Neither side “supports” libertarian conclusions. Both sides make contributions to the theories that explain such conclusions. Austrian school economics is at the a-priori end and the Chicago school is at the empirical end of what is better seen as a continuum rather than exclusive alternatives.
Does the Austrian school of Mises and Rothbard or the Chicago school provide better economic arguments for libertarianism?
If by "better" we mean "scientific" then the answer is given by Mark Blaug who says that Austrian economics is “an antiempirical undertone in the history of continental economics that is wholly alien to the very spirit of science”.
With the Austrians, one can neither evaluate the results, nor give empirical proofs, the only admissible argument is to say that it is obvious that freedom is good.
But if this evidence is not scientific (empirical), it can only be psychological.
And it's not because psychologically we see things this way that things are like that.
Post strictum: If we can't evaluate the results, I wonder how we can conclude that freedom is better?
The Austrian argument would be that the axioms, rules of human behavior, are observed, if only in ourselves, which is a reason to support conclusions deduced from them. Mathematics is not empirical — does that make it unscientific?
Mises argues that the empirical method can't work for economics because there are too many uncontrolled variables to do controlled experiments. I think he was mistaken, but it is worth at least looking at his arguments.
"Can't work" is too strong but the Austrians have a strong point. The history of standard economics, especially as seen in its models and predictions, suggests that economics in the 21st century is still unable to make reliable predictions. Economists' forecasts of even very near-term variables such as GDP growth, unemployment, and inflation are no better than random, maybe worse.
There are predictions economics cannot make reliably, other that it can, such as the effects of price controls or of printing lots of money. An economy is a very complicated system.
“rules of human behavior, is a reason to support conclusions”
A reason, yes, necessary but not sufficient.
Science is logic + empirical refutation.
Mathematics is not scientific; it is only a logical language. The logic of action (rules of human behavior) is not scientific, it is not sufficient to predict anything. Empirical hypotheses must be added, for example an empirical hypothesis on the convexity of the production set and preferences, a budgetary constraint, an assumption on the concrete intensity of competition, on the absence of externalities, etc… Only then can we support conclusions, show that freedom leads to the optimum. The logic of action alone cannot do this.
The "logic of action" is insufficient to deduce empirical implications, and therefore to build an economic science. It is insufficient, it is only when combined with other empirical hypotheses that we can deduce refutable empirical predictions and make it a science.
The "logic of action" is insufficient to deduce the normative implication that freedom is a good thing. To say that freedom is good, we need to have a way of evaluating, a norm. Of course, we cannot make inter-individual comparisons of utility, but we can use the Pareto criterion to conclude that the competitive equilibrium is optimal.
"but we can use the Pareto criterion to conclude that the competitive equilibrium is optimal."
I don't think so. The proposition "this outcome could not be improved for one person without making someone else worse off," which is what Pareto-optimal means, is true of lots of outcomes — a world with laissez-faire plus one exception, say a steel tariff, cannot be improved without making the beneficiaries of the exception worse off. To get a stronger criterion you need to replace Pareto with Hicks/Kaldor aka Potential Pareto. But since the compensating payments cannot actually be made, that's just a way of pretending to do Pareto while actually doing economic efficiency, introduced by Marshall as economic improvement and depending for its justification on interpersonally comparable utility.
I have always thought that Pareto-optimality was a bizarre and, in practice, useless concept. Maybe even worse than "perfect competition" with its impossible conditions.
Yes, you are right, very often, "compensating payments cannot actually be made". But this is a practical problem, not a theoretical one.
Whereas the impossibility of evaluating theoretically prohibits saying that freedom is better. This is the heart of your problem. This is a fundamental reason why I think the Chicago school provides a better argument than the Austrian school.
A theoretical impossibility seems to me to be more problematic than a practical difficulty, especially since sometimes compensatory payments are possible. For example, in Europe, we want to move from unanimous voting (which satisfies the Pareto criterion) to coalitions (which requires negotiations that take place "as if" compensatory payments are made).
If compensating payments can be made than you make them and the change is an actual Pareto improvement. The only situations where you need Hicks/Kaldor to support a change are ones where they cannot be made, which is virtually all large number situations.
It is a theoretical problem because if you are going to make compensating payments that affects the incentives of the actors, moving you away from the H/K optimum. Better to admit that you are adding up utilities as per Marshall than to pretend you are not.
Can you offer any other convincing reason why a H/K improvement without compensating payments is a good thing?
A lump sum payment does not affect the recipients' incentives and it is financed by a Pigouvian tax on the others. For example, petrol is taxed to reduce CO2 emissions and an "energy cheque" is given to the poorest.
Inter-individual comparisons of utilities are impossible, OK, but I will add: for the moment. Here again, a practical problem, not a theoretical one.
One reason why an H/K improvement without compensatory payments is a good thing? No, because it is not a normative criterion but a descriptive one.
Austrians emphasize that you cannot compare interpersonal utility. This makes it more difficult to show that a policy makes things better or worse. But surely Austrians also emphasize the point that people in a free market expect to benefit from a voluntary transaction. Therefore, any policy that restricts that freedom can be expected to make them worse off.
I used to be well steeped in Austrian economics in the 1980s but am very rusty now, but this seems correct to me.
X bidding up the price of Y's house makes Z, who also wants it, worse off, yet the transaction is voluntary. To show that it produces net benefits you have to be able to add up the costs and benefits of all three. Without some way of comparing personal utilities, or some equivalent, you cannot show that a change is an improvement as long as it makes at least one person worse off.
I was talking about a two-party exchange. You are right, of course, that effects on third parties could be positive or negative.
To dismiss the Austrian points about gains from exchange, you would require an assumption that third-party effects are more negative than positive generally. And to push that point, you would have to be anti-market.
Whether influenced more by Chicago or Austria, I think the vast majority of libertarian-adjacent people would agree that the presumption should be that free exchanges are mutually beneficial, and likely to be overall beneficial -- in the absence of genuine and substantial externalities.
Personally, I don't believe that you can add up costs and benefits objectively or accurately, but you can get a sense of what they are and their general magnitude. Much more than than is "the pretense of science" as some obscure person once said. This is one reason (not the only one) why I am not a utilitarian. (Note: David, I know you have explained that you don't consider yourself a utilitarian -- but you do place consequences as extremely important in decision making. So I'm not implying that you are here!)
"To dismiss the Austrian points about gains from exchange, you would require an assumption that third-party effects are more negative than positive generally. "
What does that mean? If effects on different people can't be added up, how can you say whether the net third-party effect is positive or negative?
The economic argument is to show that the gain to the gainers, measured by the dollars they would pay to get it, is greater than the loss to the losers, similarly measured. You then argue that in most contexts there is no reason to expect a dollar to represent more utility to one random person than to another. But that breaks down in contexts where there is a reason to expect the losers to have a higher marginal utility of income than the gainers, most obviously if the losers are poorer than the gainers.
There are a variety of problems with that approach, which I discuss in Chapter 2 of _Law's Order_ (http://www.daviddfriedman.com/Laws_Order_draft/laws_order_ch_2.htm), but it is a more coherent argument than "to push that point, you would have to be anti-market", since part of the point of the analysis is showing if you should be pro-market.
"Hence, an Austrian economist understands that it is impossible to show that a tariff, a minimum wage, or rent control makes the U.S. worse off. A Chicago economist doesn't."
My claim is that a Chicago economist can show a reason to believe that those things lower total utility, although not a proof that they do. Do you agree and object that one cannot show that lowering total utility makes the US worse off, indeed cannot show that anything short of making some people worse off and nobody better off, should count, or disagree?
“Since Austrians means economists are committed to the idea that economic conclusions must be based on theory not evidence, an Austrian economist qua economist cannot use the fact that policies have had bad effects in the past as evidence that they will in the future. “
The categories here are ambiguous. Economic conclusions include theoretical conclusions, estimates, predictions, policy recommendations, and historical reports. My steelman would be that even if empirical data tempted an Austrian theorist to revise a theory, the revised theory would not be based on the data, but on a priori propositions; and if estimates or predictions were to be made, past data might be usefully employed in some way. But such predictions are profoundly conjectural, and data alone can’t refute a theory.
This can still be interpreted as making the theory tautological, which is the basic objection. A more charitable view is that a theory can only be criticized in comparison with an alternative theory. So the framing seems to matter. Austrians are not the only economists who say it takes a theory to beat a theory.
So what would it take to get Austrians to revise their theories, or abandon them for new ones? I’m not sure. But I note that even non-Austrians, including quite a number of economists with a more leftish bias, have not abandoned some basic theoretical approaches that are not well supported by data.
Milton Friedman was quite willing to make use of theoretical assumptions that contradicted empirical reality. Of course, he wanted to use these to generate interesting testable hypotheses, and then test them. That project seems feasible, but it isn't clear what the universality of such hypotheses is. They are either compatible with the data or not, but a large (infinite?) number of such hypotheses exist, and no usable theory could include an exhaustive list of potentially relevant factors. Inevitably, exogenous factors can swamp a particular historical instance.
But this is out of my depth. Perhaps our host can clarify.
The connection between theory and policy recommendations is particularly opaque for me. Some people seem to genuinely believe that libertarianish policy recommendations have trended toward disaster, while the alternatives muddle through depending on the honesty and competence of the implementers. I'd like to see a real test where the subjects get to choose somehow, but It is less about refuting false hypotheses and more about letting people try things out and stick with what they like. I realize this is not possible in all cases, but I suspect it could be more widely practiced than it is.
Humans can make mistakes in their reasoning — Rothbard makes many, Mises at least one I have spotted. If your theory produces a prediction that is inconsistent with the data that could mean that you made a mistake in getting from your axioms to your conclusion.
Sure. But the inconsistency being caused by a logic error is one possible explanation. Another is that the data are flawed. A third is that one of the premises is false. Maybe there are more.
But if I really had my finger on this, I’d be able to come up with a historical example when researchers stuck with an old theory in spite of anomalous data and later were vindicated. Or at least make my point more briefly and clearly. I’ve never been able to make up my mind which side I agree with, or whether I think they’re both wrong.
By the “Chicago School” do you just mean mainstream neoclassical economics and freshwater macroeconomics (new classical/RBC)?
Or do you mean the specific “Chicago approach” which might include price theory, public choice theory, preference for markets over intervention, and emphasis on structural models relative to reduced-form approaches?
For the first debate I wanted to make it "Marshallian economics." Gene argued, probably correctly, that most of the audience at Porcfest would not know what that meant, so I agreed to make it Chicago economics.
In this post, I think the only point which might be limited to Chicago economists is footnote 4, since Chicago economists are more likely to be pro-market than other Marshallians. But I was taking advantage of Gene's terminology in the first debate, "The Austrian economics of Mises and Rothbard," to limit the other side to the Misesian branch of Austrian economics, both because I don't know enough about other variants and because Hayek arguably has been influential, mostly via _The Road to Serfdom_.
On the issue of Austrian vs. Chicago economics: although I recall not loving the book, Mark Skousen's "Vienna and Chicago" might be a helpful resource. https://www.amazon.com/Vienna-Chicago-Friends-Foes-Free-Market/dp/0895260298
It does seem like Hayekian economics might be more useful than Misean or Rothbardian. When I think “Austrian Economics” I think Hayek, at least.
Neither side “supports” libertarian conclusions. Both sides make contributions to the theories that explain such conclusions. Austrian school economics is at the a-priori end and the Chicago school is at the empirical end of what is better seen as a continuum rather than exclusive alternatives.
https://jclester.substack.com/p/critical-rationalism?utm_source=publication-search
Your problem is:
Does the Austrian school of Mises and Rothbard or the Chicago school provide better economic arguments for libertarianism?
If by "better" we mean "scientific" then the answer is given by Mark Blaug who says that Austrian economics is “an antiempirical undertone in the history of continental economics that is wholly alien to the very spirit of science”.
With the Austrians, one can neither evaluate the results, nor give empirical proofs, the only admissible argument is to say that it is obvious that freedom is good.
But if this evidence is not scientific (empirical), it can only be psychological.
And it's not because psychologically we see things this way that things are like that.
Post strictum: If we can't evaluate the results, I wonder how we can conclude that freedom is better?
The Austrian argument would be that the axioms, rules of human behavior, are observed, if only in ourselves, which is a reason to support conclusions deduced from them. Mathematics is not empirical — does that make it unscientific?
Mises argues that the empirical method can't work for economics because there are too many uncontrolled variables to do controlled experiments. I think he was mistaken, but it is worth at least looking at his arguments.
"Can't work" is too strong but the Austrians have a strong point. The history of standard economics, especially as seen in its models and predictions, suggests that economics in the 21st century is still unable to make reliable predictions. Economists' forecasts of even very near-term variables such as GDP growth, unemployment, and inflation are no better than random, maybe worse.
There are predictions economics cannot make reliably, other that it can, such as the effects of price controls or of printing lots of money. An economy is a very complicated system.
“rules of human behavior, is a reason to support conclusions”
A reason, yes, necessary but not sufficient.
Science is logic + empirical refutation.
Mathematics is not scientific; it is only a logical language. The logic of action (rules of human behavior) is not scientific, it is not sufficient to predict anything. Empirical hypotheses must be added, for example an empirical hypothesis on the convexity of the production set and preferences, a budgetary constraint, an assumption on the concrete intensity of competition, on the absence of externalities, etc… Only then can we support conclusions, show that freedom leads to the optimum. The logic of action alone cannot do this.
The "logic of action" is insufficient to deduce empirical implications, and therefore to build an economic science. It is insufficient, it is only when combined with other empirical hypotheses that we can deduce refutable empirical predictions and make it a science.
The "logic of action" is insufficient to deduce the normative implication that freedom is a good thing. To say that freedom is good, we need to have a way of evaluating, a norm. Of course, we cannot make inter-individual comparisons of utility, but we can use the Pareto criterion to conclude that the competitive equilibrium is optimal.
"but we can use the Pareto criterion to conclude that the competitive equilibrium is optimal."
I don't think so. The proposition "this outcome could not be improved for one person without making someone else worse off," which is what Pareto-optimal means, is true of lots of outcomes — a world with laissez-faire plus one exception, say a steel tariff, cannot be improved without making the beneficiaries of the exception worse off. To get a stronger criterion you need to replace Pareto with Hicks/Kaldor aka Potential Pareto. But since the compensating payments cannot actually be made, that's just a way of pretending to do Pareto while actually doing economic efficiency, introduced by Marshall as economic improvement and depending for its justification on interpersonally comparable utility.
I have always thought that Pareto-optimality was a bizarre and, in practice, useless concept. Maybe even worse than "perfect competition" with its impossible conditions.
Yes, you are right, very often, "compensating payments cannot actually be made". But this is a practical problem, not a theoretical one.
Whereas the impossibility of evaluating theoretically prohibits saying that freedom is better. This is the heart of your problem. This is a fundamental reason why I think the Chicago school provides a better argument than the Austrian school.
A theoretical impossibility seems to me to be more problematic than a practical difficulty, especially since sometimes compensatory payments are possible. For example, in Europe, we want to move from unanimous voting (which satisfies the Pareto criterion) to coalitions (which requires negotiations that take place "as if" compensatory payments are made).
If compensating payments can be made than you make them and the change is an actual Pareto improvement. The only situations where you need Hicks/Kaldor to support a change are ones where they cannot be made, which is virtually all large number situations.
It is a theoretical problem because if you are going to make compensating payments that affects the incentives of the actors, moving you away from the H/K optimum. Better to admit that you are adding up utilities as per Marshall than to pretend you are not.
Can you offer any other convincing reason why a H/K improvement without compensating payments is a good thing?
A lump sum payment does not affect the recipients' incentives and it is financed by a Pigouvian tax on the others. For example, petrol is taxed to reduce CO2 emissions and an "energy cheque" is given to the poorest.
Inter-individual comparisons of utilities are impossible, OK, but I will add: for the moment. Here again, a practical problem, not a theoretical one.
One reason why an H/K improvement without compensatory payments is a good thing? No, because it is not a normative criterion but a descriptive one.
Austrians emphasize that you cannot compare interpersonal utility. This makes it more difficult to show that a policy makes things better or worse. But surely Austrians also emphasize the point that people in a free market expect to benefit from a voluntary transaction. Therefore, any policy that restricts that freedom can be expected to make them worse off.
I used to be well steeped in Austrian economics in the 1980s but am very rusty now, but this seems correct to me.
X bidding up the price of Y's house makes Z, who also wants it, worse off, yet the transaction is voluntary. To show that it produces net benefits you have to be able to add up the costs and benefits of all three. Without some way of comparing personal utilities, or some equivalent, you cannot show that a change is an improvement as long as it makes at least one person worse off.
I was talking about a two-party exchange. You are right, of course, that effects on third parties could be positive or negative.
To dismiss the Austrian points about gains from exchange, you would require an assumption that third-party effects are more negative than positive generally. And to push that point, you would have to be anti-market.
Whether influenced more by Chicago or Austria, I think the vast majority of libertarian-adjacent people would agree that the presumption should be that free exchanges are mutually beneficial, and likely to be overall beneficial -- in the absence of genuine and substantial externalities.
Personally, I don't believe that you can add up costs and benefits objectively or accurately, but you can get a sense of what they are and their general magnitude. Much more than than is "the pretense of science" as some obscure person once said. This is one reason (not the only one) why I am not a utilitarian. (Note: David, I know you have explained that you don't consider yourself a utilitarian -- but you do place consequences as extremely important in decision making. So I'm not implying that you are here!)
"To dismiss the Austrian points about gains from exchange, you would require an assumption that third-party effects are more negative than positive generally. "
What does that mean? If effects on different people can't be added up, how can you say whether the net third-party effect is positive or negative?
The economic argument is to show that the gain to the gainers, measured by the dollars they would pay to get it, is greater than the loss to the losers, similarly measured. You then argue that in most contexts there is no reason to expect a dollar to represent more utility to one random person than to another. But that breaks down in contexts where there is a reason to expect the losers to have a higher marginal utility of income than the gainers, most obviously if the losers are poorer than the gainers.
There are a variety of problems with that approach, which I discuss in Chapter 2 of _Law's Order_ (http://www.daviddfriedman.com/Laws_Order_draft/laws_order_ch_2.htm), but it is a more coherent argument than "to push that point, you would have to be anti-market", since part of the point of the analysis is showing if you should be pro-market.
"Hence, an Austrian economist understands that it is impossible to show that a tariff, a minimum wage, or rent control makes the U.S. worse off. A Chicago economist doesn't."
Fixed, no charge.
My claim is that a Chicago economist can show a reason to believe that those things lower total utility, although not a proof that they do. Do you agree and object that one cannot show that lowering total utility makes the US worse off, indeed cannot show that anything short of making some people worse off and nobody better off, should count, or disagree?
My objection is that "total utility" is incalculable.
You don't have to be able to calculate it to use economics to show a reason to believe that a tariff reduces it. For an explanation, see:
https://daviddfriedman.substack.com/p/the-connection-between-economic-efficiency
Having "reason to believe" something is not "showing" something.
“Since Austrians means economists are committed to the idea that economic conclusions must be based on theory not evidence, an Austrian economist qua economist cannot use the fact that policies have had bad effects in the past as evidence that they will in the future. “
The categories here are ambiguous. Economic conclusions include theoretical conclusions, estimates, predictions, policy recommendations, and historical reports. My steelman would be that even if empirical data tempted an Austrian theorist to revise a theory, the revised theory would not be based on the data, but on a priori propositions; and if estimates or predictions were to be made, past data might be usefully employed in some way. But such predictions are profoundly conjectural, and data alone can’t refute a theory.
This can still be interpreted as making the theory tautological, which is the basic objection. A more charitable view is that a theory can only be criticized in comparison with an alternative theory. So the framing seems to matter. Austrians are not the only economists who say it takes a theory to beat a theory.
So what would it take to get Austrians to revise their theories, or abandon them for new ones? I’m not sure. But I note that even non-Austrians, including quite a number of economists with a more leftish bias, have not abandoned some basic theoretical approaches that are not well supported by data.
Milton Friedman was quite willing to make use of theoretical assumptions that contradicted empirical reality. Of course, he wanted to use these to generate interesting testable hypotheses, and then test them. That project seems feasible, but it isn't clear what the universality of such hypotheses is. They are either compatible with the data or not, but a large (infinite?) number of such hypotheses exist, and no usable theory could include an exhaustive list of potentially relevant factors. Inevitably, exogenous factors can swamp a particular historical instance.
But this is out of my depth. Perhaps our host can clarify.
The connection between theory and policy recommendations is particularly opaque for me. Some people seem to genuinely believe that libertarianish policy recommendations have trended toward disaster, while the alternatives muddle through depending on the honesty and competence of the implementers. I'd like to see a real test where the subjects get to choose somehow, but It is less about refuting false hypotheses and more about letting people try things out and stick with what they like. I realize this is not possible in all cases, but I suspect it could be more widely practiced than it is.
"and data alone can’t refute a theory. "
Humans can make mistakes in their reasoning — Rothbard makes many, Mises at least one I have spotted. If your theory produces a prediction that is inconsistent with the data that could mean that you made a mistake in getting from your axioms to your conclusion.
Sure. But the inconsistency being caused by a logic error is one possible explanation. Another is that the data are flawed. A third is that one of the premises is false. Maybe there are more.
But if I really had my finger on this, I’d be able to come up with a historical example when researchers stuck with an old theory in spite of anomalous data and later were vindicated. Or at least make my point more briefly and clearly. I’ve never been able to make up my mind which side I agree with, or whether I think they’re both wrong.
By the “Chicago School” do you just mean mainstream neoclassical economics and freshwater macroeconomics (new classical/RBC)?
Or do you mean the specific “Chicago approach” which might include price theory, public choice theory, preference for markets over intervention, and emphasis on structural models relative to reduced-form approaches?
For the first debate I wanted to make it "Marshallian economics." Gene argued, probably correctly, that most of the audience at Porcfest would not know what that meant, so I agreed to make it Chicago economics.
In this post, I think the only point which might be limited to Chicago economists is footnote 4, since Chicago economists are more likely to be pro-market than other Marshallians. But I was taking advantage of Gene's terminology in the first debate, "The Austrian economics of Mises and Rothbard," to limit the other side to the Misesian branch of Austrian economics, both because I don't know enough about other variants and because Hayek arguably has been influential, mostly via _The Road to Serfdom_.