Transcript INF

Course title
Economics of Innovation
Prof. Davide Infante
Department of Economics and Statistics
University of Calabria - Italy
[email protected]
Notes on lecture 3:
Technical progress and economic growth
INF
Introduction_1
•
Why do economists study technical change and innovation?
•
1. 1. Macroeconomics - to understand economic growth
•
2. 2. Microeconomics - to learn how the process works and
what motivates the actors
•
3. 3. Strategy (business economics) - to help choose strategies
•
4. 4. Policy for choosing and designing government policy
INF
Introduction_2
In this lecture we focus on economic growth
Improvements in our standard of living come from productivity
growth, getting more output per unit of input or more output per
unit of input
How?
- Investment leading to more capital per worker
- Scale or size effects (e.g., from commercial expansion), due to
fixed costs, specialization expansion)
- Increases in the stock of knowledge via technical and
institutional change
How do we model and measure the magnitude of these of these
effects?
We start from some growth indicators
INF
On convergence 1
•
•
•
We start from an analysis of cross-country data from:
- table 1 - Size of the economy in World Development Report
(2005) of World Bank
- Annex table 1 Economic Outlook (December 2005), OECD.
•
On the base of this analysis we deduct the following facts:
INF
On convergence 2
•
•
•
•
•
•
•
•
Fact # 1
There is a strong variation in the income per capita between
economies. The poorer countries have income per capita that
are less then 5% of income per capita in the richest countries.
Fact # 2
The rates of growth change substantially between countries.
Fact # 3
The rates of growth are not necessarily constant over time.
Fact # 4
The relative position of a country in the world distribution of
income per capita is non immutable. Countries can change
their position from poor to rich and viceversa.
INF
On convergence 3
•
•
•
Lucas practical rule (1988): a country that grows at a g rate per
year will double it income per capita every 0.7/g years.
We can explain this rule as the following:
Let’s assume y(t) as the income per capita at time t and let’s
take y0 as some initial value of the per-capita income. Hence:
y(t )  y0e gt
•
The time span to double a per-capita income is given by
y(t)=2y0. Henceforth:
2 y0  y0 e gt
 t* 
*
log2
g
INF
On Convergence 4
•
The rule derives from the fact that the log of 2 is =.7
•
Hence if a country grows at a rate of 1.4% per year than it will
double its pro-capita income in about 50 years, while a country
that grows at a rate of 2.9% will double its pro-capita income in
about 24 years.
INF
On convergence 5
In the current literature the term "catching up" has been often
identified with the “controversy on convergence"
The hypothesis convergence states the existence of a
convergence between countries income per capita.
This means that there is a negative relationship between
productivity and relative income per-capita.
Mechanism: being backward in the productivity level allows to
a strong potential for productivity.
INF
On convergence 6
The convergence hypothesis works between 24 OECD countries
(relation between rate of growth in income per-capita between 1965 and 1993 and starting
relative income per-capita (USA 1965=100)
GDPC6593
Scatterplot
4.5
4
3.5
3
2.5
2
1.5
1
.5
0
20
40
60
80
100
120
USA65100
INF
On convergence 6 bis
INF
On convergence 7
INF
On convergence 7 bis
The convergence hypothesis does not work at global level
(relation between rate of growth in income per-capita between 1965 and 1993 and starting
relative income per-capita (USA 1965=100)
GDPC6593
Scatterplot
10
8
6
4
2
0
-2
-4
-6
-10
10
30
50
70
90
110
USA65100
INF
On convergence 8
In this version the catch-up hypothesis seems a sort of long-run
growth insurance.
To avoid this criticism Abramovitz qualifies his definition
introducing from one side some conditions:
1) rapid productivity growth
2) dynamic economies of scale
3) modernization of technologies
4) structural change
On the other side he defines the necessary characteristics for
catching-up:
1) social capability (competences, education, institutions, firms)
2) adaptability (capacity of a country to exploit the power of
technologies)
INF
On convergence 9
It comes back the concept relative backwardness.
Abramovitz does not explain because and why new technologies
transfer from leader to follower countries.
Baumol explains it with the nature of public good of technical
progress and investment.
If a leader countries introduces innovations this, in the long run,
will turn out positively on lagging countries.
Through the following mechanisms:
1) The Schumpeterian competition
2) The theorem of factor prices equalisation
INF
On convergence 10
The catching-up idea of Abramovitz and Baumol can be
synthesised as the following:
1) the technological progress is exogenous
2) the technological progress is a non-excludible good
3) the technological progress depends on the learning capacity of
the catching-up country
4) existence of decreasing returns of capital.
INF
On convergence 11
Limits of convergence hypothesis
Three group of countries:
1)
OECD countries convergent club.
2)
Countries that present some similar characteristics to OECD
Countries and are able to converge to the club.
3) Countries that present very low rates of growth.
They do not have external inflow of technologies and their
learning capacity is so low that it is not so profitable to
introduce technical progress incorporated in machineries.
INF
On convergence 12
The growth model of convergence hypothesis is based on the
following production function equation:
Y i  A (t )K i L1i 
1
y i  Ak

i
 y i 
k i  

 A 
A = a catch-all parameter excluded by the assumption of
constant returns to scale (technical progress, returns to
scale, learning by doing, managerial efficiency).
INF
On convergence 13
The hypothesis underlining this growth model are:
1) technology is exogenous (is independent from capital and
labour)
2) constant returns to scale
[hence factors decreasing return (<1)]
The existence of factors decreasing return, full information and
factor mobility are the basis for catching up-convergence
theory.
The investment in capital is much more productive where is
scarce.
This makes possible the convergence hypothesis.
INF
On convergence and NGT 14
The new growth theory: the persistent divergence between
advanced and backward countries leaves the neoclassical
model of growth without credibility.
Romer (1990) states the existence of five principal factors for the
growth of a economic system:
1.
existence of many firms on the market
2.
absence of rivalry in information
3.
possibility to replicate activities (firms operate at the Minimum
Efficient Scale)
4.
the rate of new discoveries is linked to what firms do
5.
firms can have market power and gain monopoly profits.
INF
On convergence and NGT 15
The neoclassical growth model can satisfy conditions 1-3 but
not 4 and 5 that presume the endogeneity of technical
progress and the appropriability of innovation benefits
(innovation is an excludable good)
The facts 4 e 5 are those that permit either the catching-up
(convergence) or the lagging-behind (divergence).
For Romer the divergence between neoclassical growth theory
and the empirical evidence of low rates of growth of less
developed countries is due to:
a) existence of endogenous technical progress, that derives
from knowledge spillovers that add on normal returns of
capital and labour.
b) to the uprising of a monopoly power (patents, product
differentiation).
INF
On convergence and NGT 16
It is just the assumption of increasing returns to scale that makes
possible the continuous divergence in per capita GDP level
between countries.
To construct its first growth model Romer (1986) recurs to the
idea of Arrow´s learning by doing:
the growth of productivity depends on the accumulation of capital
in the industry (economy).
This is because new knowledge is discovered as investment and
output increase.
INF
On convergence and NGT 17
Romer (1986):

yi  k i A

where
y = output per worker
k = capital per worker
n
A  ki
i 1
Romer demonstrates that the decreasing return of capital of a
single firm 0<a<1 is counterbalanced by external spillovers.
(    1)
(   1)
(  
1)
INF
On convergence and NGT 18
For Lucas (1988) the neo-classical model is not able to
explain, during time, the differential of growth between
countries.
The neoclassical model is able to explain only the
convergence between some rich countries.
For Lucas the growth of income per capita can be explained
through the accumulation of human capital.
INF
On convergence and NGT 19
The Lucas (1988) model has the same cultural background of
Romer´s model. Both consider the idea of spillovers
through learning by doing.
Lucas assumes that the growth of income per capita is
determined only by the accumulation of human capital.
 con
y i  h
A
i
dove
y = outpu per worker
h = human capital
n
A   hi
i 1
In (1993) Lucas introduces the stock of human capital at world
level in j country
n
A   hj
i j
INF
On convergence and NGT 20
The Romer and Lucas´s models solve point 4 but not point 5,
because technology is treated as public good.
On the contrary one of the primary sources of the productivity
differentials is exactly the property of excludability that very
often is presented by knowledge.
Appropriability
Monopoly power
Innovation
The neo-schumpeterian approach to the growth theory: growth is
determined by endogenous innovations.
The development of new intermediate goods leads to a higher
specialisation in the use of resources and to a higher
productivity.
INF
On convergence and NGT 21
Romer (1990):
Three economic sectors: final goods, intermediate goods,
research


Y  Hy L Z
1   
Y = final output
Hy = human capital
L = labour
Z = aggregated measure of intermediate goods
INF
On convergence and NGT 22
Romer uses a model of product differentiation (Dixon e Stiglitz),
assuming a continuum of intermediate goods in the interval
(0, A), hence equation (1) becomes:
A
Y(Hy ,L, x)  H y L x(i)1    di
0

The sector presents increasing returns because of incomplete
property rights (part of the generated knowledge can be
utilised by other firms).
Hence an increase in the stock of human capital HA increases
the rate of growth.
INF
On convergence and NGT 23
HA
Ho
H
INF
On convergence and NGT 24
The specification of the technology that holds in the R&D sector
is of the following form:
YA  H A A
where H denotes the stock of human capital used in reaserch
and A a measure of the general accumulated knowledge,
freely available for all firm in the research sector. Romer
assumes that if the research firm j invest an amount of
human capital Hj and has access to a portion Aj of the total
amount of knowledge already existing in the research sector,
the rate of production of new designs by the research firm j
will be H j A j

where delta stands for the research productivity parameter.

INF
On convergence and NGT 25
In the Romer model each intermediate good is produced by a
“local monopolist”.
This hypothesis is coherent with the existence of growing or
complement markets, but it is not with the Schumpeterian
creative destruction.
This aspect is faced in 1992 Aghion and Howitt´s model (see
Infante 1995).
INF
On convergence and NGT 26
In Aghion and Howitt (1992) model the final output depends on
the input of an intermediate good, x. It does not take into
account capital accumulation.
In their interpretation, the model is changed according to:
Y = AF(x)
where Y is the final output, and x is the intermediate good.
The labour stock is used in two activities:
a) The production of intermediate goods (i.e., one unit of labour
produces a unit of intermediate good x);
b) The creation of innovations in the research sector.
Innovation is the invention of a new intermediate
product x, that replaces the existing ones and raises the
technology parameter, A.
INF
On convergence and NGT 27
When n units of labour are employed in the research sector,
innovations occur randomly according to the Poisson arrival
date, , where  > 0 is given and represents the productivity
research sector parameter.
When the innovation is introduced, it takes over the entire market
for the production of the intermediate good.
Each successive innovation raises productivity forever, because it
allows the productivity parameter A to increase by the same
factor  > 1 and there is no lag in the introduction of
technology, so that:
At 1  At   (At 0 )
where t (=0, 1,....) represents subsequent innovation.

INF
On convergence and NGT 28
Aghion and Howitt model differs from Romer's model because it
embodies the Schumpeterian idea of "creative destruction".
In their model, the discounted expected value for the
innovative firm is:
 t 1
Vt 1 
r   (n t 1 )
where t+1 is the flux of monopoly profits generated by the t+1st
innovation, r the interest rate, n the amount of labour applied
to R&D, and
 (n )

t 1
the Poisson arrival date of innovation after the t+1th innovation.

INF
On convergence and NGT 29
The firm is induced to innovate because it can capture the entire
market, so gaining monopoly profits.
Some negative effects:
a) a firm that introduces innovation has no incentive for further
research.
The "business-stealing effect" works either
against rivals or the innovator's expected profits that would
be
t 1
t
b) assuming free entry in the research sector, only outside firms
will do new research.
In order to innovate, they need to increase the number of
research people, thus:
(i) increasing the wage rate
(ii) raising the rate of creative destruction, hence, decreasing the
expected profits of innovation,
(iii), increasing n will accelerate the expected date of discovery
thus decreasing the time interval for monopoly rents.
INF
V

V
On convergence and NGT 30
The Aghion and Howitt's model is a strong advance on the socalled neo-Schumpeterian approach of the New Growth
Theory.
Some of the essential features ("creative destruction" and
persistence of monopoly power in the innovative sectors) of
Schumpeter's approach in this model are endogenised.
Some reservations:
- “memoryless” mechanism: the assumption that innovation
occurs according to a Poisson probability distribution is
fairly strong.
INF
On convergence and NGT 31
- In addition, the memoryless mechanism assumption leads to a
situation in which the monopolist has no incentive to further
research
whilst outside firms have because they have the same probability
of discovering a new intermediate good and becoming a monopolist.
- The absence of experience (Poisson assumption of independence),
weaken the Schumpeterian idea of innovation as a source of
monopoly rents.
- Another problem is that regarding , the factor affecting the
productivity parameter A, which is assumed to be constant; i.e. the
rate of increase in the parameter A is proportional to the amount
reached in the current time (technology), this means that the
following relationship exists between
dA A and x
dx
 A
INF
On convergence and NGT 32
, therefore, is a proportional constant that is completely independent
of the variable x.
It does not depend on the type (and numbers) of innovation (radical,
incremental, cross-sector) discovered.
- If  varies its impact on A, it would increase (decrease) the flow of
monopoly profits generated by the t+1st innovation, matching the
decrease in its expected value, generated by an increase of n (the
labour stock dedicated to research activities).
INF
References
Abramovitz, M. [1986], Catching Up, Forging Ahead, and Falling Behind, in «Journal of Economic
History», June 1986, 46, pp. 383-406.
Aghion, P., and Howitt, P. [1992], "A Model of Growth Through Creative Destruction", Econometrica,
March 1992, 60:2, 323-51
Baumol, W. J. [1986], Productivity Growth, Convergence and Welfare: What Long-Run Data Show, The
American Economic Review, 76 (5), pp. 1072-1085.Dixit, A. K e Stiglitz, J. E. [1977], Monopolistic
Competition and Optimum Product Diversity, «American Economic Review», 67, 3, pp. 297-43.
Gerschenkron, A. [1962], Economic Backwardness in Historical Perspective, Belknap Press of Harvard
University Press, Cambridge.
Infante, D. [1995], Catching-up, Growth and Innovation Processes, Research Report No. 2/1995, Centre
for Southern European and Mediterranean Studies, Roskilde University, Roskilde.
Infante, D. [2005], Crescita e catching up nell’Unione Europea allargata, in Crescita e prospettive
nell’Unione Europea allargata (a cura di D. Infante), Centro di Documentazione Europea,
Università della Calabria, pp. 29-32
Jones, C.I. [1998], Introduction to Economic Growth, W.W. Norton and Company, New York.
Lucas, E.R. [1988], On the Mechanics of Economic Development, in «Journal of Monetary Economics»,
21, pp. 3-42.
Lucas, E.R. [1993], Making a Miracle, in «Econometrica», 61, 2, pp. 251-72
Romer, P.M. [1986], Increasing Returns and Long-Run Growth, in «Journal of Political Economy», 94,
pp. 1002-37.
Romer, P.M. [1990], Endogenous Technological Change, in «Journal of Political Economy», 98, pp. S71S102.
INF