Born in a downturn: why firms started in a recession stay small
A recession is counted in the year it happens. The firms founded during it carry it for decades, and the reason lies mostly with their customers, not their bank.
The record kept on this site measures a downturn in the usual units: quarters of falling output, jobs lost, points off the policy rate. Each of those comes back. One cost of a recession does not show in any of them, because it is paid by firms that did not exist before the recession started and is collected over the twenty years after it ends.
The cohorts
The clearest evidence comes from following American firms by the year they were born. The Census Bureau’s Business Dynamics Statistics are built from administrative records covering about 98 per cent of private employment, and they follow every firm from the year it was founded. Petr Sedláček and Vincent Sterk did that for every cohort from 1979 onward, in a paper published in the American Economic Review in October 2017.
Three findings came out of it. Employment at startups rises and falls with the cycle. The swing persists: a cohort that starts small is still small years later. And the difference between good cohorts and bad ones lies mainly in how big the firms are, not in how many of them there are. A recession does cut the number of new firms, but the larger loss is that the ones founded anyway are built smaller and never grow into the size a boom cohort reaches.
Sara Moreira, working from the same Census microdata, found the same shape and looked for the point where it fades. It does not. In her account of the evidence, businesses born in downturns start on a smaller scale and remain smaller over their entire lifecycle, with no sign that the gap closes long after entry.
Demand or credit
The obvious suspect is the bank. Lending standards tighten in every recession. If credit were the cause, though, the damage ought to fade as credit loosens, and in these data it does not.
Both papers point elsewhere. In Sedláček and Sterk’s estimated model, the force that matters most is a demand shock that raises the cost of building a customer base. When customers are scarce, the firms that start are disproportionately the ones aimed at a small, particular market, and a firm built for a niche stays a niche firm when the recovery comes. Moreira attributes the persistence to selection at entry and to demand: a firm grows by accumulating customers, and a firm that opens into a year when nobody is buying starts that accumulation from further back.
That makes a recession’s damage to new firms a national matter, decided by who is buying where the firm opens. Poland is the reminder. It was the only economy in the European Union whose output grew in 2009, so a firm founded in Poznań that year opened into a market that was still spending, while one founded in Madrid or Dortmund opened into a contraction. The banks in all three cities had read the same news from New York. Their customers had not lived through the same year.
The part credit does play
None of this makes the lender irrelevant. What a lender imposes on a young firm is simply the same in every phase of the cycle, and it has nothing to do with rates.
A bank’s usual answer to a young firm is to ask for the track record it does not have yet. Some online services try to read something else instead. Wandoo Biznes, which arranges loans for Polish one-person firms, says it does not require years of trading, and that when a firm is new it looks more at the owner’s own financial history and the turnover on the bank account. None of that makes a firm born in a slump any larger. It shortens one wait, the wait for working capital, which a track-record rule imposes on every young firm whatever the cycle.
A rule built on history taxes every young firm, in good years and bad, because the history is exactly what a young firm lacks, and loosening it helps a firm founded in any year get started. It does nothing about the reason a 2009 firm stayed smaller than a 2006 one. Faster credit cannot sell anything. A firm that borrows quickly into a market with no customers has working capital and the same empty order book, now with a repayment date attached.
The cohorts of 2008 and 2009 are in their late teens now. On the evidence above they will reach middle age smaller than the cohorts on either side of them, and the gap will not show up in any quarter’s GDP figure. It shows up only when someone counts by birth year.
Counting that way also shows what a recovery cannot repair. Rates can be cut and lending standards eased, and the firms that open afterwards will benefit from both. The firms that opened during the slump keep the size the slump gave them.