The proper scope of the firm

One of the most interesting topics in econ/business/finance is the question of the proper scope of the firm. That is, what functions (like IT or human resources or logistics or industrial design), stages in the value chain (like retail, manufacturing, or raw materials sourcing), and businesses/product lines (like a textile manufacture considering shoe-making or a grocer considering an in-store medical clinic), does it make sense for a firm to own as part of its own internal corporate structure, and which ones should it contract out to the rest of the market? I find this question so fascinating for three main reasons: (1) It takes us back to really fundamental, basic economic questions like, ‘When are markets efficient, vs. when are there market failures such that a hierarchical command structure is better?’ (2) It’s relevant to every M&A deal we read about in the news and the value of the companies we all invest in, and managers sometimes use clearly fallacious reasoning to justify costly decisions to expand their companies’ scope; and (3) The business/econ theory of scope of the firm yields a lot of insights on the proper scope of other institutions, including those for whom profit is not the main desideratum. So in this post, I’ll talk through my understanding of the proper scope of the firm, in the hopes that some readers will learn from me and others will improve my understanding in the comments.

***

It’s often said, particularly by those with a capitalist disposition, that the major insight of economics is that decentralized, competitive markets yield better outcomes than centralized controls. Markets, this line of reasoning goes, aggregate the distributed knowledge and comparative advantages of widely dispersed individuals and give each actor strong, direct incentives to meet the needs (as signaled through prices) of the others. Centralized command economies fail because human beings (who give the commands to the command economy) cannot aggregate information as well as a market-clearing price could, and, in a large hierarchical system like a bureaucracy, every individual is distant from the consequences of his/her own actions and so lacks urgency and strong incentives. That’s the argument. And while on a political/social level this may be true, it clearly cannot be true on the microeconomic level of the firm. After all, a corporation is, internally, a centralized command economy that does many different things that theoretically could be contracted via an open, competitive market. The continued existence of actual firms — as opposed to flexible market arrangements among contract CEOs and freelance legal and HR departments, with intellectual properties and factory facilities constantly trading to the highest bidder like a stock — is evidence that command structures are better at this level.

What do I mean? Well, let’s start with an example that might seem silly: Why did Henry Ford own a factory with an assembly line full of workers who had long-term contracts to build Model T’s? Why didn’t Ford instead just go into a public square with his patented Model T design and blueprints and announce that he would buy 2,000 Model T’s at the end of the month from whoever offered the lowest price? After all, classical economic theory sees all markets as working like this — constant auctions between individuals seeking the best deal at that moment. Since Ford owned the rare, unique asset, the intellectual property on the design of the Model T, and since unskilled labor is a ‘commodity,’ economics would suggest that Ford would still earn all the profits from the sale of the Model T through this competitive auction process. So why did Ford Motors own a factory with regular workers (as opposed to just earning rents on its design and contracting the building out to craftsmen)? Well, an obvious reason is that the most efficient way to build a Model T was along an assembly line, which required an enormous capital investment and cooperation among a large number of individuals. 30 people working one day on this assembly line could produce many more Model Ts than one worker could 30 days. That is, a Model T factory had enormous economies of scale as compared to individual craftsmen building Model T’s. Clearly, the benefits to be gained from this economy of scale were greater than the costs of substituting a command hierarchy for a market (costs like workers having weaker incentives and less sensitivity to price changes). So given that, why didn’t Ford outsource the manufacture of Model T’s to another company with a factory? (This is a less ridiculous question — today, after all, many major industrials outsource their manufacturing to companies like Flextronics.) There are a couple of justifications I can think of: Since automobiles were a new, nascent technology at the time, there probably wasn’t a competitive market of potential outsourcers; thus, if Ford hired an outsourcer, it could have exploited its position, claiming that new complications and cost overruns justified ever-higher prices, and Ford would have no alternative but to accept. Also, since automobiles were a nascent industry, there was probably still a lot of “learning by doing,” and Ford could use its experience assembling the Model T in-house to develop its next profitable innovation.

If I’ve belabored this, it’s to illustrate that often some firm scope seems laughably, obviously necessary, but when we dig deeper it’s actually challenging to articulate why. In the perfectly competitive market of Econ 101, in which every individual is constantly auctioning her skills and assets to the highest bidder, there’s zero advantage to owning an asset or hiring an employee long-term per se, as opposed to licensing and renting them. The price of any asset or resource, in this competitive market, is equal to the time-discounted value of the profits it would generate. So by buying an asset (and asset here is used to include resources like mines, entire companies and business units, and intellectual properties), a firm isn’t doing itself any favors unless it can add value to that asset to make it worth more internally than it is in the rest of the market. Fundamentally, firm scope has to be justified by market failure, or, to be more precise, a market failure that has costs that are greater than the costs associated with a control hierarchy. So I ask again, why do we see the level of integration and firm scope that we do see in the world? Why don’t firms hire temp CEOs for specific tasks on short-term bases? Why did Facebook buy WhatsApp instead of doing a joint venture? Why do firms have R&D divisions — why don’t they just buy or license intellectual properties as needed directly from external labs and research scientists? Why would a retailer clean or own its own building? Why are universities in the real-estate business, owning their own student housing? Etc.

Well, obviously there are many reasons. Here’s my preliminary list of some market failures that explain the firm scope we typically see in the real world:

  1. Negotiated, written contracts are costly and time-consuming and cannot fully capture everything a firm needs: This is the most basic reason for corporate scope. For example, imagine you are a firm that sells hand-woven baskets made in India. Hand-weaving baskets is not capital intensive, so theoretically, you could just continuously buy from independent basket-weavers on an as-needed basis. But calling up the weavers for each new project and writing a new contract would be costly; and, particularly for a firm hoping to develop a brand, ensuring that the independent weavers all meet quality standards, or fighting over payments on baskets that do not meet standards, etc., would also be costly. In this situation, it could be more efficient to take the weavers in-house, to a single factory floor, where the weavers are managed continuously, where hours can be planned well in advance to plan inventory to match demand, and where consistent quality can be ensured in the process. These benefits would outweigh the probable costs of paying the weavers for downtime, renting space, and reducing the weavers’ own individual incentives. A lot of firm employment relationships are arguably analogous to this. Firms don’t temporarily hire their CEOs for particular tasks or services, giving the job to whoever asks the lowest salary, because no contract could specify everything a CEO is supposed to do (CEO’s valuable actions are most unobservable); instead, CEOs are given long term contracts and lots of stock ownership of the firm and its profits.
  2. Economies of scale: This is another pretty basic reason for business scope, already discussed in the Ford example above. (Some would make a distinction between ‘scale’, as size specifically, and ‘scope’ as the range of businesses a corporate entity operates in, but I’m including scale as a subset of scope.) Economies of scale can explain why company-owned factories beat market arrangements among artisanal craftsmen. And they can also, perhaps, give advantages to conglomerates like Maersk (a single corporate entity that owns many relatively distinct lines of business that operate separately for the most part). Maersk “shares” a few key functions across its businesses, including bulk purchases. Because suppliers will offer discounts on higher-volume orders, Maersk can make its otherwise pretty distinct business units better off by making single bulk purchases of oil and other commodities as a single unit, on behalf of the whole.
  3. Monopoly/oligopoly exploitation: A classic example of this is the steel-making industry, which requires two distinct processes: metal heating and then setting the hot metal to make steel. Could these two processes be done by separate firms? Probably not. Given that transporting hot metal is costly and dangerous, the two firms would have to co-locate to make this arrangement efficient. And once they had co-located, each would effectively be a monopolist of the other’s business. The metal-heater could demand a pay raise whenever the steel-maker faced new demand, and the steel-maker would have no realistic alternative (and vice versa). As such, separate ownership, and the chance to opportunistically renegotiate contracts, could lead to market failure here. And this fear has meant that there has usually been integration in this business and similar spaces (e.g., coal mines and co-located power plants). (As a side note, interestingly, it’s often claimed that concern over oligopolistic exploitation varies across cultures and this can explain some differences in industrial structure. For example, it’s believed that Japan’s business world is characterized by more cooperative, long-term arrangements among firms and their suppliers and so this particular kind of integration isn’t as ubiquitous in Japan.)
  4. Capital markets failures: This is a broad term that can encompass a number of different things. (a) Managers may believe, correctly or incorrectly, that external investors cannot identify good investment opportunities for the firm as well as they can and that, as a result, the capital markets will not always give them the financing they need for their growth strategies. In this case, a firm could add value by expanding in scope to be able to achieve an ‘internal capital market’ — i.e., getting to the point where it can plow cash from one business into investments in another, instead of relying on the external capital markets to finance its internal investments. Conglomerates often justify their existence using this ‘internal capital market’ argument, but many investors and academics are skeptical. (b) A business can increase its value by buying another business (or another asset more generally) that is very simply underpriced. For example, a company in an esoteric, niche market may be able to identify the value of a new competitor before the capital markets see its value and price it correctly; a company that buys up this new, underpriced firm adds to its value, but essentially does so as a stock picker. (c) Investors may have too-limited time horizons: For example, an independent research lab that yielded important new insights every two decades or so might not be sufficiently supported as a standalone corporate entity by “short-term focused” capital markets. Thus, firms with long horizons (such as drug makers) tend to bring R&D in-house, and standalone, publicly-traded research labs are not (to my knowledge) common.
  5. “Synergy”: This widely-ridiculed term just means that the value of several distinct things together is greater than the sum of the values of those things separately. In business terms, this would have to mean that Company A is worth $100 million and Company B is worth $100 million, but if they were packaged together under one ownership, A-B-Corp would be worth $220 million or so. How can business entities be worth more together than they are separately? A pretty mundane example of synergy would be this: Each individual corporation has to file a number of mundane legal disclosures every year, so when two companies combine they halve the total number of these legal documents and the associated legal costs. A more interesting example of synergy is this: Disney can use the movie studio that it owns to feature Disney characters, which then gets more kids hooked on the Disney universe and drives up demand for other Disney products. A more abstract example of synergy would be this: Since the process of manufacturing new technologies — the trials and errors and failures, etc. — often yields new insights, creators and owners of intellectual properties who are “forward-integrated” into the manufacturing of the associated technologies are more likely to generate new and better intellectual properties. Thus, in this case, keeping the creative R&D work and the manufacturing/implementation work packaged together in one firm (and even one physical location) can be more valuable than compartmentalizing the creatives/innovators vs. the manufacturers in separate firms and stages. A marginal example of synergy is this: Merging companies will often say things like, ‘we have a a great distribution network and they have an innovative new product, so we can use our distribution to market their product.’ It’s not clear that’s really synergy, since the one company could have just used the other’s via a joint venture or something.
  6. Monopoly power/customer exploitation: This one probably needs the least explanation. If I’m one of fifty lemonade stands on the block, I’m a “price taker.” If I buy out all the other lemonade stands, I can raise prices above the competitive market rate and claw back some of the value that consumers would have gotten in a competitive market.

So I hope this outline helps explain the most ubiquitous examples of firm scope. But clearly there is a limit here. There must be a reason why conglomerates have gone out of favor, why the Soviet Union failed — there must be a point at which marginal increases in the scope of a command economy do worse than a competitive market. Indeed, in actual practice, most researchers and savvy investors think that firms have a bias toward going too far — on average, acquiring firms’ stock prices decline immediately after the acquisition is announced. Investors think that managers tend to overpay for acquisitions and cannot realize the value they expect from the acquisition. Why is this? Well, partly it must reflect the fact that our capital markets are working decently well and not obviously mispricing too many firms. (Indeed, if acquisitions usually clearly, unambiguously improved the value of firms, then that would mean that acquiring firms were getting a steal, which would reflect poorly on our capital markets.) And partly it reflects the fact that increases in scope bring their own costs, as referenced above. In a conglomerate-style corporation, the Vice President of one of the internal business will be partly compensated with options that depend on the performance of the conglomerate stock as a whole; thus she will have weaker incentives than she would as CEO of a standalone company. Vertically integrated firms where, e.g., the manufacturing division has to buy from the raw-materials division, can miss out on the information that is communicated by price changes in a competitive market. On the whole, economic theory would predict that rational firms would tend to grow right up to the size where the marginal costs of increased scope begin to surpass the marginal benefits.

***

So given all that theory, what are some applications? What most interests me is the fallacious arguments that managers often give to justify acquisitions. I’ll give a couple of examples:

  1. “Diversification and risk”: Managers (of, say, acquiring Company A) sometimes claim that by acquiring unrelated businesses (of, say, eaten Company E) with uncorrelated earnings, they can smooth their own earnings, and thus reduce their risk.  Doing so obviously should reduce stock price volatility, but it doesn’t actually add value for shareholders, for a very simple reasons: Shareholders of company A could just as easily buy shares of company E themselves and achieve reduced risk through diversification on their own.
  2. “We’re moving to the higher-margin stage of the value chain”: Hardware firms sometimes justify their acquisitions of software firms by noting that software is now the higher-margin stage of the technology business. But this justification is fallacious, or at least incomplete, for a simple reason: The owners of software firms know that they have high margins (and profits) and thus, like anyone else, shouldn’t sell out for a price that doesn’t reflect the time-discounted profits they expect to earn. So high-margin businesses have high value and they accordingly have high prices and so there’s no free lunch in buying into a high-margin industry. Now, there could be other good reasons for acquiring these higher margin businesses; for example, if the capital-markets undervaluing them or there is some ‘synergy’. But buying a high-margin business isn’t a free lunch per se.

Rather, to justify an acquisition, a firm has to pass at least these four tests (some of these “tests” are borrowed from Prof. David Collis of Harvard Business School):

  1. The acquirer has to be able to add value to the acquired entity, to make it better off. This added value has to be greater than the acquisition premium, obviously.
  2. The acquired business unit should be worth more inside the firm than in any other possible ownership structure. Otherwise, the acquiring firm can best profit by selling the unit to the most valuable ownership.
  3. There is no market-based way to realize the value of the acquisition — i.e., flexible market contracts and joint ventures will not realize the same value.
  4. The benefits to the acquisition have to outweigh the costs of expanded scope, in terms of internal coordination and information problems and individual incentives/motivation.

***

Ever since I got interested in corporate scope, I’ve been trying to apply the theory to understanding all of the institutions around me. It’s fun and sometimes I feel like I see a lot of surprising logic and illogic. One institution I debate about is the University. For example, why do universities own housing and also own cool facilities like rock-climbing gyms and heavily subsidize these facilities? Why can’t college students just rent their own apartments like other twenty-somethings, and pay for time at private rock-climbing gyms? Why do some business schools, after charging admits very high up-front tuition, then pay to send those MBAs abroad for ‘immersive experiences’? Why can’t the MBAs pay lower tuition and buy their own flights abroad? What is the ‘synergy’ that justifies packaging all these things together?

One hypothesis I have is that the University is sort of like the managers who think they’re smarter than the capital market and can improve on it with an ‘internal capital market’: The University is an institution that has a particularly paternalistic (and I really don’t necessarily mean that in a bad way) attitude towards its ‘customers’. The University may believe that the external capital market tends to underinvest in human capital and that it can add value through an internal capital allocation process that nudges students to try new things like rock-climbing, going abroad, living in comfortable housing in close proximity with peers, etc. That is, the University charges a high upfront fee to let you in in the first place, but, once you’re inside, all these cool experiences are heavily subsidized so that even the financially-anxious student will try things that the University thinks are worth spending money on, but that cash-constrained young people might normally eschew.

This is abstract, though, and in practice I think universities should consider unbundling many of their products. I’m curious if readers have other examples of surprisingly intelligent or stupid scope in firms or non-profit institutions.