Tune in, drop out

In the 1930s Frederick Terman, then a Stanford professor, helped two of his students, Bill Hewlett and Dave Packard, get started; he arranged a research post for the one and a fellowship for the other.^{7HPHP Garage timeline} On 1st January 1939 the pair formalised a partnership in a garage in Palo Alto with $538 of cash and kit.^{7HPHP Garage timeline} Note the order of events, though. Both men had graduated from Stanford first, in 1934.^{7HPHP Garage timeline}

Eighty years on, the question has flipped. The fashionable debate is no longer about where graduates should go but whether founders need to graduate at all. The Thiel Fellowship, launched in 2011, pays young people $100,000 over two years to work on their ideas instead of attending college; its 2018 class numbered 20.^{6BetaKitThree Canadian companies among 2018 Thiel Fellows} The pitch has a ready audience, and the reason became plainer on 13th August, when the Federal Reserve Bank of New York published its quarterly tally of household debt.^{1Federal Reserve Bank of New YorkHousehold Debt and Credit press release, Q2 201913 Aug 2019}

The other Big Short. America, student loan balances in the second quarter, $trn

	Student loans
2004	0.26
2005	0.37
2006	0.44
2007	0.51
2008	0.59
2009	0.68
2010	0.76
2011	0.85
2012	0.91
2013	0.99
2014	1.12
2015	1.19
2016	1.26
2017	1.34
2018	1.41
2019	1.48

Source: Federal Reserve Bank of New York Consumer Credit Panel/Equifax¹²

Americans owed $1.48trn in student loans at the end of June.^{1Federal Reserve Bank of New YorkHousehold Debt and Credit press release, Q2 201913 Aug 2019} That is a small dip on the quarter,^{1Federal Reserve Bank of New YorkHousehold Debt and Credit press release, Q2 201913 Aug 2019} but more than five times the $0.26trn owed in 2004^{2Federal Reserve Bank of New YorkHousehold Debt and Credit Report, Q2 2019, data workbook (Total Debt Balance and Its Composition)13 Aug 2019} (see chart). The share of student debt sliding into serious delinquency rose to an annualised 9.9%, from 9.4% in the first quarter, and the New York Fed warns that the true figure is probably worse, because about half of these loans are in deferment, grace periods or forbearance and so are not in repayment.^{1Federal Reserve Bank of New YorkHousehold Debt and Credit press release, Q2 201913 Aug 2019} A young person looking at those numbers, and at a cheque from Peter Thiel, can be forgiven for doing the sums.

The sums are less flattering than they look. Economists at the same New York Fed reckon that a bachelor’s degree still returns about 14% a year on average.^{3Liberty Street Economics, Federal Reserve Bank of New YorkDespite Rising Costs, College Is Still a Good Investment5 Jun 2019} That figure has slipped a little, mostly because of rising costs rather than falling wages,^{3Liberty Street Economics, Federal Reserve Bank of New YorkDespite Rising Costs, College Is Still a Good Investment5 Jun 2019} but it remains roughly double the long-run return on shares (7%) and well above bonds (3%).^{3Liberty Street Economics, Federal Reserve Bank of New YorkDespite Rising Costs, College Is Still a Good Investment5 Jun 2019} Few asset classes look that good, and the loan book that frightens parents is partly a measure of how many people bought in.

The bigger problem for the dropout thesis is age. The mythology runs on a handful of famous cases. A study^{5National Bureau of Economic ResearchAge and High-Growth Entrepreneurship (Working Paper 24489)} by Pierre Azoulay, Benjamin Jones, Daniel Kim and Javier Miranda, using Census Bureau records on 2.7m founders of firms that went on to hire, finds that the mean founder is 42.^{4NBER DigestWhen It Comes to Entrepreneurs, Youth Isn't Everything27 Jun 2018} For the fastest-growing one in 1,000 new firms, the mean age at founding is 45.^{4NBER DigestWhen It Comes to Entrepreneurs, Youth Isn't Everything27 Jun 2018} Even in the fields with the youngest founders, computing among them, the mean ranges from 38.5 to 40.8.^{4NBER DigestWhen It Comes to Entrepreneurs, Youth Isn't Everything27 Jun 2018} Founders with three or more years in their industry hit the top tail at roughly twice the rate of those with none.^{4NBER DigestWhen It Comes to Entrepreneurs, Youth Isn't Everything27 Jun 2018} The authors’ conclusion is pointed: to the extent that venture capital targets younger founders, early-stage finance appears biased against those most likely to succeed.^{4NBER DigestWhen It Comes to Entrepreneurs, Youth Isn't Everything27 Jun 2018}

Defenders of the fellowship have a fair reply. Averages describe the typical founder; venture investors live off the outliers, and a programme that picks 20 people a year needs only one big hit to justify itself. Some ideas, particularly in software, do have a window that closes before a degree is finished. A 19-year-old with traction and no debt risks two years and a deferred place. If it fails, he or she goes back to campus a little wiser. That is a cheap option, and options are worth more when the underlying asset is volatile.

We accept the logic, as far as it goes. The fellowship is a sensible bet for the tiny minority who already have a product and users. It is a poor guide for everyone else. The data say that what predicts startup success is experience,^{4NBER DigestWhen It Comes to Entrepreneurs, Youth Isn't Everything27 Jun 2018} and experience comes from time spent working inside an industry, which a degree tends to speed up. Debt is a real burden, and the answer to it is cheaper, shorter and more useful degrees. Telling 19-year-olds that the average winner is their age is simply wrong.

Terman’s insight was that talent should stay put and build. Hewlett and Packard spent $538 and had their diplomas, and nobody has yet shown that the diplomas slowed them down.