Why Large Corporations?

In a comment on an earlier post, Ajay asks,

what is the point of a large corporation in the first place?

Some possibilities:

1. Governments want them. Surely, from a “seeing like a state” perspective it is better to have large corporations that are dependent on favors than small firms that are not.

2. There are genuine economies of scale and scope, including network effects.

3. Workers believe that they are more secure working for large corporations, and they are willing to take less compensation as a result. Note that this sort of belief could be self-fulfilling. Note also that it is not terribly consistent with the data: compensation appears to be higher at large firms, although that comparison assumes that the investigator’s idea of objective value of workers is more meaningful than their actual choices.

Think about Google. It needs to retrieve, store, and process huge amounts of data. There are scale economies. Once you have that data, you can benefit from other data, so you want to expand into email, location services, social networking, phones, and anything else that generates data. So there are economies of scope as well.

Maybe that is an exceptional case.

My tendency is to think that economies of scale are fairly common, but economies of scope are relatively rare. I understand big companies that specialize in a relatively narrow capability–something like Fedex, for example. I am less convinced about organizations that branch into many functions, like universities or large financial firms.

When I re-read what I wrote on this topic fifteen years ago, I see that my views have not moved very much.

This entry was posted in business economics, links to my essays. Bookmark the permalink.

13 Responses to Why Large Corporations?

  1. Robert Easton says:

    One explanation that seems appropriate in the tech sector is that it is just inefficient reinvestment of capital. Most of the biggest companies in the tech sector (such as Google, Microsoft, my employer Cisco, and others) seem to have been extremely successful in one area very quickly (web search, operating system, network equipment) and then instead of returning the profits to original investors, they find more and more areas to diversify in. Some of these attempts clearly lose money. Some probably make just enough to carry on, but perhaps would not have earned the trust of investors on their own. Google have diversified into all sorts of web services that don’t make money on their own. Microsoft and Cisco have both had plenty of “flops”.

    I don’t know how well this story works outside the tech sector with say Mcdonalds or Walmart. But it may be similar. Perhaps their first few franchises were profitable enough that they had plenty of capital that they can then use for what is perhaps not really the best interest of the shareholders. But it’s just good enough to sustain itself.

  2. Daublin says:

    Reading your old essay, I’m surprised you would write then–much less believe nowadays–that proprietary software is going away in favor of reusable components.

    Perhaps you mean that Windows is giving way to Linux, or that IE is giving way to Chrome. That’s just the lowest layer of the systems, however. My impression is that companies today are writing all manner of internal software.

    Basically, part of corporate DNA nowadays is embodied in code. Procedures that previously would have involved mechanical paper pushing, nowadays involve mechanical bit pushing.

    • Arnold Kling says:

      but back then, every company wrote its own database software. every company wrote its own communications software. The layers of proprietary software have gotten much, much thinner

  3. Joss Delage says:

    Could be lower cost of financing too.

  4. BillD says:

    Any thoughts on subscription tv and/or broadband service? There are obviously economies of scale for the providers. But, how does that impact consumers?

    I don’t know of anyone who believes they get a good deal or believe their provider provides good service. (Providers include cable, satellite, telco).

    From international benchmarks, US seems to have high prices/low speeds for broadband internet access.

  5. Jonathan Bechtel says:

    I think it’s useful to distinguish between how they get big and how they stay big.

    Getting big is probably due to an intersection of black swans and recursion. They fill a hole in the right place at the right time and self-reinforcing equilibria perpetuate until it grows to its rate-limiting size.

    Staying big is a combination of that same recursion allowing them to use economies of scale to create defensive moats and engage in political rent-seeking.

    I generally agree that big corporations and big government are complimentary and usually need each other to survive.

  6. Alfred Duncan says:

    Some transactions require a lot of trust. Do you think that size, and with it the ability to grow a reputation, can act as a commitment technology and therefore relax moral hazard problems?

    This would help explain the existence of large diversified firms with seemingly few economies of scope. Here in the UK the “new banks” include Tesco Bank, M&S Bank, Virgin Bank. Subsidiaries of firms with a lot of reputational capital.

    And large services firms with seemingly few economies of scale (big four audit firms / credit rating agencies). These firms are worldwide. A firm in New Zealand looking for an auditor knows that the global firms “stand behind” their local branches, and may therefore choose a big four auditor over a local audit firm.

  7. Worth noting that big businesses pay their employees better than small businesses:


    This is why Matt Yglesias argues that regulatory barriers to business growth and consolidation lead to depressed wages in Southern Europe.

  8. jseliger says:

    I basically like the Peter Thiel answer:

    Companies exist because they optimally address internal and external coordination costs. In general, as an entity grows, so do its internal coordination costs. But its external coordination costs fall. Totalitarian government is entity writ large; external coordination is easy, since those costs are zero. But internal coordination, as Hayek and the Austrians showed, is hard and costly; central planning doesn’t work.

    The flipside is that internal coordination costs for independent contractors are zero, but external coordination costs (uniquely contracting with absolutely everybody one deals with) are very high, possibly paralyzingly so. Optimality—firm size—is a matter of finding the right combination.

    Corporations also solve or ameliorate succession and other problems.

    You could also say that corporations help align the interests of a lot of people in approximately the same direction. This mechanism obviously isn’t perfect, but it’s better than alternatives.

  9. Bryan Willman says:

    I have directly observed some reasons that at least software companies get big.

    1. The process of writing software is so slow and inefficient that large numbers of people are required. Intellectual property concerns, risk, and so forth made it quite hard to organize this except with something like “The Firm”. Can one imagine a truly small maker of airplanes? Cars? Oil tankers? And whether employees or contractors or subcontractors, they all involve large numbers of people.

    2. There’s a lot to the size and reputation issues. One problem that “open source” faces was summed up by the remark in the press (20 years ago) “My board of directors will never allow us to buy critical software from any entity we couldn’t sue and get real money from.” Small firms and individuals are often “judgement proof” (broke.) Open source is the ultimate judgement proof vendor.

    3. There are cultural factors a bit different from #2 – for example, to be taken seriously for certain transactions, such as going public, you basically need a “big name” accounting firm. The fact that there are fewer big names now because some of them were destroyed by their misbehavoir is irrelevent. It’s just part of the rules of signaling. If you bought a critical system from IBM/Intel/Microsoft and added your own app, you look serious. If you built your system from things you found for free on the web, you might do as well, but you look low-rent part-time fly-by-night. This is a huge deal to investors, management, etc. (And has of course changed over time. Your application must have “cloud” in the name now, no matter what it does.)

    4. Returns to scale with respect to management of regulation and mundane operations. I saw this big time when going from writing patents for Microsoft to writing patents for new unrelated ventures with only 1 partner. When you work in the big org, the infrastructure that manages patent filings, office cleaning, unemployment and worker’s comp, etc. etc. is all hidden away by groups of people who just deal with that. When it’s a small crew in a startup, suddenly precious time gets sucked into filing the state sales tax report and sending out 1099s.

  10. Ajay says:

    Finally finished reading your dense article from 1998, so thought I’d comment. To be fair to me, I was only posing this as a rhetorical question, one I could easily knock down with the claim that all these large orgs would fail, without much justification. ;) It was funny to read your analysis of in-house software vs outsourced component libraries, something you don’t see talked about as much anymore, even though there is still far too much in-house software being written to this day. I’ve used pure internet banks for almost a decade now, though I’m still in the minority.

    You were not the only one who was skeptical “that ‘one-stop-shopping’ is the holy grail for financial services firms,” so was John Gutfreund, when he was asked to opine on the Charlie Rose show about the 1998 Citigroup merger. As he said then, “because of the size of the new Citicorp they’ll be bailed out again when the cycle goes the other way, too big to fail will be the thesis as it was last time.” That was 10 years before it happened.

    It is interesting that you based your claim of increasing economies of scale on logistics and communication, not the standard manufacturing argument. You presuppose that management knows how to make use of that increased communication to manage from HQ. You mention that large firms could organize as a collection of small firms, reminded me of how Google is supposed to be somewhat decentralized, with many projects starting from the engineers’ 20% time to tinker with their own ideas.

    I think economies of scale have been highly overrated for the last century, ignoring the fact that they are lumpy, ie bringing a new factory on-line greatly increases your costs until you get utilization up into the sweet spot. Further, scale and the resulting hierarchy/bureaucracy have always been poisonous to innovation, because you don’t want to rock the boat when you just spent millions on capital equipment or because of the inevitable heavy hand of Kafkaesque middle and executive management.

    But we live in a service/knowledge economy today, manufacturing is a small part of GDP. The notion that these large orgs make any sense in an information economy is just silly. Yet we see tech companies trying to recreate the wheel all over again, ie Apple, Microsoft, Google, Amazon, Ebay, etc. Part of the blame goes to credulous investors, who throw money at crazy theories, then lose their shirt. As Robert Easton notes, Google makes almost no money off all their side bets, they make the vast majority of their money from search ads to this day. Android has no business model and will be wiped away as a result.

    That’s the demand side, the supply side is the natural drive to power of all management, get big at any cost, so you lord over a larger fiefdom. How else do you explain Bezos and his silly drive to expand at all cost, damn the profits? Trial and error is paramount in the information age, which is why all these large orgs will fail.

    • Ajay says:

      Huh, right after I wrote this long comment, I stumbled across this recent talk by Gabe Newell of Valve, where he says that, “It seems fairly obvious that the Internet does a better job of organizing a bunch of individuals than General Motors or Sears does. Corporations tend to be pre-internet ways of organizing production.” His gaming company is famous for being completely flat, without any corporate structure. The last 10 minutes of that talk are worth watching, well, the whole thing is, but the last 10 are most relevant to my point. :)

    • Ajay says:

      Sorry, I can’t let this go, a fascinating quote about what happens in countries where corporations get too big:

      “C: So I’m out here kind of working with my translator trying to get the language right, I mean the Japanese digital publishing industry… the Kindle just came out a month ago.

      S: It’s weird to me that such a technologically and book literate country is really behind on those things. Why do you think that is?

      C: It’s strange, because Sony released their e-reader, their E Ink based e-reader, back in like 2003, and I remember being out here in the early 2000s, going to Bic Camera, picking up a Sony e-reader, and thinking it was a cardboard display model, and then you hit a button and the screen flashed — it was just like “holy shit, this is amazing, I’ve never seen anything like this before.” They did nothing with it.

      One of the problems is is that the industry is so tightly controlled by so few people. It’s like a lot of the industry out here. Dentsu controlling a lot of the advertising, a lot of the publishing happens in backrooms and there’s a few individuals making those decisions, I think that makes innovation a little bit more difficult.” – Verge interview with writer/developer Craig Mod

Comments are closed.