What is Optimal Policy?
What is Optimal Policy?
Economists use the phrase all the time. But what is it?
You are reading Economic Forces, a free weekly newsletter on economics, especially price theory, without the politics. You can support our newsletter by signing up here:
What should we do about student loans? I don’t know.
But this policy dispute has again raised the more interesting question of the role of economists in policy disputes. Some economists, including those that supported and opposed Biden’s executive order, responded by saying economists need to get more involved in policy and politics. This shows why economists need more power in politics, so the argument goes.
Then Jennifer Doleac triggered Twitter by expressing this sentiment.
Most people glossed over the first part, which seems obviously true. As Stuart Ritchie has written, it's a bad thing when science is political. Academic economists have no expertise in politics.
The “pushing in the direction of optimal policy” boiled everyone's blood. I think it will be helpful to clarify what economists often mean by things like "optimal."
Positive vs. Normative Misses Much
The standard distinction in economics is between positive and normative economics. Positive is “what is,” and normative is “what ought to be.” Through this lens, Jennifer’s tweet is normative since she mentions optimal policy. Optimal means you are making a normative statement which means it's just your preference of what policy should be, or so goes the argument.
Yes and no. Economists mean different things by optimal, so I wrote my first paper in grad school, which ended up in the Journal of Economic Methodology.
Unfortunately, despite what you heard in Econ101, the positive/normative distinction fails to capture how economists use models, especially optimal policy models. For example, consider the following two uses of normative. Kocherlakota, in his book on the New Dynamic Public Finance, says
This book is normative... It tries to figure out what taxes we should have... The ultimate goal of (New Dynamic Public Finance) is to provide relatively precise recommendations as to what taxes should be. (emphasis in original).
Normative = should be in this formulation. We could add optimal = should be.
But this isn’t the only way economists use normative. Atkinson and Stiglitz have a classic graduate public finance textbook. In their introduction to the section ‘Normative Analysis’ they write
The aim of the Lectures that follow is not to argue the case for particular policies; it is not their intention to provide an answer to the question, ‘what ought the government to do?’... [O]ur concern here is with the structure of arguments rather than with the arguments themselves. The aim is to explore the relationship between specified objective and the policy recommendations to which these objectives lead... (emphasis in original)
It’s explicitly not about what ought policy to be but the structure of arguments. That’s really central to economics. It is about the structure of arguments.
We can see economists’ focus on the structure of arguments through their emphasis on models. After all, models are just a bunch of if-then statements, which are just a series of arguments. For example, the basic supply and demand model can be used to say, "if a price control below the equilibrium price is imposed, then there will be a shortage.”
Instead of positive-normative, following Fritz Machlup, I separate models into “positive,” “normative,” and “instrumental.” Models can be positive: A causes B in the real world. Models can be normative: B is good. Models can be instrumental or means-ends: if you want B, A will get you there. Atkinson and Stiglitz are talking about instrumental analysis.
Instrumental analysis is much more in economists' wheelhouse than normative analysis. Economists have lots of tools to study if-then statements. They have no tools to study purely normative questions.
The confusing part is that the models we use for policy (not purely positive) are not purely normative or instrumental either. They fall in between. For a more concrete formulation, take standard optimal taxation results as you’d see in Piketty and Saez’s handbook chapter on optimal labor income taxation. The argument is something like
If a government wants to maximize a weighted sum of individual utilities, and
if a government has access to linear taxes and no lump-sum taxes, and
if a government has information on wage earnings, and
if each citizen has a given productivity and utility function,
then a government should set taxes at τ .
The crucial role of economics is to derive 5 from step 1 through 4. Notice Jennifer uses the word “transparently.” Models push us to lay out each step as clearly as possible, so we can debate them.
Notice what Piketty and Saez are not doing. They are not simply stating their policy goals. As I say in the paper
The deduction does not mean that the person doing the deduction thinks that any part of the deduction should happen. The deduction and the recommendation are connected, but conceptually distinct. To take a more extreme example, if the government’s goal is to kill all puppies, no logical deduction about the means and ends has any reasonable connection to what the economist thinks should happen. The instrumental aspect of a model is conceptually distinct from the normative aspect.
The paper may reflect the author’s policy goals, but that’s not where the debate is. The debate is about the logical or empirical heft of the rest of the argument. On those grounds, economists decide whether to be interested in the model or not.
In practice, there is basically no debate on part 1 but tons of focus on how reasonable 2 through 4 are. That’s maybe not ideal, but that’s where economists are. In the paper and this newsletter so far, I am making a claim about how economists use these models, not how they should use them. I’m making a positive claim about normative instrumental models. It gets very meta very quickly.
Beyond the Model to Actual Policy Advice
So far, I have focused on well-defined optimal policies of the type you will find in the optimal taxation literature. That’s the easy case since the whole point of the academic paper is to be explicit about the mapping from assumptions to conclusions.
But economists are (allegedly) human. We can’t help but slip our own beliefs into discussions. I don’t work with models that start with step one of “If we want to kill all the kitties.”
When we are discussing among economists skipping steps is less problematic. Economists are like the joke about the old men who sat around calling out numbers “3” and “28.” In the joke, the old men had heard each other's jokes so many times they just gave them all numbers to save time. Economists can get like that when talking about models. When an economist says something like “free trade is optimal,” I can be fairly confident of the type of model to think about (pure comparative advantage) and in what sense they are using optimal (Pareto efficiency or a weaker notion).
Is it true that in all models and for all definitions of optimal it follows that “free trade is optimal?” No. Of course not. It’s easy to come up with a model where reducing trade barriers can lower some people's wages, and those people likely don’t think reducing trade barriers is optimal.
Some people will jump on that example as proof that economists' claims about optimal policy are just them imposing their neoliberal agenda on the rest of society. We are just jumping to set 5 in the deduction. People responding to Jennifer’s tweet above thought they were dunking by suggesting that “optimal” meant “arbitrary preference of the economist.” Sorry. No basket.
Let me get a bit philosophical for a second. “This is my newsletter. I can do what I want,” he screams, wondering why only five people read him.
We must recognize that human communication will never allow us to outline every possible case. Finding an exception to a statement doesn’t always destroy the statement.
My philosopher friends here will help me with some reference to Wittgenstein or something, but I often talk casually about “meaningful” statements. Suppose we recognize that communication and information processing are scarce resources and that any communication is partly about two people trying to convey information to each other. In that case, we must rely on meaningful statements, since we can't lay out everything.
The sentence “professional basketball players are tall” is not strictly true for all definitions of professional basketball players and all definitions of tall. If you include retired players, Muggsy Bogues was 5 foot, 3 inches. If I tweet, “professional basketball players are tall,” and you reply tweet, “what about Muggsy Bogues?” I’m going to ignore you. You’re not wrong, just annoying.
But you’re not right either. The opposite statement, “professional basketball players are short,” is clearly a silly premise, so what are you suggesting?
The statement wasn’t rendered arbitrary by finding a counter-example. My original statement conveyed meaning, was likely sufficiently precise for the context, and you’re a pedant. I understand that “sufficiently precise” is context-dependent. Again, this communication thing is tricky. In the same way, the optimal taxation model is not arbitrary because you can find a counter-example. It is a way of laying out a structure or reasoning.
For another example of what I mean by “meaningful statements,” suppose I say, “donating malaria nets is altruistic.” Then an overly eager econ student comes along and says, “no one does anything altruistically; all behavior is self-interested.” I point out that I understand we can redefine all behavior to be self-interested and put “stopping malaria” in the person’s utility function. Nevertheless, there is a meaningful sense in which I can say that donating malaria nets is more altruistic than buying a yacht for oneself. You know what I mean. I know you know what I mean. If you want to debate that claim outside of a super narrow context in philosophy departments, I’m simply leaving. Time is scarce.
The same is true when economists say “free trade is optimal.” There are exceptions, but it is still a meaningful statement. It is more complex than the Muggsy Bogues example, and it is up to economists to clarify the exceptions, but it’s not an arbitrary statement. It’s not simply the economist voicing their policy preferences as if they are saying, “I like pizza.” It’s not politics. Yes, it bakes in normative assumptions, but that’s not the whole pie. Economists tend to think that giving people more of what they want is, all else equal, a good thing. You can debate that. What about children who wish to consume drugs? Sure. But 99% of the time, I have no problem glossing over some people’s paternalistic goals when I say “optimal.”
I also think that many policy debates, especially the ones that economics can help us answer, are not mainly disagreements about the goals. When we disagree about trade barriers, I don’t think the other person wants them because they hate when people generate more value. We don’t disagree at the core about what is desirable but about how to achieve the desirable. There are exceptions. (Damn those exceptions.) You and I could agree on the likely effects of a minimum wage law but disagree about how to trade off workers vs. customers vs. employers, etc. That’s a normative, political dispute. That doesn’t imply that all debates, even all debates about minimum wage, are about how much to value gains or loses to different groups. There is still plenty of dispute about the facts.
And if the debate is about the facts or the best means to achieve some ends, that’s a debate that can rely on the science of economics. I refuse to concede that it’s just politics or preferences all the way down. No. The science matters and can help us live more prosperously together. There. My cards are on the table. I like prosperity. That’s normative.