Today in my antitrust and regulation class we talked about natural monopoly theory and what drives the natural monopoly cost structure. A lot of times in practical conversation with regulators and industry we talk about economies of scale, the decrease in average cost of production as the quantity produced increases, as being the main factor behind the natural monopoly structure and the associated theory on which utility regulation is based.
But that’s wrong, or at least mistaken when talking about regulating industries like electricity. Economies of scale only create that natural monopoly cost structure when the firm produces a single, well-defined product (and is facing a static demand curve that doesn’t increase into the portion of the average cost curve that has diseconomies of scale). Many firms in the economy, and indeed many firms in industries that have traditionally been heavily regulated (e.g., electricity, telecom, railroads) produce multiple products and find that to be a profit-maximizing and cost-minimizing strategy.
Why can that be profit maximizing? Consider the early history of the (competitive) electricity industry. Firms entered initially to provide electric lighting to residential and commercial customers (homes, hotels, saloons, etc.). They built generation and distribution capacity to provide that good in that market. But that meant they had generation and distribution capacity available during the day when lighting wasn’t needed.
With Tesla’s invention of the electric induction motor in 1888 and subsequent commercialization R&D while he was at Westinghouse, the induction motor was ready for business by 1900. At that time 5 percent of mechanical power in US factories was powered using electric induction motors, and its penetration increased to 50 percent by 1920. Induction motors in factories created a new market for electricity, and one that was dominantly a daytime market, which meant that electric companies could use existing generation and distribution capacity to serve an additional market. At that point electric companies become multi-product firms, and they did so because of economies of scope.
Economies of scope arise when a firm can lower average costs by producing more than one type of good, rather than the economies of scale that arise from lowering average costs by producing higher quantities of the same good. This Economist brief provides a short primer on the distinction between them.
My students offered a great example of economies of scope in a non-electricity setting: automobile manufacturing. The same factory and machinery, and design team, and finance and accounting and HR, can be used to produce different goods for different markets.
Economies of scope surround us, but get overlooked a lot of the time. They can drive a lot of merger activity (and in fact were the main driver of a lot of conglomerate mergers in the 1970s, which was the last time economies of scope got much attention). One underexplored question is how extensive economies of scope are in a digital economy — there are manufacturing examples, of course, such as Apple reaping economies of scope across laptops, iPods, and iPhones. But is there a meaningful sense in which digital technologies change economies of scope in other ways? I’m thinking about that, and interested in hearing suggestions.
Uber could be one. They are now starting to branch out into the food delivery service. During times of low transportation demand (the working hours) an excess supply of Uber drivers can now deliver food, or conceivably other tasks (do my laundry too?). It could be that businesses in the sharing economy will increasingly look to expand their service offerings while maintaining the same labor force.
How about mobile technology in rural Africa, used by farmers? More communication = better knowledge about crop options and fertilisers etc = ability to use land, labour in a diversified way or use downtime more efficiently = more fertile land and more crop diversification, so better protected against shocks related to a particular crop (for instance a fluctuating world price).
Great post, Lynne. Two small comments. First, scope economies are probably not that relevant for 1970s-style conglomerates, because the product range is so great; if they created value it was by creating some kind of financial synergies, i.e., exploiting advantages of internal capital markets. (Amar Bhide has a great piece on this in J Applied Corp Fin, “Reversing Corporate Diversification.”) Second, for scope economies to drive firm boundaries (rather than production boundaries), we need to invoke transaction costs, right? As David Teece points out in his classic 1980 paper (http://www.sciencedirect.com/science/article/pii/0167268180900025), absent transaction costs, two or more firms could collaborate (e.g, by setting up a joint venture) to exploit the scope economies without merging into a single firm. Of course, we typically do see scope economies exploited within firms because the transaction costs of renting out excess capacity are usually prohibitive. But not always — e.g., stores in a shopping mall are in some sense exploiting scope economies by sharing maintenance, security, utilities, parking, etc. even though they are separate firms.
I was just thinking about some of this stuff today. Synchronicity! Aside from the transactions costs point Peter mentioned, I think greater precision about scope and scale economies is desirable.
First, it is a theorem that without scale economies there can’t be any scope economies–check Panzar’s review article in the Handbook of I.O. Second, the way the concept is described above somewhat confounds capacity utilization, multiproduct scale economies, and scope economies. Just spreading fixed costs out over more products isn’t really scope economies–you have to first specify the product vector being produced (with more variety generally increasing average costs relative to making more of the same thing), and then compare the joint production to a set of optimally scaled specialists. Finally, there is a problem with defining what is the “same” product: If I put a serial number on each otherwise-identical ball-bearing coming off the line and define different serial numbers as different products, voila, I’ve turned scale or volume economies into scope economies. It’s a lot like the market-definition problem that come up in antitrust cases, with similar definitional solutions and defects to those solutions, e.g. the cross-price-elasticity definition runs into the problem with left-handed and right-handed golf clubs and the technological difference definition runs into heat-pumps versus gas heaters.