A recent item on the BBC web site briefly looked at the question of whether start ups could be the answer to Spain’s economic problems. Given that the example they chose was yet another cup cake seller, you could be excused for dismissing this as hyperbolic nonsense. It isn’t though.
Out of curiosity I created a very simple spread sheet model that looked at what might be the position five years after the creation of a notional cohort of new businesses. For simplicity I assumed each of these would be a single person at the outset, but it would not be too difficult to refine things to allow for a range of business sizes from the outset. I found some data which suggested that, in the UK at least, 45% of startups were still trading after five years. That is actually better than I expected, but I took that, and constructed a year on year table of survival rates that led up to that figure. New starts also tend to grow more rapidly so I also created another table of year on year growth rates. I’m looking for some data to make this a bit more robust.
From these two simple tables I set up a spread sheet to see what the situation would be after five years. The proponents of big business, politicians and the apologists for the big state will all use the relatively high failure rates of startups as an argument for putting large amounts of state money into supporting the work of large firms. What I already knew , and which seems to be confirmed by the figures from my simple model is that it isn’t the failure rate that counts, although if that can be reduced it is a bonus. What makes the difference is that because of the higher growth of new starts it doesn’t take long before the employment in surviving businesses more than makes up for the failures. In fact even on quite pessimistic assumptions about growth, the increased employment in surviving businesses almost replaces the failures, while on moderately pessimistic ones after five years there has been no reduction. On the sort of growth historically seen, the employment generated can be quite startling.
If we add to this an assumption of fresh cohorts of new startups in each subsequent year, a not unreasonable assumption even for a small town, then the numbers begin rapidly to accumulate. This isn’t just a model though. The small town of Fairfield in Iowa, popln 9000, managed to create 2000 jobs in 15 years.
The problem is that we tend to look on a new start that doesn’t succeed as a failure, whereas it isn’t, it is an experiment and sometimes experiments give us different answers to the ones we expect. We need to look for ways to avoid the stigma of failure attached to unsuccessful business ventures. Government intervention in support of large scale projects hasn’t exactly got an unblemished reputation either and the money lost in fiascos like De Lorean, or in shoring up the banks vastly outstrips what is lost in small business failures.
If anyone is interested I can send you a copy of the spread sheet to try for yourself. I would love to do an interactive version but don’t have the skills.
Previous posts on related topics:
http://www.withoutthestate.com/panchromatica/2004/07/africa_is_not_y.html
http://www.withoutthestate.com/panchromatica/2004/06/economic_garden.html
EDIT: This paper from 2004 published by Warwick University looks at the issue of firm growth from a number of perspectives.
Start-up size and subsequent growth: English and Spanish new businesses compared