It turns out that there's a field of research that directly addresses these questions about went vertically integrate, which is called transaction cost economics, or sometimes called the theory of the firm. The main propositions of this field have strong statistical validation, although, of course, there's always room for new perspectives. For example, I'm currently working on research that examines how new digital technologies like AI and cloud computing might affect vertical integration. Transaction Cost Economics, or TCE, is a theory about the scope of the firm, which can be applied to vertical integration. And also, the horizontal scope, as we shall see shortly. One of the most prominent names associated with the theory is that of Oliver Williamson, who won the Nobel Prize in Economics in 2009 for this work. TCE deals with both the cost of transactions, or economic exchanges, which can include things like negotiating monitoring, and enforcing contracts. As well as administrative costs, which are associated with organizing activities in a more hierarchical command and control fashion. Such hierarchies can be costly, in terms of bureaucracy, weak incentives, and sclerotic response to changing environments. In general, one might describe market exchanges or outsourcing as more transaction costs and vertical integration as involving more administrative costs. But there may be some overlaps in these costs, depending on the exact way in which outsourcing, or vertical integration is done. Now, Oliver Williamson's key insight about the optimal scope of the firm can be summarized in the idea of comparative organizational analysis. Essentially, Williamson asks us to separate out two questions that are often commingled when people think about vertical integration. First, what is the objective of integration? What market power, or efficiencies, or control, or coordination is being sought? You can similarly think about the objective of outsourcing, such as access to specific capabilities, etc. And the key is to separate this question from the optimal organizational form. For example, outsourcing, or vertical integration that best achieves this objective. The main point here is not to short-circuit the analysis, and jump from the first stage to the conclusion that a particular organizational form is needed. So to reinforce this key idea, I want to be clear that any time a firm has an objective, such as better control, or better efficiency, etc. There is a way to potentially achieve it by vertical integration making it, or by the market essentially buying it. Now, something I didn't tell you earlier was that Oliver Williamson was a professor of mine in my PhD program, and I even worked as his research assistant for a while. And here's something that Williamson often said, which I think goes to the core of his way of thinking, all organizational forms are flawed. What he means by this is that the two alternatives make or buy are different, in terms of the relative advantages or disadvantages they provide. So our task must be to compare relative advantages and disadvantages, and decide which alternative on balance is better for us at the most basic level. Williamson's core message, his main mantra is that using markets, or outsourcing provides better adaptation. Whereas, hierarchies, or doing things inside a firm provides better coordination. Now typically, we find higher transaction costs in markets, and higher administrative costs within firms. So let's try to understand where these costs come from. There are, of course, many mundane transaction costs and markets, such as search costs and haggling costs to find and reach a deal with the right partner. But there are also some strategic costs that are important to recognize. One of these is adverse selection, which arises due to information asymmetries that exists between firm and its transaction partner, say, a supplier. This concept is best illustrated by George Akerlof in his famous Market for Lemons paper, for which he won the Nobel Prize in 2001. Akerlof was also one of my professors at Berkeley. He is easily the most brilliant, and simultaneously humble person I've ever met. Akerlof's paper explains why transactions in the used car market are problematic. He notes that the sellers of used cars know more about the car than the buyer. So there is asymmetric information, which makes it more likely that the owners of bad cars, so-called lemons, are more likely to want to sell their car. In turn, this should make buyers more suspicious, and less willing to pay more for a used car. The result is the well-known phenomenon of a new car losing value the moment you drive it off the lot. And it also explains why many car companies have to step in to offer certified pre-owned cars to address buyer concerns about used car quality. I explain this more general idea here, so that you can think about how it might apply to buyer seller transactions in the vertical integration context. Essentially, one party may know something that the other doesn't, which then affects how it transacts. For example, a supplier might not put its best people something that only it knows on, and supporting a buyer's outsourcing needs. Or it may save the latest technological improvements for its own competing product, etc. Similarly, there's another economic principle called moral hazard, which may also make market transactions costly here. The transaction partner has private information about its performance that is not measurable, or contracted for which leads to abuse. Economists often talk about the moral hazard arising from governments being willing to bail out banks, for example, because this willingness may lead banks to take on too much risk. In the vertical integration context, a downstream dealer or service provider may take risks, or reduce service quality with customers. For example, if the only thing that's measured and rewarded in their contract is total sales. Finally, there is this important issue that Williams that emphasizes in his discussion of transaction costs, which is often called the holdup problem. The key idea here is that after the parties enter into a transaction, something changes. So that one party gains an upper hand, and can therefore engage in what Williamson calls opportunism. Williamson defines opportunism as self-interest seeking with guile. Then, due to inherent uncertainty in the environment, that party has to only wait for the next opportunity to change the terms of the contract it originally entered into. One common situation that leads to opportunism is if one of the parties has to invest in assets that are specific to the other. So for example, if I get my supplier to buy machinery that's customized to my needs, or to build an IT system that's tied into my IT system and so on. I essentially have the supplier at a disadvantage, because these investments can only now be used to service my needs. Now, I will say that there are good reasons not to abuse such a relationship, especially because the long-run consequences and reputational effects if you do. But if you are that supplier, you might want to also think about how you safeguard against the hazards posed by such specific investments. Another example of transaction hazards that can result in opportunistic behavior is exposing your company's core knowledge, or technology to your supplier, or buyer after some time. They may no longer need you, or pass that knowledge to another supplier who can then compete with you. So with outsourcing, there are these three major sources of transaction costs to worry about, adverse selection, moral hazard, and the holdup problem. Just as with market transactions, hierarchical organization within a single company also has costs, which we will call administrative costs. One big source of costs is week incentives, which simply means that an internal unit may be less laser focused on delivering the very best performance as an external supplier, maybe. Why, the supplier knows that if it doesn't perform, it will lose your business. But with an internal unit, there's often an expectation of continuity. And this may occur in part because internal units are often governed by fiat. That is, within a permissible zone, orders can be given and changed on what is expected from the unit, which an internal unit will generally have to comply with. Now, this can be an advantage. But in exchange for this Fiat, one typically gives the internal units more slack on performance, which means that their incentives are less strong. Also, another problem with internal units is the so-called principal-agent problem, which is often applied in the owner/manager context, as you might have seen in a finance course. But it can also be applied in a manager subordinate context. Say if a CEO wants one division in a vertically integrated company to supply to another, and wants that unit's performance to be top-notch. Perhaps some aspects of performance are unobservable. Then it's likely that the unit will underperform on those dimensions. You may recognize that this is akin to the moral hazard problem. But with a manager at the other end, rather than another company. Last but not least, vertically integrated hierarchies may also exhibit a lack of dynamicism, and less responsiveness to market or technology trends. In part, you could see this as a result of firms being unable to selectively intervene what this means. Is that if your organizer's a large company, with all its rules and procedures and bureaucracy, it is hard to selectively go into specific divisions and functions and say on this thing, and this supply arrangement, I'm going to treat you like an outside supplier. You either deliver the performance we are asking for, or get cut off. Now, some firms are good at this, at bringing market-like pressure to bear on their internal divisions and units, but it has limits inside the firm. So when should firms vertically integrate? Here's the overall takeaway. There is no general prescription that outsourcing, or vertical integration is always better. Rather, the important thing is to align the attributes of their transaction with the governance mode vertical integration or outsourcing that is appropriate for these attributes. So for example, if one needs investment in specific assets, or there are technology leakage hazards, or it's difficult to measure the performance of the partner, it might be better to vertically integrate. On the other hand, if you really need strong incentives for performance, or more adaptation over time to market or technology trends, then outsourcing may be better.
