LI Macroeconomics
Topic 3: Economic Growth
Dr. Mary Dawood
Associate Professor of Economics
1 / 64
Outline
Facts of Economic Growth
The Solow Model
Solow Model with Population Growth
Solow Model with Technological Progress
The Golden Rule
Romer Model of Endogenous Growth
Growth Accounting
2 / 64
Facts of Economic Growth
3 / 64
Important Facts
Growth is a characteristic of nearly all economies, particularly of
the last 200 years, although there are still different living standards
across the globe
Growth is associated with more international trade and higher
labour productivity; implies workers are better off
Growth also means changes in the structure of the economy
Innovation and the creation of ideas is an integral part of growth
Growth makes people become healthier; but inequality can
increase, and the environment can deteriorate
4 / 64
Important Facts
Standards of living have consistently increased over time for many
economies
5 / 64
Important Facts
International differences in the standard of living, with a few rich
countries and many poorer ones
Country Income per captia (2016)
United States $57,467
Japan $41,470
Russia $23,163
China $15,535
Indonesia $11,612
Egypt $11,132
India $6,572
Nigeria $5,867
Pakistan $5,249
Bangladesh $3,581
Ethiopia $1,735
6 / 64
Important Facts
The relation between growth in GDP and GDP per capital suggests
that high GDP countries grow more slowly
7 / 64
Important Facts
The impact of the industrial revolution on England’s GDP
8 / 64
Important Facts
Growth typically means changes in industrial structure
9 / 64
Important Facts
Increased international trade is associated with higher growth
10 / 64
Important Facts
Growth is associated with higher productivity of workers
11 / 64
Important Facts
Workers are becoming better off
12 / 64
Important Facts
And are having more leisure time
13 / 64
Important Facts
Innovation is becoming increasingly important
14 / 64
Important Facts
And more knowledge being created
15 / 64
Important Facts
Citizens become healthier as their countries become richer
16 / 64
Important Facts
But high income per person can be costly in terms of
environmental degradation
17 / 64
Important Facts
And the benefits of growth are not shared equally
18 / 64
Important Facts
GDP per person, top 0.1%, and bottom 99.9%.
19 / 64
The Solow Model1
1
Mankiw, Chapter 8
20 / 64
Determinants of Economic Growth
Solow model, by Nobel prize winner Robert Solow, examines how
growth in capital stock, growth in labour force and advances in
technology interact to affect the total output
Assumptions
▶ K is no longer fixed: investment causes it to grow, depreciation
causes it to shrink
▶ L is no longer fixed: population growth causes it to grow
▶ G and T are held constant at zero, but will be introduced later
We start by examining how the supply and demand for goods
determine the accumulation of capital
21 / 64
Per Worker Production Function
The supply of goods is based on the production function
Y = F (K , L)
We assume constant returns to scale for simplification
zY = F (zK , zL)
This allows us to write output per worker (i.e. productivity) by
setting z ≈ 1/L
Y K
=F ,1
L L
which is closely related to output per capita
Output per worker is a function of capital per worker, i.e. the
amount of capital that each worker can produce with
22 / 64
Per Worker Production Function
Denoting quantities per worker with lowercase letters, and ignoring
the number 1, being a constant, we get
y = f (k)
where f (k) = F (k, 1)
The marginal product of capital is then
MPK = f (k + 1) − f (k)
This is the extra output that labour produces when given an extra
unit of capital
23 / 64
Per Worker Production Function
The per worker production function exhibits diminishing MPK, i.e.
becomes flatter as capital increases
Output per Output, f (k)
worker, y
MPK
1
Capital per worker, k
24 / 64
The National Income Identity
The per-worker demand for goods comes from consumption and
investment (remember no G )
y =c +i
where y = Y /L, c = C /L and i = I /L
The consumption function per worker is given by c = (1 − s)y
where s is the marginal propensity to save
Thus we can write: y = (1 − s)y + i
Which in turn implies: i = sy = s × f (k)
Note: national income identity still implies that I = S, and
investment still depends on real interest rate, but is removed here
for simplicity of presentation
25 / 64
The National Income Identity
Output per Output, f (k)
worker, y c = f(k) − sf(k)
c
y
Investment, sf (k)
Capital per worker, k
26 / 64
Accumulation of Capital
Thus, capital per worker is the key determinant of output level,
and increases in capital stock will lead to economic growth
Two forces influence the capital stock
▶ Investment causes the capital stock to rise
▶ Depreciation and wearing out causes capital stock to fall
Changes in capital stock are defined by the Equation of Motion
∆k = i − δk = sf (k) − δk
where δ is the fraction of capital that wears out each period (rate
of depreciation)
27 / 64
The Steady State
Investment and
depreciation
Depreciation, δk
δk2
i2 Investment, sf (k)
i ∗ = δk ∗
i1
δk1
k1 k∗ k2 Capital per
worker, k
28 / 64
The Steady State
At k1 : investment exceeds depreciation ⇒ capital stock grows
At k2 : depreciation exceeds investment ⇒ capital stock falls
At k ∗ : investment equals depreciation ⇒ ∆k = 0, called the
steady state level of capital
Notes
▶ an economy at steady state will stay there
▶ an economy not at steady state will go there
⇒ Steady state represents long-run equilibrium of the economy
29 / 64
Increase in Savings
An increase in saving rate from s1 to s2 shifts the sf (k) curve
upward, causing capital stock (and output) to grow toward a new
steady state level, k2∗ , due to increased investment
Higher saving raises output growth only until the economy reaches
a new higher steady state ⇒ temporary economic growth
Investment
δk
s2 f (k)
s1 f (k)
Capital per worker
k1∗ k2∗
30 / 64
Solow Model with Population Growth2
2
Mankiw, Chapter 8
31 / 64
Growth in Labour Force
Solow model shows that capital accumulation and increased
savings cannot explain sustained economic growth
Next, we consider two sources of growth
1. population growth
2. technological progress
First, suppose that population grows at a constant rate, n, such
that
∆L
=n
L
The growth in the number of workers causes the amount of capital
per worker (k) to fall
32 / 64
Growth in Labour Force
Thus, the change in the capital stock per worker becomes
∆k = i − (δ + n)k
The term (δ + n)k is defined as break-even investment, i.e. the
amount of investment required to keep k constant
▶ δk is investment required to replace capital as it wears out
▶ nk is investment required to provide new workers with capital
Then, the equation of motion is written as
∆k = sf (k) − (δ + n)k
where the first term is actual investment and the second term is
break-even investment
33 / 64
Growth in Labour Force
Investment
Break-even investment, (δ + n)k
Investment, sf (k)
Capital per worker, k
k∗
steady state
34 / 64
Growth in Labour Force
Investment increases the capital stock; depreciation and population
growth decrease it
Investment must continually increase capital stock (K ) at the rate
(δ + n) to keep k constant at steady state
Since y = YL = f (k), we obtain a steady state that implies Y
grows at the rate n
Thus, in steady state, k and y are constant, while K and Y are
growing at rate n
Population growth does not explain the sustained growth in
standard of living ( YL ), only the sustained growth in total output
35 / 64
Mathematical Derivation
of Equation of Motion
Differentiating the capital stock per worker (k = K /L) with
respect to time (t) using the quotient rule gives
d KL dK
L − dL dK dL dK
dk dt K dt K
= = dt = dt
− = dt
− nk
dt dt L2 L L L L
where the growth rate of labour dL
dt /L is n
dK
Since dt = ∆K = I − δK , we can write
dk I − δK
= − nk = i − δk − nk
dt L
Thus
∆k = i − (δ + n)k
36 / 64
Impact of Population Growth
A rise in population growth rate from n1 to n2 shifts depreciation
line upwards, leading to a lower steady state level of capital, k2∗
Investment
(δ + n2 )k
(δ + n1 )k
sf (k)
Capital per worker, k
k2∗ k1∗
Economies with higher rates of population growth will have lower
levels of capital per worker and hence lower income levels
37 / 64
Solow Model with Technological Progress3
3
Mankiw, Chapter 9
38 / 64
Efficiency of Labour
To model technical progress, we define the production function as
Y = F (K , L × E )
where E represents labour efficiency
As technology improves, the efficiency of labour increases at a
constant rate, g , such that each hour of work produces more
output
∆E
g=
E
Increases in the efficiency of labour (E ) are analogous to increases
in the labour force (L), as both have the same effect on output
This is referred to as labour-augmenting technological progress
39 / 64
Efficiency of Labour
Changing the notations
▶ y = Y /LE : output per effective worker
▶ k = K /LE : capital per effective worker
Thus, we can write
Y K
=y =F , 1 = f (k)
LE LE
By the same analogy used before, the change in the capital stock
per effective worker is given by
∆k = i − (δ + n + g )k
Hence, the equation of motion is written as
∆k = sf (k) − (δ + n + g )k
40 / 64
Efficiency of Labour
Investment
Break-even investment, (δ + n + g )k
Investment, sf (k)
k∗ Capital per effective worker, k
41 / 64
Mathematical Derivation
Differentiating capital stock per effective worker with respect to
time using the quotient rule gives
dK dL dE dK
dk
K
d LE LE − KE − LK dK dL dE
dt K dt K
dt
= = dt dt = dt
− − = dt −nk−gk
dt dt (LE )2 LE L LE E LE LE
dK
Since dt = I − δK , we can write
dk I − δK
= − nk − gk = i − δk − nk − gk
dt LE
Thus
∆k = i − (δ + n + g )k
42 / 64
Mathematical Derivation
Detailed steps using chain rule
▶ rewrite: K
LE = K · (LE )−1
▶ using dot to denote dK /dt
d K d
= K̇ (LE )−1 + K · (LE )−1
dt LE dt
▶ focusing on the second term
d d(LE )
(LE )−1 = −(LE )−2 · = −(LE )−2 · (L̇E + LĖ )
dt dt
▶ substituting back
!
d K K̇ L̇E + LĖ K̇ K L̇ Ė
= −K = − +
dt LE LE (LE )2 LE LE L E
43 / 64
Impact of Technological Progress
Variable Symbol Steady-State Growth Rate
K
Capital per effective worker k= 0
LE
Y
Output per effective worker y= = f (k) 0
LE
Y
Output per worker =y ×E g
L
Total Output Y = y × LE n+g
The model now can finally explain the sustained increases in
standards of living observed in the data
By contrast, a high rate of saving leads to a high level of output,
with growth only occurring until a new steady state is reached
44 / 64
Policy Implications
Solow model provides a guide for policies to promote growth
▶ increasing the rate of saving by reducing fiscal deficit (negative
public saving) and increasing incentives for private saving
(reducing taxes on investment)
▶ encourage investment in private capital stock, public
infrastructure and human capital (knowledge and skills of
workers)
▶ establishing the right institutions to ensure that resources are
allocated to their best use (e.g. legal inst, financial markets,
promote competition)
▶ encouraging technological progress (e.g. patent laws, tax
incentives for R&D, research grants)
45 / 64
The Golden Rule4
4
Mankiw, Chapters 8 & 9
46 / 64
Optimal Saving Rate
We saw how different values of s lead to different steady states,
since rate of saving and investment determines levels of k ∗ and y ∗
Optimal s rate would produce a k ∗ level which maximises the
well-being of individuals, i.e. maximises the level of consumption
The steady state value of k that maximises consumption is called
∗ )
the Golden Rule level of capital (kgold
Steady-state consumption per worker (from national identity)
c ∗ = y ∗ − i ∗ = f (k ∗ ) − δk ∗
where break-even i = δk to make ∆k = 0
47 / 64
Optimal Saving Rate
The optimum can be found by calculating the first derivative with
respect to k ∗ and setting it equal to zero
max c ∗ : f ′ (k ∗ ) − δ = 0
Since the slope of the production function is f ′ (k ∗ ) = MPK , we
can write the Golden Rule equation as
MPK = δ
At the Golden Rule level of capital, the marginal product of capital
net of depreciation equals zero
This implies that the additional unit of capital per worker is adding
as much to total output as it is to total depreciation
48 / 64
Optimal Saving Rate
Steady-state output,
investment per worker δk ∗
f (k ∗ )
∗
cgold sgold f (k ∗ )
∗
igold
∗
kgold Steady-state capital per worker, k ∗
∗ : increases in k ∗ raises output more than depreciation,
Below kgold
causing consumption to rise
∗ : increases in k ∗ raises depreciation more than output,
Above kgold
causing consumption to fall
Economy does not automatically gravitate towards kgold∗ ; rather we
need a particular s to produce it, which requires policy
49 / 64
Modified Golden Rule
In case of population growth
c ∗ = y ∗ − i ∗ = f (k ∗ ) − (δ + n)k ∗
Thus the Golden Rule equation can be written as
MPK = δ + n
In case of technological progress with population growth
c ∗ = y ∗ − i ∗ = f (k ∗ ) − (δ + n + g )k ∗
Thus the Golden Rule equation can be written as
MPK = δ + n + g
50 / 64
Romer Model of Endogenous Growth5
5
Mankiw, Chapter 9
51 / 64
Endogenous Growth Theory
Romer model explains the process by which growth takes place
endogenously, rather than by exogenous technological progress
It is based on Schumpeter’s concept of Creative Destruction, where
the economy is in a permanent state of flux, with old firms and
market dying out and new ones springing up
The economy grows because of new innovations: new products,
new ways of doing things, new markets
Since ideas are nonrival, innovation moves us away from the
assumption of diminishing returns used in Solow model
If we consider knowledge and physical capital as a single input, a
rise in capital stock can generate constant returns
52 / 64
Basic Romer Model
We start with a linear production function (constant MPK)
Y = AK
where A is a constant measuring amount of output produced per K
This makes the equation of motion
∆K = sY − δK = sAK − δK
Since A is a constant, we can write
∆Y ∆K
= = sA − δ
Y K
So, output grows permanently (no steady state) when sA > δ, and
investment is the heart and engine of growth
53 / 64
Two-Sector Romer Model
There are two sectors in the economy:
1. Manufacturing firms produce goods using production function
Y = F [K , (1 − µ)LE ]
where (1 − µ) is the fraction of labour force working in
manufacturing; E measures labour efficiency
2. Research institutions produce knowledge that increases labour
efficiency in manufacturing
∆E = z × µL × E
where z is a parameter measuring research productivity; µ is the
fraction of labour in research
54 / 64
Two-Sector Romer Model
Setting g = zµL since they are all constant parameters, we can
write
∆E
=g or Et = E0 (1 + g )t
E
As the stock of knowledge grows at rate (g ) so does output, even
holding K and L constant
Thus, Romer Model explains output growth endogenously, where
▶ s affects level of income but not its growth rate (same as Solow)
▶ g affects level and growth rate of income, but is endogenous
There is constant growth in output along the Balanced Growth
Path (substitute for steady state) due to constant returns in
capital, which includes knowledge, as well as physical capital
55 / 64
Balanced Growth Path
A rise in total labour force causes g to increase, because more
researchers contribute to the production of knowledge
At t0 , the economy moves immediately to a higher growth rate
Evidence suggests that population and ideas have grown quickly in
recent history, but without increasing growth rates as much
Y
time
t0
56 / 64
Balanced Growth Path
If the proportion of workers in research (µ) increases at time t0 , g
will also rise
Initially, this leads to a fall in y , since fewer workers are employed
in manufacturing, then y grows more quickly as E increases
Y
time
t0
57 / 64
Growth Accounting6
6
Mankiw, Chapter 9 (Appendix)
58 / 64
Sources of Economic Growth
We start with the aggregate production function: Y = F (K , L)
We can totally differentiate this function to give
∂F ∂F
∆Y = ∆K + ∆L = (MPK × ∆K ) + (MPL × ∆L)
∂K ∂L
Thus, increases in output comes from two sources
▶ rise in K multiplied by extra output from each unit increase in K
▶ rise in L multiplied by extra output from each unit increase in L
We can express the equation in terms of growth rates
∆Y MPK × K ∆K MPL × L ∆L
= +
Y Y K Y L
×K
where MPKY is the capital share in income, and MPL×L
Y is the
labour share of income
59 / 64
Sources of Economic Growth
If the production function is Cobb Douglas [Y = K α L(1−α) ], then
∆Y ∆K ∆L
=α + (1 − α)
Y K L
With technological progress, we can write PF as: Y = AF (K , L)
where A represent total factor productivity
Then
∆Y ∆K ∆L ∆A
=α + (1 − α) +
Y K L A
∆A
Endogenous growth models can provide an explanation of A
∆A ∆Y ∆K ∆L
= −α − (1 − α) = Solow Residual
A Y K L
But the Solow Model leaves it as unexplained
60 / 64
Explaining the Facts
Growth Fact Solow Model Romer Models
Capital accumulation, Physical capital accum,
Growth has been a feature Labour force growth, Growth of knowledge,
of economic life over the last Technical progress Labour growth/allocation
200 years
Efficiency of Labour, Human capital and ideas,
Increases in human capital Human capital share of in- Shift of workers to research
and economic growth hap- come increases
pen together
Efficiency of labour Working methods changing,
Income per capita has risen, Industrial structure changing
even as workers work less
hours
Capital accum may in- Creative destruction is the key
Structural change runs volve industrial changes
alongside growth
Technological progress is Knowledge creates innovation
Innovation leads to new exogenous
products and markets
Households care only Same as Solow model, but re-
Costs to growth, e.g. env about consumption search can be directed to solv-
degradation, inequality ing such problems
61 / 64
Conditional Convergence
Panel data shows that there seems to be convergence of growth
rates across countries of similar structure, called conditional
convergence, but not general convergence across the globe
The models tell us that differences in income per capita are due to
▶ Capital (physical/human)
▶ Technology (productive efficiency)
▶ The lack of diminishing returns
▶ Innovation
Conclusion: differences across countries are due to the differences
in factors and the way they are utilised
62 / 64
Finance and Innovation
Long-run models of macroeconomics see the financial sector as
playing a passive role ⇒ classical dichotomy
Endogenous growth models provide a mechanism for the financial
sector to play an active role
▶ creating funding mechanisms for new innovative activities
(microfinance, small business funding)
▶ facilitating the trading of knowledge to ensure it is allocated
efficiently (e.g. market for patents)
However, financial markets can also push the economy in directions
it would not otherwise go
▶ speculative investment leading to resource misallocation, e.g.
dotcom bubble
63 / 64
Policy Implications
In addition to policy implications of Solow model, endogenous
growth models highlight the importance of promoting the
development of knowledge
▶ Education
▶ Scientific progress
▶ Market driven innovation e.g. patenting
▶ Industrial policy
A key issue for policy-makers is that knowledge is a public good,
while innovation is typically driven by the private sector
This can raise issues with regard to externalities that may imply
the private sector will either produce too few or too many ideas
64 / 64