A little over a year ago, I posted a series of posts (here, here, here, here, and here) that came together as a paper (“Hayek and Three Equilibrium Concepts: Sequential, Temporary and Rational-Expectations”) that I presented at the History of Economics Society in Toronto in June 2017. After further revisions I posted the introductory section and the concluding section in April before presenting the paper at the Colloquium on Market Institutions and Economic Processes at NYU.
I have since been making further revisions and tweaks to the paper as well as adding the names of Hicks and Radner to the title, and I have just posted the current version on SSRN where it is available for download.
Here is the abstract:
Along with Erik Lindahl and Gunnar Myrdal, F. A. Hayek was among the first to realize that the necessary conditions for intertemporal, as opposed to stationary, equilibrium could be expressed in terms of correct expectations of future prices, often referred to as perfect foresight. Subsequently, J. R. Hicks further elaborated the concept of intertemporal equilibrium in Value and Capital in which he also developed the related concept of a temporary equilibrium in which future prices are not correctly foreseen. This paper attempts to compare three important subsequent developments of that idea with Hayek’s 1937 refinement of his original 1928 paper on intertemporal equilibrium. As a preliminary, the paper explains the significance of Hayek’s 1937 distinction between correct expectations and perfect foresight. In non-chronological order, the three developments of interest are: (1) Roy Radner’s model of sequential equilibrium with incomplete markets as an alternative to the Arrow-Debreu-McKenzie model of full equilibrium with complete markets; (2) Hicks’s temporary equilibrium model, and an important extension of that model by C. J. Bliss; (3) the Muth rational-expectations model and its illegitimate extension by Lucas from its original microeconomic application into macroeconomics. While Hayek’s 1937 treatment most closely resembles Radner’s sequential equilibrium model, which Radner, echoing Hayek, describes as an equilibrium of plans, prices, and price expectations, Hicks’s temporary equilibrium model would seem to have been the natural development of Hayek’s approach. The now dominant Lucas rational-expectations approach misconceives intertemporal equilibrium and ignores the fundamental Hayekian insights about the meaning of intertemporal equilibrium.
Uneasy Money 
Excellent paper. Is Hicks (1939) the best starting point to get to grips with the idea of temporary equilibrium?
LikeLike
James, Thank you. Hicks was a very clear writer, so you will get a good intro from that. He also discussed temporary equilibrium again in Capital and Growth.
LikeLike
David,
What do you think of Ackerman’s notion that GET is dead?
LikeLike
David,
You associate GET with macroeconomics – I think you would say that you can do macroeconomics with GET.
GET is a theory of optimal resource utilization founded on relative prices. It assumes a given budget constraint (level of income) and full employment (the optimal solution sits on the PPF).
It avoids the question of how the economy gets to sit on the PPF in the first place.
Isn’t macroeconomics about the level of resource usage? Doesn’t it ask the question how is a given level of resource utilization maintained even if it is below full employment of resources?
The other the question it asks is, what can motivate an economy at a sub-full employment level to move to full employment?
GET can’t answer any of these questions, would you agree or not?
LikeLike
Henry, I’m not familiar with Ackerman, Do you have a specific citation? I think that GET does serve a limited but useful function in orienting our attention to the ways in which the real world doesn’t correspond to the various alternative ideal models of how a GE system would operate. As I point out in my paper, there is more than one way to model GE, and some ways are more relevant to real problems than others. But I certainly agree that a simple-minded adoption of GE theory — especially the Arrow-Debreu-McKenzie version –does not provide a basis for macroeconomics but subverts it.
LikeLike
David,
Ackerman’s paper can be found at:
Click to access StillDead02.pdf
A fairly straight forward read with a basic summary of the relevant literature post SMD.
I would be interested on your thoughts regarding my second comment.
LikeLike
Henry, Thanks for the link. GET is at the interface between micro and macro. GE theory may or may not be useful in understanding how economies may operate at less than full efficiency and less than full or optimal employment of available resources. If such an explanation is possible, it will require some sort of strategic change in the assumptions underlying the typical GE model. In my paper, I am suggesting that in different ways, Hayek, Hicks and Radner all provided important insights and strategies for reworking the GE model in ways that can help us understand how economies get into trouble and how they can get out of trouble. In particular, I suggest that the temporary equilibrium method may be the most promising strategic change to make in the GE paradigm. But that is just a conjecture.
LikeLike
David,
I have trouble understanding how a theory like GET, given it’s design and purpose, i.e. explaining and understanding the allocation of resources, can be applied to explaining and understanding how a particular level of resource usage is determined. Most economists who talk and think about GET do so in a way that is related to macroeconomics. I can’t see the connection. I can’t see what I am missing.
The theories of Hayek, Hicks and Radner are about explaining how price expectations are related to relative prices and allocative efficiency, across space and time, not to the general level of output and spending and the general level of prices per se.
Do you see it that way at all?
LikeLike
Henry, I don’t accept your premise that the purpose of GET to explain and under the allocation or resources. The purpose and design of GET is to demonstrate that neoclassical economic theory is a logically coherent and consistent explanation of how an economy works. If there were no equilibrium, i.e., if the equations characterizing the system did not have a solution. The theory would be logically untenable. The two welfare theorems that show that the equilibrium is pareto optimal and that any pareto optimal allocation could conceivably be attained by some general equilibrium solution were corrolaries of the existence proof, not the motivation for proving the existence of a general equilibrium.
LikeLike
David you said:
“I don’t accept your premise that the purpose of GET to explain and under the allocation or resources.”
I have to say I am utterly surprised by this statement.
Just a few quotes:
“General competitive equilibrium above all teaches the extent to which a social allocation of resources can be achieved by independent private decisions coordinated through the market.”
Kenneth Arrow, General Economic Equilibrium: Purpose, Ananlytic Techniques, Collective Choice”, Nobel Memorial Lecture 1972.
“General equilibrium analysis addresses precisely how these ‘vast numbers of individuals and seemingly separate decisions’ referred to by Arrow aggregate in a way that coordinates productive effort, balances supply and demand, and leads to an efficient allocation of goods and services in the economy.”
Jonathan Levin, “General Equilibrium”, 2006
“General equilibrium theory offers the best available answer to the fundamental questions of economics: What determines relative value? Under what conditions do competitive markets lead to an efficient allocation of resources?”
Darrell Duffie, Hugo Sonnenschein, “Arrow and General Equilibrium Theory”, Journal of Economic Literature, Vol. XXVII, pp.565-598.
I am sure I can find hundred more.
LikeLike
Henry, Life is full of surprises! But didn’t I say that the two welfare theorems show that a GE is a Pareto optimal allocation and that any Pareto allocation could be achieved through a competitive equilibrium given the appropriate initial endowments? Those two theorems were not the reason why Walras attempted to prove the existence of a GE by counting equations and unknowns. Walras never heard of Pareto optimality. There is a deep connection between neoclassical economic theory and efficiency because neoclassical theory is based on the assumption of individual optimization. What is not obvious is that individual optimization can lead to a social optimum. Obviously that is not a necessary result one has to prove that the equilibrium resulting from individual optimization results in a social optimum. Most Nash equilibria are not Pareto optimal, so the result that a GE equilibrium that is Pareto optimal is quite remarkable, but it is not true of all GE equilibria. A temporary equilibrium — a subset of all GEs — is not Pareto-optimal. So there is a lot of work to be done in exploring when and why GEs are Pareto optimal. But we also know that under some conditions, the existence of a GE cannot be proved, and those cases need to be explored in greater depth, which is one of the points I have tried to make in my paper.
LikeLike
David,
My point is not so much about Pareto optimality. It’s a side issue.
My point is that GET is about the allocation of resources based on relative prices. It’s design is based on maximizing an objective function (essentially given bundles of resources) subject to a constraint (given level of income). That is, the level of resource usage is given and assumed.
Whereas, macroeconomics is about how an economy sets the level of resource usage, i.e. the level of output/income (thereby determining the level of employment). In macroeconomics, the level of output is the variable unlike in GET where is it assumed as given and it is the relative price ratio that is the variable.
So GET, as it is designed (assumed level of resource usage, relative prices), cannot be a theory about how the level of output is determined.
LikeLike
Henry, I’m sorry the level of resource usage is not assumed, it is determined in equilibrium. The question is whether the allocation of resources is optimal or not optimal and that has to be determined by the individual decisions of the agents in the model. Under certain assumptions that allocation may be optimal, but that optimality is not a foregone conclusion as you imply, whether it is or not depends on teh nature of the assumptions made in setting up and then solving the model. GE is not the same as macroeconomics, but it is a special case and macroeconomics in some of its manifestations is concerned with how some variation of the standard GE model can operate with less than full efficiency and with less than full or optimal employment of all available resources. That is certainly not the only way in which a macroeconomic model can be set up, but it is one way. Let a thousand flowers bloom!
LikeLike
David,
“I’m sorry the level of resource usage is not assumed”
There are given bundles of resources assumed, the production possibility frontier. That is what I meant (and said) by a given level of resource usage. GET allows the equilibrium point (a particular bundle of resources) somewhere on the PPF given by the income constraint and relative price ratio.
“……but that optimality is not a foregone conclusion as you imply,…”
I don’t imply anything about optimality. It is irrelevant to the point I am making.
“GE……….is a special case and macroeconomics….”
It might look that way, but a GET equilibrium is obtained by a process of relative price change. Whereas macroeconomic equilibrium is given by the level of spending.
“…macroeconomics in some of its manifestations is concerned with how some variation of the standard GE model can operate with less than full efficiency and with less than full or optimal employment of all available resources.”
Then it is no longer a GET model. Relative prices can no longer drive the equilibrium point when the economy is operating below the PPF, i.e. where resource scarcity is no longer an issue.
“Let a thousand flowers bloom!”
: -) . Yeah, time to smell the roses.
LikeLike
I would like to add a couple of other points.
I would argue that when an economy is operating below the PPF, relative prices might direct the deployment of resources but there is no way of knowing whether an allocative equilibrium has been obtained. Relative prices also would have no effect, by and large, on the level of output. The overall level of output is given by the level of spending.
In such a situation, neoclassical economics can provide no direction and we are firmly ensconced in a Keynesian world.
LikeLike
Henry, At this point, I am not really sure what we are arguing about except that I don’t think you appreciate that general equilibrium analysis can be a flexible analytical tool and may potentially contribute to the understanding of non-equilibrium states. I agree that in situations with unemployed resources of all kinds an adjustment of relative prices is unlikely to lead to a spontaneous restoration of full employment. I agree that aggregate spending is a critical variable for macroeconomics, but general equilibrium analysis may have something to say about why aggregate spending falls short from the rate required to maintain full employment.
LikeLike
David,
” I don’t think you appreciate that general equilibrium analysis can be a flexible analytical tool and may potentially contribute to the understanding of non-equilibrium states”
You are correct , I don’t appreciate this. However, I can’t see that it is logically possible. Happy to be disabused of this opinion. And it’s not only about non-equilibrium states. Non-equilbrium states can occur at the PPF (particularly as transitional states in the real world). And, according to SMD, multiple equilibria can occur at the PPF. These are all issues for GET to deal with. These are all of interest from the point of view of resource allocative outcomes. However, for me, what is of primary interest is, how can GET be applied to states below the PPF, if at all.
“….but general equilibrium analysis may have something to say about why aggregate spending falls short from the rate required to maintain full employment.”
Have you explored this aspect anywhere? Your paper, as far as I understand it, deals with expectations about relative prices (otherwise it doesn’t make sense to me in its context of GET).
Is there a resource you can direct me towards that deals with such considerations?
LikeLike
“a flexible analytical tool and may potentially contribute to the understanding of non-equilibrium states”
Do you mean non-equilibrium states on the PPF or below it?
Do you believe that allocative equilibrium states (i.e. states given by relative prices) can occur below the PPF?
I can imagine a macro equilibirum state below the PPF that may not be in allocative equilibrium.
LikeLike
Not every equilibrium is pareto-optimal. Being on the PPF is a necessary but not sufficient condition for pareto-optimality. There can be an equilibrium with externalities which is not on the PPF. A temporary equilibrium need not even be on the PPF.
LikeLike
“Equilibrium was not an actual state that an economy could achieve, it was an end state that economic processes would move toward if allowed to play themselves out without further disturbance …
The failure of the equilibrium to be sustained may be attributed to a change in information rendering the formerly optimal plans sub-optimal given the change in information. But until the new information became available, the mutual consistency of optimal plans at a (perhaps fleeting) moment signified an equilibrium state …
The EPPPE, even if it ever exists, is momentary, and is subject to unraveling whenever there is a change in the underlying information upon which current prices and expected future prices depend, and upon which agents, in choosing their optimal plans, rely …
The relationship between information and equilibrium can be described as follows: differences in information or differences in how agents interpret information lead to disequilibrium, because information differences lead agents to form divergent expectations of future prices.”
It seems to me that the concept of temporary equilibrium would be the logical conclusion of such a sequence of thought.
But what puzzles me is why anyone would think that even temporary equilibrium would be a “state” that an economy could achieve. It would seem to imply that it’s possible for there to be some time interval where no new information arrives, or where there are no price surprises. This seems highly unrealistic, forcing the entire exercise toward a point in instantaneous time, which doesn’t seem to produce a very practicable concept.
LikeLike
David,
“A temporary equilibrium need not even be on the PPF.”
I find this very puzzling.
Are you saying that there can be an allocative equilibrium below the PPF driven by relative prices?
How would you characterize the logic behind such an equilibrium?
LikeLike
Henry, I have done so in my previous posts which I link to. A full equilibrium requires a common set of relative prices which reconciles all the optimal plans of the agents. In a sequential Radner equilibrium with an incomplete set of markets so that prices are determined only in markets for current delivery, in contrast to the ADM equilibrium in which there is a complete set of future and state contingent markets, a full equilirium requires that all agents expect the same set of future prices. A temporary equilibrium involves inconsistent expectations of future prices so that optimal plans are inconsistent. However, current markets “clear” as prices in current markets adjust in response to excess demands and supplies. Under certain conditions, a temporary equilibrium can be shown to exist despite inconsistent expectations of future prices, but that equilibrium is not pareto-optimal (which can be interpreted as a point within not on the PPF. I hope that helps.
LikeLike
David,
“However, current markets “clear” as prices in current markets adjust in response to excess demands and supplies.”
Here’s the problem. How can a producer discern whether his increase in stocks is due to, say, a change in consumer preferences or an general decline in consumer confidence owing to general consumer concerns about employment prospects, for instance. The first circumstance describes a micro consideration and the second circumstance describes a macro consideration.
This same producer might have also detected that his competitors are dropping prices. Is this due to a change in consumer preferences unique to the industry or is there a decline in the general level of prices for the reasons mentioned above?
Irrespective of the reason, the producer will have to reduce his prices to match his competitors. His sales will probably continue to decline, but at a slower rate, because consumers are generally cautious and reducing spending across all industries. But his stocks will continue to increase above desired levels.
Eventually, he may get to a point where he decides that he has to reduce output and lays people off. His competitors probably will do the same. Adding to the overall malaise. And eventually, output might be reduced to a point where the general level of stocks stabilize. I would say this is the point of macroeconomic equilibrium.
Micro (allocative) equilibrium may have been in place through the whole process. Whether this is in any sense an optimal equilibrium, I’m not sure. Still thinking about that one. Perhaps you have a view on this.
One way of looking at this notion of sub-PPF allocative equilibrium might be the following. There is always a budget constraint in operation. There is always a given level of income and spending. In the case described above the budget constraint is well below the PPF. This budget constraint defines the locus of possible production/consumption bundles. The PPF is no longer the ultimate constraint. So where along the budget constraint might the economy rest? If the consumer preference map was superimposed on the budget constraint, an optimal point along the budget constraint might then be described at the tangential intersection of the budget constraint and the highest indifference curve possible (subject to the budget constraint). This would define a stable allocative equilibrium, would it not even though the economy is well below the PPF? (And it might be considered Pareto optimal at the given level of income?)
The points I am trying to make are the following:
Excess demand/supply is necessarily a sign of allocative (i.e. one given by relative prices) disequilibrium when the economy is operating at the PPF.
Excess demand/supply may or may not be a sign of allocative disequilibrium when the economy is operating below the PPF.
When the economy is operating below the PPF, there may be allocative equilibrium but not necessarily macroeconomic equilibrium.
Relative prices changes will not bring about macroeconomic equilibrium.
Relative price changes will not lift the economy to the PPF.
The economy will only be lifted to the PPF (i.e. full employment) with actions which increase the general level of spending.
So it looks like a temporary allocative equilibrium can exist below the PPF, but it may not be a macroeconomic equilibrium while the economy adapts to changes in general (i.e. economy wide) conditions.
And I return to a previous point, that while neoclassical general equilibrium economics might describe allocative equilibrium, it cannot describe a macroeconomic one.
LikeLike
JKH, I agree that it is hard to imagine that the even a state of temporary temporary equilibrium is ever achieved by any real world economy. The justification for being interested in such a concept is that it is a simplification that allows us to handle some aspects of a complex reality in a tractable way. Whether it is in fact a useful concept is, I agree, open to serious question, but I personally don’t see a better alternative.
Henry, The scenario that you are describing might or might not be possible in the context of a temporary equilibrium model. The scenario that you are describing might also correspond to a situation in which a temporary equilibrium does not exist for the reasons I mention in my paper. Obviously if there is a sequence of temporary equilibria that result in steadily declining output and rising unemployment, those equilibria would not be pareto-optimal. A PPF is a social construct and it has nothing to do with the budget constraint that individuals are subject to in carrying out their trading plans. However, in an intertemporal context where the future prices on which people are basing their plans are merely expected not actual prices, it is easy to imagine situations in which ex post the budget constraints are violated, and bankruptcies — clearly violations of the budget contraints — occur with very substantial repercussions that can preclude the existence of an equilibrium solution.
You are using the term allocative equilibrium in some vague sense which I cannot understand. In the context of a temporary equilibrium approach markets are in equilibrium if the price in each market in that period is such that there is no excess supply or demand in each of those markets. That’s all that market-clearing means. The point to understand is that condition — even if it exists — doesn’t tell you anything about whether resources are being used efficiently or not because the equilibrium is not necessarily pareto-optimal and may occur at a point that is off the PPF. A temporary equilibrium does not imply that an increase in total spending would not move the economy toward the PPF.
LikeLike
David,
“A PPF is a social construct and it has nothing to do with the budget constraint that individuals are subject to in carrying out their trading plans.”
What do you mean by “social construct”?
Them PPF and the budget constraint are two separate intellectual constructs which when married together yield a notional allocative equilibrium. Individuals cannot carry out any trading plan. They are constrained by what is possible, i.e. the PPF and their budget constraint.
“You are using the term allocative equilibrium in some vague sense which I cannot understand.”
I apply the term to an equilibrium given by the PPF/budget constraint/relative prices. It is an equilibrium that defines an optimal (theoretically) allocation of resources under scarcity.
I use the term to emphasize that it is not an equilibrium given by macroeconomic factors, e.g. inadequate demand, expectations about the general level of economic activity etc..
“A temporary equilibrium does not imply that an increase in total spending would not move the economy toward the PPF.”
Yes, I agree. Changing relative prices and expectations regarding them won’t move the economy to another level of resource usage. Only a change in the level of total spending will. The latter is a macroeconomic consideration. The former is a GET/microeconomic consideration. This is the point I am trying to make – GET and macroeconomics deal with entirely different economic phenomena.
LikeLike
Henry, “Social construct” was an inapt term to use. I meant to say it is an theoretical construct relating to an aggregate of individuals constituting a closed economic system in contrast to a budget constraint which is a theoretical construct pertaining to an individual agent. A PPF and a budget constraint do not automatically yield an equilibrium when “married together;” an set of prices and price expectations that is results in compatible individual decisions about their optimal plans is also required.
The equilibrium resulting from a PPF, budget constraints, and a set of equilibrium prices and price expectations may or may not be social optimal and it may or may not lead to a point on the PPF. As I have pointed out repeatedly not every equilibrium is pareot-optimal.
GET and macroeconomics do not deal with different economic phenomena, they deal with overlapping sets of economic phenomena, but from different perspectives, just as chemistry and physics deal with overlapping set of economic phenomena, but do so from different perspectives. The theoretical and empirical challenge is to explore how to establish connections between the different theoretical perspectives. Establishing those connections is an enormous scientific advance, but that doesn’t mean we should assume that one perspective must be automatically subsumed under the other.
LikeLike
David,
“…an set of prices and price expectations that is results in compatible individual decisions about their optimal plans is also required.”
Thank you for making that explicit. I will reread your paper in that light.
“…but that doesn’t mean we should assume that one perspective must be automatically subsumed under the other.”
When a consumer is deciding whether it is a good time to spend a portion of his income on a good, that is a macro decision. If he is deciding to choose one good over an other, that is a micro decision. If a businessman asks himself whether it is a good idea to make an investment at a point in time, he is making a macro decision. If he is deciding which of a multitude of opportunities to invest in, he is making a micro decision. The macro decision has to be made first and independently of the micro decision. The decisions made are based on a range of completely different considerations.
LikeLike
Henry, I don’t accept your distinction between micro and macro decisions; decisions about how much to buy of anything are made based on expectations of the future, and those expectations are necessarily implicit in any GET model; the only question is whether fully determined by a complete set of markets in the Arrow-Debreu-McKenzie version, or they are left as contingent expectations that may or may not turn out to be correct in the Radner version, or in which expectations are incorrect in the Hicksian temporary equilibrium version.
LikeLike
David,
Again I am surprised by your response.
Anyway, thanks for taking the time to discuss these matters with me.
LikeLike
David,
“….decisions about how much to buy of anything are made based on expectations of the future..”
Yes I agree. But there are different kinds of expectations.
There are expectations about prices of goods relative to the prices of other goods. These expectations are a micro consideration.
Then there are expectations about prices in general as well expectations about income, profit and employment prospects. These expectations are a macro consideration.
LikeLike
Henry, You are imposing your own arbitrary distinctions on what is allowed in a GE model and what is not allowed. WADR, you don’t get to decide what expectations I am allowed to put into a model and what expectations I am not allowed to put into a model.
LikeLike
David,
I don’t think they’re arbitrary distinctions
I would say if you intend mixing macro expectations with GET, then that is arbitrary and inconsistent with the design of GET models. GET models are about resource allocation in the face of scarcity, not about the level of resource usage.
Of course, there is nothing to stop you from introducing macro expectations into a GET model other than that would be totally incongruous with the inherent logic of GET.
LikeLike
Henry, At this point, I don’t see anything to be gained by continuing to discuss general equilibrium and micro vs. macro. We might as well revisit the insane Bank of France.
LikeLike
“We might as well revisit the insane Bank of France.”
🙂
Thanks for the chat.
LikeLike