Sections

Commentary

The Metropolitan Nature of IPOs

Digital billboard announces Facebook's initial public offering in New York.

Facebook’s IPO (Initial Public Offering) is projected to value the company at $104 billion. Reportedly, only Visa has had a larger IPO. Only time will tell if Facebook is really worth such an astronomical sum, but one thing about it is not all extraordinary: Its location in the Bay Area. From 1996 to 2006, 9 percent of all U.S. IPOs were headquartered in the San Francisco metropolitan areas–where Facebook is located–and another 10 percent came from the San Jose metro area.

The data come from University of California-Davis professor Martin Kenney and his colleague Don Patton. They have put together an impressive database on U.S. IPOs for all new independent companies from 1996 to 2006, using SEC records (Firm Database of Initial Public Offerings from June 1996 through 2006) and recently shared it with me. It has over 2100 IPOs for companies with U.S. headquarters.

From an urban economics standpoint, the most striking thing about the data is that IPOs are almost exclusively a phenomenon of large metropolitan areas. Between 1996 and 2006, 92 percent of all IPOs were from companies headquartered in one of the 100 largest metropolitan areas (which are home to roughly 65 percent of the U.S. population). One can also look at how valuations vary by location by multiplying the initial share price by the number of shares outstanding. Adjusting for inflation, 93 percent of all IPO market value over that time period was headquartered in the 100 largest metro areas. Less than two percent of the number or value of IPOs took place outside metro areas.

While 159 metro areas had at least one IPO over the time period, they are extremely concentrated in large metros. From 1996 to 2006, the top five metro areas for inflation-adjusted IPO valuations accounted for 45 percent of total valuation, and the top 10 were home to 62 percent. These ten metros are San Jose with $134 billion, New York ($107 billion), San Francisco ($102 billion), Atlanta (largely because of UPS), Boston, Los Angeles, Washington D.C., Dallas, Chicago, and San Diego.

The metro nature of IPOs is consistent with the trends in metro patenting I documented in a recent blog finding that 93 percent of the world’s patent applications are filed by inventors living in metro areas. Indeed, the number of patents applied for by metro inventors from 1996-2000 is very highly correlated with the number and value of IPOs that took place from 2000 to 2006. This is true controlling for metro population size and educational characteristics.

Interestingly, though the enormous attention given to Facebook, Twitter, and Google suggest to some that the United States is headed towards a post-manufacturing economy, companies going public are disproportionately in the manufacturing sector. Indeed, nearly 30 percent of companies going public from 1996 to 2006 were in the manufacturing sector, though the sector accounts for less than 10 percent of all U.S. jobs. While Business Services was the largest source of IPOs, three of the top five industries were in the manufacturing sector and four of the top 10, including Electronics, Chemicals, Instruments, and Industrial Machinery and Equipment. Together these four industries provided 490 IPOs over the 11 year period. As my colleague Howard Wial and I recently noted, the manufacturing sector accounts for a disproportionate share of innovation on a variety of other metrics too.

From 1996 to 2006, IPOs raised an inflation-adjusted $100 billion per year. Public companies contribute massively to the economy and to tax revenues, and they do so almost exclusively in metros. Super-performers like San Jose, San Francisco, San Diego, and Boston are special because they combine a highly-skilled labor force, stellar research institutions, and agile capital markets. Mark Zuckerberg and his creation are undoubtedly exceptional, but the fact that his innovative company was conceived in Boston and developed in San Jose and San Francisco is decidedly not.