Chapter 3

Todaro & Smith (2015), Economic Development

NeoClassic Theories of Economic Growth

and Development

Lucia Piscitello, 16th March, 2021

Dipartimento di Ingegneria Gestionale

 Robert Merton Solow,
Neoclassical Growth Model

Solow, R.M. 1956. "A contribution to

the theory of economic
growth". Quarterly Journal of
Economics (Oxford Journals) 70 (1):
65–94.

Nobel Prize, 1987

Dipartimento di Ingegneria Gestionale

The Solow Neoclassical Growth Model

- The best known model of economic growth

- In the long run: economies will conditionally converge

to the same level of income if they have the same rates
of savings, depreciation, labor force growth, and
productivity growth

- Basic framework for the study of convergence across

countries
The
TheSolow
SolowNeoclassical Growth
Neoclassical Model
Growth Model

- Key modification to the Harrod Domar (AK) growth model is

substitutability between capital (K) and labor (L)
- While in HD: no substitutability (K is the only factor)

- Diminishing returns to the use of K and L

Y Y Y
In Harrod-
Domar model:
K/Y = c

K L K

- The aggregate production function Y = F(L, K) is assumed

characterized by constant returns to scale:

Y = AKαL1-α
The
TheSolow
SolowNeoclassical Growth
Neoclassical Model
Growth Model
- The aggregate production function of the economy can be
written as follows:

Y = Af(L, K)
Y is the national income
A is the TFP
K is capital
L is labor
F is the typical production function with decreasing return for each factor
separately, and constant return to scale [Y = Af(L, K) Af(γL, γK) = γAf(L, K) = γY]

Thus:
Y/L = A f(K/L) = Af(k)
y = Akα

Y/L is the productivity of labor (or income per capita)

k = K/L is capital intensity (or the stock of capital per worker)
 Factor deepening Productivity
growth due to
Technological
change

Young Alwyn. 1995. The tyranny of numbers: Confronting the statistical realities of the
East Asian Growth Experience. Quarterly Journal of Economics 110(3): 641-80.
Solution to the Solow growth model (and implications for
The Solow Neoclassical Growth Model
growth)

Cobb-Douglas function:

NB. Because of constant returns to scale, the

α 1-α
Y = AK L sum of the exponents is 1

α α
Y/L = A(K/L) or y = Ak
Using the savings-investment equation (sY = I = ∆K) from the generic growth model, we
find:
Kt+1 = (1-δ)Kt+sYt
Or, in terms of K intensity: Kt+1/Lt = (1-δ)(Kt/Lt)+s(Yt/Lt)
as Lt+1 = Lt (1+n):
(Kt+1/Lt+1)(1+n)= (1-δ)(Kt/Lt)+s(Yt/Lt) or: kt+1(1+n)= (1-δ)kt+syt
Hence:
∆k = kt+1-kt = syt-(n+δ)kt
The growth of k depends on:

• savings sf(k), after allowing for the amount of capital required to service
depreciation, δk, and after capital widening, i.e. providing the existing
amount of capital per worker to the new workers joining the labour force, nk.

Let’s assume for simplicity that A = constant (i.e. no technological change). There
will be a state in which
Δy= Δk=0

Steady state
Figure A3.2.1 Equilibrium in the Solow Growth Model

Δk<0
Δk>0 Δk=0
The
TheSolow
SolowNeoclassical Growth
Neoclassical Model
Growth Model

- The level at which the steady state is achieved:

• Increases with s
• Increases with A (that is the TFP)
• Increases with the share of K in national income
• Decreases with δ
• Decreases with n

- TFP (A) shifts the steady state upward

- The rate of savings increase per capita income at the steady state
y1/y2 = (s1/s2)(α/1- α)

The country with the higher s has the higher per capita income
The Solow
• The capital Neoclassical Growththe
per worker k* represents Model
steady state. If k is higher or
lower than k*, the economy will return to it, i.e. k* is a stable equilibrium.

• NB. In Harrod-Domar, there is no equilibrium income per worker. This is

because f(k), and therefore sf(k), does not exhibit diminishing returns (it is
a straight line). Thus, growth continues as long as line sf(k) stays above the
line (δ+n)k.

y
f(k)

sf(k)

(δ+n)k

k
Figure A3.2.2 The Long-Run Effect of Changing the Saving
Rate in the Solow Model
f(k) ↑ when there
is a jump in A

s’>s
 Two crucial drawbacks:
Endogenous Growth Theory

1. Using the neoclassical framework, it is impossible to analyze the determinants of

technological change, because it is completely independent of the decisions of
economic agents

2. The theory fails to explain large differences in residuals across countries with similar
technologies

How to stimulate growth through technological change is we cannot influence technological

change?

Endogenous Growth Theory

 Paul Michael Romer, Endogenous Growth Theory

Romer P.M. 1990. "Endogenous

Technological Change“, Journal of
Political Economy

Nobel Prize in 2018!

Dipartimento di Ingegneria Gestionale

 Endogenous Growth Theory (EGT)
• Economic growth is generated by factors within the production process (e.g.
increasing returns or induced technological change)

• EGT seeks to explain the factors that determine the size of λ (i.e. ΔA/A), the rate
of growth of GDP (Y) that is left unexplained and exogenously determined in the
Solow neoclassical model (i.e. the Solow residual)

• The most significant theoretical difference stems from discarding the

neoclassical assumption of diminishing marginal returns to capital
investments, permitting increasing returns to scale in aggregate production.
• Knowledge can function as a driver of long term growth
• The result is indeed sustained long term growth (and divergence across
countries)
 Endogenous Growth Theory: The Romer model

• Romer’s model endogenizes the reason why growth may depend on the
rate of investment (as in the Harrod – Domar model)

• Firm’s capital stock includes its knowledge

• Knowledge is a public good, like A in the Solow model, that is spilling

over instantly to the other firms in the economy.

Definition of public good:

- Non rivalrious, i.e. the goods do not decline in supply as

more people consume them
- Non excludable, i.e. the good is available to all citizens

- ex. Air, lighthouses, public defense, …

 Endogenous Growth Theory: The Romer model

• Role of technological spillovers (one firm or industry’s productivity gains

lead to productivity gains in other firms or industries)

• The model begins assuming that growth processes derive from the firm
or industry level

• Each industry individually produces with constant returns to scale (as in

Solow); however, the economy-wide capital stock positively affects output
at the industry level, so that there may be increasing returns at the
economy wide level
The Romer model

Assume symmetry across industries, so that each industry i will use the
same level of capital and labor.

The aggregate production function ∑iYi would be the following:

y
If β>0, there are spillovers and increasing returns
The Romer model

Assume that A is constant rather than rising over time; that is,
assume that there is no technological progress.

With a little calculus, it can be shown that the resulting growth rate
in the economy would be:

Where g is the ouput growth rate, and n is the population growth

rate. Without spillovers, as in the Solow model with constant returns
to scale, β=0, and therefore per capita growth would be zero
(without technological progress).
The Romer model

• However, with Romer’s assumption of a positive capital externality (β>0), we

have that g-n>0, and Y/L is growing.

• Endogenous growth here is not driven exogenously by increases in

productivity.

• Firms’/industries’ benefit from others’ research discoveries (this calls for

government intervention, as individual firms will tend to underinvest in
research)
 Structural determinants of growth

• Endogeneous growth theory (Romer, 1990): Investing in knowledge and increasing returns

It extends the Solow model by introducing a mechanism that explains the origin of
technological change (i.e. A becomes endogenous)

• Increasing returns to scale rule out convergence.

As knowledge generates more knowledge, advanced countries that already have a higher
stock of knowledge growth faster than developing economies

• Because knowledge is protected by IPR (necessary to create incentives to innovate),

technological advances are trapped in developed countries where they originate
 Structural determinants of growth (ctd)

• The Endogeneous growth model may, however, be too pessimistic about reality
– While frontier technology is privately costly, accessing technology that has become
generic after 20 years is free for latecomers, allowing technological leapfrogging

• The model shows that productivity gains accelerate as knowledge begets more knowledge.
What the model does not tell us, however, is how rich countries became rich, i.e. how did
they start acquiring the knowledge advantage?

• Productivity growth has large spillover effects across firms, as firms benefit from others’
research discoveries → role for the state to induce higher level of investment in research
 Structural determinants of growth (ctd)

• The policy implications of this growth model are important, but quite different for developed
and less developed countries

• The role of geographical clusters of firms: given spillover effects across firms, location can be an
important determinant of productivity growth.

– Agglomeration economies: Spillovers and network externalities (local labor pools with
specialised skills, joint-use infrastructures, shared community resources)

• Medieval cities
• Industrial districts in Italy; Silicon Valley
• Clusters in China (success due to proximity of hundreds of firms in the same industry)
Criticisms of Endogeneous Growth Theory

• It remains dependent on a number of traditional neoclassical assumptions

that are often inappropriate for developing countries, i.e. there is only one
sector, or all the sectors are symmetrical; this does not permit the crucial
growth-generating reallocation of labor and capital among the sectors
(structural change)
• It overlooks crucial factors like (poor) infrastructures, (inadequate)
institutional structures, (imperfect) capital and goods markets
• Too pessimistic model… as knowledge generates more knowledge, advanced
countries will always be advantaged.. However… evidence of leapfrogging
• It does not provide a theory of development for initiating growth
• It is a model for developed countries Leapfrogging may accelerate development by
skipping inferior, less efficient, more expensive or
more polluting technologies and industries and
move directly to more advanced ones.
 TAKE HOME MESSAGES

Development economics has no universally accepted paradigm

• Each theory has some strengths and some weaknesses

• Linear stage model: crucial role that saving and investment play
in promoting sustainable long-run growth
• The Lewis two sector model of structural change: importance
of transfers of resources from low-productivity to high productivity
activities
• Agreement about the role of technological change

Insights and understandings are continually evolving

 Lessons from exogenous growth
• The classical and neoclassical growth model proposed by Harrod-Domar and Solow
identify technology as a major determinant of aggregate economic growth: a low ICOR in
Harrod Domar and a high rate of TFP in Solow.

• Yet, they do not explain where technology comes from nor how the flow of new
technology come from.

• In recent years new thinking has emerged to address the issue of the role of technology in
growth.

– Endogenize technological change

– Explain conditional economic convergence
– Explain why the availability and use of technology differ across countries
– Explain the role of the state in promoting technological innovation
 Conditional economic convergence
Multiple attempts at testing empirically the predictions of the Solow model on convergence in
per capita income levels.

git, t+T = α +βyit + εit If there is convergence: β is negative

If convergence is conditional on a set of initial (presumed exogenous) country characteristics

Xit, then the equation to be estimated is:
git, t+T = α +βyit + γ Xit + εit

Finding out the X variables that allow you to reveal convergence is a policy question that has
motivated a large number of studies
→ Extended Solow model
(Mankiw, Romer & Weil, 1992, A contribution to the Empirics of Economic Growth, Quarterly
Journal of Economics, 107(2): 407-437)
 Convergence vs. divergence

Dipartimento di Ingegneria Gestionale

 Convergence vs. divergence

Dipartimento di Ingegneria Gestionale