in the OECD on paper (a top marginal rate of 56%), but the real-world burden is lighter than the OECD median at every tier - creating an invisible ¥100 million wall that blocks the flow from the middle class up to the elite.
(1995: #4)
(1980: 18%)
(OECD: 2.5×)
tax bends downward
when wage rises 5×
Act 1 · The Statistical Fall
There's a persistent rumor about Japan: "this country has no millionaires because the culture dislikes flaunting wealth," "because taxes are sky-high," "because society hates the rich." Richard Katz opens with a twist: the rumor is wrong. Japan does have millionaires - roughly 5.2 million adults hold at least ¥100 million (~$1 million PPP) including their primary residence, equal to 5% of the adult population in 2023, up from 3% (3.1 million) in 1995 and 4% (4.2 million) in 2012. If real estate and stock prices keep recovering, that share could reach 6% within a few years.
Moreover, within the already-wealthy group, assets are still growing fast. Over the past decade, the average wealth of 1.5 million households holding ¥100–500 million more than doubled, to ¥221 million ($2.2M PPP). The wealth of the top 90,000 households (¥500M or more) even tripled, averaging ¥1.5 billion ($15M PPP). Japan isn't short on rich people - Japan has rich people who keep getting richer.
So where's the problem? It isn't the count or the current level. It's the pace relative to the rest of the world. While other rich countries doubled the average wealth of their top 3% over three decades, Japan grew just 26%. In nominal yen, that's still growth. By international standards, this is the longest ranking slide in the peacetime history of a developed economy. Real wages rose in only 3 of the 13 years through 2023 - a run no other democratic society would tolerate if it lasted more than a single cycle.
Katz borrows the "Goldilocks" fable to frame the whole series: a healthy economy needs just the right number of millionaires - not too few (starving startup drive, venture capital, and high-end consumer demand) but not too many either (inequality drags growth down). The OECD estimates that from 1985 to 2005, average member-country GDP grew 28%; had inequality not widened, it would have been 33%. Japan is getting both ends wrong: the upper-middle class is too thin, yet the ultra-wealthy are still favored by the tax structure. That's the story the next four acts will dissect.
Here's a more bitter data point: Japan's 1,000 richest people. In 1995, their average wealth was 45% above the median for the equivalent group in rich countries. By 2023, it had fallen to just 54% of that median - a 91-percentage-point slide in a single generation. Japanese billionaires who once led the world no longer sit near the top of the Bloomberg Billionaires ranking. An interesting detail: 11 of Japan's new billionaires as of 2021 are almost all founders from the past 25 years - Uniqlo, SoftBank, Keyence, Nitori. Not the grandchildren of the old zaibatsu.
The depth of the slide becomes clearer when broken down by tier. The threshold to reach the top 10% (i.e. Japan's 10 million wealthiest adults) is $673,000 - 86% of the rich-country median. At the top 0.01% threshold (10,000 wealthiest), Japan sits at 95% of the median - nearly even. But once you reach the top 0.001% (1,000 people), the threshold is only $426 million - 66% of the international median. The higher you climb, the wider the gap. This matters for what follows: Japan's absolute upper class isn't just smaller in numbers - each individual's average wealth is also far smaller. This isn't "equality" - it's a net loss for the economy when large-scale venture capital is missing.
One more interesting detail from Katz's data: among those earning above ¥100 million a year, a third are business owners. In other words, the path to wealth in Japan today isn't climbing the corporate ladder - it's building the company. This detail resurfaces in Act 4, when we meet this group again through the lens of tax: a flat 20% on investment income, no matter how rich you are.
Act 2 · The Savings Trap and the Wrong Portfolio
Katz lays out four conditions for a society to produce lots of millionaires: (1) income high enough to save, (2) interest rates high enough for balances to compound, (3) investment in higher-yielding risk assets, and (4) asset prices trending upward over the long run. Japan - spectacularly - fails all four.
Real wages rose in only 3 of the 13 years through 2023. Average disposable income is ¥5.4M ($57K PPP) - 17% below the OECD median.
From 1996 to now: average deposit rates under 0.2%. ¥10 million earns just ¥20,000 a year. For cash holders, 30 years of compounding magic simply vanished.
Only 18% of household assets sit in stocks. Among the wealthy, 70% of risk assets are real estate - the reverse of the global ratio (real estate ~15%).
The TOPIX lost 70% from 1990 and only regained its 1989 level in 2024. Commercial real estate lost 80% before doubling over the past decade. The "lost decade" really means 30 years of lost compounding.
Why do the Japanese prefer real estate? Partly because the memory of the 1990 bubble left stocks with a reputation as a "casino." Partly because the tax system favors real estate: the effective tax on real estate income is just 15–17%, on dividends/capital gains a flat 20%, while wages face progressive rates up to 45–56%. Notably: real estate inheritance gets an assessed value discount of 30% below market price - a substantial freebie for wealth passed down through generations.
Someone earning ¥20 million a year invests in a rental property. By recording ¥3 million a year in depreciation (writing down the structure's value), their taxable income drops correspondingly - saving roughly ¥1.5 million in income and residence tax every year. After 10 years, that's ¥15 million flowing back onto their balance sheet - exactly 15% of the price of a central Tokyo apartment. This is why wealthy Japanese both buy real estate and mortgage it: to generate depreciation, cut taxes, and accumulate appreciating assets at the same time. Stocks don't give you that combo.
Stock concentration unchanged - while America's grew
Digging deeper into asset composition: the 18% of average Japanese household wealth sitting in stocks is just the headline number. Broken down by income decile, the picture sharpens considerably. The top 1% by income puts 54% of assets into risk classes (stocks + investment real estate). The top 10% is 42%. But the middle class - deciles 5–6 - invests only 24–26%. The rest of the average Japanese person's wealth: interest-free deposits.
On deposits - Japan's largest "dry powder" - Katz offers a telling comparison. Japan's top 1% wealthiest hold 33% of assets in bank deposits. The EU's top 1% is 34% - nearly identical. But the U.S. top 1% holds only 14% in deposits. Wealthy Americans put their money into stocks and private equity; wealthy Japanese and Europeans put it into real estate and savings. That's why, during the same decade of the U.S. stock boom, wealthy Japanese didn't grow their assets as fast - they simply weren't in that lane.
The result: NISA - the tax-free investment account expected to "democratize the market" - after 10 years (2014–2024) has left household market participation stuck at 17.3%, exactly where it started. Money flowing into NISA is mostly money shifted from taxable accounts into tax-free ones, not new money from new savings. In other words: NISA cut taxes for people who were already investing, without convincing anyone to start investing. Katz returns to this point in Act 4 to propose a new version of NISA.
Act 3 · Skilled Workers, Cheaply Paid
This is the second answer to "why so few millionaires": to build ¥100 million in wealth, you first need a wage high enough to save. But Japan's wage ceiling is too low. Only 7% of Japan's 108 million adults earn ¥10 million a year ($105K PPP). In the U.S., the equivalent figure is 21%.
More specifically, for the professions that mint "millionaires" elsewhere:
The consequence isn't just "fewer millionaires." Japan is running dry on high-skilled talent. Japan's PhDs per million people in 2023 is lower than 20 years ago - the only rich country where that's true. South Korea now has triple Japan's rate. METI forecasts a shortage of 450,000–800,000 ICT specialists by 2030. The best people leave for abroad; those who stay have no incentive to invest in further education when the reward is so thin.
The 1992 → 2023 flip: from leader to last place
Katz spends a long passage on this: if you were an average worker in Japan in 1992, your income was 12% higher than your counterparts in Germany, France, Britain, or Italy. Three decades later, that same person (or their children) earns 45% less. The same thing happened at the top: Japan's 2023 90th-percentile wage sits below the level other rich countries had already reached back in 1992 - 8% below where Germany stood 31 years ago. This is the "lost decade" quantified at the level of the individual.
The stagnation wasn't evenly spread over time. The 90th-percentile wage grew slowly until 2007–09, then collapsed during the financial crisis and never fully recovered. The median wage stayed nearly flat for all 30 years - a literal "three lost decades" for the middle class.
Another telling comparison: at the top 0.001% income level, Japan has 1,000 people earning over ¥774 million a year ($8.15M PPP). The U.S. at the equivalent threshold has 24,000 people - 24 times more, even though the U.S. population is only 2.6 times Japan's. This isn't the U.S. being "more unequal" in a bad sense - rather, the U.S. has a far thicker ultra-wealthy tier in absolute terms, and that tier funds VC, buys large homes, and spends on luxury goods all at once. Japan lacks that thrust at the top.
"When skilled pay isn't much higher than average pay, you don't just lose millionaires - you lose human capital. The upper-middle class isn't broad enough to sustain domestic demand, doesn't have the resources to take startup risk, isn't confident enough to buy stocks. The whole ecosystem dries up."
National proof: South Korea has overtaken
Reading the four-part "Missing Millionaires" series, there's an implicit question readers will ask themselves: "Is Japan really falling behind? Or is this just the natural growth slowdown of an already-rich society?" Katz answers with a separate 2022 piece - "Korea Has Surpassed Japan in Per Capita GDP" - in which he shows that Koreans' real living standards overtook the Japanese back in 2018, and the gap keeps widening.
Katz explains why Korea overtook Japan in a single sentence: "Korea shared productivity gains with workers; Japan did not." Over the three decades 1990–2020, Japan's productivity still grew slowly - but nearly all of that growth was retained by corporations as profit and cash on the balance sheet. Korea, by contrast, pushed most of its productivity gains into wages.
This is exactly why Korea now has a 90/50 wage ratio near 2.5× (matching the OECD median), while Japan is stuck at 1.8×. When high-skill wages don't rise, it isn't just millionaires that disappear - national productivity also fades, because the best people either leave or lose the incentive to invest in themselves. The "Missing Millionaires" picture at the micro level, and "Korea overtakes Japan" at the macro level, are two sides of the same coin.
Act 4 · The ¥100 Million Wall
This is Katz's final essay - and the twist ending. Everyone has heard that "Japan has the OECD's highest marginal income tax rate: 56%." The number is correct. But here's what that number doesn't tell you:
- The 56% rate applies only to income above ¥40 million ($420K PPP) - that is, 0.7% of the population.
- Averaged across the first ¥40M: just 33%.
- Someone earning 2.5× the average (¥13.5M): 36% tax, ranking 25/31 OECD - lower than 24 other countries.
- When wages rise 5× (from ¥5.4M to ¥27M), after-tax income still rises 4×. There's no "90% eaten by tax" as the rumor claims.
And here's the bend: the richer you are, the more of your income comes from investment rather than wages. Investment income (dividends, capital gains) is taxed flat at 20%, below the OECD median of 26% - ranking 21st lowest. Someone earning ¥115 million a year (Japan's top 3,000 adults) gets over 70% of their income from investment. The result: the effective tax curve climbs up to ¥100 million, then bends down.
Japan is not a country that taxes the rich hard. Japan is a country that taxes skilled labor hard - because skilled workers have only wages, not yet enough accumulated wealth to shift into investment income. Those who are already rich escape through the 20% window. Those still trying to become rich must climb the 33–45% progressive ladder.
This is why Katz's series has such a subtle structure: Act 1 poses the question, Acts 2–3 rule out the usual suspects (savings, wages), and Act 4 names the culprit - the tax system isn't "high," but it locks the upper tier shut, drying up the flow from the middle class into the elite.
Three concrete scenarios - "tax eats your raise" is fake news
Katz opens Act 4 by directly rebutting a claim common in Japan's financial press: "your wage rises 4%, tax eats 90% of it," or "70% of the raise belongs to the tax office." He shows that all of these claims are wrong on the arithmetic. Below are the three standard scenarios he uses to prove it.
Scenario 3 is a direct punch at the "tax eats everything" rumor: a 5× wage rise leads to a 4× after-tax rise. If tax "ate 90%," that number would have to be 1.1–1.3× - not 4×. Katz argues the rumor spreads because people feel the marginal rate (every extra yen taxed at 33–45%) without calculating the average tax rate across their total income.
The social-insurance loophole: a regressive tax at the top
This is a detail Katz saves for the end of Act 4 that most press coverage skips - and it reinforces the ¥100M wall argument. Japan's social insurance has an income cap:
This is a clearly regressive tax. A low earner pays 14.15% total (9.15% pension + 5% health) on their entire income. Someone earning ¥100 million pays an effective rate of just 1.4% for the same two items - because everything above the cap goes untaxed. Add the flat 20% investment tax, and you get two layers of hidden subsidy for the ultra-wealthy: the insurance cap + the 20% dividend window.
Katz's proposed fix: NISA 2.0 and the buyback question
Katz doesn't stop at description. He closes the series with two proposals. Proposal 1: expand NISA more aggressively - remove the current ¥3.6 million/year cap, simplify enrollment, and possibly allow asset classes beyond stocks. The goal: pull the middle class up to the 42% or 54% risk-asset share held by the top 10–1%, instead of staying stuck at 25%.
Proposal 2 - more interesting: reconsider the rationale behind the flat 20% tax window on investment income. The historical justification was "encouraging companies to raise new capital through stock issuance." But Katz points out that, in reality, Japanese corporations are generating far more cash than they need for investment, and now use that cash to buy back shares rather than issue new ones. If companies don't need new capital, why still give investors a tax break as if encouraging capital-raising? The 20% investment-tax window, in Katz's view, has become an irrational subsidy for the very top - and it is precisely the ¥100M wall.
"No other rich country lets the ultra-wealthy pay a lower effective tax rate than skilled workers trying to climb up. Japan doesn't do this because of 'encouraging investment' - Japan does this because nobody fixed it for three decades."
Act 5 · A Mirror - And an Open Question
Katz's essays are about Japan. But when you hold his checklist up against an emerging Southeast Asian economy, a few columns start to look familiar.
Vietnam in 2026 stands at a crossroads remarkably similar to Japan's around 1985–1990: assets heavy in real estate, high-skill wages low by international standards, generous investment-tax breaks, but no major pull yet drawing the middle class into the stock market. A real estate bubble may or may not arrive. But the question Katz poses for Japan applies here too: when high-skill wages can't escape the 1.5–2× median ceiling, can a country produce enough millionaires to sustain a broad middle class, or does it only feed a very thin group - mostly business founders and landowners?
This isn't a forecast. It's a structural comparison. Vietnam has a 10–15 year window to decide whether it wants to repeat a Vietnamese-language version of "Missing Millionaires."
- Katz, Richard. "The Myth of Japan's Missing Millionaires? Part 1" - Japan Economy Watch, 01/09/2025.
- Katz, Richard. "How Did Japan Fall Behind in Creating Personal Wealth?" - Japan Economy Watch, 15/09/2025.
- Katz, Richard. "Low Pay For the Highly Skilled Starves Japan of Human Capital" - Japan Economy Watch, 22/09/2025.
- Katz, Richard. "Are Hikes in Compensation Eaten Up by Taxes? (Missing Millionaires Part 4)" - Japan Economy Watch, 29/09/2025.
- Katz, Richard. "Korea Has Surpassed Japan in Per Capita GDP" - Japan Economy Watch, 29/01/2022. (supplementary piece, referenced in Act 3)
03 Discussion
Leave a note
A considered space for questions, counterpoints, and useful additions. Civil, on-topic, signed.
Reader notes
...Loading notes...