Why Are Developing Countries So Slow in Adopting New Technologies?
Why Are Developing Countries So Slow in Adopting New Technologies?
Why Are Developing Countries so Slow in
Adopting New Technologies?*
Raphael Bergoeing Norman V. Loayza Facundo Piguillem
University of Chile The World Bank University of Minnesota
January 2009
Abstract
We analyze the process of technological innovation from the perspective of developing
countries. Specifically, we explore how developmental and regulatory impediments to
the process of resource reallocation and firm renewal limit the ability of developing
countries to adopt new technologies. First, we study how the availability of personal
computers and incidence of internet usage ‐‐as proxies for technological progress‐‐ are
related to regulatory freedom, governance, and schooling in a large cross‐section of
countries. We find that these characteristics not only exert an independent effect on
technological innovation but also complement each other in this regard. We then build a
stochastic general equilibrium model with heterogeneous firms subject to idiosyncratic
shocks. Technological innovation is modeled as adoption of exogenous productivity
shocks requiring firm renewal. Then, we analyze the independent impact of
developmental and regulatory barriers and the complementarities of their effects on firm
dynamics and the process of technological adoption. As expected, when the process of
firm dynamics is undistorted, firms quickly incorporate the advances from shocks to the
technological frontier. However, when government‐imposed regulations deter the
ongoing process of firm destruction and creation, then technological adoption becomes
sluggish and the economy fails to generate enough growth to close the developed‐
developing gap.
* We are thankful to Naotaka Sugawara and Rei Odawara for excellent research assistance. We
gratefully recognize the financial support of the World Bank’s Knowledge for Change Program. The
findings expressed in this paper are entirely those of the authors. They do not necessarily
represent the views of the institutions to which they are affiliated.
1
I. Introduction
One of the most important developments in the world economy in the last three decades
has been the process of globalization. From the poorest to the most advanced regions,
countries are opening up not only to external trade and finance, but also to the flow of
new people, new policies, new ideas, and new technologies. Few countries, however,
are taking full advantage of the opportunities brought about by globalization. Indeed,
there is large diversity regarding the rate of adoption and adaptation of even
inexpensive technologies generated elsewhere in the world. This leads to the central
question of this paper: why are some countries so slow in adopting new technologies?
The answer we propose here focuses on the Schumpeterian process of firm renewal and
the potential restrictions to its optimal functioning.
Starting with the work of Hopenhayn and Rogerson (1993), Caballero and Hammour
(1994), and Davis, Haltiwanger, and Schuh (1996), a large body of literature shows the
key role of firm dynamics in increasing microeconomic productivity and, consequently,
macroeconomic growth. The entry and exit of firms, involving the reallocation of
resources from less to more efficient firms, explain a substantial share of productivity
improvements in the economy. Resource reallocation, however, implies costly
adjustment: It requires the adoption of new technologies and the assimilation of
production inputs by expanding firms, and the shredding of labor and capital by
declining firms. Without this costly process, economies would be unable to both reap
the benefits of an expanding production possibilities frontier ‐‐the source of long‐run
growth‐‐ and absorb and accommodate negative shocks ‐‐the antidote to protracted
recessions.
This paper studies how firm dynamics can serve as the mechanism through which
technological innovation occurs. If firm renewal is not restrained, domestic enterprises
are able to incorporate the advances of a rising technological frontier made available
through, for instance, trade and financial liberalization. In contrast, if there are obstacles
to the process of resource reallocation, a countries’ ability to adopt new technologies can
be severely handicapped. Some of these impediments are related to the development
status of the economy such as lack of human capital and poor governance, both of which
exacerbate the contractual, financial, and adaptation costs of new technologies (see
Caballero and Hammour, 1998; and Acemoglu and Zilibotti, 2001).
Other impediments, however, result from government’s distorting interventions in
private markets, such as excessive labor regulations, subsidies to inefficient sectors and
firms, barriers to the establishment of new plants, and burdensome bankruptcy laws.
These distortions, and their corresponding misallocation of resources, have been blamed
for the observed differences in growth experiences across countries. In their influential
book, Parente and Prescott (2000) argue that gaps in total factor productivity between
2
countries are produced as they enact regulatory barriers to the efficient use of available
technologies. Bernanke (2005) points out to heavy regulatory burden as the reason why
Europe lags behind the U.S. regarding productivity growth. Likewise, Nicoletti and
Scarpetta (2003) conclude that the presence of government‐owned firms with a degree of
monopoly power, together with restrictions on the entry of new firms, diminishes
competitive pressures that foster innovation and greater efficiency in the OECD. Also
focusing on industrial countries, Gust and Marquez (2004) present empirical evidence
that economies with highly regulated labor and product markets face greater difficulty
in adopting information technologies and suffer from lower productivity growth.
Our contribution in this paper consists of analyzing the process of technological
innovation from the perspective of developing countries. For this purpose, we explore
how impediments to the process of resource reallocation and firm renewal limit the
ability of developing countries to adopt new technologies. In contrast to the papers
focused on rich nations, we not only take into account the policy‐induced regulatory
obstacles to firm dynamics but also the shortcomings inherent to underdevelopment,
such as poor education and faulty governance, which also exacerbate the costs of firm
renewal. Moreover, we analyze how these two types of impediments ‐‐developmental
and regulatory‐‐ interact with each other to affect firm dynamics and, consequently,
technological adoption.
The paper is organized as follows. We first present some motivating evidence on the
importance of developmental and regulatory characteristics for the purpose of
technological innovation. Using the availability of personal computers and incidence of
internet users as proxies of technological progress in a country, we study how they are
related to regulatory freedom, governance, and schooling in a large cross‐section of
countries. We find that these characteristics not only exert an independent effect on
technological innovation but also complement each other in this regard.
To understand these relationships, we then construct a stochastic general equilibrium
model with heterogeneous firms. Firms differ with respect to their level of productivity,
which is determined by their initial technology and history of idiosyncratic shocks. Old
firms tend to become less productive and eventually leave the market. In doing so, they
release resources that are then used to form new firms, which acquire the leading‐edge
technology and enter the market. The technological frontier expands according to a
stochastic and exogenous process. This intends to capture the way developing countries
relate to technological advances, that is, as takers and users rather than developers of
new technologies. Developmental barriers ‐‐such as poor education and governance‐‐
are modeled as a parameter that affects the adjustment cost of investment. Regulatory
barriers are, in turn, modeled as either subsidies to inefficient firms or taxes to labor.
Using this framework, we conduct simulation exercises to analyze, first, the independent
impact of each developmental and regulatory barrier and, second, the
complementarities between the two. We are particularly interested in their effect on
3
firm dynamics and the process of technological adoption. Finally, we offer some
concluding remarks and policy implications.
II. Some Motivating Facts
The differences across countries regarding technological adoption are quite large.
Studying 115 technologies in 150 countries, Comin, Hobijn, and Rovito (2006) conclude
that “within a typical technology, the dispersion in the adoption levels across countries
is about 5 times larger than the cross‐country dispersion in income per capita.” What
explains these technological gaps? Most technologies have quite long gestation and
adaptation processes, which makes it hard to identify the causes underlying their cross‐
country variation. The technologies related to the information revolution, however,
offer an interesting exception: Only two decades ago, they were practically nonexistent
almost everywhere; since then, they have been adopted at different rates throughout the
world. This may allow us to identify the effect of certain initial conditions on recent
adoption levels of these technologies.
Before proceeding to this econometric exercise, let’s consider some stylized facts about
these technologies. To maximize data quality and coverage across countries, we work
with two indicators: The number of personal computers per 1,000 people as proxy for
technological progress in production and management processes; and the number of
internet users per 1,000 people as proxy for the advance in telecommunications and
information gathering. Figure 1 presents some regional comparisons on these
technologies, both in levels as of 2003‐04 and in changes between 2003‐04 and 1994‐96.
Since these technologies are rather recent, the comparisons based on levels and changes
are quite similar. The OECD far surpasses any developing region. The typical OECD
country has 5 times more personal computers per capita than the typical East Asian
developing country, 10 times more than the typical Latin American or Middle Eastern
countries, and about 50 times more than the typical Sub‐Saharan African country.
Regarding internet usage, the gaps are smaller but still considerable. The typical OECD
country leads the typical East Asian developing country by 25%, while the other regions
‐‐Latin America, the Middle East, and specially Sub‐Saharan Africa‐‐ lag much farther
behind.
These regional differences are clearly related to income levels, providing some evidence
on the importance of the developmental barriers mentioned above. What about
regulatory barriers? Figure 2 shows some evidence that they are also potentially
important. Using the Fraser index, we divide countries in three groups on the basis of
regulatory freedom. For each of them, we plot the group average of both personal
computers and internet users per population for each year in the period 1990‐2004.
Countries in the top quartile of regulatory freedom have much higher levels and speeds
of adoption of both technology indicators. Countries in the middle (inter‐quartile) range
4
of regulatory freedom also experience an increase over time but, having started their rise
much later, show levels of technology adoption in the mid 2000s that are between one‐
third and one‐half of those in the top quartile. Finally, countries in the bottom quartile
of regulatory freedom start the adoption process much later and slowly than the others,
resulting in enormous technology gaps with respect to the leaders.
The evidence presented above is suggestive that both developmental and regulatory
barriers play a role in explaining technological differences across countries. For more
rigorous empirical support, we now turn to cross‐country regression analysis. This will
allow us to ascertain whether each proposed determinant of technological adoption
exerts a statistically significant effect, even after controlling for the effects of the other
determinants. Moreover, this analysis will help us understand to what extent the
proposed determinants complement each other regarding their effect on technological
adoption.
Table 1 presents information and results of the cross‐country regression analysis. The
dependent variables are the two indicators of technological innovation, that is, personal
computers and internet users, both normalized by population. The explanatory
variables are intended to capture the most relevant developmental and regulatory
characteristics ‐‐the first represented by governance and schooling, and the last one, by
regulatory freedom.1 Specifically, for the quality of public institutions and governance,
we use an index based on International Country Risk Guide indicators on the rule of
law, bureaucratic quality, and absence of official corruption. For education and human
capital, we use the Barro and Lee (2001) measure of average schooling years of the adult
population. For regulatory stance, we use the Fraser Institute Index of Regulatory
Freedom. Appendix 1 provides additional information on definitions and sources of the
three explanatory variables.
The explanatory variables are all measured at the period 1994‐96, while the dependent
variables are measured at 2003‐04. We lag the explanatory variables sufficiently to be
able to consider them exogenous while still connected to the dependent variables. The
resulting samples consist of 83 and 90 countries for the regressions on personal
computers and internet users, respectively.
For each dependent variable, we run three analogous regressions. The first estimates
only linear effects, while the second and third allow for multiplicative interactions. The
linear regression results show that all explanatory variables carry positive and
statistically significant coefficients, indicating that each of them independently is a
relevant determinant of technological innovation. More regulatory freedom, better
1
We make this selection based on the received literature on growth and technological innovation --see De
Soto (1989) and Parente and Prescott (2000) on regulatory freedom; Olson (1982) and Acemoglu, Johnson,
and Robinson (2005) on institutions and governance; and Lucas (1988) and Glaeser et al. (2004) on human
capital and schooling.
5
governance, and higher schooling all lead to faster technological adoption. Moreover,
the results suggest that both developmental and regulatory barriers should be
considered in any attempt to explain cross‐country differences of technological
innovation and, in particular, the backwardness suffering some developing countries.
Since we are interested in assessing not only the direct effects of developmental and
regulatory characteristics but also whether they complement each other, we now
consider the regressions where the regulatory freedom variable is interacted with, in
turn, the governance and schooling variables. These regressions show a clear and robust
pattern: the coefficients on all the multiplicative interactions for both dependent
variables are positive and statistically significant. That is, governance and schooling
complement regulatory freedom, compounding each other’s effect on the availability of
personal computers and internet usage.
We can use the estimated coefficients to evaluate the effect of an increase in regulatory
freedom on technological innovation. For this purpose, we need to consider the
coefficients on both the interaction term and regulatory freedom itself.2 Given that the
total impact depends on the values of the variables with which regulatory freedom is
interacted, it is not really informative to provide a single summary measure of the effect.
Instead, it is best to show how this effect varies with governance and schooling. We do
so in Figure 3. Specifically, the figure presents the total effect on both personal
computers and internet users of a one‐standard‐deviation change in the regulatory
freedom index for the range of sample values of either governance or schooling. In
addition to the point estimates, the figure shows the corresponding 90% confidence
bands (constructed from the estimated variance‐covariance matrix of the corresponding
parameters).
Figure 3 shows that if governance and schooling are very low, an increase in regulatory
freedom may be counterproductive, leading to lower levels of technological innovation.
2
The regression with interaction terms implies the following formula for the point estimate of the
total effect of a change of regulatory freedom on either proxy of technological innovation,
ΔTech = (βREGFREE + βINT DEV) ΔREGFREE
Where Tech represents either personal computers or internet users, DEV represents either
governance or schooling, REGFREE indicates regulatory freedom, the symbol Δ denotes change,
and the parameters βREGFREE and βINT are the estimated regression coefficients on, respectively,
regulatory freedom and the interaction between regulatory freedom and governance or
schooling.
Note that ΔREGFREE is an arbitrary constant, which we set to equal one sample standard
deviation of the regulatory freedom index, and DEV corresponds to any point in the set of values
given by the sample range of either governance or schooling. ΔREGFREE and DEV can thus be
treated as constants. Then, variance of the point estimates (of the effect of a given change of
regulatory freedom on personal computers or internet users) can be obtained as follows,
Var[ΔTech] = {Var(βREGFREE) + DEV 2 Var(βINT) + 2 DEV Cov (βREGFREE, βINT)} {ΔREGFREE}2
Using this variance, we construct the confidence intervals shown in Figure 3.
6
Only countries that are sufficiently advanced in governance, institutions, schooling and
human capital can take advantage of larger regulatory freedom to speed up the process
of technological adoption. Considering the samples in our study, between 60% and 95%
of countries in the world are in a position to benefit from regulatory freedom (the
percentages depending on the proxies for technological adoption and developmental
characteristics under consideration). The rest would require improving their
government institutions and population education as they encourage larger regulatory
freedom. What explains this pattern of complementarity between regulatory and
developmental characteristics to achieve technological innovation? The mechanism we
propose in this paper is based on the incentives and costs of firm renewal. We develop
this mechanism in our model, to which we turn now.
III. A model
We use a general equilibrium model with heterogeneous plants, vintage capital, and
idiosyncratic shocks, as in Bergoeing, Loayza and Repetto (2004), and based on
Hopenhayn (1992) and Campbell (1998). We assume that there exists a distribution of
plants characterized by different levels of productivity. In each period, plant managers
decide whether to exit or to stay in business. If a plant stays, the manager has to decide
how much labor to hire. If the plant exits, it is worth a sell‐off value. New technologies
are developed every period. Plants face two types of shocks: an idiosyncratic shock to
productivity and a shock to the leading edge production process.
In this context, the economy is characterized by an ongoing process of firm entry and
exit, and job creation and destruction. Plants may decide whether to exit in order to gain
access to the leading edge technology ‐Schumpeter’s process of creative destruction‐,
although at the cost of receiving a scrap value for its capital. These investment
irreversibilities, as modeled by Caballero and Engel (1999), together with idiosyncratic
uncertainty, generate an equilibrium with plants rationally delaying exit decisions. Also,
plants may decide whether to exit forever if the economic prospects loom negative.
Within this setting, firms can become more productive over time for two reasons:
because they are exposed to better methods of production or because they thrive while
others disappear.
Finally, we allow for government policies and other general distortions. In particular,
output and investment taxes, labor market policies, and entry costs can be studied. In
what follows we consider a specific policy, a subsidy to incumbent firms, to illustrate the
role of distortions in our model economy.
A. Consumer´s problem
7
There is a representative consumer that lives forever and orders consumption and
leisure in every period according to log ( Ct ) + κ (1 − N t ) . At the beginning of the period
the consumer sells her portfolio of plants (at the price q t0 (θ t ) ) and supplies labor to the
firm. At the end of the period she consumes, buys productive plants (at the price q t1 (θ t ) )
and new plants (at the price q t1i ). Between periods both, productive and new plants
receive an idiosyncratic shock. That is, the consumer solves
⎡∞ ⎤
max E 0 ⎢∑ β t [log (C t ) + κ (1 − N t )]⎥
⎣ t =0 ⎦
subject to
∞ ∞
C t + I tc q t1i + ∫ q t1 (θ t )k t1 (θ t )dθ t = ω t N t + ∫ q t0 (θ t )k t0 (θ t )dθ t
−∞ −∞
1 ⎛ θ t +1 − θ t ⎞ 1 1 ⎛ θ t +1 − z t
∞
⎞
k t0+1 (θ t +1 ) = ∫ σ φ ⎜⎝ ⎟k t (θ t )dθ t + I tc φ⎜ ⎟⎟
−∞
σ ⎠ σ z ⎜⎝ σ z ⎠
B. Firm´s problem
At the beginning of the period the firm buys plants from the consumer ( k t0 (θ t ) ), hires
labor ( N t ) and allocates labor among plants ( n t (θ t ) ). During the period the firm
produces, and capital depreciation takes place. At the end of the period the firm decides
which plants to scrap ( θ t ) and which plants to keep operational. If a plant is scraped, a
proportion s < 1 of capital goods is lost. On the other hand, if the plan is kept
operational, it is sold to the consumer at the price ( q t1 (θ t ) ). Finally, the firm sells the
final output and the goods recovered from scraped plants to the consumer, C t , and to
the investment firm, I t . Thus, the firm solves
⎧⎪ ∞ θt ∞
Max ⎨∫ t t
k 0
(θ ) e θ t 1−α α
n (θ ) d θ + ( 1 − δ ) s k 0
(
∫−∞ t t t
θ ) d θ + (1 − δ ) ∫θ qt (θt ) kt (θt ) dθt
1 0
8
Suppose now that the firm can update each plant’s technology after buying it and before
the labor allocation and selling decisions are made. Let the function γ (θ t ) be the
increase in productivity over θ t that the firm chooses for each plant with technology θ t
and let c(θ , γ (θ ), z ) be the cost function of technology investment. Assume that
c(θ , γ (θ ), z ) is strictly increasing, differentiable, strictly convex and that c(θ ,0, z ) = 0 .
Now the firm solves
θt + γ (θ t )
⎧⎪ ∞ ⎡θ +γ (θ )⎤
⎨∫ t ( t)
θ t t ( t)
θ θ ( ) ∫ kt0 (θt ) dθ t +
δ
t ⎦1−α α
Max0 e ⎣ t
n − w n d + 1 − s
{ t t t t t t } ⎪⎩ −∞
t
γ (θ ) , n (θ ) , k (θ ) ,θ
−∞
∞ ∞ ∞
⎫
(1 − δ ) ∫ qt1 (θ t + γ (θ t ) ) kt0 (θ t ) dθ t − ∫ c (θ , γ (θ ) , zt ) kt0 (θ t ) dθ t − ∫ q (θ ) k (θ ) dθ ⎬⎭
0
t t t
0
t t
θ t + γ (θt ) −∞ −∞
C. Investment sector and market clearing
In addition to the final output firm there exists an investment firm that transforms the
aggregate good into productive plants at zero cost. To construct a new plant an
investment firm requires one unit of aggregate good; that once constructed, it is sold to
the consumer at the price q t1i . Since the transformation technology exhibits constant
returns to scale, the zero profit conditions implies q t1i = 1 .
Finally, market clearing implies that the amount or resources devoted to investment by
the investment firm, I t , has to be equal to the investment undertaken by the consumer,
I tc .
In addition, consumption and investment exhaust the total resources in the economy,
total output and scraped capital. That is
∞ θt
∫ k (θ )e n t (θ t )dθ t + (1 − δ ) ∫ k t0 (θ t )dθ t = K t N tα + S t
θ t 1−α α 1−α
Ct + I t
C
= t
0
t
−∞ −∞
∞
∫ k (θ )e dθ
θ
where K t = t
0
t
t
t
−∞
and
θt
S t = (1 − δ ) ∫ k t0 (θ t )dθ t
−∞
9
If we allow for technology updating there is an additional investment that must be
consider in the aggregate market clearing condition.
∞
Let I tT = ∫ c(θ , γ (θ ), z )k (θ )dθ
0
t t t t t ,
−∞
then the new market clearing condition is
1−α
C t + I tC + I tT = K t N tα + S t = Yt + S t
Definition of the equilibrium: For any fiscal policy {Tt ,τ te ,τ tw ,τ ts } , a Competitive
∞
t =0
t =0 t t
{ }∞
prices ωt, qt1, qt0, qt1i t =0 of labor, plants at the beginning of the period, plants at
the end of the period, and construction projects, and a vector {θ t }t∞=0 such that,
given contingent prices, production and government stochastic processes {zt, θt, τ
t}, at each period t:
1) The representative consumer solves
⎧ ⎡∞ ⎤⎫
max0 ⎨ E0 ⎢ ∑ log ( ct ) + κ (1 − nt ) ⎥ ⎬
{ct , nt , kt (θ ), kt (θ ), Itc }
⎩ ⎣ t =0 ⎦⎭
1
∞ ∞
ct + I tc qt1i + (1 − τ te ) ∫ qt1 (θt )kt1 (θt ) dθ = (1 − τ tw )ωt nt + ∫ q (θ )k (θ ) dθ − T
0
t t t
0
t t
−∞ −∞
∞ 1 ⎛ θt +1 − θt ⎞ 1 ⎛θ − z ⎞ c
kt0+1(θt +1 ) = ∫ φ ⎜⎜ ⎟⎟kt (θt )dθt + φ ⎜ t +1 t ⎟I
⎜ σ ⎟ t
−∞ σθ ⎝ σθ ⎠ ⎝ ⎠
2) The producer of the consumption good satisfies
nt (θ ) = N tα eθt / K t
1−α
⎛ Kt ⎞
ωt = α A ⎜ ⎟
⎜ Nt ⎟
⎝ ⎠
( )
q θ t + γ (θ t ) = s (1 − τ ts )
1
t
10
−α
⎛ Kt ⎞
qt0 (θ t ) = (1 − α ) ⎜
⎜ Nt
⎟
⎟ { } { }
eθt + (1 − δ ) ⎡1 θ t < θ t s (1 − τ ts ) + 1 θ t > θ t qt1 (θ t ) ⎤ − c(θ t , γ (θ t ), zt )
⎣ ⎦
⎝ ⎠
−α
⎛ Kt ⎞ ⎡ ∂qt1 (θ t + γ (θ t ) ) ⎤ ∂c(θt , γ (θ t ), zt )
(1 − α ) ⎜⎜
N
⎟
⎟ { }
eθt +γ (θt ) + (1 − δ ) ⎢1 θ t < θ t 0 + 1 θ t > θ t { ∂γ (θt )
} ⎥=
∂γ (θ t )
⎝ t ⎠ ⎣ ⎦
∞
I tu = ∫ c(θ , γ (θ ), z )k (θ ) dθ
0
t t t t t
−∞
3) The intermediary satisfies
I ti = q1t i I tc
4) The government satisfies
∞ θt
τ te ∫ qt1 (θt )kt1 (θt ) dθ + τ twωt Nt + τ ts ∫ skt1 (θ ) dθ = Tt
−∞ −∞
5) The market clearing restriction is satisfied
ct + I t + I t = Yt + St
c u
Calibration
Most parameter values used in the simulations are standard for the literature. In what
follows time will be measure in years, therefore we will use a discount factor of 0.95,
consistent with a net real interest rate of 5% yearly. Depreciation rate is set at δ=0.06.
Most estimations on the literature fluctuate around this value. The share of labor
incomes to GDP, α, used in the following simulations is 0.7. There has been certain
disagreement about the right value for this parameter, especially when analyzing
developing countries, where National Accounts Calculations seem to be lower.
However, this problem we follow Gollin (2002) and assign this differences to
measurement problems.
To solve for the numerical equilibria we use a three‐step strategy. First, we compute the
non‐stochastic steady state values for the model variables. Second, we linearize the
system of equations that characterize the solution around the long‐run values of the
variables. Third, we apply the method of undetermined coefficients described in
Christiano (1998). To solve the model we scale the variables by the long‐run growth rate
such that they converge to a steady state. Then, a mapping takes the solution from the
scaled objects solved for in the computations to the unscaled objects of interest.
11
Simulations
Figure 1.
5.0
4.5
4.0
% deviation respect to old SS
3.5
3.0
Entry tax
2.5
Eficient
2.0
Labor Subsidy
1.5 Exit tax
0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Years
Figure 1 shows the impulse response of output to a shock of 10% to the leading
technology under different policies. All changes are detrended, thus each curve shows
the transition from the original steady state (which is different for each kind of policy) to
the new one. Therefore, we are measuring the rate of technology absorption as the time
that it takes to the economy to reach a new steady state. Notice that the magnitude
(proportional to the original steady state) of the final jump in each economy is the same;
the main differences are in the time that it takes to reach the new steady state and the
speed at which this happens.
Since Figure 1 can be misleading as to the extent in which the distortions affect the
economy because in any case sooner or later the permanent (proportional) increase is the
same. In order to shed light on this issue we perform a different exercise. Instead of
considering different initial steady states (due to different initial distortions), we perturb
the efficient economy (i.e., without distortions) not only with the leading technology
shock but also with a simultaneous shock on each of the policies consider on Figure 1.
The magnitudes of the shocks are half of the assumed above. Now since the original
steady state is the same in every case the comparisons regarding steady states are
meaningful.
12
Figure 2. (change figure to make red line positive!!)
5.0
Eficient
Entry tax
4.0
Labor tax
Exit Tax
% deviation respect to initial SS
3.0
2.0
1.0
0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
-1.0
-2.0
Years
Figure 3 attacks the complementarities issue. Now, each line is the impulse response (the
same as in figure 1) to a shock of 10% to the leading edge technology with the same
entry tax (set at 10%) but with different scrap values of capital.
13
Figure 3
4.0
% deviation respec to original SS
3.0
2.0
S=0.9
S=0.86
S=0.91
1.0
S=0.77
0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
-1.0
years
14
Figure 4 is similar to Figure 3. Now, each line is the impulse response (the same as in
figure 1) to a shock of 10% to the leading edge technology with the same exit tax
(development barriers, set at 10%) but with different entry taxes.
Figure 4
3.5
% deviation respec to original SS
Te=0.0
2.5 Te=0.05
Te=0.10
1.5 Te=0.15
Te=0.17
0.5
-0.5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
-1.5
-2.5
years
15
Figure 5
10.0
% deviation respec to original SS
8.0
S=0.90
S=0.86
6.0
S=0.81
S=0.79
4.0
2.0
0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
years
Figure 5 is simulation 4 proposed by Norman, each line assumes a different steady state
value of S and the same entry tax set at 15%. Then, the economy is perturbed with a
leading edge technology shock of 10% and simultaneously the entry tax is reduced to
10%.
16
Exercise: shock of 10% on leading edge technology. Then compare different economies
with different steady state distortions. Efficient economy has no taxes. Other economies
have no taxes except by one. All taxes are 10%. Example, Entry tax means all taxes zeros
except by entry tax that is 10%.
Interpretation: each number is the percentage of the total jump that happens in that year.
For example, is after a shock the steady state of the economy jumps 5% and in the first
year the GDP jumps only 0.5%. Then 10% of the jump happens in period 1. Going
further, if for the second year the economy is already 2.5% above the pre-shock situation,
then 50% of the jump happens in the first two years, and so on. Those are the numbers is
the table bellow
Next table is like the one above (shock on z of 10% and the same interpretation) but now
there are two distortions at the same time. The entry tax is set at 10%, then we see what
happens when there are different development barriers (s).
17
Same as the last (shock on z of 10% and the same interpretation) but now we set s=0.81
(10% below the efficient) and we see what happens when there are different entry taxes.
The last exercise is explained in figure 5 that I sent before. Simulation 4 proposed by
Norman, each case assumes a different steady state value of S and the same entry tax set
at 15%. Then, the economy is perturbed with a leading edge technology shock of 10%
and simultaneously the entry tax is reduced to 10%.
18
Table 1. Technological innovation
Method of estimation: Ordinary Least Squares with Robust Standard Errors
No. of observations 83 83 83 90 90 90
R-squared 0.78 0.82 0.81 0.75 0.77 0.75
Notes:
1. t‐statistics are presented below the corresponding coefficients. * and ** denote significance at
the 10 percent and 5 percent levels, respectively. Constant terms are included but not reported.
2. Dependent variables are measured as average of the period 2003‐04. Explanatory variables are
measured as of mid 1990s.
3. Data on dependent variables are from World Development Indicators and on explanatory
variables, as indicated below each variable.
19
Figure 1. Technological Innovation per World Region
Personal Computers (per 1,000 people) Internet Users (per 1,000 people)
500
500
Level Change Level Change
300
400
400
400
300
200
300
300
200
200
200
100
100
100
100
0
0
LAC SSA MNA EA OEC LAC SSA MNA EA OEC LAC SSA MNA EA OEC LAC SSA MNA EA OEC
Notes:
1. LAC: Latin America and the Caribbean; SSA: Sub‐Saharan Africa; MNA: Middle East and
North Africa; EA: ASEAN+2 (South Korea and China) excluding LICs; OEC: OECD.
2. Median value of each country group.
3. Level indicates an average of 2003‐2004 by country, and change means the difference
between the periods 2003‐2004 and 1994‐1996.
4. Data on Personal Computers and Internet Users are from World Development Indicators.
20
Figure 2. Technological Innovation and Regulatory Freedom
Personal Computers by Level of Regulatory Freedom
450
400
350
Personal Computers
(per 1,000 people)
300
250
200
150
100
50
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
400
350
300
(per 1,000 people)
Internet Users
250
200
150
100
50
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
21
Figure 3: Innovation Effect of an Increase in Regulatory Freedom as Function
of Governance and Schooling
Personal Computers
120 90
Change in Pers. Computers
0 0
2 2.5 3 3.5 4 4.5 5 5.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 11.5
-40 -30
-80 -60
Governance (ICRG index) Schooling (average years)
Internet Users
120 90
Change in Internet Users
80 60
Change in Internet users
(per 1,000 people)
40 30
0 0
2 2.5 3 3.5 4 4.5 5 5.5 2.3 3.3 4.3 5.3 6.3 7.3 8.3 9.3 10.3
-40 -30
-80 -60
Governance (ICRG index) Schooling (average years)
Notes:
1. The solid lines show the effect of a one‐standard‐deviation increase in the index of regulatory
freedom on personal computers and internet users, respectively.
2. The x axis represents the .05‐.95 percentile range of, respectively, governance and schooling in
the sample.
3. Dotted lines are 90% confidence bands.
22
References
Acemoglu, D. and F. Zilibotti (2001). “Productivity Differences.” The Quarterly Journal of
Economics, 116(2): 563‐606.
Acemoglu, Daron, S. Johnson, and J. Robinson (2005). “Institutions as the Fundamental
Cause of Long‐Run Growth.” In Handbook of Economic Growth, ed. Philippe Aghion and
Stephen Durlauf, 385‐472. Amsterdam: North Holland.
Barro, R., and J.‐W. Lee (2001). “International data on educational attainment: updates
and implications.” Oxford Economic Papers, 3: 541‐63.
Bergoeing, R., N. Loayza and A. Repetto (2004). “Slow Recoveries.” Journal of
Development Economics, 75, 473‐506.
Bernanke, B. (2005). “Productivity.” Remarks at the C. Peter McColough Roundtable
Series on International Economics.
Caballero, R. and E. Engel (1999). “Explaining Investment Dynamics in U.S.
Manufacturing: A Generalized (S, s) Approach.” Econometrica, 67, 783‐826.
Caballero, R. and M. Hammour (1994). “The Cleansing Effect of Recessions.” American
Economic Review, 84(5), 1350‐1368.
Caballero, R. and M. Hammour (1998). “The Macroeconomics of Specificity.” Journal of
Political Economy, 106(4), 724‐767.
Campbell, J. (1998). “Entry, Exit, Embodied Technology, and Business Cycles.” Review of
Economic Dynamics, 1, 371‐408.
Comin, Diego, Bart Hobijn, and Emilie Rovito (2006). “Five Facts You Need to Know
About Technology Diffusion.” NBER Working Paper No. W11928.
Davis, S. J. and J. Haltiwanger (1992). “Gross Job Creation, Gross Job Destruction, and
Employment Reallocation.” Quarterly Journal of Economics, 107, 819‐863
Davis, S. J., J. Haltiwanger, and S. Schuh (1996). Job Creation and Destruction. Cambridge,
MA: The MIT Press.
De Soto, H. (1989). The Other Path: The Invisible Revolution in the Third World. New York:
HarperCollins.
Easterly, W. (2001). The Elusive Quest for Growth, MIT Press.
23
Glaeser, E., R. La Porta, F. Lopez‐de‐Silanes, and A. Shleifer (2004). “Do Institutions
Cause Growth?” Journal of Economic Growth, 9(3): 271‐303.
Gollin (2002), JPE.
Gust, C., and J. Marquez (2004). “International Comparisons of Productivity Growth:
The Role of Information Technology and Regulatory Practices.” Labour Economics, 5, 33‐
58.
Gwartney, J., and R. Lawson (2006). Economic Freedom of the World: 2006 Annual Report.
Vancouver: The Fraser Institute. www.freetheworld.com.
Hopenhayn, H. (1992). “Entry, Exit, and Firm Dynamics in Long Run Equilibrium,”
Econometrica, 60, 1127‐1150.
Hopenhayn, Hugo, and Richard Rogerson (1993). “Job Turnover and Policy Evaluation:
A General Equilibrium Analysis,” Journal of Political Economy, 101(5), 915‐38.
Jovanovic, B. (1982). “Selection and the Evolution of Industry,” Econometrica, 50, 649‐670.
Lucas, R. (1988). “On the Mechanics of Economic Development.” Journal of Monetary
Economics, 22(1): 3‐42.
Nicoletti, G., and S. Scarpetta (2003). “Regulation, Productivity, and Growth: OECD
Evidence,’” OECD Economics Department Working Paper No. 347.
Olson, M. (1982). The Rise and Decline of Nations. New Haven: Yale University Press.
Parente, S., and E. Prescott (2000). Barriers to Riches. Cambridge: MIT Press.
24
Appendix 1. Definitions and sources of explanatory variables
Variable Definition and Construction Source
25