Transcript Document

Chapter 17
Macroeconomic Policies
and Long-Term Growth
© Pierre-Richard Agénor and Peter J. Montiel
1




Wide dispersion of output growth rates across countries.
Table 17.1: growth performance of developing
countries.
Traditional neoclassical approaches: incapable of
explaining the wide disparities in the pace of economic
growth across countries.
“New growth” literature: existence of “endogenous”
mechanisms that foster economic growth, and new
roles for public policy.
2








The Neoclassical Growth Model.
Externalities and Increasing Returns.
Human Capital, Knowledge, and Growth.
Effects on Financial Intermediation.
Inflation Stabilization and Growth.
Government size and Growth.
Commercial Openness and Growth.
Exchange-Rate Unification and Growth.
3
The Neoclassical Growth
Model



Solow (1956) and Swan (1956): neoclassical growth
model.
Assumptions:
 Production function: aggregate, constant-returns-toscale, and combines labor and capital in the
production of composite good.
 Savings: fixed fraction of output.
 Technology improves at exogenous rate.
Cobb-Douglas production function:
Y = AKL1-,
0 <  < 1,
(1)
Y: total output; K: capital stock; A: level of technology;
L: workers employed in the production process.
5

Output per worker, y = Y/L, is given by
y = Ak,


k: capital-labor ratio.
Capital accumulation is given by
.
k = sy - (n + )k,

0 < s,  < 1,
(2)
s: propensity to save;
n > 0: exogenous rate of population growth;
: rate of depreciation of physical capital.
(2) incorporates equilibrium condition of goods market
or, equivalently, equality between investment I and
6
saving, I = sy.


Suppose that A is constant over time.
Substituting (1) in (2) and dividing both sides of the
resulting expression by k yields growth rate of capitallabor stock:
.
gk  k/k = sAk-1 - (n + ),
(3)
from which the rate of growth of output per worker can
be derived as
.
.
gy  y/y = kAk-1/Ak = gk.


Figure 17.1: behavior of capital stock per worker.
Horizontal line at n + : depreciation line.
7
F
i
g
u
r
e
1
7
.
1
C
a
p
i
t
a
l
A
c
c
u
m
u
l
a
t
i
o
n
i
n
T
h
e
S
o
l
o
w
S
w
a
n
G
r
o
w
t
h
M
o
d
e
l



s
A
k
A
p
g
k
B
r
g
k
E
n
+

~
1
/
(
1
r


p
k
k
k
=
(
s
A
/
n
+


0
0
k
8






Curve sAk-1: savings curve.
Savings curve is downward-sloping due to assumption
of decreasing marginal returns to capital.
As implied by (3), gk is the difference between the two
curves.
Point of intersection of the two curves: steady-state
value of k.
If technology grows at a constant rate, steady-state
values of output per effective worker and
capital/effective labor ratio are
 constant;
 proportional to the rate of technological change.
Although s has no effect on growth rate per capita in the
long run, it affects level of per capita income in steady
9
state.





Model implies that countries with similar production
technologies, and comparable saving and population
growth rates should converge to similar steady-state
levels of per capita income.
Figure 17.1: “poor” country starts with capital stock of k0p
has higher initial growth rate than “rich” country starting
with k0r .
Poor country grows faster during the transition.
But, if both countries possess the same level of A, s, ,
n, they will both converge to the same steady-state level
~
of the capital stock, k.
Convergence occurs because each increment to capital
stock generates large additions to output when capital
stock is initially small with diminishing marginal returns
to capital.
10




“Sources-of-growth” approach: empirical methodology
to analyze determinants of changes in output.
It uses aggregate production function to decompose
growth into “contributions” from different sources plus
residual.
Residual: “technical progress,” or more adequately
growth in total factor productivity.
Assume: production function is y = Af(k, n).
11

In terms of percentage changes:
.
.
.
.
g  y/y = (A/A) + Afk(k/y) + Afn(n/y)
. A + kgk + ngn,
=g


h = fhh/y (for h = k, n): elasticity of output with respect
to input h;
gA: rate of growth of total factor productivity and is
derived as a residual.
Under conditions of competitive equilibrium, factors are
paid their marginal products: k (n) is equal to share of
capital (labor) income in total output.
In the presence of constant returns to scale, sum of all
12
share coefficients must be equal to unity.
With Cobb-Douglas production technology as in (1),
assuming that factors of production are paid their
marginal products implies that
 k = 1-n, and
 labor's share corresponds to the parameter .
 Even though hypotheses of constant-returns-to-scale
production function and competitive factor markets are
restrictive, there are studies based on the model.
Chenery (1986):
 Studies based on sources-of-growth methodology in
the 1960s and 1970s.
 Average capital share is about 40%, which indicates
that production function exhibits diminishing marginal
returns to capital.

13
Growth in capital stock had limited effect on output
growth.
 Average contribution of the residual was less than in
developed countries.
 Most countries had high growth rate of labor input.
 Estimates of capital share vary across countries,
ranging from 26% for Honduras to more than 60% for
Singapore.
 Effect of capital accumulation on growth varies across
countries.
 Contribution of total factor productivity to growth also
varied across countries.
Elías (1992):
 Growth process of Argentina, Brazil, Chile, Colombia,
14
Mexico, Peru, and Venezuela during 1940-85.









He considers different kinds of labor and capital inputs,
and defines gross and quality component for each of
them.
For labor: gross component is arithmetic sum of
employment across characteristics.
For capital: it is arithmetic sum of different categories of
capital.
Quality component captures changes in composition of
factors of production.
Output growth averaged 5.3% for the group as a whole.
Quality of labor rose on average by 1.4%, and quantity
of labor by 2%.
Quality of capital fell by 0.4%, its quantity grew at 4%.
Given average labor share of 40%, labor contributed
15
1.3% to average growth rate.
Capital's contribution was 2.5%.
 Technological progress was therefore 1.5% of rate of
growth.
 Thus, capital made the highest contribution to output
growth (47%) because of its quantity and its share.
 Quality of labor played more important role in growth of
labor input.
Table 17.2:
 Decomposition of trend or potential output growth for
developing countries during the 1970s and 1980s.
 Contribution of capital to potential output growth was the
most important.
 Total factor productivity accounted for about the same
share as labor in its contribution to growth.

16
Differences across regions: total factor productivity
 accounts for negligible share of growth in Africa and
the Middle East,
 provides substantial contribution to growth in Asia.
Limitations of neoclassical growth model:
 Capital assumed to exhibit diminishing marginal returns.
 This prevents it from providing an explanation
 for the wide variations across countries in either per
capita income or growth rates, and
 for the fact that poor countries do not grow faster
than rich ones (Figure 17.2).
 It is assumed that output growth is independent of
saving rate and is determined only by demographic
factors and technological progress rate.
17

F
i
g
u
r
e
1
7
.
2
A
v
e
r
a
g
e
R
a
t
e
o
f
G
r
o
w
t
h
P
e
r
C
a
p
i
t
a
a
n
d
P
r
o
p
o
r
t
i
o
n
o
f
G
N
P
P
e
r
C
a
p
i
t
a
t
o
U
S
G
N
P
P
e
r
C
a
p
i
t
a
0
.
1
0
.
0
5
AveragtofGNPgrwthpecait(197-5)
0
0
.
0
5
0
.
1
0
0
.
0
5
0
.
1
0
.
1
5
0
.
2
0
.
2
5
0
.
3
G
N
P
p
e
r
c
a
p
i
t
a
r
e
l
a
t
i
v
e
t
o
U
S
G
N
P
p
e
r
c
a
p
i
t
a
I
n
1
9
7
0
(
i
n
U
.
S
.
d
o
l
l
a
r
s
)
S
o
u
r
c
e
:
W
o
r
l
d
B
a
n
k
.
18



Since population growth and technological change are
assumed exogenous, the model does
 not explain the mechanisms that generate steadystate growth,
 not allow evaluation of mechanisms through which
government policies can influence growth process.
Assumption that rate of growth of output is independent
of saving rate is at variance with the evidence; highgrowth developing countries have higher saving rates.
New growth literature addresses these limitations by
proposing variety of channels through which steadystate growth arises endogenously.
19
Externalities and Increasing
Returns




Two approaches were followed to relax assumption of
diminishing returns to capital:
First approach views all production inputs as some
form of reproducible capital, including
 physical capital,
 human capital (Lucas, 1988) or
 “state of knowledge” (Romer, 1986).
Simple growth model along these lines: AK model
proposed by Rebelo (1991).
It results from setting  = 0 in (1):
y = Ak,
where k = K/L as before, but K includes both physical
21
and human capital.


Thus, production function is linear and exhibits constant
returns to scale, but does not yield diminishing returns
to capital.
Using the capital accumulation (2), steady-state growth
rate of capital stock per worker:
gk = sA - (n + ).

Steady-state growth rate per capita:
gy = sA - (n + ).

Growth rate is, for sA > n+, positive (and constant over
time) and level of income per capita rises without
bound.
22


Implication of AK model:
 Increase in saving rate raises growth rate per capita.
 Poor nations whose production process has the
same technological sophistication as other nations
grow at the same rate as rich countries, regardless of
initial level of income.
 Thus it does not predict convergence even if
countries
 share the same technology;
 are characterized by the same pattern of saving.
Rebelo (1991):
 Implications of considering separately the production
of consumption goods, physical capital, and human
capital goods.
23
Endogenous steady-state growth obtains if “core” of
capital goods is produced
 according to a constant-returns-to-scale
technology;
 without nonreproducible factors.
Second approach: introducing spillover effects or
externalities in growth process.
Externalities: if one firm doubles its inputs, productivity
of inputs of other firms will also increase.
Introducing spillover effects relaxes assumption of
diminishing returns to capital.
Mostly externalities take the form of general
technological knowledge that is available to all firms,
which use it to develop new methods of production.





24




Exceptions.
 Lucas (1988): externalities take the form of public
learning, which increases the stock of human capital
and affects productivity of all factors.
 Barro (1990): externalities associated with public
investment.
Externalities is associated with increasing returns to
scale in production function.
But, important implication of models exhibiting spillover
effects and externalities is that sustained growth
 does not result from the existence of external effects,
 rather result from assumption of constant returns to
scale in all production inputs.
Rebelo (1991): increasing returns are neither necessary
25
nor sufficient to generate endogenous growth.
Human Capital, Knowledge,
and Growth


The Production of Human Capital.
The Production of Knowledge.
27
The Production of Human Capital





One of the sources of externalities: accumulation of
human capital and its effect on productivity of the
economy.
Lucas (1988):
 Spillover effects of human capital accumulation.
 Individual workers are more productive, if other
workers have more human capital.
Simplified version of Lucas' model is examined here.
Human capital is accumulated through explicit
“production”: part of individuals' working time devoted to
accumulation of skills.
Let k denote physical capital per worker and h human
28
capital per worker (“knowledge” capital).

Production process:
y = Ak[uh]1-, 0 < u < 1,


u: fraction of time that individuals devote to producing
goods.
Growth of physical capital depends on saving rate.
Growth rate of human capital depends on time devoted
to its production:
.
h/h = (1-u),

 > 0.
Long-run growth rate of both capital and output per
worker is  (1-u).
29
Rate of human capital growth, and ratio of physical to
human capital converges to a constant.
 In the long run, income is proportional to the economy's
initial stock of human capital.
 Saving rate has no effect on growth rate.
Implication of the model:
 Under purely competitive equilibrium there will be
underinvestment in human capital since private agents
do not take into account external benefits of human
capital accumulation.
 Equilibrium growth rate is thus smaller than optimal
growth rate.
 Growth would be higher with more investment in human
capital.

30

Thus government policies are necessary to increase the
equilibrium growth rate.
31
The Production of Knowledge




Romer (1986): source of externality is stock of
knowledge.
Knowledge is produced by individuals.
But since newly produced knowledge can be partially
and temporarily kept secret, production of goods and
services depends on both
 private knowledge, and
 aggregate stock of knowledge.
Since individuals only partially reap rewards to
production of knowledge, market equilibrium results in
underinvestment in knowledge accumulation.
32



Romer (1990):
 Explains endogenously decision to invest in
technological change;
 uses a model based on distinction between research
sector and rest of the economy;
 firms cannot appropriate all the benefits of knowledge
production;
 tax and subsidy can be used to raise rate of growth.
Simplified version of Romer's (1990) model is presented
here.
Two production sectors:
 goods-producing sector uses physical capital,
knowledge and labor in the production process;
 knowledge-producing sector (same inputs are used).
33




cL (cK): fraction of labor force (capital) is used in the
knowledge-producing sector.
1–cL (1-cK): fraction of labor (capital) in the goodsproducing sector.
A: total stock of knowledge that can be used in both
production activities.
Assuming Cobb-Douglas technology, output in goodsproducing sector:
Y = [(1-cK)K][A(1-cL)L]1-

0 <  <1. (9)
Constant returns to both capital and labor.
34

Production of new knowledge (changes in A) is
determined by generalized Cobb-Douglas form:
.
A = B(cKK)(cLL)A,



B > 0,   0, ,  0,
(10)
B: shift parameter.
There is either diminishing returns in production of new
ideas or increasing returns, depending on , , and .
 can be equal to unity, or strictly greater or smaller than
unity.
Assuming s is constant and there is no depreciation of
capital stock, then
.
K = sY,
0 < s < 1.
(11)
35

Population growth is exogenous:
.
L = nL,

Begin analyzing te model by substituting (9) in (11):
.
K = KKA1-L1-,

n  0.
K  s(1-cK)(1-cL)1-.
Dividing both sides of this expression by K:
.
gK  (K/K) = K{AL/K}1-.

Its rate of change:
.
gK = (1-)(gA + n - gK).
(14)
36




gK is rising (falling) if gA+n-gK is positive (negative), and
remains constant if gA+n = gK.
Curve KK in Figure 17.3: combinations of gA and gK for
which gK is constant over time.
Slope of KK is unity; above (below) KK, gK is falling
(rising).
Dividing both sides of (10) by A:
.
gA  A/A = AKLA-1,

A  BcKcL.
This implies that
.
gA = gK + n + (-1)gA.
(15)
37
F
i
g
u
r
e
1
7
.
3
D
y
n
a
m
i
c
s
o
f
C
a
p
i
t
a
l
a
n
d
K
n
o
w
l
e
d
g
e
:
C
a
s
e
I
+
<
1


g
K
A
K
~
E
g
K
K
n
0
A


~
g
A
g
A
n
/
S
o
u
r
c
e
:
R
o
m
e
r
(
1
9
9
4
,
p
.
1
0
7
)
.
38








gA is increasing (falling) if right-hand side of (15) is
positive (negative), and constant if it is zero.
Curve AA in Figure 17.3: combinations of gA and gK for
which gA is constant over time.
Slope of AA is (1-)/, which is ambiguous in sign.
Figure assumes that  < 1, so that the slope is positive.
Above (below) AA, gA is rising (falling).
(9) exhibits constant returns to scale in K and A.
Thus whether there are on net increasing, decreasing,
or constant returns to scale to A and K depends on
whether (10) exhibits constant returns to scale.
This equation can be rewritten as
.
A = KA(qL),
q  BcKcL.
39
Degree of returns to scale to A and K in production of
new knowledge is  + .
 Consider the three separate cases, depending on
whether  +  is less, equal, or greater than unity.
If  +  < 1:
 (1-)/ is greater than unity and AA is steeper than KK.
 This case is illustrated in Figure 17.3.
 Regardless of initial values of gA and gK, they converge
to equilibrium point E.
.
~
~
 Equilibrium values gA and gK are obtained by setting gA
.
= gK = 0 in (14) and (15), and are given by


+

gA =
n,
1 – (+)
~
~
~
gK = n + gA.
40

From (9), aggregate output and output per worker are
growing at rates given by
gY = gK + (1-)(n + gA) = gK,
~
~
~
~
~
~
gY/L = gK – n = g~ A.
Thus, economy's growth rate is endogenous: increasing
function of n and is zero if n is zero.
 cL,cK and s have no effect on growth rate.
If  +  > 1:
 AA and KK diverge (Figure 17.4).
 Regardless of economy's initial position, it enters the
region between two curves.

41
F
i
g
u
r
e
1
7
.
4
D
y
n
a
m
i
c
s
o
f
C
a
p
i
t
a
l
a
n
d
K
n
o
w
l
e
d
g
e
:
C
a
s
e
I
I
+
>
1


g
K
K
A
K
n
0
A


g
A
n
/
S
o
u
r
c
e
:
R
o
m
e
r
(
1
9
9
4
,
p
.
1
0
8
)
.
42
Once this occurs, growth rates of A and K increase
without bound.
 There cannot be steady-state growth.
If  +  = 1:
 (1-)/ is equal to unity and AA and KK have the same
slope.
 If n is positive, KK lies above AA:
 upper panel of Figure 17.5;
 there is no steady-state level of growth.
 If n = 0, AA and KK are identical:
 lower panel of Figure 17.5;
 regardless of initial position of the economy,
balanced growth path is reached;
 this path is unique;
43

F
i
g
u
r
e
1
7
.
5
a
D
y
n
a
m
i
c
s
o
f
C
a
p
i
t
a
l
a
n
d
K
n
o
w
l
e
d
g
e
:
C
a
s
e
I
I
+
=
1


g
K
K
A
K
n
0


A
g
A
n
/
S
o
u
r
c
e
:
R
o
m
e
r
(
1
9
9
4
,
p
.
1
0
9
)
.
44
F
i
g
u
r
e
1
7
.
5
b
D
y
n
a
m
i
c
s
o
f
C
a
p
i
t
a
l
a
n
d
K
n
o
w
l
e
d
g
e
:
C
a
s
e
I
I
+
=
1


g
K
A
A
a
n
d
K
K
g
A
S
o
u
r
c
e
:
R
o
m
e
r
(
1
9
9
4
,
p
.
1
0
9
)
.
45
economy's growth rate on that path depends on all
the parameters of the model including s.
Existence of knowledge-producing sector may explain
positive correlation between s and rate of economic
growth (Figure 17.6).


46
F
i
g
u
r
e
1
7
.
6
S
a
v
i
n
g
R
a
t
i
o
a
n
d
O
u
t
p
u
t
G
r
o
w
t
h
P
e
r
C
a
p
i
t
a
(
1
9
7
1
9
5
)
0
.
1
0
.
0
8
0
.
0
6
0
.
0
4
0
.
0
2
AveragtofGNPgrwthpecai
0
0
.
0
2
0
.
0
4
0
.
0
6
1
0
0
1
0
2
0
3
0
4
0
5
0
6
0
G
r
o
s
s
D
o
m
e
s
t
i
c
S
a
v
i
n
g
a
s
a
p
r
o
p
o
r
t
i
o
n
o
f
G
D
P
S
o
u
r
c
e
:
W
o
r
l
d
B
a
n
k
.
47
Effects of Financial
Intermediation

Introduce financial factor, following Pagano (1993), to
assume that 1- of saving is “lost” as a result of financial
disintermediation activities:

sy = I,

0 <  < 1.
Assuming that production technology is constant returns
to scale to capital, steady-state growth rate per capita:
g = sA - .

How financial development affects economic growth:
 raise s;
 raise A (marginal productivity of the capital stock);
49
 increase in  (“conduit” effect).




Effects on the Saving Rate.
Effects on the Accumulation of Capital.
The “Conduit” Effect, Financial Repression, and Growth.
Financial Development and Growth: Empirical
Evidence.
50
Effects on the Saving Rate




Early development literature: existence of positive effect
of financial development on s.
New growth literature: direction of this effect is not
consistent.
Jappelli and Pagano (1994):
 development of financial markets offers households
possibility of diversifying their portfolios and
increases their borrowing options;
 this affects proportion of agents subject to liquidity
constraints, which may affect s.
Financial development also
 reduces overall level of interest rates;
51
modifies structure of interest rates by reducing
spread between rate paid by borrowers and that paid
to lenders.
In each case effect on s is ambiguous.
Ambiguous effect of financial intermediation on s may
be compounded when all partial effects associated with
financial development are taken into account.
Bencivenga and Smith (1991): direct effect of banking
activities may be reduction in s.
But, if positive effect of financial development on
productivity of capital and efficiency of investment is
taken into account, net effect on growth may be
positive.





52
Effects on the Allocation of
Capital



Figure 17.7: investment and output growth are
positively correlated in developing countries.
Role of financial intermediaries: facilitate efficient
allocation of resources to investment projects that
provide the highest marginal return to capital.
Financial intermediation increases average productivity
of capital A in two ways:
 by collecting, processing, and evaluating relevant
information on alternative investment projects;
 by inducing entrepreneurs, through their risk-sharing
function, to invest in riskier but more productive
53
technologies.
F
i
g
u
r
e
1
7
.
7
I
n
v
e
s
t
m
e
n
t
a
n
d
O
u
t
p
u
t
G
r
o
w
t
h
P
e
r
C
a
p
i
t
a
(
1
9
7
1
9
5
)
0
.
1
0
.
0
8
0
.
0
6
0
.
0
4
0
.
0
2
AveragtofGNPgrwthpecai
0
0
0
.
0
2
0
.
0
4
0
.
0
6
5
1
0
1
5
2
0
2
5
3
0
3
5
4
0
G
r
o
s
s
D
o
m
e
s
t
i
c
I
n
v
e
s
t
m
e
n
t
a
s
a
p
r
o
p
o
r
t
i
o
n
o
f
G
D
P
S
o
u
r
c
e
:
W
o
r
l
d
B
a
n
k
.
54
Greenwood and Jovanovich (1990):
 Link between informational role of financial
intermediation and productivity growth.
 Capital may be invested in safe, low-yield technology or
risky, high-yield one.
 Return to risky technology is affected by
 aggregate shock;
 project-specific shock.
 Financial intermediaries with their large portfolios can
 identify the aggregate productivity shock; and
 induce their customers to select technology that is
most appropriate for current shock.
55
Efficient allocation of resources channeled through
financial intermediaries raises productivity of capital and
thus growth rate of the economy.
Pagano (1993):
 Another function of financial intermediation: it enables
entrepreneurs to pool risks.
 “Insurance” function: financial intermediaries allow
investors to share uninsurable and diversifiable risk
from rates of return differences on alternative assets.
 Risk sharing affects saving and investment decisions.
Liquidity:
 In the absence of banks, households can guard against
idiosyncratic liquidity shocks by investing in productive
assets that can be liquidated.

56

Bencivenga and Smith (1991): banks increase
productivity of investment by
 directing funds to illiquid, high-yield technology;
 reducing investment waste due to premature
liquidation.
57
The “Conduit” Effect, Financial
Repression, and Growth




Financial intermediation operates as a tax in
transformation of saving into investment.
Financial intermediation thus has growth-deterring effect
because intermediaries appropriate share of private
saving.
Costs associated with financial intermediation represent
payments that are received by intermediaries in return
for their services.
In developing countries:
 Such absorption of resources results from explicit
and implicit taxation and by excessive regulations.
58
This leads to higher costs and thus inefficient
intermediation activities.
If financial system reforms reduce cost and
inefficiencies associated with intermediation process,
growth rate will increase.
Role of financial repression in growth models:
In countries where collecting conventional taxes is
costly, governments choose to repress their financial
systems to increase revenue.
Roubini and Sala-i-Martin (1995): inflation is viewed as
proxy for financial repression.
Courakis (1984): constraints on bank portfolio choices
may reduce volume and productivity of investment by
 reducing funds channeled to deposit-taking financial
59
intermediaries;







causing less efficient distribution of any given volume
of such funds.
60
Financial Development
and Growth: Empirical Evidence.
Recent research has explored empirical relationship
between financial “deepening” and economic growth.
King and Levine (1993a, 1993b):
 Four alternative measures of financial depth:
 ratio of liquid liabilities of financial system to GDP;
 share of total credit allocated by banks;
 share of total domestic credit received by private
sector;
 ratio of credit to private enterprises to GDP.
 Contributions of such indicators in explaining
61
 long-term real GDP growth;

share of investment in GDP;
 rate of growth of total factor productivity.
All of financial depth indicators are statistically
significant with large positive effects on variable being
explained.
This association did not reflect reverse causation from
growth to financial indicators.
Evidence linking financial depth to long-term economic
growth, both through
 incrementation of resource accumulation, and
 enhancement of productivity growth,
is strong in cross-country record.




62
Inflation Stabilization and
Growth



High rates of inflation can be expected to reduce
economic growth through variety of mechanisms which
can influence both
 rate of capital accumulation;
 rate of growth of total factor productivity.
Fischer (1993): government which tolerates high
inflation is one which has lost macroeconomic control,
and this deters domestic investment in physical capital.
Other arguments: high inflation
 means unstable inflation and volatile relative prices;
 reduce information content of price signals;
 distort efficiency of resource allocation, affecting
growth of total factor productivity.
64


Simplified version of De Gregorio’s model (1993) is
presented here.
Assumptions:
 Closed economy consists of households, firms, and
government.
 Households hold no money but hold indexed bond
issued by government.
 Capital is only input in production process, which
takes place under constant returns to scale.
 Firms hold money because it reduces transactions
costs associated with purchases of new equipment.
 Capital mobility is precluded, so that domestic
investment must equal domestic saving.
 Inflation is exogenous.
65

Representative household maximizes present value of
utility stream


0
c1- -t
e dt,
1-
0 <  <1,
(18)
subject to flow budget constraint
.
b = (1 - )(y + rb) – c - ,
(19)
  1/: elasticity of intertemporal substitution;
b: real stock of government indexed bonds;
0 <  < 1: income tax rate;
r: real rate of return on bonds; y: total factor income;
66
: net lump-sum taxes paid by households.

Maximization of (18) subject to (19) yields
.
c/c = [(1-)r - ].

Production exhibits constant returns to scale:
y = Ak.



Firms require money to purchase new capital goods.
Cost of investing I units is thus equal to I[1+(m/I)],
where m is firms' real money holdings.
 < 0 and  > 0: holding money reduces transactions
costs but entails diminishing returns.
67



Representative firm maximizes present discounted
value of its cash flow, net of opportunity cost of its
holdings of money balances.
This opportunity cost is equal to (r + )m, where  is
inflation rate.
Thus firm maximizes:


0
m
.
Ak - 1 + ( ) I – (r+)m – m e-rtdt,
I
.
subject to k = I.
68

Solution yields
(23)
m
-(
) = r +   m = (r + )I,  = 1/ < 0,
I
.
(24)
q/q = r – (A/q),
m
m
m
q = 1 + (
)( ),
I
I
I
(25)
q: shadow price of capital.


(23): firm's demand for money.
Since cash flows are not subject to direct taxation,
opportunity cost of holding money is sum of before-tax
real interest rate plus inflation rate.
69



(24): shadow price of capital is equal to present
discounted value of marginal product of capital.
(25): q exceeds unity due to existence of transactions
costs incurred in buying new unit of capital.
Substituting (23) in (25) yields
q = 1 + [()] + (r + )() = q(r + ),


~ if  is constant.
(26): q is constant (at q)
From (24, real interest rate is:
~
q > 0.
(26)
~
r = A/q.
70

Government budget constraint:
.
.
m + b = g - y -  - m,
g: public expenditure, which is taken to be a constant
fraction of output.
.


Assume b = 0, and government adjusts lump-sum taxes
to maintain fiscal equilibrium.
Aggregate resource constraint of the economy:
y = c + 1 + ( m ) I + g.
I
71

Consumption, output, capital, and real money balances
grow at constant rate in the steady state:
g = [(1-)r - ].
~


(30)
Model has no transitional dynamics; that is, economy
grows continuously at rate given by (30).
This model generates an inverse relationship between
output growth and  due to negative effect of inflation on
profitability of investment.
 Higher  raises “effective” price of capital goods,
which incorporates opportunity cost of holding money
to facilitate purchases of capital goods.
72
Increase in transactions costs raises shadow value of
installed capital, dampens investment, and reduces
growth rate.
Barro (1997):
 Cross-country evidence on relationship between
inflation and growth.
 Data set consists of 100 countries, with annual
observations on macroeconomic data during 1960-90.
 Three periods: 1965-75, 1976-85, and 1986-90.
 Other things equal, 10% increase in  reduces long-run
growth by about 0.025% per year.
 It is level of , rather than its variability, that affects
growth adversely.
 Results are robust with respect to exclusion of few high73
inflation outliers.






Interesting aspect of Barro's work: introduction of some
novel instruments for inflation.
Barro uses prior colonial history: these are uncorrelated
with innovations in recent growth experience, but
correlated with long-term inflation performance.
Using these as instruments for  leaves previous results
in place.
Transition from high to low  may not be associated
with contemporaneous acceleration in economic
growth.
Favorable growth effects from disinflation materialize
with a lag, so that growth may slow during transition,
and perhaps for some time thereafter.
74
Bruno and Easterly (1998):
 Evidence about growth effects of transition from high to
low .
 Methodology:
 compiling a sample of countries that had experienced
successful stabilization over 1961-92; and
 comparing their growth rates relative to world
average before, during, and after, their inflationary
episodes.
 Growth fell by an average of 2.8% during high-inflation
episode, but rose by an average of 3.8% during
successful stabilization.
 This pattern was repeated for growth of total factor
productivity.
75
 But investment ratio did not rise above world average.


Conclusion: growth accelerates during and just after
stabilization when initial level of inflation is high.
Inflation stabilization component of market-oriented
reform policies should be growth-enhancing.
76
Government Size
and Growth



Inflation stabilization implies the need for reduction of
fiscal deficits that can be reduced by decreasing
expenditures or increasing revenues.
Difference between two approaches: resulting size of
government sector.
Both level and composition of government expenditures
may matter for long-term growth:
 Holding fiscal deficit constant, larger government
expenditures imply need for additional revenues.
 But such revenues would be raised through
distortionary taxation.
 This would reduce rate of growth through adverse
effects on efficiency of resource allocation.
 Some government expenditure may be productive.
78
Expenditures on health and education may be
interpreted as investments in human capital.
 Other expenditures may represent investment in
“social capital” in form of institutions that
safeguard property rights.
Barro (1991):
 Examines coefficients of government spending
variables when other long-term growth determinants are
controlled for in the regression.
 Government expenditures are disaggregated into
 government investment;
 government consumption excluding spending on
defense and education;

79
spending on defense and education separately;
 spending on transfer payments.
Government consumption net of defense and education
and transfers may affect growth adversely through
distortionary effects of taxation.
Government investment and defense and education
spending add to productive resources and thus would
have ambiguous effects on growth.
Empirical results are mixed:
 Government investment has positive and statistically
significant partial correlation with growth.
 Government consumption net of defense and
education is negative and significant.




80
Neither education nor defense spending is related to
long-run growth.
 Spending on transfers is positively related to growth,
but Barro interprets this as reverse causation.
Barro (1997) confirms negative effect of government
consumption on long-term growth.
However, interpretation of these results remains open to
question, due to potential for reverse causation.
To identify separate effect of government size on
economic growth appropriate instrument is required.
Such instruments have not been easy to find.
Thus, interpretation of negative partial correlation
between government consumption and growth remains
ambiguous.






81
Commercial Openness
and Growth




Under conditions of financial openness, increased
commercial openness may reduce risk premium that
external creditors require.
Under neoclassical assumptions, this may result in
larger steady-state capital stock and thus more rapid
accumulation-driven growth during transition.
Endogenous growth models: exporting and importing,
by increasing economy's exposure to new technologies,
 facilitate their adoption;
 thus increase rate of growth of productivity.
Implication: trade liberalization, which promotes
commercial openness, should induce
 increase in the level of income;
 increase in its rate of growth.
83
Dollar (1992):
 Relevant definition of openness: one that combines
liberal trade regime with stable real exchange rate.
 To measure outward orientation of trade regime, he
uses deviations of Summers-Heston price levels from
values predicted from regression of price levels on
 per capita GDP, and
 measure of population density.
 Distorted trade regime would result in appreciated real
exchange rate, and thus high price level.
 Results:
 Asian developing countries had the most liberal trade
regimes.
 African countries are the least liberal.
84
Latin American countries in between.
 Increased trade distortions and increased real
exchange rate variability have significant and large
negative effects on economic growth.
Sachs and Warner (1995):
 Factors that determine whether countries with low
income per capita will achieve convergence.
 Two conditions are critical: preservation of private
property rights and commercial openness.
 Their methodology involves classification of countries
into two groups:
 those which safeguarded property rights and
maintained commercial openness (“qualifiers”),
 those which did not (“nonqualifiers”).

85




Trade openness was defining characteristic of two
groups, since almost all countries that failed to qualify
on openness criterion also did so on political criterion.
Qualifiers grew more rapidly, and both political and
trade variables had significant partial effects on growth.
No country which maintained substantially opened trade
failed to grow by at least 2% per year during 1970-89.
Conclusion: safeguarding property rights and
maintaining open trade regime
 are conducive to growth, and
 constitute sufficient conditions for attainment of
rapid economic growth.
86
Frankel, Romer, and Cyrus (1996):
 Both of the previous studies leave open direction of
causality between growth and openness.
 They addressed this issue by using gravity model to
instrument for openness in cross-country growth
equation.
 They found strong positive correlation between
exogenous component of openness and economic
growth.
87
Exchange-Rate Unification
and Growth
Restrictions on financial trades involves foreign
exchange transactions.
 Most common form involved capital account
transactions in balance of payments.
 Such restrictions have been intensified when domestic
economic distortions have created incentives for
residents to remove funds from the country.
 Private agents have sought to circumvent restrictions by
trading foreign exchange outside official markets.
 This gives rise to parallel exchange market at which
foreign exchange trades at substantial premium over its
official value.
Effects of removal of restrictions:
 Removal of restrictions on capital inflows can generate
89
resources for investment.

Removing restrictions on outflows may do so as well, by
 assuring foreign creditors that they will be able to
repatriate their funds when desired, and
 reassuring both domestic and foreign investors that
their capital will be less subject to taxation.
 Enhanced liquidity provided to domestic residents may
induce them to undertake less liquid but more
productive investment projects.
 Financial integration may affect growth indirectly by
fostering deeper domestic financial markets, thus
reinforcing growth benefits of financial deepening.
Evidence:
 Evidence on effects of easing of foreign exchange
restrictions on economic growth is of two types:

90
Use of premium on foreign exchange in parallel
markets as a proxy for capital-account restrictions in
cross-country growth regressions.
 Assessing whether international financial integration
affects economic growth through indirect channel of
promoting domestic financial depth.
Levine and Zervos (1996):
 Provided evidence of the first type.
 Used cross-country sample of 119 countries.
 In testing for effects of parallel market premium, they
investigated robustness of its role since large premium
may reflect variety of policy distortions.
 Result: premium had robust negative partial correlation
with long-term growth.

91
Implication: foreign exchange restrictions exerted
independent negative effect on growth.
De Gregorio (1992):
 Provided evidence of the second type.
 Explained cross-country differences in measures of
financial depth on the basis of
 set of control variables (initial GDP per capita,
average rate of inflation, and measure of commercial
openness), and
 measures of degree of international financial
integration.

92

Results:
 Three of his measures of international integration had
statistically significant partial correlation with
measures of financial depth.
 This is interpreted as evidence in support of indirect
effect.
 No evidence of direct effect of openness on growth is
found.
93