When Businesses Fail

So far, these simulations have made a quiet, generous assumption. Every business the community funds pays its expected return, on schedule, forever. No duds. No bad years. That is not how enterprise works. Some ventures fail outright. Many of the rest come in above or below plan. A model that only balances when nothing ever goes wrong hasn't been stress-tested. It has been coddled.

So we introduced failure, deliberately and honestly. Then we watched what the machine does when the money it invests sometimes simply doesn't come back.

Two dials. The failure rate is the chance any funded business is a total loss. Its owners took their turn, and it returned nothing. The return spread lets the survivors vary: most near their expected return, some better, some worse. At zero, both vanish and the model reproduces its earlier results exactly.

One rule we held to: failures genuinely lower the total return. We did not quietly inflate the winners to paper over the losers. If one business in five is a total loss, one-fifth of that capital really is gone. We wanted to see the model carry the loss, not hide it.

Below, the model runs at rising failure rates. Each bar is when the community reaches full independence. The whisker is the spread across runs — the band of possible finishes.

Two things happen as failure climbs, and one thing conspicuously doesn't:

• The finish moves out. A one-in-ten failure rate pushes independence from about eleven years to twelve. One-in-three pushes it to about fifteen. Lost capital is lost time.
• The band widens. With no failure, the finish is almost a point. With heavy failure, it becomes a range. Uncertainty about the return becomes uncertainty about the date.
• No one is excluded. This is the striking part. Across every run, not a single member is left behind. That holds even at failure rates as brutal as one business in two. The cost of failure is measured in time, never in people.

And pure variance, without outright failure, barely registers. Let survivors swing widely around their mean, and the graduation date hardly moves. Averaged across hundreds of businesses, the ups and downs cancel. Which points at why the community is so hard to derail.

The reason is structural, and it is worth seeing clearly. The capital the community invests each month comes from its members' own spending — the surplus on what they were going to buy anyway. It does not come from the returns of past businesses. So the investment engine never starves. A wave of failures cannot cut off the fuel supply. The fuel is the community's own everyday commerce, arriving again next month regardless.

So failures are simply replaced. A business dies; the pool funds another. And every business that does succeed pays a permanent income stream. Income here only ever accumulates, never reverses. Losses don't compound the way they do for a lone investor, who can be wiped out. They just slow the climb. Give the community enough time and the surviving streams always add up to cover everyone. Failure changes the schedule. It cannot change the destination.

If the risk is the width of the band, the useful question is: what narrows it? The answer the model gives is diversification. The community gets it for free, by being large.

Hold the failure rate fixed and change only the size of each business. Fund a handful of big ventures, and the finish is a wide, nervous range. A few bad losses swing the whole outcome. Fund many small ones instead, for the same total capital, and the band tightens sharply. At a one-in-three failure rate, spreading the money across four times as many businesses cuts the spread of outcomes by roughly a factor of three. No single failure matters much when it is one of hundreds.

This is ordinary portfolio logic — the law of large numbers. But notice where it lives. A lone family cannot diversify across hundreds of businesses. A collective of five hundred does it automatically, simply by pooling. The community's scale isn't just what makes it fast. It is what makes it safe. Earlier chapters showed that reinvestment and role-mixing buy speed. They turn out to buy certainty too. They are the model's built-in insurance.

A design that only works when every bet pays off is a fantasy. Any real economy fails sometimes. The interesting question is what failure costs, and who pays it. Here the cost is time, and the whole community shares it. No one is dropped for being unlucky enough to have backed the venture that didn't make it.

That is a different relationship with risk. It is not the promise that nothing will go wrong. It is the design goal that when things do, the group absorbs it together — and in every run of the model, everyone still arrives.