The House Always Wins

In a casino the bank earns on every game, whoever wins. The stock exchange is no different. The only question is who the bank is.

I. The pattern of 2002

The pension reform of 2001 lowered the level of Germany’s statutory pension. The gap was to be closed by private, state-subsidised provision, with allowances, tax advantages and the promise that contributions paid in would be preserved. The name of the labour minister of the day became a brand: Riester.

By the end of 2023 more than 20 million Riester contracts had been concluded. 4.6 million of them, almost a quarter, no longer exist. In March 2026 the consumer organisation Finanzwende analysed the model calculations of 18 Riester tariffs. On average up to 35 per cent of the money paid in goes to the provider as costs, with the most expensive up to 46 per cent. Two thirds of the tariffs do not even reach a return of two per cent over the term. Anyone who wants their contributions back including compensation for inflation must on average live to about 99, depending on the tariff between 89 and 106. The insurers themselves assume a life expectancy of 96 for men and 100 for women. The typical lifelong pension is around 139 euros a month. The insurance industry disputes this calculation. But nobody disputes that a quarter of savers have left.

What was sold was a bridge across the pension gap. What people got was a product on which the provider earned with every contract, even every cancelled one, while the saver bore the risk of not living long enough.

II. The new offensive

Now comes the successor, this time with shares. On 27 March 2026 the Bundestag passed the retirement savings depot, meant to replace Riester from 2027. Shares, funds and exchange-traded index funds are subsidised; the effective costs of the standard product are capped at one per cent a year. On 16 October 2026 the “early-start pension” followed: for every child from the age of six to eighteen the state pays ten euros a month, invested exclusively in broadly diversified global shares or equity funds, as a rule not payable before 65.

Others supply the mood. In 2025, 14.1 million Germans owned shares or equity funds, more than ever. The neobroker Trade Republic has more than ten million customers. In “The Fool and the Donkey” we described how in 1872 anyone who left his money in a savings account was considered a backward fool, and how in 2026 anyone who waits for his pension is called a donkey. It is not the risk that is advertised; it is caution that is shamed.

To understand who earns from this offensive, one must first understand which game is actually being played.

III. Three games in one share

Imagine you own a share in a bakery. The bakery bakes bread, sells it at a profit, pays out part of it and puts the rest into a second oven. That is value creation. It arises from work, flour and customers, and nobody loses by it. That is the first game, and it is a good one. According to the Global Investment Returns Yearbook, equities worldwide returned on average 5.2 per cent a year in real terms from 1900 to 2024, bonds 1.7 per cent.

Now you sell your share to your neighbour at a higher price than you paid yourself. What you gain, your neighbour has paid. If the price keeps rising, the neighbour was right; if it falls, you were. Either way one gains what the other loses. That is the second game, a zero-sum game. Daily trading on the stock exchange consists almost entirely of it. We described this in “The Effects of Zero-Sum Games”.

And with every change of hands someone takes a fee: the broker, the notary, the manager. This turns the zero sum into a negative sum. Buyer and seller together have less than before. That is the third game.

In 1991 the economist William Sharpe showed what necessarily follows. Those who trade actively in order to beat the market achieve before costs, on average, exactly what the market achieves, because active investors together are part of the market. After costs they fall short. That is not theory; it is arithmetic.

IV. Where it becomes a negative sum

The figures confirm it. According to the SPIVA report by S&P Dow Jones Indices, over the ten years to the end of 2025 around 97 per cent of actively managed European equity funds and more than 98 per cent of global equity funds performed worse than their benchmark.

It becomes even clearer with the products that speed up trading. BaFin, the German financial regulator, examined the behaviour of around 543,000 private investors who traded turbo certificates between 2019 and 2023. 74 per cent made a loss, on average 6,358 euros, together more than 3.4 billion euros. Of those who traded up to ten times, 70 per cent lost; of those who traded more than a thousand times, 91 per cent. With contracts for difference, according to European supervisors, 74 to 89 per cent of retail accounts lose money. The faster the game, the more surely the house wins.

V. The bank in the casino

Who earns from this game without playing it? The guiding question is: who earns from the movement, and who from the result?

Brokers earn from the movement. Trade Republic reported commission income of 732.6 million euros for the financial year to September 2025 and increased its profit sevenfold. Until June 2026, German brokers were allowed to be paid for routing their customers’ orders to particular trading venues. Since the European ban on this practice, Trade Republic acts as the counterparty to its own customers in its standard execution. The house now sits at the table.

Fund providers and insurers earn from the stock of assets. Their ongoing fees are due whether the fund rises or falls. A simple calculation shows how much this matters: for anyone who leaves a sum invested for forty years, costs of one per cent a year take almost a third of the final amount. One per cent is the new cap for the retirement savings depot.

Index providers and index funds earn from the growth of what is already big. An index fund buys every share in proportion to its market value. The more expensive a share becomes, the more of it the fund buys, and the more expensive it becomes. In the MSCI World, the best-known global equity index, American shares account for around 70 per cent, Nvidia alone for more than five.

Issuers and existing shareholders earn at the initial public offering. They sell at the price that enthusiasm will bear to those who come last. In “The Forced Purchase” we described how in the largest IPO in history pension savers were made into buyers without being asked.

The companies themselves buy back their shares. The 500 largest American companies spent almost a trillion dollars on this in the twelve months to June 2025, and 654 billion on dividends. Buybacks support the share price, and the money flows to those who sell their shares in the process.

Finfluencers earn from attention. Since January 2026 BaFin has required them to disclose payments and their own holdings. That such a requirement is needed says enough.

And the state earns from relief. With every reform it shifts a piece of the risk it once bore itself onto the individual: the risk of bad markets, the risk of high costs, and with Riester also the risk of dying too early.

VI. Where the children’s money goes

The early-start pension must be invested in broadly diversified global shares. Global today means around 70 per cent United States and more than five per cent a single chipmaker. The money of German six-year-olds thus flows to a large extent into the valuations of large American corporations, at a time when the cyclically adjusted price-earnings ratio of the S&P 500 stands at around 40 according to YCharts, close to the highest values in its history. In “Is It ’29 Again?” we described that a boom does not end when prices are too high, but when the last buyer has bought.

A child who starts at six and draws at 65 has time to sit out a crash. That is the strength of the early-start pension, and it must be said fairly. But whoever enters a depot at fifty and needs the money at sixty does not have that time. Losses in a crisis do not hit those who can hold on, but those who have to sell.

VII. The thesis

A share is not a zero-sum game. The stake in the bakery brings bread and profit, and whoever buys a broadly diversified index cheaply and holds it for decades receives most of that value. But the business around the share is a zero-sum game or worse. It lives not from the bread but from turnover.

The offensive now meant to bring even the last saver to the stock exchange is carried by those who earn from movement rather than from results: brokers, sales networks, product providers, influencers. Their interest is not that the saver becomes rich, but that he comes, stays and trades. Riester showed how it ends when they are given free rein. The house always wins. The only question is whether the saver knows that he is not the house.

Further reading: The Fool and the Donkey · The Effects of Zero-Sum Games · Is It ’29 Again? · The Forced Purchase