Collusion and welfare enhancement
Stanley Hoff
Supervised by Andro Rilovi
ć
Abstract: This paper writes about the possibilities in which collusion can enhance total welfare. It shows that
welfare can be enhanced when externalities are present. This paper will also show how cartels can be welfare enhancing after firms have joint up in a research joint venture. Finally, the paper shows that cartels can enhance welfare when firms that collude have different costs of production. In this paper, current regulation will be looked at and possible changes to the regulation are proposed.
Statement of Originality
This document is written by Student Stanley Hoff who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
Introduction
In current textbook literature, cartel formation is seen as welfare-reducing (Brod & Shivakumar, 2003). Cartels are often formed to reduce output of the product in order to be able to set a higher price of the product. When a cartel is formed, a part of the consumer surplus is absorbed by the producer surplus, whereas the other part of the consumer surplus is lost because of a reduction in sales. This is called the deadweight loss (Pepall, Richards & Norman, 2010). Current legislation does not allow cartels when these cartels reduce output to increase the prices. (European Commission, 2003). The only form of cartel that is legal is when it improves distribution of the product or when it is beneficial to society through research and development, which will be discussed later in this paper. Total welfare is defined as the total of producer surplus and consumer surplus. That is, producer surplus is seen as equally valuable as consumer surplus.
There are, however, a lot of reasons why the formation of a cartel can be total welfare enhancing. A good example, described by Andrew Torre and Daniel Morgan (2012), is the Australian beer industry. Beer is a good that has a lot of negative externalities. Alcohol can lead to multiple explicit externalities like assault, drunk driving and vandalism, but also has a lot of implicit externalities like decreased work performance. Because a cartel fixes quantities and therefore increases prices, less beer is consumed and the negative externalities are reduced (Torre, 2010). This is just one of many examples that show that collusion can be beneficial for both the consumer and the producer.
In the next section I will present an overview of the current regulation of cartels and dominance-abusing monopolies. Section III will be an examination of the welfare-increasing possibilities of cartels because of externalities. Section IV will examine the increase in welfare because of collusion after a research joint venture. Section V will examine how differences in total costs of companies can increase total welfare through cartelization. In the sections III, IV and V I will also present options other than the formation of a cartel and show why these options do or do not hold in certain cases. Finally I will pose some suggestions to adjust the current antitrust law in section VI and show my conclusion in section VII.
II. Current regulation
The current regulation of cartels and monopoly-abusing companies is called the antitrust law. In Europe, this is Article 101 and 102. Article 101 and 102 of the Treaty on the Functioning of the European Union has been in place since 2012. Before the treaty they were listed as Article 81 and 82 of the European Commission in 2003. The reason for these articles is listed on the website of the European Commission (2003):
‘Competition encourages companies to offer consumers goods and services at the most favorable terms. It encourages efficiency and innovation and reduces prices. To be effective, competition requires companies to act independently of each other, but subject to the competitive pressure exerted by the others.’
Article 101 states that companies are considered to abuse their power when they:
1. Directly or indirectly fix purchase or selling prices or have any other trading conditions 2. Limit or control production, markets, technical development, or investment
3. Share markets or sources of supply
4. Apply dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage
Article 101 has some exceptions. The article is invalid if companies that have agreements use these agreements to either improve the production or distribution, help with the promotion of the product or help with the technological improvement, of which consumers get a fair share or the benefits. If the European Commission finds companies that violate this article, these companies are fined. This fine is representative of the damage that these companies have done to the economy, with an added extra. This fine cannot exceed 10% of the annual turnover of the firm. (European Commission, 2013).
There are multiple reasons why cartels should be made illegal. According to Wardhaugh (2012, pp. 370-371) these are the five main reasons why cartels have been made illegal:
1. Consumers pay more for their goods, ie the cartelist appropriates all or some of the consumer surplus to itself
2. Cartelists create a deadweight loss to the economy
3. In creating and protecting a cartel, the cartel participants engage in socially wasteful expenditures
4. The cartelist is inefficient, stunting the development of new products and hindering the development of more efficient production processes; and
5. The existence of cartels exacerbates any existing managerial ‘slack’ (or, in the parlance, ‘X-inefficiency’). Thereby hindering efficient, productive and profitable growth for both a firm and the economy.
These are all very good arguments of why cartels should not be allowed. Some of these arguments, however, are not always applicable. In my paper, I assume that producer surplus and consumer surplus are equally important. Therefore, Argument 1 would not be important. If you would consider consumer surplus as more important, however, Argument 1 still holds. Argument 2 is a very
commonly known reason of why cartels should be banned. Though this holds in theory, sometimes negative externalities are present in the market, and then it can be proven that the deadweight loss can be compensated. This will be shown in section III of this paper.
If the wasteful expenditures mentioned are very large, this would be a very good reason why cartels are a loss to total welfare. Currently, governments spend money trying to prevent collusion from happening. If these costs are lower than the wasteful expenditures of the cartels, controlling for cartels will be beneficial as will be shown in Section V of this paper.
Argument 4 does not necessarily have to hold. Though some cartels will not have the urge to innovate because they actively drive out competition, some companies need to innovate in order for to be able to form a cartel. Therefore cartels could be a result from an earlier innovation. This will be addressed in Section IV.
Finally argument 5 could be a reason why cartels are inefficient. Unfortunately, the author did not write any proof to back up this argument, so the reason why this is true is very vague. It is also a part that is outside the scope of this paper, and should therefore be researched more thoroughly.
The European Commission has also adapted the leniency policy. With this policy, the first firm to that turns in the cartel to the European Commission gets a reduced fine or no fine at all. In Europe, this can be every firm in the cartel, including the ringleader (Bos & Wandschneider, 2013). This is different in the United States, where almost the exact same rules apply, but ringleaders are not allowed to apply for the leniency policy. The ringleader is seen as the leader of the cartel. In one out of five cartels there exists a ringleader that is dominant within the cartel. The idea behind the exclusion of the ringleader is that the other firms have a higher chance of being the first to turn themselves in, because there are less firms within the cartel that are able to apply for leniency. This, in turn, makes it harder for a ringleader to form a cartel (Bos & Wandschneider, 2013). Bos & Wandscheider find in their article that with ringleader exclusion, other members of the cartel will
turn themselves in faster and it is ambiguous whether it is less likely that the ringleader starts a cartel.
In the 2003 paper of Crandall and Winston (2004), they find that antitrust policies often accomplish the opposite of what the policies are intended to do. They find four reasons why the antitrust policy is ineffective: The cases against monopolies and cartels take too long, and by the time the issue is resolved the companies have already become less of a problem to the welfare, Antitrust policies cannot keep up with the fast speed at which the industries are growing in terms of technical development which results in an outdated policy, the difficulty of proposing remedies against cartels and antitrust policies cannot effectively differentiate which cartels are welfare decreasing (Crandall & Winston, 2004). Especially the last argument is important to see where the policy lacks in its basic set of rules. The authors also find that the cartels normally do not raise their prices a lot. In the most extreme cases the prices are raised by 10 percent, whereas the vast majority of the cartels only raise their prices by 3 to 7 percent (Crandal, 2004). If regulation would condition collusion instead of banning it, a minimal and maximal price increase could be discussed between the firm and the regulator. For these reasons the question arises whether a better policy is possible. In the remainder of this paper the focus will be on which kinds of cartels should be excluded from the current
regulation, or how the regulation should be changed to better accommodate certain cartel formations.
III. Externalities
One of the reasons cartels can be beneficial to total welfare is because of externalities. Currently the welfare from trade is measured through consumer surplus and producer surplus (L. Pepall et al., 2010). However, a lot of companies have externalities, both positive and negative. There is evidence that a lot of companies set their prices too low because not all costs of the product are internalized (Torre & Morgan, 2010). An example of this is the petroleum industry. Oil is used all around the world for cars, motorcycles, machines and a lot of other devices. The use of petroleum generates a lot of externalities like air pollution, noise and driving accidents (I. Mayeres, 2013). Because these costs are not internalized by the companies that produce petroleum they will not be taken into account when pricing the petroleum. In the petroleum market, a cartel already exists, namely the OPEC cartel. Prices are much higher than they would be in a competitive equilibrium because of this cartel. People are unable to get as much petroleum and they have to use the petroleum they have more efficiently because of these higher prices. This way, less petroleum is being used, there is less air pollution and less noise on the streets. It might therefore be good that this cartel exists. (Torre & Morgen, 2010)
sufficient to display the total welfare effect. This is why I introduce another surplus into the equation: the social surplus. This surplus measures the costs of use of the product that are not reflected in the prices. To present how this surplus affects the total welfare, I will use an example of the beer industry because beer has negative externalities like alcohol abuse, violence and assault. (Torre & Morgan, 2010). The model of this industry without the social costs is presented in Figure 1.
Figure 1: the equilibrium of the market without social costs. Reprinted from ”Is all cartel activity evil?” by A. Torre and D. Morgan, 2010, a journal of applied economics and policy. 29(4) p. 438. Copyright 2011 by the Economic society of Australia
In this Figure, the demand function D equals the marginal social benefit function of the consumer. The marginal costs equal the marginal social costs because of the absence of the marginal social welfare in this model. In a competitive equilibrium the price will equal the marginal costs and will therefore result in a price of P2 and a quantity of Q2. In a monopoly or cartel, the profits of the firms will be maximized, which is where the marginal revenue of the companies equals the marginal costs. This is at price P1 and quantity Q1. The producer surplus will increase but the consumer surplus will decrease by a larger amount, leaving a deadweight loss of A. In this case, having a cartel reduces total welfare by A. This is the main reason cartels are illegal at this very moment (L. Pepall et al., 2010). Now the social costs of beer will be introduced. This can be seen in Figure 2 (Torre & Morgen, 2010).
Figure 2: the equilibrium of the market with low social costs. Reprinted from ”Is all cartel activity evil?” by A. Torre and D. Morgan, 2010, a journal of applied economics and policy. 29(4) p. 438. Copyright 2011 by the Economic society of Australia
In this figure, the social costs might look simplified. If you think about it, the production of beer might also have marginal social benefits. To see that the figure still holds, consider that a lot of the surplus of beer is actually consumer surplus. Having a few beers might make you feel more confident, for example, or more fun at a party. These are purely benefits to the consumers of the drinks. The socials benefits are that others might have more fun with you because you consumed some beers, for example. This is just a very small portion of all the benefits that having a few beers has. And even when these social benefits would be large, as long as the marginal social costs are higher than the marginal social benefits, the figure above holds.
The marginal costs, demand and marginal revenue line in Figure 2 (Torre & Morgen, 2010) are the same as that of Figure 1. The difference is that the marginal social costs are now different from the marginal costs that the firm faces. These marginal social costs reflect the negative
externalities that come from the consumption of beer. In a competitive equilibrium the price equals the marginal costs and is therefore P3 with quantity Q3. The producer surplus is then 0, the
consumer surplus is the triangle that is formed with the marginal cost line, the demand line and the Y-axis. The social costs that are not accounted for by the firms are represented by the marginal social cost minus the marginal cost line times the quantity. The social costs that are not compensated for by the consumer surplus are given in triangle A. This is because the marginal social costs minus the marginal costs is the square that has the corners P2, P3 and Q3. The consumer surplus in this square is everything except the triangle A.
It would be optimal for the total welfare if firms would produce at a cost equal to the marginal social costs. This is represented by price P2 and quantity Q2. If this would be the price, the
loss of consumer surplus would be more than compensated for by an increase in producer surplus and a reduction of social costs.
If companies were allowed to demand monopoly prices, they would ask price P1 with quantity Q1. This would create a new deadweight loss that is equal to B. even though this is not the optimal quantity and price and for the highest total welfare, the deadweight loss is compensated by an even larger reduction in social costs and is therefore a better solution than the competitive equilibrium (Torre and Morgen, 2010).
Consider also a third case in which the marginal social costs are even higher. This is represented in Figure 3 (Torre & Morgen, 2010).
Figure 3: the equilibrium of the market with high social costs. Reprinted from ”Is all cartel activity evil?” by A. Torre and D. Morgan, 2010, a journal of applied economics and policy. 29(4) p. 439. Copyright 2011 by the Economic society of Australia
In this figure, the optimal price that would account for all social costs is P2 with quantity Q2. If monopoly pricing would be allowed, firms would set a price equal to P1 with quantity Q1. This price, even though it is a monopolized price, is still too low to account for all social costs. However, it does compensate a lot of the social costs and is therefore a better solution than the competitive
equilibrium.
An output reduction because of the formation of a cartel is strictly forbidden within the European antitrust laws. Therefore, the monopolized price cannot be reached and will be at a competitive equilibrium (European Commission, 2003). If these cartels would be able to get their profit maximizing prices if they prove they can compensate, at least to some extent, for the social costs that their product causes, a higher total welfare would be generated.
intervention. This is by raising taxes on those products that cause negative externalities. This is a so-called Pigouvian tax. A Pigouvian tax is focused on increasing the price so that the optimal amount of the product is consumed. This can be done by the government. With a tax, all of the producer surplus goes to the government instead of the companies. The government can then decide to give the surplus back to the public, the ones that would have had all the surplus with perfect competition in the first place. Also, if the government knows the market that produces the negative externality perfectly, it can set a tax that will equalize the price and the marginal social costs.
This kind of a solution actually comes very close to being the perfect alternative for allowing cartels. There are, however, some problems with a Pigouvian tax. The first problem is that the government does not know all the details of every market (Bork, 1993). When the price of a product suddenly decreases, this could be because of fierce competition, or because a monopolist or a cartel are trying to force other firms out of the market (Crandall & Winston, 2003). Governments might not make a good decision on how high the tax should be because it is not as well informed as the firms are. This also implies that the government has to incur costs to get knowledge about the market. If the government would impose a tax that would result in a quantity where the marginal social costs line equals the demand line, they would still have to pay for the costs of finding all the information. If you consider Figure 2 (Torre & Morgen, 2010), and assume that the government will know exactly what tax to impose after it has done its research, the tax would only be beneficial for the market if the costs of the research by the government is lower than triangle B. Because firms want to make a profit, they would want to be allowed to collude in the market. If firms can apply for collusion by sending all their information to the government, the regulator can find out about the costs and social costs without having to incur high costs to find it themselves.
Also, producer surplus implies that the company takes that part of the profit. However, A lot of companies that form a cartel do not keep the profit for themselves. Evidence shows that members of a cartel often invest their earnings in the expansion of their product line or invest in research, to drive down the marginal costs or create a new product (Bos & Pot, 2012). In the next section of this paper, I will show how cartels can be beneficial to the economy through research.
A last thing about the Pigouvian taxes is that it can be proven that they are not efficient in the long run. (Carlton & Loury, 1980) This is because in the long run, firms will try to earn back their average costs and not their marginal costs. Because a tax reduces the output of the firms in the market, they will thereby also raise the average costs of the firms. In the long run, some companies will have to go out of the market and the other companies will increase their supply to reduce the average costs. So, in the long run, there will be the same amount of output with less companies that supply the product. Therefore, it does reduce welfare but since it does not reduce output, the externality still exists.
Another alternative to the Pigouvian taxes and the cartels are lump-sum taxes (Carlton & Loury, 1980). In their article, they prove that a lump sum entry tax or subsidy combined with a Pigouvian tax can restore the output to the optimal output level. The regulator should give a subsidy when the average damage of the externality is rising at the optimal output of the firm. If the average damage is falling, a lump sum tax should be levied. This also ensures that the industry welfare is maximized (Carlton & Loury, 1980). Though this seems like an intensive task for the regulator, it does solve the welfare maximization problem completely. It is therefore a very viable alternative to allowing firms to collude in the market. It might be better than the market equilibrium because it can completely maximize welfare. However, it is more intensive for the regulator to assess how high the taxes should be than it is to simply allow a cartel. This is why I would suggest allowing cartels only when they are able to get very close to the social optimum. This is when triangle B in Figure 1 (Torre & Morgen, 2010) and Figure 2 is very small or non-existent.
IV: Research joint ventures
A second reason that cartels can be beneficial for the total welfare is with research joint ventures. This is when companies cooperate with the research of a new product or the improvement of an already existing product. Conducting research is often shared because companies can share their knowledge and because they can share the costs of research. To conduct research, the companies will have to do very large investments hoping they will find something that will pay back the investments (Qiu, Cheng & Fung, 2009). Article 101 states that companies are excluded from the collusion article if they contribute to the technological welfare, so most of the research joint ventures are already allowed.
However, once the product is invented, the collusion has to stop because a collusion through price fixing is illegal. Now, the companies that colluded to invent the product have to compete on the market. Because they have to compete, their profits will be lower than when they were allowed to form a cartel and set monopoly prices. Firms know this before they sink their investments. Therefore, it could be that firms do not think they will earn back their investment when they have to compete with the other company, but do think they will earn back their investment if they can collude with the other company. The research might never start because this collusion is not allowed.
This can also have consequences for the consumers. Companies that share their information are more efficient and market efficiency can lead to higher welfare (Myagiwa, 2009). Also, if a new product is invented, this product might increase the consumer welfare as well, even with monopoly prices.
Firms that start a research joint venture do have a better option to start a cartel after their research because of an increase in the stability between the two firms (Seldeslachts, Duso & Penning,
2012). The chance of the firms colluding legally or illegally increases after a research has been done because the stability between the firms has increase. Therefore, firms might still start their research because they know they will collude after the research has been conducted. Allowing the firms to collude will not have any effect In this case, except that the regulator has more insight in the process and can stop the collusion at a certain period. This could be a reason why firms will never apply for collusion in the first place and rather perform the task illegally.
A way I believe this can be prevented is by giving a fine that is sufficiently high to scare the companies away from illegal activity. A test should be conducted to see whether firms can be pushed into reporting their plan to collude.
At this moment, patents are being used to protect companies that invent new products. These patents allow the company that has invented a new product to be the only one that is allowed to produce the new product. This way, they can monopolize the market for a specified time so they can earn back their investments. After the specified period, new entrants can enter the market, which enables perfect competition and therefore welfare maximization. However, if two or more companies started a research joint venture, the patent would hold for both companies, so even though other companies cannot enter the market, the incumbents still have to compete against each other.
To show why this is unfair mathematically, consider first a single company that has invented a new product. I assume that πn <πc < πm, where πn is the competitive equilibrium profits, πc the
Cournot equilibrium profits and πm the monopoly profits. I assume this because profits of a firm is 0
when the market has perfect competition. Cournot competition results in profits for the competing companies, so these profits are higher than 0. Monopoly profits are considered the highest, because firms are allowed to set their quantity and prices any way they want, so they set their prices to maximize their profits. I therefore assume it is higher than Cournot profits because it is the highest amount of profits the firms can make. To keep the equation simple, I assume that the market becomes perfectly competitive as soon as the patent ends. I also assume that companies have a chance of to find a new product when they start their research. Also, I assume the patent has a duration of x years. The company will invest in a new product if the net present value of the investment is positive. So they will invest if:
𝐶 < ∝ (𝜋𝑚+ 𝜋𝑚 1+𝑟+ 𝜋𝑚 (1+𝑟)2… 𝜋𝑚 (1+𝑟)𝑥) (1)
Where C is the fixed cost of the investment of a single firm.
Now consider a research joint venture. As is explained above, companies that share
tasks. If one company has machines that can test the new invention and the other company has the machines to develop the new product for example, neither company has to buy machines that the other company already owns. This would be the case when the firms would invent the new product themselves. Therefore, I assume that their combined sunk investments are lower than that of a single firm. The efficiency bonus will be described as . The efficiency bonus will be between 0 and 1. If the firms were allowed to collude, they would start a research joint venture if:
(1 − 𝛽)𝐶 <∝ (𝜋𝑚+ 𝜋𝑚 1+𝑟+ 𝜋𝑚 (1+𝑟)2… 𝜋𝑚 (1+𝑟)𝑥) (2)
Because (1- )C < C, the probability that the companies will invest increases as long as is not 0. Also, if is larger than 0, their costs will decrease, which can be seen as an increase in producer surplus as compared to the innovation of a single firm.
Thirdly, consider a research joint venture that has to compete a la Cournot. The companies will start a research joint venture if:
(1 − 𝛽)𝐶 <∝ (𝜋𝑐+ 𝜋𝑐 1+𝑟+ 𝜋𝑐 (1+𝑟)2… 𝜋𝑐 (1+𝑟)𝑥) (3)
For these companies, their probability to invest will always be lower than when they could collude because the costs of production are higher and the profits are lower. Also, they will rather produce solo if the profits of (1) is higher than half the total profits of Cournot competition that is described in (3), so: 12(𝛼 (𝜋𝑐+ 𝜋𝑐 1+𝑟+ 𝜋𝑐 (1+𝑟)2… 𝜋𝑐 (1+𝑟)𝑥) − (1 − 𝛽)𝐶) < 𝛼 (𝜋𝑚+ 𝜋𝑚 1+𝑟+ 𝜋𝑚 (1+𝑟)2… 𝜋𝑚 (1+𝑟)𝑥) − 𝐶 (4)
If we rearrange this equation, this shows that companies rather invent the product themselves if: (12+ 12𝛽) 𝐶 < (𝜋𝑚− 𝜋𝑐 2) + 𝜋𝑚−12𝜋𝑐 1+𝑟 + 𝜋𝑚−12𝜋𝑐 (1+𝑟)2 … 𝜋𝑚−12𝜋𝑐 (1+𝑟)𝑥) (5)
This shows a couple of things:
1. If the information sharing efficiency bonus is high, The probability that firms tend to start a research joint venture is higher.
2. If the costs of an investment is high, firms tend to start a research joint venture with a higher probability.
monopolize.
4. If the patent length increases, firms will have a higher incentive to monopolize.
Some surplus is lost when firms rather monopolize because you lose the cost reduction of a research joint venture when firms monopolize.
However, allowing firms to collude in a new market is a hard decision for a regulating body. If the new product turns out to be very popular, and the regulator allowed the firms to collude, the firms make enormous profits and a lot of consumer surplus is lost compared to a situation where new entrants were allowed. On the other side, if the regulator does not allow firms to collude, and firms know the market will not have a large demand, the firms will not invest in the first place. Both these situations are unfavorable (Qiu et al., 2009). Therefore, another alternative to a pure cartel and a free market needs to be observed. One possibility is the trigger-point mechanism. The trigger-point mechanism has been used in China for the container port entry. With the trigger-point mechanism, two firms are allowed to collude in the market. However, if demand is sufficient at one point in time, other entrants are allowed from that period to supply the remainder of the demand. This way the innovators can enter and collude once the new product has been produced. If the demand exceeds their supply, other entrants are allowed. This way a new product can be created whereas innovators wouldn’t have produced a new product if they knew that entrants could enter from the second period onward (Qiu et al., 2009).
There are also a few problems with the trigger-point mechanism. The most important one might be that firms that started in the market might actually produce an output that is lower than the monopolized equilibrium. This is because a high price ensures they have the capacity to produce all the demand (Qiu, et al., 2009). However, this is still better than no production of the good at all, so the trigger-point mechanism is still a good procedure.
It would, however, be better to find yet another conditional approach. I therefore suggest an alteration of the trigger-point mechanism that is not based on the amount of demand, but rather on the amount of investments. In this approach, companies apply for the adjusted trigger-point
mechanism by presenting their plans and costs. This way the regulator can easily assess the information. Firms could cheat by giving wrong information. This is discusses at the end of this section. The regulator could then propose a mechanism that allows the firms to collude until they have earned back their investment and earned something with their new product. Once the firms have earned back their investments, or even be allowed to make some profit, other firms can enter the market. Besides making a profit, firms will be interested in innovating because it increases the survival rate of their companies (Cefis & Marcili, 2006). Because firms are ensured they can earn back their sunk costs, innovation is still promoted and encouraged. However, once they have earned their
sunk costs and some profit, other competitors join the market, ensuring a price lower than the monopoly price, creating a situation that increases total welfare. I call this the profitable trigger-point.
There is a way to test this theory. It would however be a very complicated experiment. The experiment would exist of 21 people, ten markets of two firms and a regulator. The regulator will not have any information beforehand. The firms will have a sheet that holds an estimated demand line for a new research. They would also be given an estimation of the sunk costs of the innovation. They also get an estimation of the sunk costs if the two firms would innovate together, which should always be less than half the sunk costs of researching it alone. The combined sunk costs should be lower than the sunk costs of research by a single firm because of information sharing. In the first step of the game, the two players that play as a firm are offered to collaborate with the research, do the research by themselves, or not innovate at all, resulting in a gain of 0. After almost every choice, they can choose to talk to the regulator and after that reconsider their choice. This will be explained later. In this game, I assume the two firms to be equally efficient. If both players choose to conduct the research by themselves, they both have a 50% chance to be the first to develop the new product. The one that is chosen to be the first can enter the market, the one that is not chosen will have a loss of its sunk costs. If one firm wants to collaborate and the other chose another option, the firm that chose the other option will be informed that the other firm wants to collaborate and will be given the choice to also cooperate. If the other person still does not want to cooperate the firm that wanted to cooperate in the first place is given the option to conduct the research himself or not enter the new market. There are 4 different interesting options within each market on how the game will continue: 1. Both companies choose not to conduct research. Then, the new market will not be created, resulting in a welfare increase of 0 and a producer surplus of 0.
2. One company chooses to conduct research, the other company chooses not to conduct research. Now, the company goes to the regulator to ask a patent on the new product. The company can tell the regulator the sunk costs, the expected demand and the marginal costs of the new product. This information can be made up by the firm. The years the patent is active, the company can earn monopoly profits. After the patent, it will become a perfectly competitive market where the profits of the company are 0. The regulator can then choose to believe the information and offer a patent length accordingly or pay a fixed amount to get the real information and offer a patent length according to that information. The company can then accept or decline the offer. The goal of the company is to maximize the producer surplus and the goal of the regulator is to maximize the total welfare. Total welfare can be calculated because a demand line is given.
3. The two companies choose to innovate, but not together. One of the companies will not enter the market and lose its sunk costs, while the other company can proceed in the same way that is done in
step 2.
4. The two companies choose to innovate together. The companies can again give information to the regulator to get a patent together, which the regulator can believe or check for a fixed price.
However, the companies can now also ask for the adjusted trigger-point mechanism. This way, the companies are ensured they will not lose on their investment, and will be allowed to earn a small profit. The regulator can accept or decline the adjusted trigger-point mechanism. If the adjusted trigger-point mechanism is declined, the firms will be offered a patent and if they accept it they will earn Cournot profits for the duration of the patent. After being offered the patent in this case, the companies are given the choice to collaborate illegally. The regulator is asked if he wants to check whether this happens for a fixed fee during every round the patent was given. If the regulator finds out this is happening, the companies will be fined. The choice to collaborate should be given to the firms for multiple rounds to see whether they will collude illegally and in which rounds they will. The regulator should be able to check every market every round for a fixed price per market.
I expect to see the following results from this test:
1. Total welfare is maximized when the companies compete through Cournot competition after innovating, but firms will often not innovate after this offer, so the average welfare after the Cournot proposal will be lower than the average welfare after companies are allowed to use the adjusted trigger-point mechanism.
2. After a proposal of the firms to use an adjusted trigger-point mechanism is declined by the regulator, the regulator will more often choose to check the market with Cournot competition, knowing the firms wanted to collude.
3. There route that has the highest acceptance rate of both the regulator and the companies will be when the firms propose an adjusted trigger-point mechanism, because this will always have a positive payoff for both the regulator and the companies.
4. A regulator that often checks companies to see if they will collude will have a loss of total welfare because of the fixed price it has to pay to check, but it will increase the competition of the firms because firms are frightened to collude.
What is interesting from this experiment is to see if firms would innovate if the regulator would not allow collusion in the market. They would make a loss if they would play competitively every round, so they have to collude illegally in some rounds to earn back their investments. This experiment could be able to prove or disprove my proposal to check whether firms innovate and collude illegally if the fines are sufficiently high.
The adjusted trigger-point mechanism might have some problems in real life that do not occur in the experiment. First of all, a regulator might have problems finding the exact marginal costs and could therefore expect companies to earn more or less than they do in reality. You could
therefore think of a premium that allows the firms to earn more than their investment costs before opening up the market. If marginal costs were estimated too low, firms will still make a profit because of the premium. If marginal costs were estimated too high, it would have been best not to give the premium, but you would still be very close to a perfect scenario.
Another problem is that companies might purposely increase their marginal costs to make sure they can keep the collusion going for a longer time. These companies would still earn back their investment, but it takes them longer to earn it. This is not necessarily a bad thing. The increase of the marginal costs could mean the companies take more care of the product, which is not a bad thing in itself. It could, however, also make the company lazier and therefore less efficient. It could also spend money on frivolous things to artificially increase the costs of the research. This would allow the firms to collude for a longer period because they have higher research costs. This way, the company would be rewarded for their inefficiency. There is, however, a way to avoid this. This solution is not optimal as well, however. The regulator could pay its fixed fee to find out more about the product and agree with the firm how much money the innovating firms can earn with collusion before other firms enter the market. This way, firms will be rewarded for being more efficient because they can make profits if they do not spend too much money innovating the product. The regulating body does need to spend time and money thoroughly investigating the expected costs and profits, so this is not optimal. It would be interesting to conduct research on this matter and find a solution to this problem.
V: Company cost differences
The last part of this paper will focus on market inefficiencies because of cost differences. One of the reasons cartels are seen as welfare decreasing is because they cause X-inefficiencies. X-inefficiencies are the difference between the maximal efficiency that firms can achieve in theory and their real efficiency (Bos & Pot, 2012). Firms do not have the incentive to innovate because they do not have competition. However, collusion can also work to become more efficient. In this part of the paper I will show how companies with different marginal costs can increase total welfare if they form a cartel. Bos and Pot (2012) consider two different kinds of cartels. These are the strong and weak cartels. Strong cartels are cartels that have side-payments to a company that earns less within the cartel. Weak cartels are not giving side-payments within the cartel. In this paper they prove that both kinds of cartels can enhance welfare. The reason these cartels can increase welfare is because one of the cartel members is more efficient than the other.
First, Bos and Pot (2012, pp. 203-204) assume that the total welfare is the profits of the two firms and the consumer surplus, so:
𝑇𝑊 = 𝜋1+ 𝜋2+ 𝐶𝑆 (6)
The profits of a firm are given by the function:
𝜋𝑖= 𝑝(𝑄)𝑞𝑖− 𝑐(𝑞𝑖) (7)
Where ‘i’ can be 1 or 2, denoting the companies. The Consumer surplus is given by the function:
𝐶𝑆 = ∫ 𝑝(𝑥)𝑑𝑥 − 𝑝(𝑄)𝑄0𝑄 (8)
If we substitute (7) and (8) into Equation (6) we will find:
𝑇𝑊 = ∫ 𝑝(𝑥)𝑑𝑥 − 𝑐(𝑞1) − 𝑐(𝑞2)
𝑄
0 (9)
In the first case, Bos and Pot (2012) consider a situation where the costs of company 1 are lower than the costs of company 2. If the two companies collude, I assume they will restrict the output to a lower level than the equilibrium output, which raises the price to a level higher than the equilibrium price of non-collusion. With collusion, the companies can also shift the market share so that
company 1, which has a lower cost of production, will take more of the market share. Total welfare will then increase if the loss in consumer welfare is more than compensated by the increase in welfare of the companies, which translates to:
𝑐(𝑞1) + 𝑐(𝑞2) − 𝑐(𝑞1∗) − 𝑐(𝑞2∗) > ∫ 𝑝(𝑥)𝑑𝑥 𝑄
𝑄∗ (10)
Where 𝑞1∗ is the production of firm 1 with collusion, 𝑞2∗ is the production of firm 2 with collusion and
Q* is the total quantity produced with collusion. Within a strong cartel, this will be possible if the joint profits of the two firms are higher than before the collusion, or they will not cooperate in the first place. Within a weak cartel, not only do the joint profits need to be higher, but the profits of both the cartel players needs to be higher than before the collusion. This is because the companies cannot pay each other. Therefore, a cartel is more likely to be formed when it is allowed to make side-payments to each other (Bos & Pot, 2012). This proves a cartel can enhance welfare when inefficiencies are involved.
There are some limitations to this setting. To increase the total welfare, the cost asymmetry between the firms has to be sufficiently high. If these costs are not very different, consumers might lose a lot of welfare to increase the welfare of the firms. To know exactly how this will play out beforehand is very hard for a regulator. This is because certain synergies between firms could decrease the production costs even more, but it could also be the case that firms have a lot of replacement costs which increases the total production costs (Bos & Pot, 2012). This aspect is not used in the model and could therefore lead to different results in real life as opposed to the outcome of the equation.
Also, this setting does have some alternatives. In a market with perfect competition, the model does not apply because the firm that is less efficient will simply be driven out of the market and a more efficient firm will enter. Also, in a Bertrand competition, the less efficient firm will be driven out of the market and the most efficient firm will be able to reap all the profits. Only in a Cournot competition this shift of production could increase total welfare.
The last problem that arises is that this method keeps an inefficient firm in the market. In a normal market firms that are inefficient could be replaced by firms that are more efficient. I think this would be best for total welfare.
VI. Jurisdictional consequences
In three different cases, I have shown how the formation of a cartel could improve total welfare. However, all these forms of cartels are not yet allowed. To better benefit the economy, we should therefore make some changes to the current legislation. First I will describe the conclusions I have drawn from the case of the externalities. As was shown, a reduction of externalities through the limitation of output can be good for total welfare. There are two ways I discussed to do this: through a cartel and through Pigouvian taxes. As is shown, Pigouvian taxes do not always work. This is because of three reasons: The resources the government needs to research the market, the
conclusion of the government on cartel profits might be wrong and the taxes might be inadequate in the long run. Therefore, cartels could be a good substitution to the current implementation of taxes. However, a lump sum tax combined with Pigouvian taxes can also reach a maximal total welfare. It is debatable and situational which of the two solutions is best. My conclusion is that collusion is better when the collusion output is close to the optimal output. Since the formation of the cartel is most profitable to the producers of the good and because they know most about the market, they should present an idea to the European Commission to be allowed to collude. If the companies can show figures that prove that a collusion increases total welfare, they should be allowed to do so in the market. I think the government will incur less costs verifying the claims of the companies than it would assessing the market themselves. This does need to be researched further. Showing proof that a reduction of output will benefit society will be possible for companies that work in environments with large negative externalities. By showing simple models that show the profits and costs of companies I believe it would be hard for companies to cheat. The collusion after the research joint venture is harder to assess. Both the company and the government do not yet know how much demand there will be for the good after it is produced. This is why it is very hard to present a good law that would work in general. Also, firms are more stable in their illegal collusion when they have joined up for research, so they will have less incentives to open up to the government. A regulatory
solution could be to impose a fine that is high enough to demoralize illegal collusion. A good
alternative is the adjusted trigger-point mechanism, which allows for firms to collude as long as they have not yet earned back their marginal costs. This does require more work for the regulating body to assess the market, but it makes it safer for the companies to invest in new technologies.
Collusion to decrease the costs of inefficiency are perhaps the hardest to regulate. I think it will be very hard for a regulator to assess how inefficient each firm is and what the alternatives are to the collusion. Firms that are more inefficient than other firms might have the possibility to increase their efficiency through better leadership, better allocation of resources or through some other mechanism. Because it will be very hard for regulators that do not have a lot of experience in the market to find out where the inefficiency comes from, it will be very hard to assess what option is the best option.
A general conclusion should be that cartels should be allowed as long as they can prove they become beneficial to the total welfare beyond a reasonable doubt. This way the companies are responsible for the costs to prove the cartel is beneficial. Companies can present the idea with the lowest cost because they know most about the market. There is another reason that would make this rule beneficial. Companies that want cartelization of their market will come the regulating body. If they fail to prove they can be beneficial for the society they will no longer be able to collude illegally, since the regulating body knows their plans and will be able to spot their actions far more easily. Because firms know the government will be able to monitor their actions more closely they might be less inclined to apply for collusion. I think that firms that are absolutely sure they will improve total welfare will have no problem going the government. Only the firms that do not benefit society when they reduce their output or firms that are not sure will hesitate. Increasing fines for colluding illegally will probably reduce this hesitation.
VII. Conclusion
What I have found in this paper is that current regulation of cartels is not always effective in certain markets because the regulation is too simplified. Current regulation does not account for
externalities, markets that can only be started through collusion and cost differences between firms. In the first part of the paper, I found that negative externalities can be reduced or eliminated when a cartel is formed. This is because the cartel will reduce the output of the product, which decreases the social costs. As long as the reduction of the social costs is larger than the reduction of the consumer surplus, it is useful to have a cartel. I have also looked at other solutions that may be better in some cases, like a combination of a Pigouvian tax with a lump-sum tax, but this does not always hold true. In the case of collusion after the innovation of a new product I have proven that companies may not innovate if they are not allowed to collude after the product reaches the market. Some firms
will collude illegally but this can be reduced by imposing a higher fine to collusion. To reduce the negative effects of a cartel, I have proposed a solution that is an adjustment of the trigger-point mechanism. This allows firms to innovate without the risk of incurring a loss when they enter the market with the potential of earning a profit.
The last situation I found is when two firms compete at a market where one firm incurs higher costs than the other. A shift of production with collusion allows the firms to produce more efficiently. This can result in an increase in producer surplus that is higher than the reduction in consumer surplus. Situations are even possible where both the consumer and the producer surplus increase.
I have concluded with the statement that cartels should be allowed when the firms can prove they will increase total welfare beyond a reasonable doubt. This is different from current regulation where firms are always illegal when they limit supply in some way.
Because of these implementations, the government can more precisely assess markets because they can get information at a lower cost when firms apply for collusion.
Limitations and further research
In the case of the externalities, a lot of conclusions are drawn from a paper that is not listed on the Tinbergen list. This is probably because the paper over-simplifies certain areas. However, I believe that the paper is not in the Tinbergen list because the conclusions are too predictable. The fact that marginal social costs increase with externalities is in my opinion very straightforward and right. One over-simplification is that cartels can have social costs in itself that are not accounted for in the model. One of these social costs could be that cartels get lazy because they have no competitors. I do not believe this to be true, because one cartel member can drive a lazy cartel member out of the market when the lazy cartel member becomes too inefficient. This should be tested empirically to see whether it is true.
Another alternative that is presented in this paper is the Pigouvian tax combined with a lump sum tax. I believe this is a very good alternative that will be better as long as the costs of the
government are not greater than the triangle B of Figure 2 (Torre & Morgen, 2010) and Figure 3. This is because these are sunk costs that are not beneficial to anybody. If the producer surplus is as important as the consumer surplus, these sunk costs should be minimized in order to maximize welfare. It should be researched further which of the two costs is greater in most circumstances. With the adjusted trigger-point mechanism, the idea is novel and has not yet been tested. Therefore, the conclusions I have derived from it are not certain. Also, the test I have proposed might not be similar to real life. This is because in real life, regulators might not be able to find out the real market values even after paying to check the market. Also, in real life the regulator will be more
informed about how much the companies should be fined, making it a better regulator. The last thing that might not be representative is the fixed costs of checking the market. Normally, the search for cartels is done on a large scale, with the possibility of overlooking the market in which a cartel is active. Since there is only one market in the experiment, the regulator is aware that collusion might take place. Further research on this should be testing the experiment and its results, and find possibilities that enable the regulator to make the right judgement.
In the case of market inefficiencies because of cost differences between firms, I have already stated that is very hard to propose a way that regulators can be sure collusion is the best answer. This is because they do not have the expertise to find out where the inefficiencies come from. Further research should be conducted to enable ways to make researchers take the right choices. Also, there are not a lot of papers regarding cartels and market inefficiencies because of cost differences, so most of the conclusions are drawn from just one paper.
As a last limitation, some alternatives to cartels might not have been mentioned in this paper. I believe I have found the best and most used alternatives to cartels, but it could be that some are overlooked. To show the flaws of every alternative is also not the focus of this paper. What I have tried to show is that cartels are a very viable welfare-enhancing option in certain cases.
