Cross-subsidy: choosing which part of the business pays for the rest

The burger company that became a landlord

In 1956 McDonald's was a young franchise business that earned its money from franchisees. That year it set up a separate company, Franchise Realty Corporation, to buy cheap land along well-traveled roads.

The idea was simple. Headquarters would lease the land and the building, then sublease the package to the franchisee, who paid as much as a 40% markup on the root lease. The franchisee got a restaurant. McDonald's got a tenant.

Decades later the side business had become the main one. In 2009 McDonald's collected $4,841.0 million in rent from its franchised restaurants, and $2,379.8 million in royalties.

Rent has run roughly two to one over royalties for decades. The brand is what customers see. The real estate is what pays.

That is the whole decision in one story: inside one company, one business makes the money and another is carried by it. McDonald's chose which was which on purpose, before its future restaurants were built, and built the pipe between them. This guide is about making that choice on purpose.

What cross-subsidy is

This guide calls the part that makes the money the payer, and the part it carries the passenger. Sometimes the passenger is a bet on the future, like a new console or a streaming service. Sometimes it is the face of the company, like a newspaper or a brand, and the payer is something customers never think about: land, insurance float, cloud servers, bag fees.

Why it is so hard

Every company with more than one business has some money moving between them. Three things make doing it on purpose hard.

The ways companies did it

The decisions on record take four shapes. They differ in where the payer comes from, and in how visible the subsidy is.

What 21 companies did

Here is every decision on record, placed by the year it was made and marked by how it turned out. Hover a mark, or open the table, for the company and the outcome.

Of the 19 decisions that can be judged, 16 were never undone, 2 were walked back and 1 was kept only in part.

"Never undone" does not mean the subsidy lasted forever. The Washington Post Company's decision to sell its paper is on record as held. It held because the subsidy had already ended.

Nor does it mean the subsidy worked. CVS kept Aetna, and the insurance arm's adjusted operating income fell to $307 million in 2024. Samsung kept building memory chips through the downturn, and SK Hynix passed it in DRAM revenue share, 36% to 34%, ending a lead Samsung had held since 1992. Both decisions stood; neither payer paid as planned.

The walk-backs started on the side that was supposed to make the money. Sears bought its payer on a forecast. Philip Morris spent cigarette cash on Kraft to get away from a declining, increasingly regulated cigarette market, and Altria handed the food business back to its shareholders in 2007.

The subsidies that held had something in common too: the money moved through a mechanism that worked whether or not anyone's forecast came true. McDonald's collects rent. Ryanair collects the bag fee. Berkshire invests the float. None of these depends on customers deciding to buy something new.

Sears is the clearest contrast. Its plan depended on 36 million customer families choosing to buy financial products where they bought everything else. That was a hope, not a pipe. A subsidy you build is a contract. A subsidy you buy is a forecast.

How to run this decision

If you are weighing a version of this choice, here is what the record suggests doing, in order.

1. Name the payer and the passenger in writing, with this year's numbers

   Write the sentence out: this unit earned this much, that unit lost this much, and the difference moved between them. If nobody can write it, nobody has decided it, and the subsidy is running on habit.

   Seen in: Alphabet (Google), 2015
2. Make the transfer a mechanism, not an expectation

   A mechanism produces the money whether or not the strategy works: a lease, a fee, an internal price, a float. An expectation that customers will start buying something new is a forecast, and forecasts are what got walked back.

   Seen in: McDonald's, 1956 · Ryanair, 2006 · Sears, 1981
3. Size the passenger's loss before you launch it, and say when it ends

   Microsoft knew the Xbox would sell at a loss; Disney set its price knowing it was below cost. A loss with a size and an end date is a budget line that can be defended in a bad quarter. A loss discovered afterwards becomes an argument every quarter.

   Seen in: Microsoft, 2001 · Disney, 2019
4. Be suspicious of buying the payer

   Buying a business to make the money for the rest means paying today for profits that are only a forecast. If the plan depends on customers or synergies that do not exist yet, build the payer out of something you already do well enough to sell, as Amazon did with its data centers.

   Seen in: Sears, 1981 · CVS Health, 2018 · Amazon, 2003
5. Plan for the payer's worst year, not its best

   If the passenger's cost is fixed and the payer's profits rise and fall, the subsidy will break in some year you can predict. Work out which year that is, and decide now what you will do when it comes: cut the passenger, sell it, or borrow.

   Seen in: Samsung, 2022 · The Washington Post Company, 2013
6. Check who is paying, and whether they know

   The paying unit's managers and customers carry a cost they did not choose. Make the internal price and the transfer visible to them, and ask once a year whether the money is building something or covering for something that is slipping.

   Seen in: Samsung, 2026 · UnitedHealth Group, 2011

Warning signs

What it has looked like when this decision was going wrong.

• The subsidy shows up in a segment report rather than in a decision anyone remembers making.Alphabet (Google), 2015
• The case for an acquisition says synergy and names no fee, lease or internal price.Sears, 1981
• The passenger's loss has no size and no end date.Alphabet's Other Bets
• The payer's best year is being used to budget the passenger's next five.Samsung, 2022

All 21 cases

Each case is one company at one decision: what was on the table, what it chose, why, and what followed, with every fact sourced.

Questions

What is a cross-subsidy in business?

When one part of a company makes the money that keeps another part going. McDonald's earns more from renting land and buildings to its franchisees than from royalties on its brand, and rent has run roughly two to one over royalties for decades.

Is cross-subsidization good or bad?

Neither by itself. It lets a company fund a new business, or keep prices low where customers can see them. It goes wrong when nobody has decided it on purpose, when the payer is a forecast rather than a business, or when the payer has a bad year.

What are examples of cross-subsidies?

McDonald's real estate carrying its franchise system, Berkshire Hathaway's insurance float funding its investing, Amazon's cloud business beside its retail, Microsoft's Windows and Office carrying the first Xbox, and Disney's cable networks covering early losses at Disney+.

Why do cross-subsidies fail?

Usually because of the payer, not the passenger. On record, the walk-backs came after a company bought a payer whose profits were only a forecast, or after the payer came under pressure of its own.