A Very Brief History of Demand and Supply

I’m teaching History of Economic Thought again this year and during my progression through the material this term what has struck me is the very long road over time –literally hundreds of years - to understanding markets and value as the simultaneous interaction both supply and demand side factors culminating in the standard diagram of price and quantity determination in a market. There was recognition of separate demand and supply side factors and even adjustment to some idea of equilibrium as far back as the medieval period but the concepts were like ships passing in the night when it came to price determination.  However, what also struck me is how much of the theory of demand and supply was ultimately grasped nearly simultaneously in the nineteenth century by an assortment of economists working separately.

 

In terms of markets and commercial activity, much of medieval economic thought centered on the notion of the 'just price' of a transaction.  Medieval thinkers felt goods had two prices - a selling price and a price based on value or worth or what was termed the “just price”. The idea that a 'just price' was based on value or worth suggests that Medieval thinkers believed it was all right to reward factors for their cost of production but speculative gains were not morally correct.  However, value was also based on need as Thomas Aquinas himself argued that price varies with need or what he termed indigentia.  

As Aquinas wrote: "...if the price exceeds the quantity of the value of the article, or the article exceeds the price, the equality of justice will be destroyed.  And therefore, to sell a thing dearer or to buy it cheaper than it is worth is, in itself, unjust and illicit...The just price of things, however, is not determined to a precise point but consists of a certain estimate...The price of an article is changed according to difference in location, time, or risk to which one is exposed in carrying it from one place to another or in causing it to be carried.  Neither purchase or sale according to this principle is unjust." (Ekelund and Hébert, p. 30)

Skipping forward to the 1600s, there is a demonstration of the law of demand by Gregory King (1648-1712) that was disseminated by Charles Davenant (1656-1714). He presented a relationship between the price of wheat and the harvest which illustrates the law of demand and elasticity.  The results suggest that if the harvest falls by 50%, the price would rise by 500% (See Table 1).  Again, this of course intertwines both the supply side and the demand side as a reduction in the harvest is a supply side shift.

Richard Cantillon (1680-1734) in Essai sur la nature du commerce en general. (Essay on the General Nature of Commerce) argued the price of a product was divided into a normal price, measuring the quantity and quality of the land and labour that were used in its production, and the market price. The normal price can deviate from the market price, which reflects changes in the market.

  Table 1: Gregory King’s “Demand” Schedule

Reduction in Harvest         Increase in Price of Wheat

 

         1/10                         3/10

         2/10                         8/10

         3/10                         16/10

         4/10                         28/10

         5/10                         45/10

(Source: Spiegal, 142)        

It is left to Adam Smith (1723-1790) and the Wealth of Nations to tackle the problem using an adjustment mechanism that more closely resembles what we now teach in first year economics.  Yet he also seems to separates demand and supply into separate actors with one determining price in the short run and the other in the long run. In Chapters 5-7 of the Wealth of Nations, Smith articulates a labour theory of value as well as the concepts of demand and supply and market adjustment. According to Smith, a good can have two prices: its market price and its natural price.

The value of a commodity, according to Smith, was the sum of the amounts payable to the factors used in making a good - that is, the cost of bringing a product to market.  The constant adjustment of demand and supply produces the long run 'natural price' that just covers the cost of bringing the product to market.  When a commodity is brought to market, it is demand that sets the price but in the long run, the final price is set solely by producer costs.  That first price, in response to the increase in demand is the market price while the price after the supply adjustment is the natural price.  Of course, there were no graphs.

John Stuart Mill (1806-1873)– the great synthesizer of the Classical School –finally brings demand and supply together in a theory of an equilibrium price using verbal analysis.  He presents demand and supply as schedules showing the relation between price and quantity demanded and supplied, all other things given.  This finally resolves the past inability to see value determined simultaneously by two separate sides.  First, utility, which led to demand and then cost, which led to supply.  Mill provides an explanation similar to the modern exposition of equilibrium price determination.  As Mill writes:

Meaning, by the word demand, the quantity demanded, and remembering that this is not a fixed quantity, but in general varies according to the value, let us suppose that the demand at some particular time exceeds the supply, that is, there are persons ready to buy, at the market value, a greater quantity than is offered for sale.  Competition takes place on the side of the buyers and the value rises...At what point, then will the rise be arrested?  At the point, whatever it be, which equalizes the demand and the supply: at the price which cuts off the extra third from the demand, or brings forward additional sellers sufficient to supply it.  When in either of these ways, or by a combination of both, the demand becomes equal and no more than equal to the supply, the rise of value will stop.

The converse case is equally simple.  Instead of a demand beyond the supply, let us suppose a supply exceeding the demand.  The competition will now be on the side of the sellers: the extra quantity can only find a market by calling forth an additional demand equal to itself.  This is accomplished by means of cheapness; the value falls, and brings the article within the reach of more numerous customers, or induces those who were already consumers to make increased purchases.... Thus we see that the idea of a ratio, as between demand and supply, is out of place, and has no concern in the matter: the proper mathematical analogy is that of an equation.  Demand and supply, the quantity demanded and the quantity supplied, will be made equal.  If unequal at any moment, competition equalizes them, and the manner in which this is done is by an adjustment of value.  If the demand increases, the value rises; if the demand diminishes, the value falls: again, if the supply falls off, the value rises; and falls if the supply is increased.  The rise or fall continues until the demand and supply are again equal to one another. (Mill, 446-448)

It was left to Alfred Marshall (1842-1924) to provide a graphical illustration – or so it seems.  In Principles of Economics (1890), price is determined by demand and supply, cut, as it were by 'both blades of the scissors.' He reconciled the demand and supply side explanations of value into a single analytical framework.  Utility theory provided the formulation of the demand curve while costs determined the supply curve and the interaction of demand and supply - set price.  In the short run, it was demand that set price but in the long run, supply adjusted so that a competitive economy in the long run would tend towards the lowest possible costs of production. I suppose this is what Adam Smith had in mind when talking about market and natural prices.

Economists usually see Alfred Marshall as the originator of standard demand and supply diagrams and their use in economic analysis including welfare applications and consumer surplus.  However, it turns out that Marshall’s diagram was not the first use of demand and supply curves.  Cournot originates a scissors type diagram to illustrate tax incidence in 1838 that has a demand and a supply curve with price on the horizontal axis and quantity on the vertical axis.  Indeed, at least five economists used similar tools before Marshall published it –  Antoine Cournot (1838), Karl Rau (1841), Jules Dupuit (1844), Hans Von Mangoldt (1863) and Fleeming Jenkin (1870). According to Humphrey (1992), these individuals were largely unaware of the others work.

According to Humphrey (1992), Cournot contributed the original curves while Rau and Mangoldt provided stability analysis.  Mangoldt did the first comparative statics exercises and Jenkin applied the analysis of price determination to the long and the short runs.  Cournot, Dupuit and Jenkin use the framework for tax incidence analysis while Jenkin devises the concept of producer surplus.  To Humphrey’s knowledge, Marshall was probably aware of much of this work.  One wonders if Marshall's placing of price on the vertical axis and quantity on the horizontal axis might have been done to differentiate his work from the others – Cournot has quantity on the vertical axis and price on the horizontal (but see Gordon 1982 for another explanation of why Marshall did it – to motivate the desire to use price as a money measurement of benefits and costs in measuring surplus and welfare). Humphrey (1992) asserts that Marshall nevertheless deserves some credit because he gave the diagram its most “complete, systematic and persuasive” statement but not because he was the first to invent it.  So there you have it.

References:

Blaug, Mark (1983) Economic Theory in Retrospect (Cambridge: Cambridge University Press)

Ekelund, Robert B. Jr. and Robert F. Hébert (1990) A History of Economic Theory and Method, 3rd Edition (New York: McGraw-Hill).

Gordon, Scott (1982) “Why Did Marshall Transpose the Axes?” Eastern Economic Journal, Vol. VIII, No. 1, January, 31-35.

Humphrey, T.M. (1992) “Marshallian Cross Diagrams and Their Uses Before Alfred Marshall: The Origins of Supply and Demand Geometry”, Economic Review, March/April, 3-23.

Mill, John Stuart (1848/1969) Principles of Political Economy (New York: Augustus M. Kelley)

Smith, Adam (1776/1982) The Wealth of Nations (New York: Penguin Books)

Spiegal, Henry W. (1983) The Growth of Economic Thought (Durham: Duke University Press)

 

Comments

[bsf]

Livio: Don't overlook Xenophon, especially "Ways and Means", but also Cyropaedia.

[Livio Di Matteo]

Bsf: Xenophon-well that is really going way back but I suppose you are right! There are certainly utility strands in the Greek philosophers when it comes to economic thought.

[Nick Rowe]

Lovely post.

On reading that passage in Mill twice, there is very clearly a downward-sloping demand curve. But it's not 100% clear to me he has an upward-sloping supply curve. What he says seems compatible with a vertical supply curve, so all the adjustment is on the demand side. Not sure.

[Jason Smith]

Nice post! I don't think I've ever seen such a clear thread from the beginning to end presented in such a succinct manner.

[John Chant]

Very nice.

Must be a great course if this is representative.

[Bob Murphy]

And just think Livio, now we've come full circle with economists throwing out supply & demand and going back to just price theories...

[Gene Callahan]

Two items:

1) I recommend Mary Morgan's _The World in the Model_, which has a nice history of supply-and-demand diagrams (as well as many other models).

2) I would suggest that just price theories are not incompatible with supply and demand analysis. My reading of Aristotle (from whom Aquinas would have drawn his basic analysis) suggests to me that what he was looking at (without having the terms, of course!) was producer and consumer surplus, and the "equality" that had to hold was between these surpluses, so that if I would buy at any price under $2, and Murphy would sell at any price above $1, the just price should be around $1.50. "Unjust" prices would come about when one of us has far greater bargaining power, and "forces" the price to $1.01 or $1.99.

[Jacques René Giguère]

My own take of what people think about "just price" is something about LRAC. Hence the widespread hostility to surge pricing in case of emergencies.

[Livio Di Matteo]

Gene:
Thanks for the Morgan reference.

[rsj]

I think that supply and demand curves, in and of themselves, are trivial observations known to pretty much everyone, but not really used as the lense through which the price setting process was viewed because, operationally, they are not a good description of the price setting process and provide no useful guidance to price setters or to buyers.

However *if* you use them, you can get really far and talk about different types of competition, surpluses, effects of taxes or shortages, etc. So it was not the use of these curves per se that was influential, but pushing the whole analysis forward to show that these abstractions were useful.

I.e. the drawing of the curves is not useful, the whole book using the curves is useful. It's like a model -- the model does a bad job of describing what happens, but the model is useful. The mathematicians talk about the unreasonable effectiveness of mathematics but economists should talk about the unreasonable effectiveness of simple models. That, to me, is the revolutionary thing, rather than the curves.

And unfortunately I think textbooks should talk about this more at the very start. One of the biggest problems I've personally had is that when reading an economics text, my mind screams "But it can't possibly work like that! That is wrong!" and yet, if you can suspend the disbelief, you do end up with overall conclusions that are often right. Supply and Demand curves might be the first such example -- intentionally adopting something that is clearly wrong as a description of market behavior -- and pushing it do be able to make all kinds of quantitative insights about long run (static) outcomes.

On the other hand, "just price" was trying to answer "what price should I charge", with the theologians basically saying you should charge the price that covers your costs -- which is simple and actionable and also clearly sustainable. The two concepts really are ships passing in the night.

[Nick Rowe]

rsj: "On the other hand, "just price" was trying to answer "what price should I charge", with the theologians basically saying you should charge the price that covers your costs -- which is simple and actionable and also clearly sustainable."

Let's flip the question: "what price should I pay?" Should I pay a price that covers my benefits?

Are we talking average cost/benefit, or marginal cost/benefit?

What happens if there is a one period lag in production, and the marginal benefit this period is not equal to the marginal cost last period?

Or, to use (was it the English Opium Eater's?) example, what is the value of a musical snuff box offered for sale in a canoe in the middle of Lake Superior? Is it what it cost to produce in Birmingham? Or is it the benefit of whoever wants to buy it or sell it?

[Nick Rowe]

Thomas de Quincey switched from math to economics because he found math too hard when suffering withdrawal symptoms. He was one of the precursors of utility theory and the demand-side.

[rsj]

"What happens if there is a one period lag in production, and the marginal benefit this period is not equal to the marginal cost last period?"

This is why we don't price things at marginal cost, in practice. Producers, when asked about their marginal costs, get very confused because in general they don't know what their marginal costs are, neither is it important to know this. All firms have substantial fixed costs, so the relevant metric for firm survival and profitability would be the average cost, not the marginal cost. Firms do know their average cost. But even then, pricing is very difficult and firms screw it up all the time. This is not a reason to look down on firms, but to acknowledge how difficult the price setting decision is. So they use a number of different strategies -- typically markups traditional in their industries such as keystone (charge 2x cost no matter what), or they charge the MSRP -- let someone else determine the cost. There is a lot of data about cost determination on the part of firms, none of which supports the notion that firms price at marginal cost. But just because it doesn't happen at all like that doesn't mean it's not a useful concept. It's a very useful concept, but it's surprising that it is so useful because it so idealized and disconnected from reality.

The examples that you gave are a great reason why firms, IMO, follow a more evolutionary approach rather than an optimizing approach per se. The optimizing approach says to find the unique equilibrium and you're done. This assumes you have all the information and are able to find the equilibrium (which is an NP complete problem, so really no one can find it, even if they had all the information).

The evolutionary approach is more of an algorithm determined by competitive pressures. I.e. if you are running out of product, then flip a coin. If it comes up heads, increase the price. If it doesn't, do nothing. This type of algorithm is a bit like evolution -- over time, we grope towards overall improvements, but at any point in time, there are many different solutions when the criteria is of firm survival rather than maximum profitability. I.e. there is not one optimum living organism, but a vast variety of different ones, all imperfect in their own way, but good enough to survive. Just like all the oddball firms that somehow manage to survive. This approach is much more robust -- e.g. some disease comes along and suddenly the animals which weren't the most fit in some sense become more fit. This requires a diversity of solutions rather than a single solution, which could be wiped out by a change in in the environment. Some of those firms making bad decisions, should the environment change, will be making good decisions and their market share will grow. So it is robust to have firms take a very heterogeneous approach to price setting, and then throw in some creative destruction as competitive pressure.

Recently, Adobe has been experimenting with trying to find out what the optimal pricing is for their Creative Suite products. As is typical, the problem is not knowing the demand curve. As you may be aware, Adobe was much hated for charging far more for their products in Australia than in the U.S. -- like $1000 more. So Adobe decided to try to be scientific and test how much people were willing to pay by introducing some A/B testing. You can imagine how well that worked out. http://www.creativebloq.com/adobe/adobe-charges-100-extra-using-wrong-browser-101413289

I think its all very fascinating, but progress was made when we ignored reality and started pretending that prices are set by a mechanism which clearly has nothing to do with how prices are set. Yet this abstraction is extremely useful. The uncommon applicability of models to describing complex self-organizing behavior.

[rsj]

Also interesting is baseball ticket prices. You would think all the owners have an incentive to sell all the seats. There were different sized stadiums in different cities with different income levels. Some games featured popular teams but other games less popular teams. Some games were on the weekends but others on weekdays. But the National League kept all prices fixed at 50 cents a seat from its founding in 1876 until the early 20th century. During that time, there were significant movements in inflation as well as baseball demand. When the American League came along, they fixed all prices at 25 cents a seat. Shortly after that, teams were allowed to change prices. During the 1920s, ticket prices went up to a dollar in some teams.

During the Great Depression, Jacob Ruppert, the owner of the Yankees refused to lower ticket prices during the deflation, preferring to leave many seats unsold. When asked why he did this, he answered that because he didn't raise prices during the inflation prior to 1929, he didn't feel the need to lower them now. The Yankees survived the Depression. Another interesting problem they had to worry about was radio. The owners were concerned that by broadcasting the games for free, they would lose ticket sales. They kept radio out until 1921 and didn't put lights into the stadiums until the 1930s -- night games would be better for working people to see, but the owners weren't sure if they could recoup their costs. Comprehensive use of night games didn't arrive until much later. The Chicago Cubs played their first night game in 1988.

It's very easy to draw two curves and say "this is the price" or just to assume that someone is doing the best possible thing. Recently -- in the early 21st century -- a radical idea was introduced: "dynamic pricing". According to this theory, teams would adjust their prices in response to changes in demand up until all seats would be sold. A few teams use dynamic pricing, but even many teams that claim to do so still leave many seats unsold. I think it's best to view this as an evolutionary process rather than as a "well, just pick the best price, stupid", because pricing is hard and when firms are doing well financially, they don't need feel the need to make a lot of changes.

[Kris]

"Yet he also seems to separates demand and supply into separate actors with one determining price in the short run and the other in the long run."

Smith's discussion of the natural price is perfectly consistent with what we teach in Intro Micro. The long run competitive price is equal the minimum average total cost. Demand factors determine the size of the market but supply factors determine the price (and scale of firms).