Saturday, February 03, 2018

Networks or supply chains for ideas?


Ferguson is a prominent historian who is readable.  Here he dramatizes well-known historical episodes (who allied to go to war, to build empires, large companies, etc.) by noting the networks and networking that were involved.

We all know that networks and networking are important.  We also know that all of us are keen to find sources of useful information (Mokyr). This is why I prefer to note supply chains for ideas to networking. Ferguson cites the spread of ideologies.  Ideas can be thought of as “in the air” and as they rain down on us. But purposeful action in seeking ideas is more interesting and more descriptive. Supply chains are everywhere. In fact supply chains for ideas accompany many supply chains for things. They can be intertwined. We often learn when we transact. 

In recent work, John Cho and I have looked at pairwise co-locations of industries in the greater Los Angeles area. Using census block groups we estimated 2,991 co-location coefficients.  For all of the industry pairs we also know sales and purchase coefficients from input-output tables.  Using the latter as explanatory variables in a regression, we see that they explain just 3% of observed co-location.  What explains the rest?  There is surely noise in the data but we surmise that much of the rest must be due to the draw of information exchange.

The textbooks teach that information is a “public” good and unlikely to be traded. But only some of this is true.  Because we are keen to find useful information and because so much useful information is tacit, requiring interacting, we choose locations that help us with access specific information. The strong and the weak links are in play.

We network for many reasons to secure goods and to secure ideas. We do all this over many media, electronic as well as face-to-face. Choosing the best location for us to get all of this done becomes important and tricky. It also suggests that “agglomeration” can be many things, near as well as far. Fitting our data to Ripley-k functions shows that non-chance odds of encountering a same-industry firm, increase beyond 5km (the side of a large but square downtown). Near and far.

We agglomerate not just in tight clusters but over many geographic ranges.  New York is a financial hub but one that extends beyond Wall Street. L.A. is an entertainment hub but one that extends well beyond Hollywood (and even the San Fernando Valley). San Francisco is a tech hub that extends far beyond Silicon Valley.  High rents in all of these places suggest supply and demand forces.  Restrictions on supply have been widely noted. Strong demand is what our story is about.

All of  this illustrates once again that spatial arrangements and networks (including the paths we wear over lawns that were never laid out for us) emerge.  Jane Jacobs famously noted all this many years ago. “Their intricate order – a manifestation of the freedom of countless numbers of people to make and carry out countless plans – is in many ways a wonder” (Jacobs, 1961)