CHAPTER FOUR

INVESTMENT

INVESTMENT

INTRODUCTION
Investment makes up 20-33% of GDP, and although obviously not as large as consumption, it is
important because investment adds to the capital stock, which a key input used to produce output
and most of the fluctuations in output seem to occur because of fluctuations in investment. As a
result, it is crucial that we understand what factors affect investment and how they do so not only
because it is part of GDP but also because it is crucial in determining fluctuations and growth in
GDP.

CATEGORIES OF INVESTMENT
Before developing any theories about investment choices by people, it is important to define
what is meant by investment.
Investment: is the formation of real capital, tangible or intangible, that will produce a stream of
goods and services in the future. Investment undertaken in an economy is classified according to
the following categories:
i. Business Fixed Investment — plant and machinery, office buildings
ii. Residential Construction Investment — houses, and significant additions to
houses (that require building permits).
iii. Changes in Firm’s Inventories — as they provide revenue in the future, nottoday.
iv. Purchases of Durable Goods by Households — cars, fridges.
v. Government Purchases of Investment Goods and Buildings — roads, hospitals,
harbours.
vi. Investment in Human Capital — schooling, university study, apprenticeshipsetc.
vii. Knowledge — things we know about scientific and other areas.
The national income accounts only measure 1, 2, 3, and 5 as investment and it is not clear as to
how well they measure 2 and 5 (e.g. additions not requiring permits, anddurables purchased by
government departments). The national accounts do not measure 6 or 7.

1
Investment can be undertaken not only by firms or householders, but also by the government.
For example, looking at Business Fixed Investment of different countries, government sector
undertakes significant proportion of this category:

1.2 INVESTMENT AFFECTS FUTURE OUTPUT

The capital stock is a very important input used to produce output and the level of investment
ultimately determines the level of the capital stock and thus the growth of output. The
relationship between investment and the capital stock can be expressed in the following way:

K t 1
 K t  I t  K t

Where

K t
= stock of capital at time t.

It = gross investment at time t.

δ = the depreciation rate.
Another way of stating this relationship is looking at the net change in the capital stock, which
can be expressed in the following way,

K t 1
 K t  I t  K t  I n,t

Where I n,t is net investment at time t. This implies that if I n,t increases then K t 1
will increase

and thus Y t 1  Y t , that is, output should increase. Note that the notation used above assumes

that investment undertaken today does not really become productive until the future. This is
usually the assumption we use in modern macro.

1.3 INVESTMENT FLUCTUATES A LOT

A well-established fact of modern economies is that I fluctuates a lot more than C or Y (we can
also see that Y fluctuates more than C, probably because of the large fluctuations in I which
helps to make up Y). Furthermore, inventory investment, which makes up only a small
percentage of GDP, is extremely volatile and in a recession, more than half the fall in spending
for a typical industrialized country comes from a decline in inventory investment as firms run
down their inventories. As a result, if we want to explain fluctuations in output, then we need to
understand what causes fluctuations in an economy’s investment behaviour. This has been
recognized for quite a long time, including by Keynes himself who put it down to ―animal

2
spirits‖, or that fluctuations of firm’s investment is due to seemingly random fluctuations in
firm’s expectations about future profitability. There is likely an element of truth in this view,
since we have seen that people’s consumption choices depend on future income, but it is in some
sense a pessimistic view since it essentially says that investment is what it is and that is all we
can say. For the rest of the topic we shall study three categories of investment and in the process
see if Keynes was right.

2. BUSINESS FIXED INVESTMENT

Many people are intensely interested in business fixed investment and it is not difficult to
understand why; it constitutes the basic equipment used by firms to produce output whether it be
computers, factories, machine tools, or offices. We will learn about two theories developed to
explain this form of investment.

2.1 THE ACCELERATOR THEORY OF BUSINESS FIXED INVESTMENT

The first coherent theory for explaining the level of aggregate investment and why it behaves as
it does is called the accelerator theory of investment. Its initial focus is on what firms might do
and then applies the results to aggregate investment.

Theory behind the Accelerator Model

The theory’s line of argument goes as follows. Firms produce output and they need capital to
produce this output. It is likely that a firm has some desired stock of capital in mind when
producing any given amount of output. If the desired capital stock differs from what the firm
actually has, then it must change its capital stock and how much it changes it is the amount of
investment. When would a firm change how much capital it desired? Following this line of
reasoning the answer is easy, when output changed. Investment is thus the result of changes in
output.

The Formal Model

We will now write the basic idea described above as a formal model. To do this, define the
following variables:

k t
= the actual capital stock of the firm at time t.

3
 d
k = the desired capital stock of the firm at time t.
t

y = output produced by the firm at time t.

t

i = the amount of net investment by the firm at time t.

t

 = the desired capital to output ratio by the firm.

Next, we need to specify the two key assumptions used by the model:
1. There exists a long-run desired amount of capital related to the amount of
d
output, or, k t = α y .There may be short-run deviations from this amount
t

because of output fluctuations, but on average, this relationship holds.

2. All investment undertaken today is assumed to be ready for use today (note
that this differs from the ―standard‖ assumption that investment today only
becomes productive in the future). In symbols we can show this as:

i k k
d
t t t 1
and notice the different time subscripts as to why this is true.

This is a strong behavioural assumption that businesses know what is going on

all the time and adjust for it very quickly.
Now, put everything together. The logical implication of the second assumption is that the
capital stock each period equals the desired capital stock, or that,

k  k  y
d
t t t

This means that we can rewrite the above equation for net investment in the following way:

i k k
d
t t t 1

k k t t 1

 y  y
t 1

  y  y 
t

t t 1

  y
t

So what we find is that investment is determined by changes in output. This is called the naive
accelerator model, since all the change happens in one period.
Example: if  y t
 y  y  = $2 million and α = 1 then i
t t 1 t
= $2 million.

Testing the Accelerator Theory of Investment

4
What we have derived above for a firm is a simple specific relationship between investment and
output. Now apply this model to aggregate investment then we can see if the model is consistent
with what we observe. It is relatively easy to do as all we need is aggregate investment and
output data for each year and we can calculate A for the economy from the equation,

A I t
,
Y Y
t t 1

Where the upper case letters just mean we are using aggregate data. The above quantity clearly
does not have a roughly constant value of A nor is it always positive. The former problem can be
waived away as an aggregation problem, but the latter cannot. Note too that there are other more
sophisticated versions where the adjustment to the desired capital stock takes longer time periods
called flexible adjustment accelerator models. These assume that there are costs involved in
adjusting the stock of capital (e.g. plant shut-downs, overtime costs) and that investment changes
less than one-for-one with changes in output. While these models seem to be better at explaining
what we see happen in the real world, they still do a pretty poor job overall.

Why the Accelerator Model Is Inconsistent with Aggregate Investment – There are likely to
be two possible reasons as to why the accelerator model is not consistent with observed
aggregate business investment and knowing these helpsus to create a model that will be better at
explaining aggregate investment.
These are both related to missing explanatory variables:
1. Relative Prices
A very strong implication of the assumptions underlying the accelerator model, in the naive or
the flexible versions, is that wages, prices, taxes, and interest rates, have no direct effect on the
level of investment. They do have an indirect effect if they have an effect on output. This rules
out the existence of substitutions between factors of production. For example, it rules out
changes in the ratios K/Y and L/Y if their relative prices change. It seems unrealistic that interest
rates would not have a direct influence on investment.
2. Expectations of the Future
Another factor, which is ignored in the accelerator model, is the role of expectations about future
conditions. This seems unreasonable because if investment is productive tomorrow, then the
benefit of today’s investment depends on tomorrow’s economic conditions. But this means that

5
we have to form expectations about what tomorrow is going to be and so our expectations about
tomorrow affect our investment plans today. This tells us that a dynamic approach to figuring out
investment is really needed to explain what happens. This has been done, but is very
complicated, so well ignore this aspect for the course although you should be aware of it in real
life (you do study this if you do more economics).

2.2 NEOCLASSICAL THEORY OF INVESTMENT

Our next theory of aggregate investment, and the one that is currently used by economists.The
key difference between it and the accelerator theory is that whereas the accelerator theory did not
explore where the desired capital to output ratio came from, the theory we are going to learn
about explicitly invokes the optimization principle used in economics.
Underlying Theory
An important underlying assumption of the accelerator model is that firms do not change their
production techniques. In effect they employ the same proportions of capital and labour
independentof their prices. The neoclassical theory of investment looks at investment decisions
as being affected by output, as with the accelerator model, but on the basis of a cost-benefit
decision. That is, firms may have a desired output level they wish to produce, given the price at
which the output can be sold, but the bundle of inputs they use will depend on their relative costs
and benefits so as to maximize the profits of firms.

The Formal Model

We will now write the basic idea described above as a formal model. To begin with define the
following variables:
y = the quantity of output produced by the firm.
k = the quantity of capital used by the firm.
l = the quantity of labour used by the firm.
f [K,L] = the production function which relates how the inputs capital and labour are used
to produce outputs.
Next, we need to specify the key assumption used by the model:

6
1. Firms act to maximize their profits where profits equal revenue less costs. – If firms maximize
profits by hiring labour and capital, then the amount they choose to hire will satisfy the condition
MB=MC for each factor, or in this case their MRP equals their cost. Let,
W = the marginal nominal wage cost paid to employees.
R = the marginal nominal ―rental‖ cost that firms incur in using capital; Where it is
important to note that the implicit or explicit cost of using capital is not simply the actual price of
the capital since the capital can be resold. The cost of using the capital is the implicit (or explicit
if rented from another firm) cost of renting the capital over the period it is used (we will discuss
this in more detail soon) which we are calling R (e.g. if we build a dam and use it to produce
electricity then the capital cost per unit of electricity is not the total cost of the dam, but some
fraction of it relating to what portion of it was used to produce the unit electricity). Given our
assumption, we know that the logical implications that capital and labour are employed up to the
point where.
W
P * MPL  W or MPL 
P
And
R
P * MPK  R or MPL 
P
We will now explore more about the MB and the MC of using capital.

1. Marginal Benefit of Using Capital

The MP that the production firm earns from the output it produces using the capital gives the real
benefit of using capital. It is also worth notinthat the MP of the factors, and hence the MPK,
depends in general on the actual quantities of the factors used. That is,
y f k , l 
MPK  
k k
Where it is clear that the level of output depends on both l and k. In many cases, economists
assume diminishing returns to the factors of production, orthat as we increase the amount of a
factor used output increases but at diminishing rate. We will use this assumption here.
From following the logical implications of our assumptions, we have deduced that there are three
things that will influence the benefit of using capital:

7
 1. Stock of Capital — e.g. the higher the stock of capital, the lower the MPK. This
is just the assumption of diminishing returns to the factors of production.
2. Quantity of Labour Used — e.g. the higher the quantity of labourused, the
higher the MPK.
3. State of the Production Technology — e.g. the better the technology, the higher
the MPK.

2. Marginal Cost of Using Capital

We know that the real cost of using capital is not just the price of capital, but is instead a fraction
of this relating to the period over which the capitalist used, which we have called the real rental
cost (or R/P above). This is made up of the following factors:
i. Foregone Interest Earnings (or interest costs), say the nominal rate of interest is
it hen iPK is the interest cost, or the foregone interest cost, where PK is the
purchase price of capital. This in fact gives you the fraction of the actual
purchase price of the capital, which is relevant for the period for which it is used
to produce things.
ii. Valuation Losses or Gains, i.e. PK may go up (valuation gains) forgo down
(valuation losses). Let −ΔPK = cost of gains or loss.
iii. Wear and tear of capital hired out, i.e. the capital depreciates in value when
rented out purely from its being used to produce goods and services. Let δ = the
rate of depreciation. This implies that δPK is the nominal depreciation cost.
From i –iii, we now have the total cost of renting out one unit (or using)capital for producing
output for one period from i-iii: nominal cost of capital = rPK − ΔPK + δPK
= PK(i − ΔPK/PK + δ)and to summarize, the nominal cost of capital depends on:
1. The price of capital
2. The interest rate
3. The rate at which capital prices are changing
4. The depreciation rate
What we have worked out is perfectly correct, but we will invoke one more assumption to
simplify the expression for the cost of capital. For simplicity, assume that the price of capital

8
goods increases at the same rate as the inflation rate. This can be characterized mathematically
as,
ΔPK/PK = π where,
π = the inflation rate.
We also know that r = i−π (i.e. the real interest rate is approximately equal to the nominal
interest rate less the inflation rate — this should really be the expected inflation rate, but how do
we calculate this?!). Hence, nominal cost of capital = PK(r + δ) and dividing through by P and
we get, real cost of capital = (PK/P) (r + δ).
This tells us that we can expect the real cost of a capital to depend upon:
1. The relative price of the capital good
2. The real interest rate
3. The depreciation rate.
3. Profit Rate of Capital
Now we know what determines both the MB and the MC of using and thus owning capital. We
are now in a position to answer the basic question of how much capital should firms in total use
and own. The answer to this is that it all depends on the profit rate of capital, which equals,
Profit rate = revenue − cost
= MPK − (PK/P) (r + δ)
There are three cases to consider:
I. If MPK > cost of capital, then there are potential profits a firm could earn by increasing
its capital stock. This implies that K is increases and the amount of net investment is
positive (i.e. K t 1  K t
).

II. If MPK < cost of capital, then the firm is making losses on the last units of capital it is
using and it could increase profits by reducing its capital stock. As a result K increases
and the amount of net investment is negative (i.e. K t 1  K t ).

III. MPK = cost of capital, then the firm is maximizing the amount of profits it can earn from
using capital and does not want to change its capital stock (i.e. K t 1  K t
).

Therefore, a firm’s decision about whether to add to its capital stock, or reduce it through
depreciation depends upon the profit rate or the difference between the MPK and the cost of
capital. Note that unlike the accelerator theory, the neoclassical theory shows us what determines

9
the optimal or desired capital stock, it is the amount of capital where MB=MC, or no more
profits can be made from changing the firm’s amount of capital.
Investment and the Capital Stock
A logical conclusion from the neoclassical theory of investment is that both net and gross
investment of firms are a negative function of r because the cost of capital is positively affected
by changes in r, which negatively affects the profit rate of capital. We can show this relationship
graphically as:

Real r

I(r)

And a change in any other variable that affects the MB or MC of using and owning capital by
firms in an economy other than r can be shown by a shift of the investment function (e.g. an
increase in the MPK, say from a technological innovation, causes a rightward shift in the
investment schedule.
The amount of investment that is undertaken depends on whether or not firms are using and
owning the optimal amount of capital, that is, whether or not the profit rate is zero or non-zero.
In the long-run equilibrium, with the profit rate-equalling zero, we would expect that the amount
of net investment is zero with gross investment being positive and just offsetting the amount of
depreciated capital each period. Once a firm is in its long-run state, only a change in r, romp, or
some other variable, will affect the demand for capital and thus the amount of investment

10
demanded. This will cause net investment to stop being zero and the capital stock will adjust to
get back to a new long-run equilibrium situation.
Effects of Tax Laws
We have focused on a few key variables, which affect the amount of investment undertaken, but
in reality, anything, which affects the cost or benefit of using capital, will affect the amount of
investment undertaken by firms. One key setoff variables are tax rates set by the government.
We can amend the profit rate relationship we derived above to include the effects of government
taxes. Let τbe the tax rate on firm revenues. Then the after tax profit rate of capital equals,
Profit rate = (1 − τ) MPK − (PK/P)(r + δ)
So we can see that the tax reduces the profits earned from capital by reducing the MB of using
and owning capital. For example, three common policies, which affect investment, are:
1. Taxes on Income Earned from Owning Capital
This is simply the company tax. This tax is not directly levied on capital income but on a
company’s profits, which include income and costs formal factors, but some proportion of its
incidence is on income earned from capital and we can think of it in this way. An increase in the
company tax will effectively increase τ and hence decrease me. This sort of policy will
discourage the accumulation of K.

2. Investment Tax Credits

This is a policy where firms’ tax liabilities are reduced by how much they spend on new capital
during a year. This effectively lowers τ, reducing the cost of capital and causing an increase in I.
This sort of policy will encourage the accumulation of K.

3. Tax Treatment of Depreciation

Corporate profits are taxed and profits are just revenue minus costs. Part of the costs of
production of firms is depreciation on capital so that firms are allowed to include depreciation of
their capital as a cost when calculating the amount of profit to be taxed. But depreciation is
calculated using historical cost, not replacement cost which tends to understates the true cost of
depreciation in periods of high inflation, which overstates the profits being made. This means a
tax is levied even if the economic profits zero, which effectively increases τ, reducing I, and
discouraging the accumulation of capital.

11
3. RESIDENTIAL INVESTMENT
Another type of investment dear to the hearts and wallets of many households, as well as
involving large expenditures, and one which varies strongly with the business cycle, is
residential investment. What we want to know is what factors determine investment in residences
and how they do so.

3.1 NEOCLASSICAL MODEL OF RESIDENTIAL INVESTMENT

Underlying Theory
The neoclassical model of residential investment directly uses an important concept when
thinking about capital: the existing capital is a stock variable whereas investment is a flow
variable. This is true of any capital and investment but is made explicit in the model we will
learn about. The idea is that there are two parts to the housing: the market for the existing stock
of houses which determines the equilibrium housing prices and the supply of new housing which
determines the flow of residential investment. Note that this is again true for all forms of capital
but we are using it explicitly here in our model.

The Formal Model

First off, define the following variables:
PH = the going price of housing in an economy
P = the price level in an economy
KH = stock of residential capital or housing.
IH = investment in housing.
Determining the Equilibrium Price PH/P — the Market for Existing Houses
As with any normal market the relative price of housing, PH/P is determined jointly by demand
and supply. The stock of houses is fixed in the short-term and so supply curve of existing houses
is vertical. If we invoke the assumption of the law of demand, then the demand for housing is of
the normal downward sloping form in the relative price of housing. You might ask how, given a
fixed population, the demand for housing can change if the relative price of housing changes. It

12
is easy to see why using an example. Say for example, suppose PH/P increases then the quantity
of housing demanded falls because:
People live in smaller houses
People share residences
Homelessness

Determining Investment — the Supply of New Housing

We know that the supply of housing does change over time. We also know that the cost of
building new houses is made up of materials, hiring labour, and hiring capital (tools, machines
etc). These form the MC of building new housing. We will assume the law of supply that
supplying extra quantities of new housing is more expensive because of diminishing returns to
the factors of production used by the firms. That is, the existing resources become more
expensive as firms bid up their prices to get them, additional factors require higher payments to
induce them to supply their services, or the additional factors are just not as productive as
existing resources used. The MB for firms or developers from building new housing is the
relative price of housing. To induce construction firms to build new houses the MB, or
relativeprice, of building new housing has to be higher than the MC. This tells us that supply of
new housing curve is upward sloping in the relative price of new housing.

Changes in the Demand for and Supply of Housing

We now have a formal model of the residential housing market. Any change in the relative price
of housing just involves movements along the existing curves, as per normal supply and demand
curves. Any other change involves a shift of a curve which will have other flow on effects. We
will look at one particular example.
Say a country’s population increases either because of an increase in the reproduction rate of
because of higher net immigration. This causes the following sequence of things to happen:
1. Demand for housing curve shifts outwards
2. PH/P increases to clear the housing market
3. Supply of new housing increases, increasing the future stock of housing (this will then cause a
fall in PH/P although it will still be higher than the original level.

13
There are many other things that can affect the demand for housing including: the interest rate
(i.e. because it affects mortgage costs, or the opportunity cost of holding wealth in housing); the
tax deductibility of interest payments (i.e. if these exist they increase the benefit from owning
houses; better technology, especially in building materials and construction techniques (as it
lowers the cost of supplying new houses and thus shifts the supply curve down); incomes of
people (i.e. how much income people have affects their demand curves); and a host of other
factors.

4. INVENTORY INVESTMENT
The last category of investment that we will study is inventory investment. It is not that large
compared to the other forms of investment, but as mentioned at the beginning of this topic there
is a strong relationship between changes in inventory investment and changes in output over the
business cycle.

4.1 REASONS FOR HOLDING INVENTORIES

Before developing a theory of inventory investment, we will study the reasons why inventories
are held. Knowing the reasons for holding inventories will help us understand better the factors
which influence the level of inventory investment. There are four reasons for holding
inventories:

1. Production Smoothing
Often it is cheaper to produce goods at a steady rate, than to continually alter production runs.
This means that during slow periods inventories build up and during boom periods inventories
are run down.

2. Inventories as a Factor of Production

Firms carry inventories of spare parts in case a machine breaks down, so that the entire assembly
line does not come to a stop.

3. Stock-Out Avoidance

14
To avoid lost sales and profits when sales may be unexpectedly highe Abook retailer may carry
multiple copies of a book if they are not sure how popular the book will be.

4. Work in Process
Partly completed products are counted as inventories e.g. cheese making where cheeses are
aging.

An important point to notice about all of these reasons is that inventory investments have a short-
term focus and not a long-term focus as with the other forms of investment. This suggests that
they will not be so heavily influenced by short-run changes in relative prices but will instead be
tied closely to output. Another point to note is that reason 1is not likely to be an important reason
why inventories fluctuate during business cycles since this sort of inventory investment goes up
as output goes down where as inventory investment as a whole goes down as output goes down.
This means reasons 2-4must be the main causes of fluctuations in inventory investment during a
business cycle.

15