Lecture 24: Market Power

Click here for the quiz on this reading.

For the past few weeks we’ve been analyzing perfectly competitive markets: how they converge to equilibrium, how they don’t take externalities into account, how they can’t provide public goods.

Yet if we look around, we see lots of sectors – arguably, most of the most important ones in our daily lives – which are not characterized by perfect competition. Recall that in perfectly competitive markets, no buyer or seller is large enough to influence prices on their own; for this reason we characterize competitive markets as being populated by price takers. But most of the firms we think about on a regular basis aren’t small, price-taking firms. They’re firms which influence the markets they operate in at a minimum (Delta for air travel, Coca-Cola for soda, General Electric for appliances) or even outright dominate them (Google for search, Nvidia for AI chips, SpaceX for satellite internet). Which of the results we’ve been looking at are still true in markets dominated by one or a few firms? Are such markets efficient? Do they solve problems that competitive markets don’t, or do they cause more problems of their own?

One thing’s for sure: market power is everywhere, and any economic model that fails to take it into account will be sorely lacking. But before we examine market power, let’s look at why the competitive model says it shouldn’t exist at all.

Why profits should approach zero in competitive markets

Recall that when we solved for competitive equilibrium, one of the things that was fixed in the short run was the number of firms. But in the long run, firms can enter an industry, or leave it, causing the supply curve to shift.

Firms enter and leave markets all the time, especially in times of economic turbulence. This New York Times article from May 2025 details how many Chinese manufacturers responded to Trump’s imposition of massive tariffs by racing to establish production in countries like Vietnam, on which the US had imposed lower tariffs than it had on China. We can think of such a move as “leaving” China and “entering” Vietnam.

While it’s often portrayed as a firm “going out of business,” leaving a market can just mean discontinuing a product line. Famously, Jack Welch, the CEO of General Electric, established a “top-two” strategy: GE would only compete in sectors in which it would “number one or number two globally.” That meant not being afraid to leave markets where they were no longer as profitable. You can think nowadays of a company like Google or Microsoft “entering” the market for AI, even though they have many other businesses; at the same time, just last week Microsoft shuttered its unprofitable Skype service.

Fundamentally, the decision to enter or leave a market depends on opportunity costs. In high school Econ courses, the firm’s “shut down” price is often denoted as a “zero profit condition,” but this concept is badly misunderstood. Let’s review what is meant by this.

Review: visualizing firms’ profits in a cost curve diagram

Recall the profit-maximization model for a firm. The firm chooses its profit-maximizing quantity $q^\star$ by setting $MR = MC$. Its profit, then, may be represented by $\pi = (AR - AC) \times q$ Therefore the firm is running a positive profit when $AR > AC$, a loss (or negative profit) when $AR < AC$, and zero profit when $AR = AC$.

For a competitive firm, $TR(q) = pq$, so in this particular case both $p = MR$ and $p = AR$. So in this case the profit-maximizing condition is $p = MC$ and the “zero profit” condition is $p = AC$. Note that since it’s always true that $p = MC$, it follows that when a competitive firm is earning zero profits, we have $p = MC = AC$, which means it’s the point at which the $MC$ and $AC$ curves meet. This is also the lowest point on the AC curve, or the “minimum efficient scale.”

Try dragging the price up and down in the following diagram to see how the market price affects firm’s profits or losses:

See interactive graph online here.

Firm profits and entry and exit

Now let’s think about the interaction of this diagram with a supply and demand graph.

The left-hand graph below shows the cost curves for an individual firm. The right-hand graph shows the market supply and demand. The slider shows the number of firms in the market; note that the market quantity is the individual firm quantity multiplied by the number of firms.

Now notice that firms in this industry are making a profit. This will attract additional entrants into the market; slide $N_F$ to the right, and see what happens as firms enter:

See interactive graph online here.

As you can see, entry into the market shifts the market supply curve to the right, reducing the market price. Indeed, if you slide $N_F$ all the way up to 24, you’ll see that the market price drops so far that firms start making a loss. At this point they’ll begin leaving this industry market. At some number of firms, though, each firm in this market will be making zero economic profit; at that point we say that the market is in long-run competitive equilibrium.

Note that in long-run competitive equilibrium, since $p = MC = AC$ we know that all firms will be operating at their minimum efficient scale: that is, they’ll be producing their good at the lowest possible average cost! This is why competition drives costs, and therefore, prices, down to their minimum possible level.

The long-run industry supply curve

The short-run industry supply curve shows the quantity supplied at every price for a fixed number of firms. Where it meets the current market demand curve is the short-run equilibrium price and quantity. We motivated that notion of equilibrium by saying that when the price was at that point, there was no excess supply or excess demand; so prices would be stable there.

But what if we allow for entry and exit? If, at the current equilibrium price, firms in this industry are more profitable than firms in other industries, we would expect firms in other industries to enter, driving prices down. On the other hand, if firms in other industries were more profitable than this one, we would expect firms to leave this industry.

We can, therefore, introduce a new concept: that of a market in long-run competitive equilibrium. In such a market, two things are true: the market is in short-run equilibrium, so prices are stable; and firms in this industry are just as profitable as firms in other industries, so the number of firms is also stable.

What happens to a firm in long-run equilibrium, though, when there’s a persistent change in demand – for example, because this good became more popular, or less popular, or because consumers’ income increases, or because more consumers enter the market?

Let’s consider the case of a market which is currently in long-run equilibrium. This is shown in the diagram below. As you can see, at the current market price, $p = AC= MC$, so the firm is making zero economic profit.

See interactive graph online here.

Now suppose there’s an increase in demand. Drag the demand curve as far as you can to the right. The immediate effect of this is that there’s an increase in the market price. Each firm produces more, and now they’re making a healthy profit.

However, this profit attracts other firms, who enter the market. As each enters the market, it drives the price down, until it is back down at the point at which $p = MC = AC$. The market has achieved a new equilibrium with the higher demand; but entry has driven the price back down to where it was originally.

Why is it the same as the original price? Look at the cost curves. In this case, as firms entered or exited the market, the MC and AC curves remained the same. In other words, whatever else was happening in the market wasn’t enough to affect the input prices these firms paid. Therefore, the point at which the MC and AC curves crossed remained the same, and the height of that point must be the price at which the market settles in long-run equilibrium. In other words, as demand for this good increases or decreases, there will be short-run fluctuations in price, but in the long run the price will always be the same. Hence the long-run industry supply curve is a horizontal line at that long-run equilibrium price. (If you check the box that says “show long run supply” you can see this.)

Characterization of long-run equilibrium

Short-run equilibrium was characterized by three conditions:

Consumer optimization: each consumer was choosing to consume quantities that maximized their utility, subject to their budget constraint $(p_i = MB(x_i))$
Firm optimization: each firm was choosing to produce quantities that maximized their profit, given their production function $(MR_j(q_j) = MC_j(q_j))$
Market clearing: the total quantity supplied at the market price in each market was equal to the total quantity demanded $(S(p) = D(p))$

In the long run, we add a fourth condition: that exit and entry drive profits to zero.

Zero profit condition: firms in this industry are making the same economic profits as firms in other industries $(AR_j(q_j) = AC_j(q_j))$

Worked example

Let’s look at the math behind the diagram above, and see how to solve for long-run equilibrium.

In these kinds of problems you’ll be given the cost function for an individual firm, and the market demand function. In the case above, the cost function is the one we’ve been looking at: $c(q) = 64 + {1 \over 4}q^2$ and the demand function is $D(p) = 384 - 24p$

Part I: Solve for the equilibrium price as a function of the number of firms.

We can solve for the short-run equilibrium as usual, but for now let’s leave the number of firms $N_F$ as a variable.

The individual firm’s marginal cost is $MC(q) = {q \over 2}$.
We set $p = MC$ to get the individual firm supply function, $q^\star(p) = 2p$
We multiply $q^\star(p)$ by the number of firms to get the market supply function $S(p,N_F) = N_F \times 2p$.
We set supply equal to demand and solve for $p$ as a function of $N_F$. $\begin{aligned} S(p,N_F) &= D(p)\\ N_F \times 2p &= 384 - 24p\\ (2N_F + 24)p &= 384\\ p^\star(N_F) &= {384 \over 2N_F + 24} \end{aligned}$

Now let’s impose the zero-profit condition to find out what the long-run equilibrium price must be in this constant-cost market: The average cost is $AC(q) = {64 \over q} + {q \over 4}$ When this industry is in long-run equilibrium, we know that the firm must be earning zero profit: so we must have $\begin{aligned} AC(q) &= MC(q)\\ {64 \over q} + {q \over 4} &= {q \over 2}\\ {64 \over q} &= {q \over 4}\\ q^2 &= 256\\ q &= 16 \end{aligned}$ Therefore, we know that if the industry is in long-run competitive equilibrium, each firm must be producing $q = 16$ units.

We also know what the equilibrium price must be: the marginal cost (and average cost) of the 16th unit: $MC(16) = {16 \over 2} = 8$ So we must have $p = 8$ in long-run competitive equilibrium.

Plugging this into the expression for $p(N_F)$ above, we can solve for $N_F$: $\begin{aligned} 8 &= {384 \over 2N_F + 24} 16N_F + 192 &= 384\\ N_F &= 12 \end{aligned}$ Therefore there must be $N_F = 12$ firms operating in this market at that level of demand.

Another way to solve for $N_F$ is to do the second part first:

Solve for the long-run equilibrium price $p$ and how much each individual firm will produce at that price
Plug that price the market demand to find the market quantity
Divide the market quantity by the quantity produced by each firm to find the number of firms

For example, when you dragged the demand all the way to the right, it became $D(p) = 576 - 24p$ If we plug $p = 8$ into that function, we find that the long-run quantity in the market would be $Q = 384$; dividing by the quantity produced by each firm $q = 16$ would give us $N_F = 24$.

Why firms can persistently run profits in real life: barriers to entry

The story above can be summarized with no math and some very simple intuition:

Persistently high demand for a product raises prices
Higher prices increase the profitability of firms in that industry
If firms are earning more profits in this industry than in others, firms will gravitate toward the industries with higher profits
As firms enter the industry, the increased competition drives prices down, decreasing the profitability in the industry
The process continues until profits are equal across industries

However, this story rests on an assumption freely enter and leave industries. If corn prices are high and wheat prices are low for a few seasons, farmers can just plant corn instead of wheat. But if AI chips are suddenly insanely profitable, it doesn’t mean that anyone could suddenly go head to head with Nvidia. One does not simply enter the AI chip market.

Barriers to entry can come in many forms: it may be hard to develop the expertise to design chips for AI. It may require a huge amount of cash upfront. It may require deep knowledge of production processes, and relationships wih suppliers. But one major barrier to entry is the existence of a firm with market power in the first place.

For example, in the late 2000’s and early 2010’s, Facebook was competing with two rapidly growing companies: Instagram for photo sharing, and WhatsApp for messaging. Had those two companies continued on their own trajectories, the profits on both of those markets would have been lower for Facebook as they would be forced to compete. Instead, Facebook bought them both: Instagram for €1 billion in 2012, and WhatsApp for a whopping €22 billion in 2014.

The US Federal Trade Commission sued Facebook in 2000, alleging that these purchases were intended to create a monopoly in social media. Just a few weeks ago, Facebook (now Meta) won the case – new competition from the likes of TikTok meant that Meta, despite its ownership of these dominant apps, is not a monopoly. But an email that surfaced during the email underlined the reason for the enormous purchases, in which Mark Zuckerberg made the now-infamous assertion that “it is better to buy than compete.”

Furthermore, in a world where some firms in an industry are getting larger and larger, acquiring more and more market power, the other firms in the industry don’t want to be left behind. Even if they are profitable on their own, they too will have an incentive to either acquire or merge with other companies to remain competitive. So one firm acquiring market power provides a perverse incentive for its competitors to follow suit.

In other words, if there’s one “gravitational pull” dragging profits to zero, there’s another, potentially more powerful force pulling in the opposite direction: as firms acquire market power, one of the things they can do with that power is to consolidate their position by destroying or acquiring potential entrants. And as more firms in an industry grow large, the incentive to keep up through mergers and acquisitions becomes stronger and stronger.

The effect of increased consolidation

One of the “bombshell” economic papers in recent years was De Loecker and Eeckhout’s paper “The Rise of Market Power and the Macroeconomic Implications.” It generated quite a bit of controversy for both its methodology and its conclusions, but the core data it reported were striking. In particular, it attempted to look at the change in markups charged by firms, using firm-level data all the way back to 1950.

You might recall that a firm’s markup is the amount by which the price it charges exceeds its marginal cost. In the canonical picture of profit maximization for a firm with market power, the firm produces at the point where $MR = MC$: visually, the $q^\star$ where the marginal revenue $(MR)$ and marginal cost $(MC)$ curves intersect. The price it charges is the height of the demand curve it faces at this quantity: that is, the highest price it can charge and still sell $q^\star$ units.

See interactive graph online here.

In this case, the firm maximizes its profit at $q^\star = 8$, at which point it charges $p = 12$. Since the marginal cost at $q = 8$ is 4, its markup is $12 - 4 = 8$. Note that that’s a 200% markup over marginal cost. One way to represent that markup is that it’s 3 times the marginal cost: so De Loecker and Eeckhout would report that as a markup of 3.

Now, as with many of our “toy models,” this diagram is something of an extreme example – but what the authors of this paper find is remarkable. In particular, they report that “while average markups were fairly constant between 1960 at around 1.2” – i.e. a 20% markup of price over marginal cost – “there was a sharp increase starting in 1980 with average markups reaching 1.67 in 2014. In 2014, on average, a firm charges prices 67% over marginal cost compared to only 18% in 1980.”

As John Oliver documented in a hilarious but profanity-laden episode in 2017, this period coincided with a dramatic increase in corporate mergers and acquisitions. There are lots of examples, but this primer on the recent history of the airline industry is typical. As the authors note, “between 2001 and 2016, the number of major airlines dropped from 12 to five, with five major mergers after the Great Recession.” At the same time, profits skyrocketed: while the years leading up the Great Recession were miserable, with 14 U.S. airlines declaring bankruptcy between 2002 and 2005, the years since 2009 have seen one profitable year after another (with the exception of Covid).

The welfare effects of market power

Just as we asked with our model of perfect competition, we can ask whether all of this is actually…bad? Large companies, after all, provide a lot of value to their customers – for all the problems posed by large airlines, you’d probably rather fly a major carrier than a “mom and pop shop.” (There are some very small airlines…but their planes tend to be very small as well…)

Here we need to be a little careful, and understand what our models are telling us, and what they aren’t. It’s certainly true that a firm with market power charges a price above marginal cost; but that doesn’t necessarily mean it’s hugely profitable. Look at the diagram above. If the firm were to produce where $MB = MC$ – that is, at the intersection of the demand and MC curve – it would actually be running a loss, not a profit. That’s because there are large fixed costs, so the AC curve is pretty high. And indeed, we see a lot of firms with market power (airlines, chip manufacturers, etc) in industries with very large fixed costs. Those high fixed costs can themselves, of course, be a barrier to entry – but that doesn’t mean they don’t exist.

It’s also true that there may be economies of scale in many of these industries, which means that a single firm may be able to provide a good at a lower cost than a lot of smaller firms. In other words, while it’s true that the supply curve represents the marginal cost curve for society, the supply curve for a competitive industry may well lie above the marginal cost curve for a single firm that can take advantage of economies of scale.

Nonetheless, it’s true that a monopolist, by restricting its output to keep prices high, does not produce units for which the marginal benefit to society exceeds its marginal cost of production. And this represents a welfare loss, in the same way as we saw with a tax:

One very final model: monopsony

OK, I don’t know if we’re going to have time for this one, but here goes. One last model, though it’s really just an inversion of the one you saw above.

We’ve talked about market power in the market for outputs – but a firm with market power can also exercise that power as a buyer of inputs. Think about Amazon delivery drivers: Amazon has a huge degree of power over them. And more broadly, think about truck driving in general. John Oliver has another great episode about the economics of long-haul truckers. It’s very clear that the market power of trucking companies vastly exceeds the market power of drivers – who are usually classified as independent contractors, not employees. You don’t need to watch the whole episode to realize that truckers are paid much, much less than their value to the company.

But the show actually starts out with another fact: that there’s a “shortage” of truckers. Now, normally when there is a shortage of something, its price rises until the shortage goes away. But in this case that’s not happening; why?

For the same reason that a monopolist will produce at a quantity less than that at which the price it pays would be equal to its marginal cost, a monopsonist – that is, the only buyer in a market – will buy a quantity less than that at which the value of that input (i.e., its marginal revenue product) is equal to the price of that input.

Remember when we talked about a competitive firm buying labor – we said that its labor demand was given by the marginal revenue product of labor, and it would hire at the point where $MRP_L = w$. For example, if the value of the $L$th worker is given by $MRP_L = 24 - 0.2L$ and the firm has to pay a wage rate of $w = 8$, the firm will hire up until the point where $MRP_L = w$, or $L = 80$:

See interactive graph online here.

That was, of course, for a price taker in the market for labor. But what if the firm has market power – and instead of taking the wage rate as given, it takes the labor supply curve of potential workers as given?

Let’s assume that the firm sets a wage rate, and the number of workers it hires is given by $L(w) = 10w$; therefore, if it wants to hire $L$ workers, it has to pay them a wage of $w(L) = 0.1L$. Therefore the total cost of hiring $L$ workers is $TC(L) = w(L) \times L = 0.1L \times L = 0.1L^2$ and the marginal cost of the $L$th worker is $MC(L) = TC'(L) = 0.2L$ Notice that the wage rate they pay is $0.1L$, but the marginal cost is twice that! This is because every time you want to hire another worker, they have to raise the amount you pay all your workers!

How many workers will a monopsonist hire? As usual, they’ll choose the quantity at which marginal revenue (in this case, of labor) equals marginal cost: $\begin{aligned} MRP_L(L) &= MC(L)\\ 24 - 0.2L &= 0.2L\\ 0.4L &= 24\\ L^* = 60 \end{aligned}$ The $MRP_L$ of the 60th worker is 12, but the firm only pays them 6:

See interactive graph online here.

So, just as a seller with market power charges a markup over marginal cost, a buyer with market power pays its workers less than their marginal revenue product.

Discussion

Lots, in class. :) I think this is enough reading for a week-10 assignment. See you Wednesday!

Reading Quiz

That's it for today! Click here to take the quiz on this reading.