finalexam

History of Economic Thought

Notes for Week 12

John Maynard Keynes (1983 – 1946)

In understanding John Maynard Keynes’s major contribution to the development of

economic thought, summarized in his “The General Theory of Employment, Interest and

Money” (1936), it’s important to start out by explaining the neoclassical interpretation of a

market economy, in particular the way in which Neoclassicals, such as Alfred Marshall, would

explain the presumed tendency of the system to move, of its own accord and without the aid of

government, to a full employment equilibrium. Explaining the mechanics of this process should

make it easier to appreciate the unique contribution made by Keynes.

Say’s Law, with a detour into the Quantity Theory of Money

The neoclassical economists accepted Say’s Law and, as such, accepted the notion that

capitalism moves to a full-employment equilibrium. This “law” was first developed by the

French political economist J. B. Say (1767-1832), who saw himself as expanding upon the logic

and implications of Classical political economy, and in particular, Adam Smith’s “Wealth of

Nations”. While there are, according to Joseph Schumpeter, several different versions of this

“law” within J.B. Say’s own work, the version that’s accepted nowadays is summarized by the

aphorism “supply creates its own demand.” But what exactly does this mean? The easiest way to

think of this idea is to restate it in the following terms: no one offers anything for sale unless they

intend to use the proceeds of that sale to purchase something else. When stated this way, and

assuming a closed system with no international trade, it should be obvious that the value of

everything that’s offered for sale must be equal to the value of everything that’s demanded; but

only if we assume that the sellers spend all their sales revenue on the purchase of other things. In

the language of contemporary macroeconomics, aggregate supply must equal aggregate demand.

But while the value of aggregate supply may equal the value of aggregate demand, it is not

necessarily the case that market supply and demand are equal in each and every market.

Aggregate supply may equal aggregate demand but, at the same time, the supply of shoes might

exceed the demand for shoes, or the supply of wheat might fall short of the demand of wheat. In

short, there may be excess supplies and excess demands in numerous markets at the same time

that aggregate supply is equal to aggregate demand. But if Walras’ Law holds, then it must be the

case that the value of the sum of all the excess supplies equals the value of the sum of all the

excess demands. What’s more, and assuming an unchanged context, the excess supplies and

demands will quickly vanish as prices fall in markets with excess supply and rise in markets with

excess demand. In this fashion, the system as a whole moves toward a general competitive

equilibrium in which supply and demand are equal in each and every market, bringing about the

general competitive equilibrium Walras believed was an inevitable outcome of free competitive

market systems.

The Neoclassical economists viewed this idea as consistent with the claim, which we

have yet to explain in full detail, that the system moves toward the full-employment of labor.

But, it’s not at all clear that the Classical political economists thought of Say’s Law in the same

way. An interpretation that has emerged among some historians of Ricardian economics is that

the Classicals interpreted Say’s Law to mean the full-employment of productive capacity (i.e.,

capital), not necessarily the full-employment of labor. This would make sense, since the

Classicals always assumed that in the long run wages would equal subsistence; implying that

there’s always enough of an excess supply of labor, i.e., unemployment, to keep wages at

subsistence. The Classicals interpreted Say’s Law to mean that the income generated from the

full use of productive capacity would be spent purchasing the output generated by that same

productive capacity. Workers would spend all their wages on the purchase of consumer goods

while landlords would also spend all of their rental income on consumer goods, albeit at a higher

and more ostentatious level. In the meantime, capitalists would spend all of their profits on the

purchase of both consumer and capital goods. As a result, and for the system as a whole, national

income in the form of wages, rents, and profits, would have to equal total expenditures on

consumer and capital goods (i.e., aggregate supply must equal aggregate demand). The

assumption was that any saving (that is, non-consumption) that might take place would quickly

be transformed into investment (i.e., the purchase of new capital goods).

2

Ricardo used this reasoning to argue against Malthus who had claimed that a glut of

commodities (excess aggregate supply) was a real possibility under capitalism. In order to

understand this argument it’s important to expand a bit more on Say’s Law. Since commodities

are offered in return for money and money is the medium through which expenditures are made,

another way of stating Say’s Law is to claim that aggregate supply will equal aggregate demand

if and only if the demand for money (generated from the sale of commodities) is equal to the

supply of money (generated from the purchase of commodities). If aggregate supply is equal to

aggregate demand, then it must also be the case that the demand for money is equal to its supply.

This condition may hold true for the system as a whole even though there may be, as already

noted, excess supply in some markets counterbalanced by excess demand in other markets. Once

again, prices would move about to eliminate the excess supply and demand in specific markets

while aggregate supply remains equal to aggregate demand for the system as a whole.

But what Malthus was arguing wasn’t that there might be an excess supply in some

markets, he was instead claiming that it was possible for there to be an excess supply in all

markets simultaneously, that is, aggregate supply could be greater than aggregate demand.

What’s more, this could be a long-run scenario, an extended period of stagnation, instead of a

momentary deviation from the norm (which presumably involved the full employment of capital

– productive capacity). Ricardo disagreed with this and used the quantity theory of money to

back up his claim.

The quantity theory of money goes back to the Spanish School of Salamanca in the 16th

century, but the version that’s most well-known dates its origin to the work of David Hume, in

the middle of the 18th century. The simplest version of this theory argues that there’s a positive

relationship between the supply of money and the general price level. An increase in the

circulation of money will bring about an increase in the general price level, while a decrease in

the circulation of money will bring about a decrease in the general price level. The theory,

particularly as expressed throughout the 18th and 19th centuries, generally assumed that money

could be thought of as a commodity, more specifically a precious metal, like gold or silver. So,

when reference was made to an increase in the supply of money, they were literally thinking in

terms of an increase in the supply of gold or silver. The version that’s most frequently mentioned

3

nowadays can be traced to the work of the American economist Irving Fisher. His version of the

theory was summarized using the following equation:

M·V = P·T (eq.1)

where M represents the stock of money (gold), V represents the velocity of money (the

frequency with which the unit of money circulates throughout the system), P represents the

general price level, and T represents the volume of market transactions. It’s common in many

contemporary textbooks to substitute the symbol Q for T, to represent the volume of gross

output. Doing so does not destroy the basic idea.

The general assumption is that V is determined by institutional conditions (such as the

frequency with which income payments are made to workers and capitalists, and the frequency

with which they spend their money) while T is determined by real economic conditions (such as

the amount of capital, the employment of labor, and the productivity of labor). Thus, V

and T can be thought of as given or fixed, so that an increase in M, other things equal (i.e. V and

T unchanged), must bring about an increase in P. Because of this, the above equation is often

rewritten in the following form:

·M (eq. 2)

to underscore the idea that, with a constant [V/T] , an increase in M must bring about an increase

in P of the same magnitude.

It must be noted that not all theorists accepted this strong version of the quantity theory of

money. Some accepted the idea that changes in the supply of money can increase the level of

transactions rather than the price level, or some combination of the two. Thus, as early as the

mid-seventeenth century, Thomas Mun (a famous mercantilist of the time) argued that an

increase in the flow of money, due to a positive trade balance, will increase the volume of

transactions while keeping the price level relatively stable. From this perspective, increasing the

amount of money in circulation would increase the level of output and employment while

keeping the price level unchanged. So, it’s not necessarily the case that an increase in the supply

of money will always lead to an increase in prices, it might instead lead to an increase in output.

This is especially the case if there’s considerable excess capacity within the system.

Alfred Marshall worked with a modified version of the quantity theory of money that

P = [V /T ]

4

focused less on the supply of money than on the demand for it, what he called the demand for

real cash balances. He argued that the public has a demand for money balances which is dictated

by the ongoing level of economic activity (transactions). As the level of transactions increases

the amount of cash balances demanded will grow by some constant proportion, and if the level of

transactions decreases, then the demand for cash balances will fall by that same constant

proportion. This version of the quantity theory of money was expressed by Marshall in the

following form

M = k·P·T (eq 3)

which is non-other than equation 1 obtained by dividing both sides of the equation by V. That is,

Marshall’s k is equivalent to (1/V). Marshall used the symbol k to represent the fraction of

monetary transactions that the public wishes to hold in the form of money, a constant proportion.

He saw the real cash balance, k, as fairly constant, changing very slowly with changes in the

financial habits of the public. In the short-run it was considered fixed. So, if the price level, P, or

the volume of transactions, T, increases, then the amount of money the public will wish to hold

will increase by the proportion k. Notice also that if we divide both sides of the above equation

by the price level, P, then the demand for real cash balances can be represented as

M/P = k·T (eq 4)

Showing real cash balances (M/P), the purchasing power of money, being some constant

proportion, k, of T.

Now, coming back to Ricardo, he argued that aggregate supply could never be permanently

in excess of aggregate demand because this would imply a permanent increase in the demand for

money. Since money holds no value beyond its ability to facilitate exchange, a permanent

increase in the demand for money was impossible. It would be spent fairly quickly, he believed,

eliminating the excess demand for money and, simultaneously, the excess supply of goods.

Let’s explore the steps in the argument. If aggregate supply were greater than aggregate

demand, then prices would begin to fall causing the purchasing power of money to increase.

People would now have real cash balances that exceed what they need to carry on transactions

and will proceed to reduce those excess cash holdings by exchanging them for goods, i.e. buying

more output. In terms of equation 4, as the purchasing power of money (M/P) increases, as a

5

result of a fall in P (while nominal money (M) remains unchanged), transactions (T) will have to

increase to maintain the same constant proportion (k) between real cash balances and

transactions. As a result, the combined effect of falling prices and increased purchases will have

the effect of restoring equilibrium between aggregate supply and aggregate demand, while

simultaneously eliminating the excess demand for real cash balances. Real cash balances will be

restored to their proper relation, k, to the overall level of transactions.

The Neoclassical Theory of the Labor Market and the Loanable Funds Market

The neoclassical version of Say’s Law is very similar to the Classical version, except that

the Neoclassicals imagined that Say’s Law also implied the full employment of labor. Thus, not

only is capital fully employed (as the Classicals believed) but so too is the employment of labor

(emphasized most frequently by Neoclassicals). This is because the Neoclassicals do not see the

wage rate as being exogenously determined by subsistence, as the Classicals did. Instead, the

Neoclassicals see the wage rate as determined endogenously through the labor market.

Let’s now examine the Neoclassical conception of the labor market. To begin with it’s

important to remember that they imagined every agent (firm, worker, or consumer) to be a

maximizer and every market, including the labor market, to be extremely competitive (purely

competitive). Firms, consumers, and workers are thus price takers, forced to accept the

competitively determined price. All production is characterized by diminishing returns and labor

and capital are thought to be easily substitutable for one another. This idea is represented by a

production function showing output growing at a diminishing rate as more labor is applied to a

fixed amount of capital. It is also represented by a downward sloping marginal product of labor

curve, showing the marginal productivity of labor declining as more labor is applied to given

amount of capital. At the same time, individuals are said to experience growing disutility from

work; that is, each extra hour of work is said to have a higher marginal disutility than the

previous hour of work. This is represented by an upward sloping supply of labor curve showing

the marginal disutility of labor growing as more labor is offered on the market.

Given this context, profit-maximizing capitalists will hire labor up to the point at which the

marginal productivity of labor just matches the going real wage of labor. If the marginal product

6

of labor happens to be greater than the real wage rate, then capitalists will hire more workers.

And if the marginal product of labor is less than the real wage rate, then capitalists will hire

fewer workers. It is only when the marginal product of labor matches the real wage that the profit

maximizing number of workers is hired. At the same time utility-maximizing individuals will

offer labor up to the point at which the going real wage is matched by the marginal disutility of

labor. If the real wage happens to be greater than the marginal disutility of labor, then more labor

will be offered. And if the real wage is less than the marginal disutility of labor, then people will

offer less labor. It is only when the real wage matches the marginal disutility of labor that the

utility maximizing amount of labor is offered.

These ideas are depicted in the following figure showing supply and demand conditions in

the labor market. The downward sloping demand curve is the marginal product of labor curve,

and the upward sloping supply curve is the marginal disutility of labor curve. The Neoclassicals

think of the real wage as being determined by competition among capitalists and workers seeking

to either hire workers or find employment. If the real wage happens to be (W/P)1, then the

number of workers seeking employment will be LS and the number of workers sought by

capitalists will be Ld. That is, there is an excess supply of labor (unemployment) in the amount

equal to the difference between LS and Ld. The unemployed workers will compete with the

employed workers by offering their labor to capitalists at a lower nominal wage. As the wage rate

falls, the number of workers demanded by firms will increase, since they are now cheaper; at the

same time the number of workers seeking employment will fall (drop out of the labor market)

because the wage is now below their marginal disutility of work. Eventually the real wage will

fall to (W/P)e, where the marginal product of labor just matches the marginal disutility of labor,

demand and supply are equal, and full employment is achieved at Le.

It should be noted, or better yet, remembered, that the neoclassical belief that the system is

forever close to full employment, or in Walrasian terms - forever hovering around a general

competitive equilibrium, does not mean that the real wage that is captured at full employment is

livable. As noted in our discussion of Walras’ notion of a general competitive equilibrium, the

7

equilibrium real wage need not be consistent with subsistence. It might be above, equal to, or

below subsistence.

The neoclassicals were aware of this and, for that reason, were not unsympathetic to

Pigou’s income redistribution schemes, mentioned in last week’s notes. The welfare economics

developed by Pigou suggested that government could be used to redistribute income from the

wealthy to the poor (including, of course, the working poor). So, while the market system may,

of its own accord, move toward a general equilibrium and thus full employment, the real wages

obtained at that point may not be enough to subsist. We’ll come back to this point when

considering Keynes’s critique.

In the meantime, the point of this story is to underscore the neoclassical idea that, in a free

competitive market, unemployment is eliminated – or stated differently, full employment is

achieved – through reductions in the real wage. But the real wage is reduced through reductions

in the nominal wage. Since the real wage is nothing other than the nominal wage divided by the

nominal average price of commodities, and since workers can’t influence the nominal average

price, the only variable workers can negotiate is the nominal wage. As a result, with a given

8

W/P

L

S

D

W/P1

W/Pe

Ld LsLe

nominal average price, it must be the case that the real wage will fall when the nominal wage

falls.

Notice that this entire argument hinges on the assumption that whatever money is captured

from offering something for sale is quickly spent on the purchase of other things in the same

system at roughly the same time. But what if there are individuals who decide to not spend a

portion, or all, of the money they receive from selling the goods they’ve offered on the market?

What, in short, would happen, if there were people who saved or hoarded the money earned from

selling goods and/or labor? On the surface, it would seem that the existence of saving would

undermine the validity of Say’s Law. If there is saving taking place, and no corresponding

counter expenditure, then it would have to be the case that the value of everything offered for

sale, aggregate supply, exceeds the value of everything demanded, aggregate demand; in short, a

violation of Say’s Law.

The Neoclassicals rejected this possibility for two reasons: first they were convinced that

money was used solely for transactions purposes; and second, saving is undertaken to capture

interest income. The first assumption is straightforward: money is never hoarded (a form of

saving) since no rational person, they believed, would hoard money if there were the possibility

of earning interest from the purchase of a financial asset (i.e. lending money through the

purchase of debt instruments). A rational individual, in short, would rather lend his/her money at

interest than hoard it without earning interest.

The second assumption, however, is a bit more complicated. The Neoclassicals thought

of saving as involving disutility, since it required abstaining from the pleasures of current

consumption. As a result, they believed that saving would be undertaken only if the value of that

saving was returned in the form of greater income (and thus consumption) sometime in the

future. But this, in turn, required an active loanable funds market wherein savers would lend their

money to borrowers in return for interest income. The amount people would save (and thus lend)

would depend on the real rate of interest. And since the marginal disutility of saving grows with

the volume of saving, it had to be the case that more saving could only be motivated by a higher

real interest rate. In short, the Neoclassicals believed that people save up to the point at which the

real interest rate is equal to the marginal disutility of an extra dollar’s worth of saving.

9

In addition, since the Neoclassicals conceive of all of this in real terms, it had to be the case

that the interest income that can be captured from saving (the lending of money) must ultimately

be the result of the growth in real output. That is, saving is being channeled, through the loanable

funds market, into investment spending - which increases the system’s productive capacity. Since

the system is thought to be at full employment, an increase in investment spending (buying

capital goods) can only occur through a reduction in consumption spending (i.e., an increase in

saving). Thus, those who intend to invest must provide the lenders (savers) a real rate of return

that’s made possible by the extra real output generated by that investment; and that, in turn, is

equal to the marginal product generated from the addition of one more unit of capital to the

process of production. But since the marginal product of capital diminishes with the increased

usage of capital, more investment could only be motivated by a lower real interest rate. That is,

Neoclassicals believe that capitalists invest up to the point at which the marginal product of

capital just equals the real interest rate, so that a reduction in the real interest rate will motivate

more investment.

This whole process, they believed, was carried out through the loanable funds market (see

the figure below). The supply of loanable funds (i.e., saving) is positively related to the real rate

of interest, while the demand for loanable funds (i.e., investment) is negatively related to the real

rate of interest. What if the volume of saving happens to exceed the level of investment, as is

occurring at a real interest rate of i1? Since saving was always channeled to investment though

the loanable funds market, the Neoclassicals imagined savers competing against each other for

the chance of earning interest income. This would motivate them to offer their saving at a lower

interest rate. As the interest rate falls, some savers will drop out of the market, or offer a lower

volume of saving. At the same time, the falling interest rate will motivate some investors to enter

the market and borrow more funds (and use those funds to purchase capital goods). This process

would quickly lead to a situation where the volume of saving just matches the volume of

investment, i.e., the supply and demand for loanable funds is equal at real rate of interest ie.

10

So, from a Neoclassical perspective, the existence of saving was not viewed as a refutation

of Say’s Law because it was assumed that the loanable funds market operated in such

a fashion that it was forever equilibrating the supply and demand for saving. If saving happens to

be greater than investment, momentarily severing the presumed equality between aggregate

supply and aggregate demand, real interest rates would quickly fall helping to restore, once

again, equality between saving and investment.

11

LF

S

D

i1

ie

LFd LFsLFe

i