Home > Macro/Monetary Theory, Micro, Uncategorized > More on Walras’ Law

More on Walras’ Law

Have taken a hiatus from blogging to deal with moving, new job, weddings, etc. and trying to get back in the habit so I figure I will finish up a post on Walras’ Law that I mostly wrote a while ago.  The topic may be a little stale now but whatever.  After all, this debate seems to have been going on for years.  I have a bunch of outstanding business with Nick Rowe but am having difficulty putting it all together.  After this little warm-up, I will try to work through that backlog.

Following the latest [at the original conception of this post] installment from Nick Rowe, it is pretty clear to me that there are three distinct issues which are all mixing together in the discussion so I want to try to separate them.  I will go through them in increasing order of significance.

1.  Is Walras’ Law useless?

I say no but that’s because I’m a micro guy at heart (and in training).  And for the record, I think I got kind of a weak acquiescence out of Nick on this so I don’t think there is very much room between our views but just for the record, here is my argument.

This is the entry from the index of Mas-Colell, Whinston and Green (the standard graduate micro text).

Walras’ Law: 23, 27, 28, 30-2, 52, 54, 59, 75, 80, 87, 109, 582, 585, 589, 599, 601, 602, 604, 780

Why am I telling you this?  Because I’m trying to demonstrate that if you want to expunge Walras’ Law from the record, you will need to totally rewrite microeconomics.  You can’t solve the Walrasian model without it.  You can bad-mouth the Walrasian model all you want, I’m not saying it perfectly represents every aspect of a real economy but if you want to tear down the pillars of that model (rather than adding on to it) you are essentially taking a wrecking ball to the rock on which our church is built.  Some people will argue for doing that, for sure, but it’s a rather extreme position which I don’t think is what folks like Nick really want.

Now the real issue is some people like to misuse the law by applying it carelessly to other models without doing the necessary work to determine whether it actually makes sense or not in those contexts.  This, I think, is what Nick objects to.  I didn’t carefully go through all of the above sections but I would be willing to bet that nowhere in there does it say that Walras’ Law proves that if we observe a shortage in some market because the price mechanism is not functioning in the way specified in the model, then there must also be a surplus in some other market.

2.  What if some price doesn’t adjust?

The Walrasian model is a model of price adjustment.  If you want to hold some price constant and ration quantity somehow, you are changing the model.  That’s fine, but you can’t take a “Law” from a different model and just try to slap it carelessly onto your new model.  If you fix the price of some good and put a quantity constraint on buyers of that good, you can find a vector of prices for the other goods such that all other markets clear given that constraint.  Whether this “violates” Walras’ Law is a nonsensical question because that law can’t be stated in the same way in the new model.

If you want to have an analogue for Walras’ Law in your new model, you have to redefine things.  The way I would go about doing this would be to treat it as a model of price adjustment in the markets for the n-1 goods, since there is nothing happening endogenously in the other market (at least nothing interesting, you have a kind of “corner solution” where you run into the constraint).  Then you would get a version of the law that applies in the subset of the market where the price mechanism is functioning in the same way that it functions in the original Walrasian model.

Alternatively, if you want to get a bit more esoteric, you can define excess demand for each good in real terms (in quantities of other goods).  This will complicate your model because you will need a lot more prices, but then you can take the price vector to be all prices, including the fixed price, and you will find that even when the remaining markets “clear” given the constraint, there is still some “excess supply” (assuming a shortage in the fixed market) of those goods relative to the good whose price is too high.  This is the sense in which Walras’ Law indicates something about such a market that is true but this phenomenon will not show up if you just look at any one of those markets and see if there is a shortage or surplus at the prevailing money prices (which is another reason to keep it, but only if you use it carefully).

This is all consistent with everything Nick has said but it is worth mentioning that the issue isn’t whether we think of it as one market for n goods or n markets for goods and money.  The issue is what constraints we put on people’s behavior and how we define things like excess demand and Walras’ law in the presence of these constraints.  The original model is set up in such a way that defining this in terms of money is equivalent (at least in equilibrium) to defining it in real terms and makes the model simpler.  But the reason it is equivalent is that when all prices can freely adjust, the marginal rate of substitution between any good and any other good has to be equal to the ratio of their prices in equilibrium so the marginal value of apples measured in dollars worth of bananas has to be equal to the marginal value of apples measured in dollars worth of papayas.  This means that instead of measuring the marginal value of each good in relation to each other good and getting a price of each good in terms of every other good, we can just measure the marginal value of each good in terms of dollars and get a price of each good in terms of dollars and have only n prices rather than n(n-1)/2 prices.  The whole matrix of relative prices in equilibrium can be expressed by this vector of dollar prices because of the equilibrium conditions on all of the marginal rates of substitution.

But once you stick in a price that doesn’t adjust, this will not be the case in equilibrium.  The marginal value of a good will be equal to the same dollar amount of every good whose price is free to adjust but not of the good whose price is fixed.  So how do we define excess demand?  In real terms or nominal terms?  The answer is: it doesn’t matter, it’s just two ways of describing the thing that happens in the model.  The important thing is whether we understand what is going on in the model.  If you just memorized Walras’ Law, without really appreciating what it means and tried to clumsily apply it to every model, then you probably don’t understand.  But by the same token, if you were never taught Walras’ Law at all, then you probably never understood the original model and you still probably don’t understand.  (Neither of these is meant to apply to Nick, who, I think, completely understands what is going on in the model.)

3.  What is the role of money in all of this? (And is a general glut possible?)

While the most recent rounds of Walras-bashing have centered mainly on the issue above, the original debate (which started years ago) was mostly about general gluts.  Walras’ law seems to imply that such a thing is impossible, yet we seem to observe them.  This is a different question from the one above.  Above the question is can one market be out of equilibrium while all others are in equilibrium?  Here, the question is can all markets be out of equilibrium in the same direction (excess supply) at the same time?

This is where the role of money becomes critical.  The Walrasian model is not a model of money.  Money is used as a rhetorical device to streamline the model.  There is no attempt made in that model to characterize the demand for money, the velocity of money or anything like that.  It is assumed that people don’t care about money, they only care about “real” goods and that money is nothing more than a mechanism which somehow allows the market to work perfectly, eliminating any frictions and allowing the “Walrasian auctioneer” to call out more complicated matrices of relative prices as a relatively simple vector of nominal prices.  (Though it is worth noting that this does restrict the set of possible relative prices.)

So this begs, not the question: does Walras’ Law hold in the real world, but the question: is that really how money works?  And the answer to that is obviously no.  Since the answer is no, it is dangerous, again, to take a simple conclusion from such a model and clumsily try to apply it to the real world.  But, also again, that doesn’t make the model worthless.  Another question one might ask is does money work kind of like that sometimes?  This is sufficiently vague to admit of no concrete answer but there is room to argue in the affirmative I think.  A better question is how does the actual nature of money differ from that assumed in the model and what are the possible consequences of that difference.  It’s questions like this that allow us to climb onto the shoulders of giants like Walras and hopefully see a bit further over the horizon.

Of course, I have a lot of thoughts about that which I will mostly avoid getting into here.  But here is a question that I think is worth pondering.  If a technology were developed tomorrow that allowed barter to be carried out frictionlessly, like with the Walrasian auctioneer, what would happen to the value of money?  Would it go to zero?  (Hint: no.)




  1. October 1, 2014 at 11:30 am

    Interesting. A couple of comments:

    1. What is “n’? Is n the number of non-money goods, or is n the number of goods including money? Because if it is the latter, there are n-1 markets in a monetary economy.

    2. “If a technology were developed tomorrow that allowed barter to be carried out frictionlessly, like with the Walrasian auctioneer, what would happen to the value of money? Would it go to zero? (Hint: no.)”

    I would say it does go to zero, because the demand is zero, unless: the good that is used as money also has a non-monetary demand; it’s convertible on demand into some other good; it’s a bubble asset, like in Samuelson.

  2. Free Radical
    October 1, 2014 at 4:47 pm


    2. “unless: the good that is used as money also has a non monetary demand, it’s convertible on demand into some other good; it’s a bubble asset, like in Samuelson.”

    Yes, I agree, that’s my point. (And it’s not the last one)

    1. The short answer is that what I had in mind is n being the number of non-money goods. Whether it is appropriate to treat money as a good is, I think, a fairly deep question which is related to number two above and my question at the end. For the record though, I do see a general glut as an excess demand for money. I think the reason for it is poorly understood but whatever it is, it falls outside the Walrasian framework.

    (Also for the record, I don’t disagree any of your characterizations of markets for goods and money or your answer to the question you posed in your post, etc.)

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: