Losers and Welfare
What is the relevant welfare criterion for economic analysis?
Whenever I write about welfare, I try to do so using Pareto efficiency as the criterion. Judging by the feedback that I receive when I write about welfare, people do not like this. However, I would argue that if you want to do objective welfare analysis, this is the only criterion we have. Every other criteria that people use requires an implicit ethical judgement. What’s worse, the failure to recognize this can lead to political responses or outcomes that surprise people. There should be no reason for such surprise. Understanding winners and losers does a lot of the work when it comes to political economy. In addition, the Pareto criterion seems to be the most empirically relevant criterion as well.
Two Predominant Criteria
There are two main criteria that economists tend to use to evaluate welfare. A Pareto improvement occurs when someone or some people can be made better off without any becoming worse off. The is the Pareto criterion. Pareto efficiency occurs when no such improvements remain.
A separate approach to welfare evaluation is what is known as the Kaldor-Hicks criterion. This criterion states that something is welfare-improving if the total gains exceed the total losses.
These concepts are not unrelated. Consider any sort of policy change. It is likely to create both winners and losers. The winners are the people are who better off. The losers are the people who are worse off. If these new benefits exceed the new costs, then this is considered a Kaldor-Hicks improvement. If the winners use some of their benefits to compensate the losers, then the outcome could also be a Pareto improvement. Why? Well, if the benefits exceed the costs, then some fraction of the benefits could be transferred to the losers with some benefit still left over. Some people are better off and no one is worse off.
For this reason, many people think that Kaldor-Hicks is sufficient. After all, if something is a Kaldor-Hicks improvement, a transfer is sufficient to generate a Pareto improvement. However, I don’t think that things are this simple. There are a couple of issues that need to be addressed.
First, a Kaldor-Hicks improvement might be sufficient for thinking about this on the chalkboard, but we need to think about things in practice. Just because a transfer could produce a Pareto improvement doesn’t mean that we will see such a transfer in reality. There might be important political constraints or transactions costs that prevent such transfers from taking place. If a transfer is feasible and likely, then economists have an objective standard upon which to argue in favor of the policy. If a policy makes some people better off without harming others, it doesn’t require any ethical judgement to say that the policy should be implemented.
On the other hand, if the transfer isn’t feasible, arguing that the policy should or should not be implemented requires an ethical judgement. One can appeal to Kaldor-Hicks to argue in favor of the policy, but in doing so one is implicitly making a utilitarian argument. That’s fine, if one is a utilitarian, but one should be explicit about that ethical judgement.
This isn’t a minor point about ethics and economics. Instead, I think that this distinction is often at the heart of political economy.
The second major point here is about which criterion leads to more empirically relevant predictions about what we actually see in the world. For example, how do people handle issues related to common pool resources? How do people deal with externalities? If you use the Pareto criterion for thinking through these issues, you might get different implications and predictions than if you use the Kaldor-Hicks criterion.
Winners, Losers, and Political Economy
Price discrimination (charging different prices to different consumers for the same good) tends to be unpopular. However, this isn’t universally true. There are many types of price discrimination that people do not complain about. People who show up at the store with coupons pay a different price than people who do not. Perhaps it happens, but I’ve never seen a person complain that the person with the coupon paid a lower price.
My own casual empiricism suggests that people tend to dislike price discrimination the closer it gets to first-degree price discrimination, in which every consumer gets charged their exact willingness to pay. For example, Brian previously wrote about accusations that Instacart was varying prices across consumers to try to get a sense of each individual customer’s willingness to pay for particular goods. That turned out to be incorrect, but a lot of people seemed really concerned about what Instacart was doing. They didn’t like it.
In a Kaldor-Hicks sense, price discrimination is always welfare-improving. However, this is not always true when it comes to the Pareto criterion. Let’s consider the logic and then a couple of scenarios.
A profit-maximizing firm will set marginal revenue equal to marginal cost. The marginal revenue of a price-setting firm that faces a downward-sloping demand curve will always be below its price. Thus, a profit-maximizing, price-setting firm will always set its price above the marginal cost of production. But think about the marginal consumer. Suppose the firm sets a price of $10 and the marginal cost is $7.50. The marginal consumer might be willing to pay $9.50 for the good. This is lower than the price that the firm is charging, but higher than the firm’s marginal cost. There is a potential gain from trade here that is going unrealized. Selling the good to this one consumer for $9.50 would allow the consumer to actually obtain the good, while also increasing the profit of the firm.
Nonetheless, the firm cannot lower the price to $9.50 for everyone. By doing so, the firm would gain $2 from that one consumer, but would lose $0.50 from every other consumer who would have been willing to pay $10.
Price discrimination can improve welfare using either criteria. Suppose the firm charged $10 to everyone and $9.50 to this one particular consumer. None of the previous consumers are made worse off. The new consumer by definition is indifferent between having $9.50 and having the good. The firm earns additional profits. The firm is better off and everyone else is no worse off.
Now consider moving from the $10 price to first-degree price discrimination. The total surplus from trade is maximized. By the Kaldor-Hicks criterion, this is a welfare improvement. Nonetheless, this is not a Pareto improvement. None of the consumers are better off. Many are worse off (all of those with a willingness to pay above $10). The firm is better off.
It therefore shouldn’t a surprise that consumers tend to dislike first-degree price discrimination.
A similar type of reasoning can help us to understand arguments over international trade. In a lot of standard trade models, you get the following type of result. Removing barriers to trade allows consumption beyond the production possibilities frontier. In other words, people are able to consume more than the productive capacity of their own economy because they exploit the new gains from trade. However, at the same time, the changing nature of production following the removal of the trade restriction will tend to lead to the convergence of factor prices. This means, for example, that workers in the high-wage country will tend to see downward-pressure on their wages. Overall, the benefits exceed the costs. Nonetheless, there are winners and losers. To turn the Kaldor-Hicks improvement into a Pareto improvement requires transfers to these workers. When there are no such transfers, it shouldn’t be surprising that some workers are opposed to the policy.
In short, what the distinction between Pareto improvements and Kaldor-Hicks improvements reveals is that a lot of issues in political economy revolve around distributional effects. When there are distributional effects that produce winners and losers, it is not surprising that the losers are not persuaded by the Kaldor-Hicks argument that the overall benefits exceed the overall costs. Furthermore, distributional effects make it difficult for economists to make normative arguments about the policy because it necessarily requires using some ethical standard.
Empirical Relevance
I view the main job of an economist as an attempt to understand and explain the world as it is. Prior to making any recommendations to improve the world, one should have some understanding of how things work.
Which criterion we use can matter for thinking about how people solve problems. For example, when we are thinking about markets as I described in my example above, property rights are well-defined and we have a sense of how people would behave. However, there are a lot of interesting questions that we think about where property rights are undefined or ill-defined. (As a testament to how interesting these types of issues are, just search this Substack for the posts that Brian and I have written on externalities.)
In a previous post, I discussed the example of externalities using the Pareto criterion. I tend to prefer this approach because it points to a Coaseian discussion. What that post was meant to point out is that even in the presence of externalities, we can think in terms of the Pareto criterion. Once we introduce all parties involved (consumers, firms, and those who suffer from the external cost), it should be clear that there are gains from trade to be had between the groups. In other words, it is possible to move in the direction of the optimal allocation even in the Paretian sense. This is a Chicago-style price theoretic approach to Coase.
What that example revealed is that the Pareto criterion can actually help us to understand how and why society tends to move in the direction of the optimal quantity. However, this approach requires a solution that is “off the demand curve.” If consumers are able to choose whatever quantity of consumption they like, such a solution is infeasible. To prevent that from taking place, you need some type of non-linear pricing scheme or you need some type of explicit agreement between the parties involved.
In that previous post, I focused on the non-linear pricing scheme, but here I want to focus on the alternative. When we think of issues related to incomplete, ill-defined, or undefined property rights, we actually see a lot of voluntary activity to mitigate the adverse effects thereof. Humans tend to form a variety of groups to solve these types of problems. These groups include families and homeowners associations and churches.
As a result, externalities don’t always require Pigouvian taxation. Common pool resources do not always lead to the Tragedy of the Commons. As the work of Nobel laureate Elinor Ostrom shows, societies turn out to be remarkably good at forming groups to sustain and manage common resources.
The Pareto criterion approach to these issues pushes one in a Coaseian direction and pushes one towards a better understanding of the institutions that make up a civil society. As a result, it seems (at least to me) to be much more empirically relevant for understanding how the world actually functions.
Some Final Thoughts
It might seem like a minor thing, but the distinction between the Kaldor-Hicks criterion and the Pareto criterion is important. Narrowing in on the differences between these criteria can help one to avoid making implicit ethical judgements and to understand the importance and significance of distributional issues for political economy.
But it is also useful to understand these distinctions for thinking about existing policies and solutions. The Pareto criterion leads one in the direction of Coase and Ostrom when it comes to externalities and property rights issues. That is an advantage, in my view, because it points to the development of the institutions that make up civil society, which seem much more empirically relevant than a lot of textbook conversation and textbook solutions to these issues.