In the previous two chapters, we have examined theories of demand for money which

have, on the whole, been developed for the purpose of macroeconomic application.
They were either explicitly macroeconomic in terms of their formulation, such as
Fisher's work, or following the example of Friedman's thesis, parts of a “typical”
behavioral function by implicitly assuming that what is true for a the individual is also
relevant for the economy as a whole. However, not all theories of demand for money
are similar to this pattern. The recent work of Baumol and Tobin, by broadening the
Keynesian analysis of the motives of transactions and speculation, presents results of
individual behavior which are not so easily transposable, by analogy, to the whole of
the world. 'economy. Nevertheless, this work draws our attention to certain factors of
the decision to hold money that we would have neglected otherwise; for this reason
alone, it should be studied.

Modern theoretical studies of the demand for currency transactions due to Baumol [2]
and Tobin [43] seek to make the analysis more rigorous and to deduce more precise
results on the variables which determine it. Getting precise conclusions from a model
usually requires making explicit assumptions. Those made by Baumol are as follows (1).
It analyzes the behavior of an economic agent, be it a company or a household, and it
assumes that this one receives an income once per period, for example once a month (2).
However, the economic agent must spread his purchases over time. We assume, for
simplicity of analysis, that it spends all of its income at a constant rate during the
period. Thus, at any time except at the final moment at the end of the month when the
last expenditure is made, the individual has a certain volume of assets, the fraction of
his income not yet spent. His problem consists in determining how to hold this asset,
given on the one hand that there are interest-bearing bonds that he can buy at the same
rate as the cash on hand and, on the other hand, that there is a fixed cost associated with
converting bonds into currency.

Obviously, it will try to minimize its costs over the entire period. This problem can be
solved as follows. Suppose that T is the real value of the income of this individual
which is also equal to the real amount of the volume of transactions carried out; r the
interest rate per period assumed to be constant over the entire period; b the real cost of
converting bonds into money (what Baumol calls “brokerage fees”) and K the real value
of bonds converted into money each time such a transfer occurs.

The costs borne by the individual have two components. First of all, every time he sells
bonds he has to pay brokerage fees and since he spends all his income and sells his
bonds in lots of equal K value, the brokerage fee disbursement will be equal to b (T /
K). At the same time, if he holds money instead of bonds, he has to give up interest and
this too must obviously be seen as a cost. Since the expenditure flow is constant, the
average value of money held during the period is K/2, i.e. half of its income from the
sale of bonds. This, multiplied by the interest rate for the period, gives us the
opportunity cost of money.
The total cost of transactions can then be written with y the cost:

Now, to find the value of K which minimizes the cost, it suffices to take the derivative
of (6-1) with respect to K, equal it to zero and solve. This gives us:

And since the amount of money held during the period has an average value of K/2, as
we saw earlier, the money demand equation that results from this analysis is:

That is, the demand for cash balances in real terms is proportional to the square root of
the volume of transactions and inversely proportional to the square root of the interest
rate (3). It can still be written

However, the reader will have noticed that in establishing this so-called “square root
rule” nothing is said explicitly about the usefulness of holding cash for carrying out
transactions, nor about the fact that 'by holding money, one gives up collecting interest.
One of the great contributions of Baumol's theory is that it does not require such
concepts. It only assumes that money is a medium of exchange in the economy, that
there is a cost involved in transforming interest-bearing assets into money, and that
there are brokerage fees. If b takes the value zero in the equation (6-4), the expression
will obviously be reduced to zero, which means that if the sale of bonds does not
involve costs, there will be no demand for money even for an economy where it is. the
only means of exchange. Without brokerage fees, it would be interesting to perfectly
synchronize the sales of bonds with the purchase of goods, so that the currency would
be held only during the moment when it passes through the hands of the person selling
the bonds. bonds and buys goods.

Brokerage fees are therefore a major variable. It is important to interpret it with care. It
would be wrong to interpret it as being, in practice, fees charged by an intermediary for
the sale of shares on behalf of a client, because it is too narrow an interpretation. In the
model, it plays the same role as any other cost resulting from the sale of interest-bearing
assets; it might as well be the trouble and inconvenience of an individual to sell his asset
himself. Simplifying to the extreme, if it is impractical to go around the corner to a
savings bank to withdraw money according to your expenses, you assume brokerage
fees if the someone was being paid to sell government securities on an organized
financial market.
In interpreting the concept of brokerage fees in this way, the unease that the reader
might have had in thinking that it is unrealistic to view them as a fixed value rather
than as a fraction of the value of the transfer made must be largely dispelled. However,
this point of view draws attention to the fact that the payment of income is generally
made in the form of cash, and not of obligations, and that it is entirely justified to
consider the existence costs for the acquisition of bonds at the beginning of the period.
As long as this cost does not vary with the amount of the purchase, it does not play a
role. Equation (6-1) then becomes:

where g is the fixed cost of buying the bonds. Since g does not vary with the amount or
with the frequency of withdrawals of money, the optimum values of these variables are
independent of g, except in the case where the purchase cost is so high that the
individual prefers keep its resources in the form of cash, if it was originally paid that
way. However, even modified in this way, the model still generally predicts that the
demand for money will grow less quickly than the volume of transactions and that
there are economies of scale for the cash held by an individual (4 ).

This assumption has two important macroeconomic points. The first is that, for the
economy as a whole, the demand for money depends on the distribution of income as
well as on the level of income. If we assume, as before, that the ratio of the volume of
transactions carried out in the economy to the level of national income is constant, then
it is obvious that the more a given amount of income is concentrated in the hands of a
small group of individuals, the lower the demand for money for a given aggregate
income. This results from the fact that the economies of scale studied above relate to the
individual carrying out transactions so that a single person carrying out a given volume
of transactions will hold less money than two persons each carrying out half of the
amount. this same volume. If the distribution of income changes, the demand for
money will also change (5).

The second remarkable point about these economies of scale is that monetary policy can
carry more weight than previous theories did not suggest. Given a certain distribution
of income, any increase or decrease in the quantity of money will have more effects on
income in a situation of underemployment than if the demand for money were
proportional to income. For a given interest rate, doubling the quantity of money, in the
case of proportionality, requires to absorb the increase a doubling of the level of income.
If the economy followed the simple “square root rule”, the result would be a
quadrupling of real income. However, in a situation of full employment, the price level
will vary in proportion to the quantity of money just as in the other models because the
nominal value of the transactions as well as that of the brokerage fees vary
proportionally with the price establishing a proportional relationship between the
request from mynaie and prices.

So that's what justifies taking this model seriously. The previous results suggest not
only that the linear money demand function used in Chapter 1 may be too much of a
simplification of reality, but also that the characteristics of a function which includes
only the level of income and the interest rate are not sufficiently complete to allow it to
be part of a model from which one expects precise forecasts on any economy. In
particular, this modern theory of demand for transaction money demonstrates that the
distribution of income is a factor to be taken into account.

However, in the Keynesian concept, the demand for transactions constitutes only a part
of the aggregate demand and as we have seen, the novelty of the Keynesian approach to
the problem of the role of money in the macroeconomics lies in his analysis of the
demand for speculative money. It is therefore not surprising that research has been
carried out in this direction. As we will see now, this analysis has been considerably
improved.

The motive for speculating money appears, on the one hand, because the capital value
of money does not vary with changes in the interest rate unlike other financial assets
and, on the other hand, because that there is uncertainty about the future variation of
the interest rate. Keynes solved these problems by posing that in choosing between
money or bonds, each individual acted “as if” he was certain of the future interest rate
and therefore held either bonds or money, as he expected. It was only by supposing that
at all times individuals will have different opinions about the variations in this rate that
Keynes arrived at a uniform relation for the whole economy between the demand for
speculation money. and the interest rate.

Modern studies on this problem, which are mainly due to Tobin [44], present a more
sophisticated analysis of individual behavior (6). This is obviously necessary since even
a quick glance shows that we do not find individuals who hold all their wealth either
only in the form of money or only in bonds or any other asset. Rather, individuals
diversify their portfolios. Such behavior cannot be explained by a theory which states
that individuals act "as if" they are certain of the future. If this were so, they would only
own the assets from which they expected the greatest benefits. It is therefore necessary
to explain why the portfolios are diversified. The theory which will now be described,
even if, here, it is limited to the study of the problem of diversification between money
and bonds, can provide a very general application of this problem.

The key to the analysis is to be found in a very simple statement about the tastes of
individuals: people see wealth as a good but see risk as a bad thing, that is, as
something that reduces satisfaction. tion provided by holding wealth. To give a concrete
example, it is assumed that individuals will prefer, for example to receive $ 100 with
certainty rather than $ 50 or $ 150 with only 50% chance. In both cases, the expected
gain is $ 100. Indeed, in the first case, the sum is certain and in the second, if the offer
was accepted several times, half the time we would get $ 50 and the other half $ 150 for
each contingency which gives an average of $ 100. However, in the second case there is
an inferential risk to the outcome, which decreases the desire to receive the dollars in
this way. If the risk was even higher, let's say that the possibilities are $ 175 and $ 25,
this choice would be even less desirable (7).

Let us now see how this simple and rich concept can be applied to the problem of the
demand for speculative money. Consider an individual who receives his income once
per period and who also saves. Between each period, he must keep this savings in some
form. Suppose, therefore, that he has a choice between money and bonds. As the price
level is assumed to be constant, the currency cannot generate interest or constitute a risk
for the holder. However, since bonds pay interest and are subject to price fluctuations,
they generate a profit, although the profit is uncertain. This income has two
components: the interest payments due to the bondholder, the amount of which is
known, and the capital gains or losses that will have to be estimated. To simplify the
analysis, we will assume that the individual, when he evaluates the probabilities of
realizing gains or losses on bonds, does so in such a way as to cancel these gains and
losses between them. Thus, the estimated value of the remuneration of the bonds
becomes just equal to the market interest rate (8). However, there is a risk to the
possible profit which can be assessed by the standard deviation - a usual measure of the
dispersion - of the probability distribution from which the individual determines his
predictions on the future evolution of bond prices (9).

The problem which then arises for the individual at the end of the period is to
determine the way in which he will allocate his 1 savings, the value of which is
supposed to be already fixed, between money and bonds in order to maximize the
usefulness that 'he hopes to get out of it. Owning a large number of bonds increases the
achievable benefits on savings and since this increases, the heritage which it hopes to
have in the next period, its usefulness tends to increase. However, this also increases the
dispersion of the possible values of its heritage in the following period.

The more his portfolio contains bonds at variable prices, the greater the possible
fluctuations in the value of this portfolio (10). And since risk decreases utility any
introduction of new bonds into the portfolio increases both the discounted value of
future wealth and the risk on the portfolio. A few geometric figures will illustrate this
and continue our analysis further.

In figure (6.1), the assets of the following period (w) are plotted on the y-axis and the
risk (J on the x-axis. The curves lu, Il> etc., are curves of indifference whose
interpretation is well known. Each curve represents the geometrical locus of the
combinations of heritage and risk between which the individual is indifferent. The
ascending slope to the right of each curve results from the hypothesis according to
which an additional wealth is a
"Good" which increases the utility and that the risk is an "evil" which decreases this
same utility. It follows that, if his wealth is increased, the individual will be satisfied
with it unless the risk also increases and maintains it at the same level of satisfaction as
before. For the same reason, the indifference curves represent levels of increasing utility
as one goes up towards the left. An increase in wealth without increased risk or a
reduction in risk without a corresponding decrease in wealth improves the situation of
the individual. The curves are convex downwards, because it is assumed that the
greater the wealth, the less an additional wealth attracts the individual and
consequently the lower the increase in the risk that he will accept to incur in order to
increase his wealth. heritage.
The line W o - W'j (1 + r) represents the budget constraint which illustrates the
combinations between risk and wealth from which the individual can choose to
constitute his portfolio. If he decides to keep all his assets in the form of money, he will
not derive any benefit from it, but he will no longer bear any risk. Thus, the budget
constraint passes through the point Wo which represents the value of its initial and final
assets if it has only cash balances. Likewise, if he chooses to hold bonds, his wealth is in
this case equal to Wo (l + r), where r represents the interest rate while (Jo is the
maximum risk that can be incurred: what the individual will have to bear if he holds all
his wealth in the form of bonds. If we assume that all bonds produce the same interest
and generate risks, any point of W o - W o (1 + r) is available at the individual's choice.
He can hold both money and bonds in his portfolio. The more bonds he will have, the
higher the expected income, while the risk he incurs will increase in proportion to the
volume of bonds held in his portfolio.

However, the concern of the individual who owns a portfolio is to obtain maximum
utility from his portfolio, for a given interest rate and possible risks. His goal is to reach
the highest possible indifference curve, i.e. the point E where the budget constraint is
tangent to the indifference curve II 'At this point, he will have a portfolio composed at
the same time. both currency and bonds. The analysis then manages to explain the
diversification of a portfolio but its interest goes further, because it can be used to
establish the link between the market interest rate and the demand for money.