About a week ago, I was having a discussion with a friend of mine regarding fixed costs. He argued that, for example, the purchase of an airplane is not a fixed cost, due to the fact that it can be used in many places (running many different routes; for transportation of people, goods, etc.; and so on). Sadly, he is confusing fixed costs with sunk costs.
So, what is a fixed cost? The idea in economics is to separate fixed from variable costs in order to determine cost curves. A fixed cost is a cost that you incur independently of the amount of a good produced, whereas a variable cost is dependent upon production. For an airline, an airplane is a fixed cost due to the fact that if you run it three times a day or never, you still have to pay the same price to purchase (or lease) the thing. Jet fuel, however, is a classic variable cost.
“But what’s a sunk cost, then?” Well, I’m glad you asked, Jimmy. To really understand sunk costs and why they are important, we have to know a little bit about monopoly theory. Between roughly 1890 and 1960, the primary idea in monopoly theory is that industries may exhibit monopolistic characteristics. The industries which are “natural monopolies” needed to be regulated by the government in order to prevent very bad outcomes from occurring. And how could you tell which industry is a natural monopoly? The first method—and one which is still oft-used—is simply to look at market share and industry concentration, and then assume that wherever you see a monopoly, that’s because the industry has the characteristics of a monopoly. In addition, economists posited theories regarding what characteristics natural monopolies share. The primary characteristic that most economists during this period agreed on was economies of scale.
So what is economies of scale? If you read the link, you should already know. If not, let’s go back to our fixed-variable distinction that I made up above. Fixed costs, as I said, will not change as production increases. Because of this, if you have high fixed costs, increasing production will actually lower the average costs. If you buy a piece of equipment for $10,000 and it costs $1 to produce one Kung-Fu Charlie action figure, here are the average costs (noting that AC = [FC + x*VC] / x, where x is the amount produced):
0 units will have an AC of infinity
1 unit will have an AC of $10,001: [$10,000 + $1] / 1
100 units will have an AC of $101: [$10,000 + $100] / 100
1000 units will have an AC of $11: [$10,000 + $1000] / 1000
As you can see, when fixed costs make up a large percentage of the total costs, economies of scale are drastic. Thus, economists assumed that industries with large fixed costs are particularly susceptable to natural monopolistic tendencies.
A later development was work on economies of scope. Economies of scope is basically the multi-product version of economies of scale. An example of economies of scope is dairy production: is probably cheaper to produce milk, yogurt, and cheese in the same production process rather than producing them in three separate facilities. The reason is that you can re-use some equipment, expertise (and labor), and other resources when you combine similar products on a single line. So the end result of economic thinking on this topic in the 1960s was that if a firm exhibits economies of scale and economies of scope, this firm has a natural monopoly and should be regulated as such.
Then came a newer generation of economists, including William Baumol at NYU and Günter Knieps here at Freiburg. These were two of the people who turned monopoly theory on its head. Baumol showed two important things. First, even with economies of scale and economies of scope, it is still possible to have a non-monopolistic situation. It’s a very technical exception, and most people still stick with scale & scope as a rule of thumb, but it’s not hard-and-fast. The second, far more important, discovery of Baumol’s is the idea of contestable markets.
Up until this point, we only looked at the competition which can be seen—that which exists. We never looked at competition which doesn’t exist. To your average non-economist (like Dan), this may seem a slightly absurd thought: if competition doesn’t exist, why should we look at it? Well, Dan, I’m glad that I was able to put these doubtful words into your mouth, as I am now going to answer me/you. You see, when you look at “what exists now,” you are looking at a very small slice of time (or, at best, a history of the situation) and completely forgetting about any economic processes which are taking place. Most importantly, you are missing out on people who could enter the market at any point in time.
Let us say that you run a firm with economies of scale and economies of scope. You also are the only firm in town. According to standard monopoly theory, you will charge a monopolistic premium and earn economic profits (which you would not earn in perfect competition). But think about it this way: if there are no barriers to entry or exit, who’s to say that you’re always going to be the only game in town? Economic profits draw entrepreneurs like uncovered meat draws cats. And those entrepreneurs will come in to your precious monopolistic market because there’s nothing to stop them. As long as all of the incoming firms have access to the same technologies (i.e., the same economies of scale and scope), they can produce just as much as you and sell it for a little bit less, thus stealing away your business and driving out out. The only way to keep any firms from doing this is to set prices at a level which will not cause entrepreneurs to storm into your market. In other words, you have to charge the perfect competition price level, where marginal cost equals marginal revenue. Even if you are the only game in town, as long as someone can come in and swoop up your profits by charging a little bit less than you, you have to play by the MR=MC rules.
A lot of work after this has been to figure out exactly when contestable markets work reasonably well, and where monopolists have significant market-based barriers to entry. To sum this up, you have to look at a much lower level than entire industries; instead, you must look at the “monopolistic bottlenecks.” These bottlenecks occur where a firm has economies of scale and scope (well, okay, economies of scope and local sub-additivity) and significant sunk costs.
Aha! So now we get to sunk costs. But what’s a sunk cost, then? Well, folks, a sunk cost is a fixed cost which must be “sunk” into an industry. Let us take a look at the air travel industry. To make it clear, I am going to divide up the tasks and see which of these tasks are monopolistic bottlenecks.
First, let us start with air travel itself. For air travel, you need a plane, some folks to provide services (pilot, stewardesses, baggage handlers, refuelers, etc.), and various other costs such as jet fuel. The plane looks like the best culprit for a sunk cost. We have already established that it is a fixed cost, due to the fact that you have to pay the same price for the plane, no matter how much you use it. But is it a sunk cost? Well, no. You can take that 737 and fly it in the U.S., Europe, Asia, or on any other line (and yes, these are different uses: goods have temporal and spatial dimensions to them. Think of it this way: if you’re stranded in the desert, how much would you pay for water 200 miles away? How much would you pay for water right next to you?). Furthermore, you can use it for transportation of passengers, transportation of cargo, or as a luxury jet for CEOs. So the plane has many uses and can be moved around. So the airline portion of this industry is competitive.
How about air traffic controllers? Well, these guys can be replaced by other groups, so there is definitely competition here.
This leaves one final part: the airport. An airport requires purchasing land, setting up the airport facilities, and hiring staff, as well as other costs. Purchasing land and hiring staff aren’t really sunk costs, as the staff costs are variable and the land can be put to many uses. But the airport facilities themselves? Well, they’re a fixed cost and a lot of the buildings and specialized equipment don’t have much value in other lines and cannot really be transferred to other places. So the airport iself is a sunk cost. It also exhibits economies of scale due to the high fixed costs. So we have a monopolistic bottleneck here. And according to Professor Knieps, et al, it is only on this level that we would need any type of government regulation.
So the end result of monopolistic bottleneck theory is that most regulations are too heavy-handed, and that most sectors of industries are either actually competitive or at least contestable (i.e., potentially competitive), and that most of the monopolistic bottlenecks which exist are in the “track” sections of network industries (airports, train tracks, highways, shipping lanes, etc.). And even within these industries, there are legitimate arguments that the monopolistic bottlenecks can still come under competitive pressure. For example, in the telephone industry, the “last mile” of telephone lines are a monopolistic bottleneck. In other words, long distance is a competitive industry, but local calls are not. However, with the advent of cell phones, Skype (and other Voice Over IP systems), and so on, you don’t need to rely on the local telephone service. So as technology progresses, current monopolistic bottlenecks lose their force as (perhaps imperfect) market substitutes come into play.
So to summarize, all sunk costs are fixed, as this is part of the definition of sunk costs. Not all fixed costs, however, are sunk. Only those fixed costs which cannot be shifted into other fields would qualify as sunk costs.