Scribd
Upload
39 views

Lecture 05 - 06 Notes - Spring 2009 PDF

The document summarizes the neoclassical growth model. It discusses two main sources of differences in GDP per capita across countries: 1) physical capital stock, which depends on investment rates and savings; and 2) productivity, which depends on technology and efficiency. It then presents the Solow growth model, which uses a production function to model output as a function of capital and labor. The model predicts convergence of countries' per capita GDP over time. However, it cannot explain sustained long-run growth. The AK model is presented as an alternative that can explain positive long-run growth through externalities from capital accumulation.

Uploaded by

6doit
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
39 views

Lecture 05 - 06 Notes - Spring 2009 PDF

The document summarizes the neoclassical growth model. It discusses two main sources of differences in GDP per capita across countries: 1) physical capital stock, which depends on investment rates and savings; and 2) productivity, which depends on technology and efficiency. It then presents the Solow growth model, which uses a production function to model output as a function of capital and labor. The model predicts convergence of countries' per capita GDP over time. However, it cannot explain sustained long-run growth. The AK model is presented as an alternative that can explain positive long-run growth through externalities from capital accumulation.

Uploaded by

6doit
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Lecture 6: The neo-classical growth model

February 23, 2009

The main sources of divergence in GDP per


capita across countries
A first source of dierence in GDP per capita is the stock of physical
capital. Physical capital consists in the all the non-labor inputs. e.g
machines, vehicles, buildings, and other pieces of equipment. For example,
in 2000 the average US worker had $148,091 worth of capital to work with;
in Mexico in the same year, the capital per worker was $42,991 and in India
it was $6.270.
Question: where do the dierences in physical capital stocks come
from?
Answer: from dierences in investment rates, which themselves come
from dierences in savings (savings rates, initial income)
Second source of dierence in GDP per capita: productivity, i.e how much
output is produced with each unit of capital.
1. First source of productivity dierence: technology, i.e the amount of
knowledge accessible to producers in dierent countries. Technology
is acquired through innovations (R&D investments) and imitation of
technologies from more advanced countries.
2. Second source of productivity dierence: eciency, which relates
more to the organization of the economy, institutions, and so on.
Of course, to get a good understanding of the dierences in GDP per capita
and of growth rates across countries, one must go further and understand
what ultimately determines savings, R&D investments, technological diffusion, productive eciency in firms and markets.

The neo-classical growth model


his lecture presents a capital-based theory of why countries dier in their
levels of income per capita, and why less advanced countries may or may
not converge (in per capita GDP) towards more advanced countries.
1

2.1

The nature of capital

Five main characteristics of (physical) capital:


1. Capital is productive: it raises the amount of output a worker can
produce
2. Capital is itself produced. The process of producing capital is called
investment. Because it is produced, capital requires sacrifice of some
consumption. In the US, in 2000, 1.76 trillion dollars, i.e 18%, were
spent on investment.
3. Capital is rival in its use: only a limited amount of people can use
a given piece of capital at a given moment in time; this distinguishes
capital from ideas, which in turn are produced through another kind
of investment: R&D investment.
4. Capital earns a return from renting it. However some capital goods
like roads and ports are built and owned by governments.
5. Capital depreciates, both due to physical obsolescence and also because the arrival of new technologies make capital goods that embody
old technologies, become obsolete.

2.2

The first equation of the Solow model

The first equation you already know:


Y = AK L1 ,

(1)

where the left-hand side is the current flow of output, and the right hand
side is equal to a technology parameter A times the capital contribution
times the labor contribution.
Remember that:

K M P K = Y,

so that

K MPK
Y
is the capitals share of output. Figure 3.3 shows that share for a sample
of 53 countries. The average is about 1/3.
=

2.3

The second equation of the Solow model

Growth rate in discrete time, between t and t + 1 :


gt (z) =

zt+1 zt
.
zt

Growth rate in continuous time:


gt (z) =

dz
z
/z = = zb;
dt
z

Second equation of the Solow model:

dK
= sY K,
dt
that is: net current capital accumulation is equal to total current investment (itself equal to total savings in equilibrium of the goods market)
minus total current depreciation of the capital stock.

2.4

Steady state

Fix labor supply at L = 1. The steady-state level of capital is simply


determined by:
dK
= 0,
dt
or equivalently
sY = sAK = K,
or equivalently again:
K = K ss = (

1
sA 1
.
)

This steady state is stable, in the sense that starting from a level of capital
K < K ss , capital will accumulate until the capital stock reaches its steadystate level K = K ss ; similarly, starting from a level of capital K > K ss ,
capital will decumulate until the capital stock reaches K = K ss .
To this steady-state level of capital corresponds a steady state level of
output:
1
s
Y ss = A(K ss ) = A 1 ( ) 1 .

Remarks:
1. This latter equation provides us with a theory of the sources of (longrun) income dierences across countries: in particular (per capita)
GDP across two countries may dier either due to dierences in productivity as captured by A, or because of dierences in savings or
investment rate rate as captured by s. Thus, suppose, that we abstract from dierences in technologies and concentrate in dierences
in savings or investment rates, and compare the predicted ratio of
income per worker in each country to income per worker in the US,
to the ratio of actual incomes between that country and the US. We
are very far from the 450 line. In particular, Uganda should have a
3

GDP per capita of around 33% of US GDP per capita, however the
true ratio is only 3%. Why? Maybe Uganda still lies very far from
its steady-state, whereas US lies close to its steady-state.
2. The Solow model predicts no growth of GDP per capita in the longrun, since output per capita converges to a fixed value Y ss . The
reason for this no-growth in steady state result, is the assumption of
decreasing returns to capital accumulation and therefore of decreasing savings to output ratio, whereas depreciation occurs at a constant
rate; so, eventually, depreciation catches up with savings.

2.5

Convergence

The Solow model predicts cross-country convergence: namely, a country


very far from its steady-state will grow very fast, whereas a country very
close from its steady-state will grow very slowly. For example, suppose
that all countries have the same savings rate, productivity parameter, and
depreciation rate, but the only dierence between them is that they start
from dierent levels of capital stocks (and therefore from dierent income
levels). The growth rate of capital stock is equal to:

thus since

b = dK /K = sAK 1 ,
K
dt
< 1,

we see that the higher the current level of capital, the lower the growth
rate of the capital stock, and therefore the lower the growth rate of output
since
b
Yb = K.

Thus the country will lower current level of its capital stock will grow faster
than the country with the higher level of its current capital stock. Note
also that the Solow model leads to the following additional predictions:
1. if two countries have the same rate of investment but dierent levels
of income, then the country with lower income will grow faster;
2. if two countries have the same level of income but dierent rates of
investment, then the country with a higher rate of investment will
have higher growth;
3. a country that raises its level of investment will experience an increase
in its growth rate of income.

While the Solow model predicts convergence to the same steady state for
all countries or regions with similar savings rates s, depreciation rates
and productivity parameter A, it does not explain why some countries
manage to converge towards the most advanced countries and why others
4

stagnate and do not converge. This phenomenon we refer to as club


convergence. The Solow model also does not explain why some countries
started to grow very fast and then stopped converging.
While the convergence prediction appears to be (partly) verified by crosscountry (or by cross-state data within the US), the prediction of zero
long-run growth is at odds with the evidence that growth in developed
countries has been sustained at 2 to 3% per year. So, how can extend the
Solow model to account for positive long-run growth?
1. allow for technical progress in the form of a permanent growth in productivity A. However, the problem is that we cannot explain how this
technological progress will be remunerated, especially since the Euler
theorem implies that once you pay capital and labor at their marginal productivities, you simply exhaust total output Y and therefore
there is nothing left to pay innovators.
2. allow for population growth; however this will aect the growth of
total output, not the steady-state result on output per capita: suppose
Y = AK L1 ,
with
L = ent
=
y = Ak ,
where y = Y /L and k = K/L. Also, one can show that equation (2)
implies:
dk
= sy k.
dt
So, we are back to the previous model, which implies that k converges
to a steady-state level kss and y converges to a steady-state level y ss .
This implies that in the long run, total GDP Y grows at same rate
as population, that is at rate n, but per capita GDP stops growing
in the long-run.

2.6

A first glimpse at the AK model

A first attempt at endogeneizing the long-run growth rate, was to say


that even though individual firms may experience decreasing to capital
accumulation at an individual level, yet, because of externalities in learning by doing across firms (or externalities in capital accumulation across
firms), at the aggregate level the production function might exhibit constant returns to capital only, namely:
Y = A0 K.
This view of the world identifies knowledge accumulation with capital
accumulation by all firms simultaneously.
5

Now, let us replace the first equation in the Solow model by this equation,
and put it together with the second equation of the Solow model; we have:
dK
= sA0 K K,
dt
and therefore:

b = sA0 .
K

We thus get a positive long-run rate of growth, which depends positively


on the savings rate and negatively on the depreciation rate. However, if
we restore positive long-run growth, we also lose the convergence result of
the Solow model since all countries no matter their current capital stock
growth at the same rate. Therefore less advanced countries can no longer
catch up with more advanced countries since they do not grow faster than
more advanced countries.
Note that the learning-by-doing externality must be exactly right in order
to obtain an aggregate AK function. More generally, externalities will lead
us to an aggregate function of the form:
Y = A0 K ,
where
> .
b increases with K), whereas if
If > 1, then we get explosive growth (K
< 1 we obtain zero growth in the long-run as we are back to the Solow
model but with a higher coecient of capital.
Early attempts at justifying the AK model on empirical grounds, relied
on the observation that the Solow model with = 1/3 predicts too fast a
speed of convergence compared to what we seem to observe across countries or across US states. However, next time we shall see that there is
a simple way to extend the Solow model so as to deal with this problem:
namely, by adding human capital on top on physical capital and labor as
a third factor of production.
In any case, from this subsection we have seen that growth models based
on capital accumulation can not explain convergence and long-run growth
simultaneously. And they do not account for several important aspects of
convergence/divergence across countries.

You might also like