Economics of Finance

September 15, 2020

Richard Robb
Lecture Notes: Consumption and Exchange in a Two-Period Model

Today, we’ll develop some surprisingly useful equilibrium models of interest rates.

A Theoretical Model

Individuals don’t just borrow to buy durable goods like houses and cars. They also
borrow (and lend) to smooth out their consumption over time.

We can gain deep insights from a very simple model of consumers allocating income
between the present and the future. We’ll employ calculus in this discussion, but don’t
worry if you haven’t studied it. The intuition should come through without much math.

Suppose consumers live for two periods, t=1 and t=2. Each consumer is endowed with a
single non-storable consumption good in each period. x1i is the amount the ith consumer
receives at t=1 and x2i is the amount he receives at t=2. We will refer to t=1 as the present
and t=2 as the future. By saying the consumer is ―endowed,‖ we mean the good comes
automatically. Our aim here is to simplify the world to make a point; we are uninterested
for now in how the consumer obtains his income.

The consumers in this model can trade present consumption for future consumption based
upon their preferences and endowments. The price at which they trade is determined in
the market. There’s no uncertainty; everyone knows at t=1 what endowment they will
receive at t=2. After t=2, the world will end. We will consider capital and social
investment in detail later in the course and briefly later today, but they play no role in this
model.

For now, this theoretical economy does not use money. (We’ll introduce this additional
complexity shortly.) Consumers instead trade claims on consumption with each other,
such as, ―I’ll give you one unit of consumption in the present in exchange for 1.03 units
in the future.‖
Utility Functions

A ―utility function‖ U(x) ranks consumption bundles for each individual where x
represents a bundle for goods. x is a vector so it can have many dimensions.

A utility function will exist if the consumer’s preferences follow two basic
axioms: (i) Consumers can compare all possible bundles—the consumer prefers x
to y, y to x, or is indifferent between them. (ii) Preferences are transitive—a
consumer who prefers x to y and y to z must prefer x to z.

A utility function maps each bundle to a single number. The higher the number,
the more the agent (i.e., the person choosing or exercising her ―agency‖) likes it:
bundles with higher numbers will be selected over bundles with lower numbers. It
is easy to show that utility functions exist given the two axioms. We could assign a
―1‖ to the bundle that is preferred to no others, a ―2‖ to the one that comes next,
and so forth. Maximizing the utility function is equivalent to picking the bundle
the agent prefers subject to whatever constraints he faces.

There is no natural unit for utility functions. If U(x) describes a person’s

preferences, so do the utility functions U*(x) = ln [U(x)] – 5, U**(x) = [U(x)]2 , or
any transformation that preserves the ranking of different bundles the agent could
Let c1i and c2i denote the amount that the ith individual consumes in each period.
select.

If a person does not engage in trade, her consumption would equal her endowment: c1i =
x1i, c2i = x2i. A utility function Ui(c1i,c2i) ranks consumption bundles for each individual,
so if the ith consumer prefers to consume 5 at t=1 and 5 at t=2 to consuming 6 at t=1 and
4 at t=2, then Ui(5,5) > Ui(6,4). The subscript i indicates that each individual can have his
own utility function.

We’ll make a simplifying assumption about the form of Ui(c1i,c2i):

U (c 2 i )
U i (c1i , c2i )  U (c1i )  .
1  i

The function U(c) describes the utility of consumption for each person in each period. At
1
t=1, the ith person discounts the future by a factor of . The psychological constant,
1  i

2
ρi, is the rate of time preference. People who care mainly about the present have high ρi
values. The ρi value is often presumed to be around 2% p.a. Sociobiologists speculate that
2% corresponded to the annual probability of dying over the span of human evolution. As
the (dubious) theory goes, maximizing the expected value of consumption while alive
conferred a survival advantage in the ancestral environment. Although incapable of
computing probabilities, early humans evolved an instinct for discounting. Whether you
like this story or not, it emphasizes that time preference is psychological.

In our model, consumers can trade current consumption for claims on future
1 1
consumption. The price at which they can trade is . That is, is the price of
1 r 1 r
future consumption in terms of present consumption (i.e., how much present consumption
a person has to give up to get one more unit of future consumption), while 1+r is the
price of present consumption in terms of future consumption (i.e., how much future
consumption a person must forego to get one more unit of present consumption) and r
defines the interest rate. We’ll call a claim on future consumption a bond that matures at
t=2. Someone who gives up consumption at t=1 buys bonds, and someone who acquires
consumption at t=1 issues bonds. (In the real world, of course, individuals don’t issue
bonds, but instead borrow from banks or with credit cards.) Remember, the bonds do not
involve the payment of money; they promise the delivery of one unit of consumption
goods at t=2.

Equilibrium in the credit market means (i) each consumer buys or sells as many bonds as
he likes at the prevailing level of r and (ii) the total supply of bonds equals the demand
for them.

Assume there are a lot of consumers in this economy and each one takes the interest rate
as given (ignoring the effect of their own transactions on the market-clearing rate). Each
consumer maximizes his utility subject to his budget constraint:

U (c 2 i ) c x
max U (c1i )  subject to c1i  2i  x1i  2i .
1  i 1 r 1 r
c1i , c 2i

3
The consumer’s solution must satisfy the first order condition:1

1 r
MU 1  MU 2 (1)
1  i

The marginal utility of consumption in the first period equals the marginal utility of
consumption in the second period discounted by one plus the individual’s rate of time
preference multiplied by the price of present consumption in terms of future
consumption.

The intuition underlying equation (1) is simple. If

1 r
MU 1  MU 2 (2)
1  i
the consumer could transfer a small amount of consumption to the first period and
become better off. His gain in utility from consuming ε more at t=1 will be MU1 × ε.
After consuming ε more at t=1, he must consume (1+r) × ε less in t=2. He will experience
(1+r) MU2 × ε less utility at t=2 as a result. At t=1, he discounts t=2 utility by (1+ρi), so
1 r
the cost in utility terms of reducing next period consumption is MU 2  . But if
1  i
equation (2) holds, the utility lost is less than that gained from consuming more in the
present. So (2) cannot possibly be true if the consumer is maximizing his well-being as
1 r
we have defined it. A similar contradiction arises if MU 1  MU 2 ; in this case, the
1  i
individual would be better off consuming less at t=1.

U (c 2 i )  c x 
1
Solve max U (c1i )    c1i  2i  x1i  2i  , differentiating with respect
1  i  1 r 1 r 
c1i , c 2i , 
1  i
to c1i and c2i, U (c1i )   and U (c2i )   .
1 r

4
In a graph of this maximization problem, the slope of the consumer’s budget line is
1
 . When r rises, he decreases his consumption in the first period from c1L to c1H and
1 r
raises it in the second from c2L to c2H. The graph describes a consumer with a relatively
large endowment in the first period—this wealth effect allows him to move to a higher
utility function when rates rise.

Market Equilibrium with Many Consumers

In general, (1) will determine a unique solution to the consumer’s allocation problem,
assuming U(c) exhibits diminishing marginal utility. (U is a concave function or,
equivalently, U (c)  0.) To see why, imagine that (1) holds for some c1*, c2*. If the
consumer increases c1, he must decrease c2 to stay within budget. Then the left-hand side
of (1) would fall (MU1 drops because t=1 consumption is up), the right-hand side would
rise (MU2 rises because t=2 consumption is down), and the equation could not hold.
Therefore, raising c1 would leave him worse off. The argument works in reverse:
lowering c1 would leave him similarly worse off. We must conclude that c1* is unique.

So far, we have considered a single consumer deciding how much to borrow or save
while taking the interest rate, r, as given. We’ll now consider whether there exists an
equilibrium level of r at which everyone will be satisfied.

5
From (1), we can construct the excess demand curve for c1i for each consumer: D1i(r)
defines the amount of first-period consumption the ith consumer would like to buy or sell
as a function of the interest rate. Excess demand is the amount he chooses to consume at
t=1 minus his endowment.
D1i(r) = c1i – x1i

Clearly, D1i(r) is a decreasing function of r: the lower the interest rate, the more
consumption the consumer chooses in the first period. If D1i(r) > 0, the consumer is a net
buyer of consumption at t=1 and therefore an issuer of one-period bonds which will
mature at t=2. Consumers who have negative excess demand at t=1 supply consumption
and buy bonds. Equilibrium in the goods market at t=1 requires that the total amount
demanded equals the total amount supplied. In equilibrium, total excess demand must
equal zero:

Goods market equilibrium at t=1

n

D
i 1
1i (r )  0 (3)

where n= the number of consumers. It is straightforward to see intuitively that a unique

value of r will satisfy equation (3). For very low levels of r, all consumers will have
positive excess demand at t=1 (r is allowed to be negative in this model as long as it is
higher than -1). As r rises, each person will gradually demand less at t=1 and switch to
supplying t=1 consumption. For very high values of r, everyone will supply consumption
in the first period. For some value of r in between these extremes, equation (3) will hold.

What about t=2? Since each consumer satisfies his budget constraint, we are guaranteed
that supply will equal demand for consumption at t=22. Rewrite the budget constraint
x c
c1i  x1i  2i  2i  D1i ( r ) (4)
1 r 1 r

Add up (4) over all consumers to reach

n n
x 2i c
 D1i (r ) 0  
i 1 i 1 1  r
 2i .
1 r

Rearranging this equation,

n

D
i 1
2i (r )  0.

2
This is called Walras’ Law.

6
Now we can use this simple model to draw interesting conclusions.
i) An individual’s consumption pattern does not depend on the timing of his
endowment. Since the individual has perfect access to capital (i.e., can borrow or
lend as much as he likes without transactions costs), only his wealth (i.e., the
present value of his endowment) matters. The present value is x1i + x2i/(1+r).

ii) In equilibrium, (1) holds for all individuals—rich or poor, patient or impatient. All
individuals are indifferent to transferring ε at t=1 in exchange for (1+r) × ε at t=2.
Consuming ε at t=1 raises t=1 utility by MU1 × ε. Consuming ε at t=2 raises t=2
utility by MU2 × ε, so consuming (1+r) × ε at t=2 raises t=2 utility by (1+r) × MU2
× ε. Since the consumer discounts t=2 utility by a factor of (1+ρi), he must be
indifferent between losing ε at t=1 and gaining (1+r) × ε at t=2. Thus he must prefer
an exchange whereby he gains more than (1+r) × ε at t=2 if he gives up ε at t=1; he
must also prefer an exchange where he gains ε at t=1 and gives up less than (1+r) ×
ε at t=2.

iii) In the special case where all individuals are the same (ρi = ρ) and they all have the
same endowments in both periods (x1i = x2i), then the interest rate equals the rate of
time preference (r = ρ). To see why this must be so, consider the case where r>ρ.
Then (1) can only be true if MU2 < MU1. This implies c2i must be greater than c1i.
But this is impossible, since the amount of consumption available in the second
period equals the amount in the first period.

iv) Suppose again that all individuals have identical preferences and endowments, but
this time each is endowed with more at t=1 than at t=2 (i.e., x1i > x2i). Everyone in
this scenario will try to lend at t=1 if r = ρ, driving the interest rate down. We can
conclude that r<ρ when consumption is expected to decline, and, conversely, r>ρ
when it’s rising. Intuitively, if overall consumption is rising, everyone attempts to
borrow to keep consumption streams smooth. Since it’s impossible for everyone to
borrow from each other at once, credit market equilibrium under these
circumstances requires a higher interest rate.

v) Assume again that all individuals are the same and x1i= x2i. If x1i and x2i increase by
the same amount, then (1) will continue to hold without any adjustment in the
interest rate. Even though output and consumption rise, r does not change, since r
determines the price of present consumption in exchange for future consumption. If
both c1 and c2 rise, the relative price stays the same.

How would interest rates be affected by a short-term war? A plentiful harvest? A major
scientific breakthrough that will take time to perfect, such as cars fueled by water? A

7
scientific breakthrough that raises consumption instantly and permanently? How do we
understand the low level of current real interest rates?

Aside: Analysis with and without Money

In the first part of this lecture, we modeled events such as a poor harvest in terms of a real
economy. Later, we introduced a nominal sector in order to consider inflation—without
money, it does not make sense to talk about the aggregate price level or inflation.
Inflation is a change in the price of goods relative to money. There is no such thing as
inflation in a barter economy.

We did not need money in the model of consumption and exchange, because we were
interested in the allocation of real resources. In fact, thinking in terms of money often
leads to confusion, as the following examples show:

“By buying securities rather than investing in property, plants and equipment, companies
harm the economy.”

Securities purchasing, a financial activity, doesn’t substitute for resource allocation. The
seller must figure out what to do with money from the sale. It is not as if capital and labor
somehow vanish into the ether. The picture becomes clearer if we strip away financial
activity and think about the consumption or investment that remains.

“Everyone is selling stock today.”

This claim, which appears in newspapers, is an extreme example of the confusion that
can arise from thinking in terms of money. To whom is everyone selling? For every
seller, there must be a corresponding buyer. ―Follow the money,‖ while perhaps a good
rule for investigative reporters, is less helpful for finding an economy’s equilibrium.

8
You may find this op-ed from the Financial Times a few years ago to be more
controversial (you don’t have to read it if you don’t want to):

If you tax the 1 per

cent it is the middle
class who will suffer
By Richard Robb

The super-rich might consume a lot but there are not very
many of them, writes Richard Robb

T he uneven distribution of material goods is one of the big political issues of

the day. But it cannot be fixed by taking income or wealth from the super-rich
and giving it to the poor. This is not an ideological precept but a matter of logic.

In ancient times, when wealth consisted purely of material possessions, rich

people could strike a blow for social equality by handing over some of their
physical possessions to the poor. After the transfer, the rich would consume less
and the newly elevated poor would consume more.

Nowadays, things are more complicated. The top 1 per cent derive most of their
worth from stakes in businesses and other financial assets. These, too, can be
given away, and the recipients will be able to enjoy more material comforts. But
their gains will inevitably come at the expense of others. The losers would not be
the super-rich, since their consumption is a drop in the bucket. Bill Gates is not
so extravagant that if he were to cut his annual shopping list in half, the savings
would make much difference to the rest of the world.

What Mr Gates could do, of course, is sell a portion of his Microsoft shares and
give the proceeds to the poor. It is tempting to assume that this would alleviate
their plight. But this is uncertain.

Would inflation follow, as the recipients of Mr Gates’s largesse bid up prices of

goods and services? If so, then Mr Gates’s living standards would not suffer, so
long as he retained a large enough part of his fortune.

9
Instead, it is the moderately rich who would end up consuming less, since the
money in their pockets would now be worth less. This, of course, would free up
resources for the poor. But then, so would imposing confiscatory taxation on the
middle class. It is the affluent rather than the super rich who will pay with lower
living standards, no matter which group is the ostensible target.

Shifting financial assets from one person to another will not automatically change
the level of output. Redistributing income or wealth will not cause new treats to
appear on the shop shelves. If the poor are to consume more, others must
consume less. True, members of the top 1 per cent each consume a lot. But there
are not very many of them. Others will have to make do with less if millions of
people are to be given better lives.

Harry Frankfurt’s clear-headed 1987 paper “Equality as a Moral Ideal” never

garnered much attention; perhaps it was published a generation ahead of its
time.

In it, he argued that our concern with equality stems from a moral intuition that
we all share: we object to the poor subsisting in material conditions that prevent
them from leading satisfactory lives.

But this objection does not extend to economic discrepancies among people who
have enough. In that case, pursuing equality becomes a harmful distraction.

Frankfurt warned that a preoccupation with the income of others “tends to divert
a person’s attention away from endeavouring to discover – within his experience
of himself and of his life – what he himself really cares about and what will
actually satisfy him, although this is the most basic and the most decisive task
upon which an intelligent selection of economic goals depends”.

We might desire a more equal distribution to accomplish some other objective:

for example, to curb political corruption or to forestall an uprising. The best way
to stop the rich from abusing political influence might be to clean up politics, or
maybe it is easier just to get rid of the rich altogether. But this does not mean that
we should care about equality for its own sake.

We can leave the debate about whether the dispersion of income is growing to
the man of the hour, Thomas Piketty. The tough question remains: “What does it
mean to have enough?”

10
Everyone should have food and housing, access to healthcare and education – but
how much, and of what quality? It is tricky to define “enough” and easy to
measure income inequality.

But that does not justify a muddled concern for numerical formulas that tell us
little about what is desirable or just.

The writer runs an investment management firm and is a professor at

Columbia University’s School of International and Public Affairs
-------------------------------------------

Introducing Money into the Economy

If we are going to introduce money, we have to motivate the demand for it—i.e., give a
reason why people want to hold money rather than spend it or invest in bonds. Assume
that each person needs to hold enough money to buy one unit of the consumption good.
To keep the model simple, assume that people use money when they go to the store,
where they buy one unit of the good at a time. They have no use for money in excess of
this—if they acquire more than one unit, they lend or spend it. In economic terms, each
person’s demand for money in both t=1 and t=2 is perfectly inelastic.

Also assume that each person is endowed with the same amount of money in the first
period, M1, in dollars. For example, at t=1 each person might have M1= $100.

Given these assumptions, in the first period, the price level equals the per capita money
supply (P1 = M1). If M1>P1, people have more money than they would like and bid up
goods until prices rise. Similarly, if M1<P1, people refrain from spending, and the price of
goods in terms of money drops. People are content to hold the money supply only when
P1=M1.

Now suppose the government increases the money supply at t=2 by giving an equal
amount to each person. Let M2 denote the money supply per person in the second period.
The future money supply is known to everyone at t=1. As before, there are no surprises in
this simple model. In the second period, the price level again equals the money supply (P2
= M2).

We can define inflation, π, as the rate of change in the price level:

11
P2
 1.
P1
In an economy with money, individuals borrow and lend nominal claims—bond buyers
pay an amount of money at t=1 in exchange for some amount at t=2; borrowers receive
money at t=1 in exchange for a payment of money at t=2.

Consider the price of a discount bond, B, that pays $1 at t=2. The ith consumer’s wealth,
in dollar terms, is
P1 x1i  B  P2 x2i  M 1  B  (M 2  M 1 ).

We convert the consumption endowments into money by multiplying by the period’s

price level. M1 is the individual’s endowment of money, and M2–M1 is the extra money
the government gives him at t=2. We multiply money coming in t=2 by B to take the
present value. In short, the expression above shows how much money the ith individual
would have at t=1 if he sold everything he has and everything he will receive.

The consumer spends his wealth on consumption, but must also set aside enough money
to facilitate transactions at t=1 and t=2.3 The present value of the money holdings
requires wealth of M1 + B × (M2–M1), which can’t be consumed. Thus the budget
constraint can be written:

P1 x1i  B  P2 x2i  M 1  B  (M 2  M 1 )  P1c1i  B  P2 c2i  M 1  B  (M 2  M 1 ).

Rearranging terms, the individual’s consumption must adhere to the budget constraint:

B  P2 B  P2
x1i  x2i  c1i  c 2i
P1 P1

Let r* be the interest rate that satisfies equation (3) from our analysis before we
introduced money. The goods market achieves equilibrium if consumers can trade future
1
for present consumption at . If the bond price, B, satisfies the relationship
1 r *

B  P2 1
 (5)
P1 1 r *

3
Think of consumers’ money holdings as cash they must keep in their pockets, rather than invest,
so they can buy goods.

12
then the consumer faces precisely the same maximization problem as in the nonmonetary
economy, and we know the previous solution will work. Therefore, (5) defines the
equilibrium bond price in an economy with money.
P
Substituting the definition of inflation,   2  1 , into equation (5),
P1

1
B .
(1  r*)(1   )

1
Define i = the nominal rate of interest, where B  . Then
1 i

i  (1  r*)(1   )  1  r *  . (6)

Equation (6) says that the nominal interest rate is approximately equal to the sum of the
equilibrium real interest rate (r*) and the inflation rate (π). In this model, inflation is fully
anticipated. Inflation that comes as a surprise in the second period could not, of course,
impact the nominal interest rate at t=1. Thus it’s customary to replace π in (6) with
expected inflation, πe. r is called the real interest rate because it defines intertemporal
exchange of actual consumption goods:

i  r  e .

Thus, in a steady state (where nothing is changing), r = real interest rate = the rate of time
preference, and π = the approximate rate of growth in the money supply.

Besides i  r   e , we have learned one more thing by introducing money that perhaps
was not obvious. As long as increases in the price level are fully anticipated, money is
neutral with respect to the real economy. People end up with the same consumption
patterns in this model as they did in the one without money.

Here is a ―Dot Plot‖ from four years ago, which will be familiar to students from
International Capital Markets. In the long run, Federal Reserve Open Market Committee
members expected interest rates to settle in around 3% (even though they were near zero
at the time). This corresponds to inflation of around 2% (the Fed’s target) and a real rate
of around 1% (a time preference of 2% seems a little high).

13
The dots have been drifting down. This presents something of a puzzle. The latest dot
plot shows the short-term rate converging around 2.5%. We’ll interpret that as 1.75%
inflation and 0.75% real rate.

14
There is some tension between the FOMC participants and financial markets. Here is data
from September 11, 2020:

The top part shows market for Treasury notes. Ten-year T-notes are bid at a yield to
maturity of 0.667% and offered at 0.666%. The yield on 30-year T-notes is only 1.416%.
That doesn’t fit with 2.5% short-term rates, unless the FOMC believes short rates will
take 25+ years to get to their long-run level. If the FOMC expects 0% for five years and
an average of 1.34% for the subsequent 5 years, we can get to an average of 0.67%. With

15
short rates at 1.73% for years 10-20, the 20-year average would rise to 1.20% and match
the 20-year T-note yield. Then if short rates are 1.85% in years 20-30, the 30-year
average would be 1.416%. How do we reconcile this? The market is more in tune with
the single FOMC member who expects long-run rates to reach 2%.

We can be more sophisticated about this calculation (and we will be soon), but this rough
estimate captures the main idea.

The real interest rate presents a greater puzzle. If we expect 1.75% inflation, the real rate
over 30 years is negative: around -0.34%. This is confirmed by the Treasury Inflation
Protected Securities (TIPS) above. This buyer pays 118% for a 0.25% real coupon, i.e.,
0.25% plus inflation. Since the 118 will redeem for 100, the investor loses 18 points over
the bond’s life. This translates to a negative real return (more on yield calculations later
in the course.)

How can the real rate go negative? Nothing in our model precludes it. If households
expect tough times ahead they want to save, as in Equation (1). Eighteen months ago, the
same calculation showed a real rate of 1.20%. It’s not hard to argue that the world’s
economic prospects have taken a turn for the worse since then.

NY Times: ―Flames overtook a fire truck near Oroville, Calif.‖ Sep 9, 2020

Alternatively, CPI may overstate inflation and so the real rate may not be as negative as it
seems.

16
Introducing Capital

Society as a whole can save by accumulating capital (physical, human or social) or

building inventories. Individual countries can also save by running current account
surpluses, thereby accumulating net claims on the rest of the world. (Of course, the entire
world cannot save in this manner at the same time, since current account surpluses must
equal deficits.)

Assume the economy can be characterized by a production function with inputs, capital
and labor. Let the amount of labor be fixed, so output depends on only capital. Defining
kt = capital in period t, any sustainable path satisfies:

ct + kt = f(kt-1) + (1–δ)kt-1.

In words, at the end of t=1, society has that period’s output plus the leftover capital stock
after it depreciates at rate δ. People eat a portion of the t=1 production, ct, and save the
rest. In equilibrium, the interest rate is the marginal product of capital net of
depreciation:

rt-1 = f′(kt-1) – δ.

If rt-1 < f′(kt-1) – δ, an entrepreneur could borrow one unit, invest it for a year, earn f′(kt-1),
replace the capital that depreciated, pay back the loan and end up with a profit. If rt-1 >
f′(kt-1) – δ, an entrepreneur could sell some of her capital, invest in the bond market at rt-1
and end up ahead.

In the steady state, ct = ct+1, so r = the rate of time preference, ρ, which defines the
steady-state capital stock, k*:

Steady State: ρ = f′(k*) – δ.

At levels of k > k*, the extra return from investing does not justify postponing
consumption.

Here’s an example of a growing economy approaching its steady state. Suppose the
economy is subject to a Cobb-Douglas production function with constant depreciation
rate δ:
f(kt+1) = ktα .

17
We can graph the paths of capital, consumption and interest rates if we know the form of
the utility function. For simplicity, assume u(c) = ln(c), so ct+1 = ct[1+α ktα-1 –δ]/(1+ρ).4
The following analysis assumes α=0.5, ρ=0.02 and δ=0.1. In this case, the steady state
level of capital stock is k=17.4, and steady state consumption is c=2.0. In a growing
economy (blue line), as capital rises over time from its starting point of 7.4, consumption
rises and interest rates fall. In an economy that starts with more than the steady state
(pink line, with initial capital of 27.3), consumers eat into the capital stock while interest
rates rise.

Capital Stock Paths with Cobb-Douglas Production Interest Rate Paths with Cobb-Douglas Production Function
Function and Logarithmic Utility and Logarithmic Utility

30 9
8
25
7
20 6

Percent
5
15
4
10 3
2
5
1
0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
-1
Period Period

In this simple version of a ―Solow growth model,‖ destruction of a portion of the capital
stock (for an economy that starts at the steady state) immediately lowers consumption,
raises interest rates and raises savings until the economy rebuilds. A gradual increase in
capital lowers the marginal product of capital as well as the interest rate.

1 1 1+r 1+r
4
From (1), we have =c so ct =ct . Substitute r = f′(k) – δ and note that f′(k) = α ktα-1 to arrive
ct t 1+ρ 1+ρ
at the formula in the text.

18
Non-Exponential Discounting

Let’s consider the multi-period problem with perfect certainty where an agent solves

U (c2 ) U (c3 )
Maximize U (c1, c2 , c3 ,...)  U (c1 )    ... subject to his budget constraint.
1   (1   )2

The agent must decide how to allocate resources to consumption in each period of his
life. We assume that U(c) has diminishing marginal utility of consumption in each period,
that the agent earns interest, r, on his savings, and that all things equal, he prefers a
smooth flow of consumption to a bumpy one. At t=1, he simply allocates fixed resources
over his lifetime, choosing current consumption c1 and forming a plan for c2, c3, c4…
Maximization implies that the marginal utility of consumption at t=1, U (c1 ) , must equal
the discounted marginal utility at t=2:

1 r
U (c1 )  U (c2 )
1  .

This is the same as equation (1) with slightly different notation.

The agent also plans consumption in t=2 than in t=3. Setting the marginal utility of
consumption at the same ratio as before,

1 r (1  r ) 2
U (c2 )  U (c3 ) .
1  (1   ) 2

If the left-hand side of the above equation were bigger than the left, he would do better
with a consumption plan that transferred from t=3 to t=2. Similarly, if the right-hand side
were bigger, he should transfer from c2 to c3 until the equality holds.

When t=2 arrives, the plan set in t=1 for c2, c3… will still seem best because the equation
above is equivalent to U (c2 )  U (c3 )(1  r ) /(1   ) . Put another way, the agent’s
preferences are time consistent.

But time-consistent preferences are only an artifact of geometric discounting (1,

1 /(1   ), 1 /(1   ) , …). The agent could not trade off the present for the future so
2

neatly using any other method. Suppose, for example, that he had such a strong concern
for the immediate present that he defined his problem as

19
U (c2 ) U (c3 )
Maximize U * (c1 , c2 , c3 ,...)  2  U (c1 )    ... subject to budget constraint.
1   (1   )2

Then, as before, at t=1, U (c2 )(1  r ) /(1   )  U (c3 )(1  r ) 2 /(1   ) 2 , or equivalently,
U (c2 )  U (c3 )(1  r ) /(1   ) . When t=2 arrives, 2 U (c2 )  U (c3 )(1  r ) /(1   ) . The
agent must then consume more than planned at t=2 to restore U (c2 ) to the level
anticipated at t=1. However, back at t=1, he would foresee his future self will stray from
the plan. Since the t=2 splurge looks sub-optimal at t=1, he might try to constrain future
consumption using a commitment device.

Examples of commitment devices include locking yourself into a year-long health club
membership, forcing yourself to invest by taking out a mortgage rather than renting your
home (forcing savings when the mortgage payments come due), or placing the alarm
clock across the room to force yourself to get out of bed. An army that crosses a bridge
and then burns it employs a commitment device to prevent retreat.

Functions like U * (c1, c2 , c3 ,...) that cannot be written as

U (c1 )  U (c2 ) /(1   )  U (c3 ) /(1   )2  ... are said to exhibit non-exponential or
hyperbolic discounting, which leads to preference reversals.

Behavioral economists worry about this type of discounting and propose schemes like
forced government savings for those who lack self-control and need an external authority
to bind them to lower consumption in each period.

Personally, I believe that hyperbolic discounting is among the least persuasive themes in
behavioral economics. We are allowed to write down U( ) in the first place because
preferences obey the axioms described on p. 2. Once we write preferences over different
consumption plans that build in contradictions like
U (c 2 )
U * (c1 , c2 , c3 ,...)  2  U (c1 )   ... then U( ) loses its meaning. In my view, the
1 
neoclassical model of intertemporal choice shouldn’t be pushed too far. I have written
about this at length in my book, Willful: How We Choose What We Do (Yale University
Press, November 2019); while I believe I’m right, of course, until my arguments become
more mainstream, we won’t really cover them in this class.

20
APPENDIX – An Episode from The Price Theory Kids, R. Robb, 2001-2002

Many years ago, when I first started teaching at SIPA and my daughter, Alice, was a little
girl, I used to tell her bedtime stories called The Price Theory Kids. The Price Theory
Kids are four middle-school girls (based on actual SIPA students)—Jennifer Katz, Parita
Shah, Veronica Davila and Yuki Taketani—from Pleasantville, New York. The girls use
economic theory to solve problems that arise in their everyday lives. When my daughter
graduated from high school, I wrote them down for her as a present. The Kids had 25
episodes in all. One episode relates to our model on consumption and exchange. Maybe
you can get someone to read it to you tonight! (Note: This is 100% optional obviously.
It’s not part of the course.)

Through the habit of thrift does man ennoble himself

The proverb chiseled into the facade of the First Bank and
Trust Company has provided moral instruction to the
people of Pleasantville for 100 years. But the Price Theory
Kids do not regard thrift as a moral issue. Parita enjoys
making the point, ―The bank wants people to save and put
their money in the bank. What if everyone saved? Then who
would borrow? What would the bank do with all that
money? The bank needs some people who will borrow and
some who will lend to the bank by making a deposit. The
bank lends at a higher interest rate than it pays – that‘s its
business. Without a balance between borrowers and lenders,
the bank will not make a profit.‖

The girls are equally unimpressed by the famous proverbs they study in ethics class.
―Measure twice and cut once.‖ That means we should be careful. ―Fortune favors the bold.‖
That means we should act quickly and take chances. Which is it? They can‘t both be true.
That‘s why the girls prefer economics. An economic theory can be tested by checking how
well it fits the data.

Although the Price Theory Kids enjoy discussing economics,

most of the time they chat about small things, just like everyone
else. The Kids were relaxing in their clubhouse after school one
afternoon when Veronica said to Yuki, ―I was thinking of getting
the new Hello Kitty purse that comes in red and black plaid.‖
―Oh,‖ said Yuki, ―the one where she has a Scottie dog?‖
Veronica nodded. ―I‘ve seen that,‖ Yuki continued. ―It‘s cute.‖

―But it costs $12. Maybe it‘s better to save my money. You know,
as Benjamin Franklin said, ‗A penny saved is a penny earned.‘‖
Parita had been listening to the two girls and joined their conversation. ―Benjamin Franklin
may have known about electricity and politics, but his economics advice seems shaky. What‘s

21
the point of earning money if you never spend it? Veronica, you save every penny you get.
The strap on your old purse broke and there is no way to fix it. If you want the Hello Kitty
purse, I think you should buy it.‖

Jennifer agreed. ―Last night at bedtime,‖ she said, ―my mother told the story of the three
little pigs to Zippy. You‘re supposed to admire the third pig who builds his house out of
bricks. This pig sacrificed while his brothers played and danced a jig all summer. But it only
makes sense to build your house with bricks when the climate is cold or there is a risk of big
bad wolves. If there are no wolves or the climate is warm, it‘s sensible to build your house
out of sticks or even straw. That way you can play while your foolish brother wastes his time
stacking bricks. Telling children they should ‗always prepare for the future‘ by saving is not
wise advice. Sometimes you should save. Sometimes you should consume what you have
saved. And sometimes you should borrow from others.‖

Veronica saw the point. ―In Colombia, near the equator, it

is hot most of the time. Even rich people don‘t have thick
walls or fireplaces. A house made of simple materials is
suitable.‖

―Did you know that the average indoor temperature is

higher in the wintertime in places in the United States with
the coldest climates?‖ asked Parita. ―It sounds surprising,
but you can easily explain it in terms of economics. In cold
climates, people have more efficient heaters that use the
cheapest fuel. It is worthwhile for them to invest in houses
with thick walls and airtight windows. The cost of making
the house a bit warmer is low for them. In a climate where
the temperature is rarely cold, no one bothers installing a
big heater. This makes it more expensive to warm the
house, but it‘s only necessary a few weeks a year, so
homeowners put up with it.‖

―Of course, I appreciate American culture,‖ said Yuki, ―but I really like Japanese children‘s
stories. The morals are reasonable. Usually, some person or animal who is cheerful and
generous will receive a reward. An animal or demon who is mean will end up being
punished. That‘s much better than saying that pigs should work all the time.‖

Veronica looked like she wanted to hear more, so Yuki summarized her favorite story.
―Everyone in Japan knows Bamboo Hats for Jizu. A poor old couple has nothing to eat on
New Year‘s Eve. The wife gives the husband her embroidered hairpins, so he can trade them
down in the village for sweet rice cakes. Somehow he trades the hairpins for bamboo hats.
On his way home, he meets six stone statues called ‗Jizu.‘ Snow is falling on the statues, the
old man feels sorry for them, and so he places the hats on the Jizus‘ heads. When he reaches
home, his gentle wife compliments him for his generosity. The humble couple goes to sleep
hungry but content. During the night, the statues come alive and reward the couple with a
great feast.‖

―Good story,‖ said Jennifer, Veronica and Parita. ―Much better than the Three Little Pigs.‖

22
Veronica decided she would buy the purse. She had also gained some insight into why
people save and why they borrow. They borrow and save so they can live comfortably over
the course of their lives. When they are young, it‘s time to acquire an education that can be
put to use for many years. College usually requires student loans. Education will result in
higher income over time; that income will pay back the student loans. While they are middle
aged, people save so they will be able to retire when they are old. If we could not borrow
and lend, we would have to consume what we earned every year. People would lead
miserable lives when they were young and old and consume a lot during their working years.
That‘s why people borrow and lend—so they can smooth out the changes in their income.

Veronica wondered why most people still have money left over when they die. Why do they
work hard and save money that they can never use? Jennifer could answer her young friend‘s
question. ―Two reasons,‖ Jennifer said. ―People care about their children and grandchildren
and charities. They want to help take care of them even after they die. Also, no one knows
exactly how long they will live, so they might die before they use their money all up.‖

―And look at the way people deal with money they come by unexpectedly,‖ Parita added.
―They spend a bit, but mostly they save it, so they can spread the benefit over their lifecycle.
But if their salary goes up and they expect it to stay high, they will respond differently. In
this case, they spend the extra money as they receive it, because they believe that their
standard of living has risen permanently.‖

―I have an idea for a better version of the Three Little Pigs,‖ Veronica suggested excitedly.
―One pig finds a huge pile of nuts. It‘s a once-in-a-lifetime windfall. He could eat them right
away, but decides he would rather raise his standard of living over time. So he lends a third
of his pile of nuts to each of the other two pigs. These pigs pay him back gradually. The pig
who found the pile of nuts is happy.

―A goat who lives in the yard next door is even happier. A magical tree grew in her yard
overnight. This tree bears an abundant harvest of nuts, fruits and other foods that goats like
to eat. The lucky goat knows that the tree will last a long time. She consumes all the food on
the tree. There is no point in saving it, or lending to others, since she can always rely on the
magical tree for more.‖

―That‘s a very instructive story,‖ Jennifer said. ―It makes more sense than the original
version. Your story might catch on more quickly if you can find a way for one of the
characters to say, ‗let me in by the hair of your chinny chin chin.‘‖

The next time she visited New York City, Veronica bought the Hello Kitty purse. She also
stopped by the Japanese bookstore, where she picked up an English translation of traditional
Japanese stories. She put her new book in her new purse. She had planned to study for her
math test on the train ride home, but there was little point. She had already worked out every
practice problem in the back of the book. Twice. So she decided to enjoy a pleasant train
ride back to Pleasantville reading the first story, ―Bamboo Hats for Jizu.‖

23