Hofman 2016

Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Telecommunications Policy
URL: www.elsevier.com/locate/telpol
Information and communication technologies and their

impact in the economic growth of Latin America, 1990–2013
André Hofman n,1, Claudio Aravena, Vianka Aliaga
Economic Commission for Latin American and the Caribbean (ECLAC) in Santiago, Chile
a r t i c l e i n f o abstract
This article examines the contribution of information and communications technologies

Keywords: (ICT) to economic growth and productivity in Latin America in the period 1990–2013.
Information and communication technologies Increased ICT investment explains an important part of the acceleration of economic
Productivity growth in the United States since 1995. An objective of this study is to verify if the same
Sectoral growth happened in Latin America. The analysis is presented at two levels. First, the study takes
Growth accounting into account the aggregate level of the economy, for a set of 18 countries in the region.
Latin America Second, it analyzes the relationship of ICT, economic growth and productivity by sector in
five countries, Argentina, Brazil, Chile, Colombia and Mexico, which are part of a project
on sectoral growth analysis using a growth accounting approach (LA-KLEMS project). The
decomposition of the factors that determine the gap in GDP per capita with the US
showed that improvements in the labor factor have helped to reduce the gap in GDP per
capita relative to the US and conversely poor labor productivity has contributed nega-
tively. The main cause that the labor productivity gap remains is the widening gap in ICT
capital that counteracts improvements in human capital in Latin America. The role of ICT
has been very low, representing less than one sixth of the total capital contribution. The
analysis by economic activity, using the LA-KLEMS data base for Argentina, Brazil, Chile,
Colombia and Mexico, shows that the contribution of capital is the main source of growth
in the fastest growing industry, the transportation industry and communications, and
went hand in hand with high investment especially in ICT.
& 2016 Elsevier Ltd. All rights reserved.
1. Introduction
This article examines the contribution of information and communications technologies (ICT) to economic growth and
productivity in Latin America in the period 1990–2013. Low labor productivity in Latin America has been widely docu-
mented, as well as the divergence of GDP per capita in the region compared to the U.S.. The literature has established that
the increase in ICT investment explains an important part of the acceleration of growth in the U.S. (U.S.) since 1995.
A primary objective of this study is to verify if the same happened in Latin America. It specifically aims to estimate
investment in ICT in a group of countries of Latin America and explain how the low level of ICT investment is one of the
main causes of the weak performance of labor productivity and an important source of the divergence of GDP per capita
between the region and the U.S.
n
Corresponding author.
E-mail addresses: andre.hofman53@gmail.com (A. Hofman), claudio.aravena@cepal.org (C. Aravena), vianka.aliaga@cepal.org (V. Aliaga).
1
Retired Staff Member of the Economic Commission for Latin American and the Caribbean (ECLAC) in Santiago, Chile and actually Professor at USACH
and UAB in Santiago, Chile.
http://dx.doi.org/10.1016/j.telpol.2016.02.002
0308-5961/& 2016 Elsevier Ltd. All rights reserved.
Please cite this article as: Hofman, A., et al. Information and communication technologies and their impact in the
economic growth of Latin.... Telecommunications Policy (2016), http://dx.doi.org/10.1016/j.telpol.2016.02.002i
2 A. Hofman et al. / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎
The analysis is presented at two levels. First, the study takes into account the aggregate level of the economy, for a set of
18 countries in the region.2 Second, it analyzes the relationship of ICT, economic growth and productivity by sector in five
countries, Argentina, Brazil, Chile, Colombia and Mexico, which are part of the LA-KLEMS3 project.
This article is organized as follows. After the introduction, Section 2 revises the literature and describes the evolution of
the level and growth rates of GDP per capita and ICT investment in Latin America. In Section 3 the methodology is pre-
sented, Section 4 presents the data sources. In Section 5 the results are analyzed and Section 6 concludes.
2. Evolution of per cápita GDP and ICT investment in Latin America
The use of digital technologies has increased dramatically in the World and in Latin America in the last quarter century
(ITU, 2015). The impact of ICT in economic growth has been documented extensively (Inklaar, Timmer, & van Ark, 2008;
Jorgenson & Vu, 2005; van Ark, Inklaar, & McGuckin, 2003; Vries de, Mulder, Borgo, & Hofman, 2010). Moreover, recent
approaches to networking, complexity of production and diffusion of knowledge as the core of economic development
underline the central role of ICT in economic growth.4 At the end of the 1990s the difference in economic growth between
Europe and the U.S. was a matter of great concern in the European Union. A possible explanation for the difference was the
greater production and use of ICT in the U.S.. The EU-KLEMS project established the role of ICT in economic growth with
analysis at the industry level of investment in ICT goods and services (Timmer, Inklaar, O'Mahony, & van Ark, 2010).
The digital economy has three main components: Communication equipment, computers and software. ICT production
contributes directly to economic growth by satisfying the increasing demands for their products and also through increased
productivity and investments in ICT using industries. Greater production and use of ICT was found in the US as compared
with European countries, especially in market services (van Ark, Gupta, & Erumban, 2011).
In Latin America analysis on the impact of ICT in economic growth are scarce. Capital formation in ICT was estimated in
Latin America and especially Brazil had relatively high levels (Vries de et al., 2010). Balboni, Rovira, and Vergara (2011) make
use of micro data to study the impact of ICT in Latin America. In Cimoli, Hofman, and Mulder (2010) the impact of ICT in
Latin America was analyzed using different approaches. ECLAC'S latest publication on the digital economy stresses the use of
ICT in the production process (ECLAC, 2015b) and in ECLAC (2013) ICT, structural change and equity was analyzed.
ICT analysis on the sectoral level in Latin America is even more limited. The most recent advance is in the LA-KLEMS
project on where ICT and economic growth were analyzed at the industry level (Hofman, Mas, Aravena, & Fernandez de
Guevara, in press). In Restuccia and Rogerson (2013) the sectoral productivity performance of Latin American countries is
compared with the U.S.. Productivity growth in the agricultural, industrial, and services sectors was below that in the U.S.
during 1950–2000. The exception is the agricultural sector in Chile, which has been catching up since the 1980s. The fall in
relative productivity was more marked in the services sector. Restuccia and Rogerson (2013) makes use of the 10-sector
database of the Groningen Growth and Development Centre (GGDC) which does not have information on capital stock by
sector.
Vergara and Rivero (2006) used capital per industry data from the Banco Central de Chile to calculate industry TFP
growth in Chile during the period 1986–2001. The average annual TFP growth was highest in the Wholesale and retail sector
and lowest in the Manufacturing industry. This high TFP growth in the service sector could be related to the of ICT adoption,
which are relatively large in this sector. Fuentes and García (2014) analyze productivity growth in nine sectors in Chile to
explain recent lower TFP growth. They indicate that the increase in the minimum wage, especially for industries intensive in
labor like manufacturing, construction and retail, and the increase in energy costs could explain lower TFP growth.
Rodríguez et al. explore econometrically, using panel estimations, the impact of ICT in the level of innovation in Latin
America and the Caribbean. Finally, Restuccia (2013) emphasizes that both low labor productivity and total factor pro-
ductivity growth in Latin America explain much of the poor performance in terms of GDP growth.
Szirmai, Naudé, and Alcorta (2013) summarize the literature with respect to the manufacturing industry and its
importance in the development process and emphasize its role in the 19th and 20th century. Szirmai and Verspagen (2010)
stress the role of the manufacturing industry sector especially during periods of accelerated growth. Timmer and de Vries
(2009) point out that the manufacturing sector has become less important in recent periods.
During the 1990s Latin America showed an economic growth of 4.2% (annual average), which represented a recovery
with respect to growth experienced by the region during the 1980s, when average growth only reached 1%. Despite this, the
Latin American region exhibits a slight deterioration in its per capita GDP compared with the U.S. In the 2000 decade, the
region grew at an annual average rate that allowed it to reduce its per capita GDP gap with US, see Table 1.
2
The 18 countries included in this part of the study are: Argentina, Bolivia (Plurinational State of), Brazil, Chile, Colombia, Costa Rica, Ecuador, El
Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Peru, Dominican Republic, Uruguay and Venezuela (the Bolivarian Republic of). In the
remainder of this article the Plurinational State of Bolivia will be named Bolivia and the Bolivarian Republic of Venezuela will be named Venezuela.
3
The LA-KLEMS project is coordinated by the Economic Commission for Latin America and the Caribbean, ECLAC of the United Nations. The project
started with 4 countries, Argentina, Brazil, Chile, and Mexico and later on additional countries, Colombia, Costa Rica and Peru, joined the project.
4
Hidalgo and Cesar (2015), and Hausmann et al. (2014) argue that ICTs lower the cost of storing and reproducing crucial codified information to
develop economic complexity and – in complement with the human capital-accelerate economic growth.
A. Hofman et al. / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3
Table 1
Latin America per capita GDP as percentage of U.S. per capita GDP, 1990–2013 (on the basis of 2011 PPPs).
Source: Prepared by authors on the basis of official sources.
1990 2000 2010 2013
Latin America 0.20 0.19 0.23 0.24

Argentina 0.27 0.29 0.37 0.40
Bolivia 0.10 0.10 0.11 0.12
Brasil 0.30 0.24 0.29 0.28
Chile 0.25 0.32 0.39 0.42
Colombia 0.21 0.18 0.22 0.23
Costa Rica 0.20 0.21 0.25 0.26
Ecuador 0.20 0.16 0.18 0.20
El Salvador 0.12 0.14 0.15 0.15
Guatemala 0.14 0.14 0.14 0.14
Honduras 0.09 0.08 0.09 0.09
Mexico 0.34 0.32 0.32 0.33
Nicaragua 0.08 0.07 0.08 0.09
Panama 0.20 0.22 0.30 0.37
Peru 0.14 0.14 0.20 0.22
Paraguay 0.16 0.13 0.14 0.15
Dominican Republic 0.13 0.16 0.21 0.22
Uruguay 0.26 0.27 0.34 0.37
Venezuela 0.40 0.32 0.34 0.34
Both the evolution of the gap of GDP per capita with the U.S., as well as the level of the latter are heterogeneous across
countries. The countries with lowest gap at the beginning of the nineties were Venezuela, Mexico and Brazil, with a per
capita GDP higher than 30% of the GDP per capita of the U.S.
Table 1 shows the evolution of GDP per capita in PPP of 18 Latin American countries analyzed in relation to the U.S.
during the period 1990–2013. Seven countries – Brazil, Ecuador, Guatemala, Honduras, Mexico, Paraguay and Venezuela –
increased their gap in GDP per capita with respect to U.S. Venezuela is the country that most increased its gap, around 15
percentage points. This fall of the GDP per capita is explained mainly by the reduction in the year 2003, around 30% with
respect to 1990, which was offset by the strong growth between 2003 and 2008.
Eleven countries reduced their distance to the U.S. with respect to the data observed twenty-three years earlier. The
decline after 1998 was offset by increases in GDP per capita from 2003, starting year of a phase of economic growth during
five years on the basis of a very favorable international context. The increase in prices of commodities and favorable external
financing conditions allowed the acceleration of growth at an annual average of 5.7%. The 2009 financial crisis in the U.S.
that caused a global meltdown ended a phase of strong growth in Latin America. After 2009, Latin America recovers its
growth, but at one percentage point lower than the previous period. At the end of the period, in 2013, those countries with a
per capita GDP above 30% of the U.S. GDP per capita were Chile, Argentina, Uruguay, Panama, Venezuela, and Mexico. A
different situation to that observed in 1990.
What has determined this evolution in the gap in GDP per capita with the U.S.? Have ICT played a role in this recovery?
ECLAC (2013) indicates that up to 2011, the ICT development in Latin America had a backlog of more than ten years
compared to the ICT development in OECD countries, mainly explained by the delay in access and use of ICT.
Fig. 1 shows the evolution from 2004 to 2014 of a set of three indicators of access and use of ICT: mobile cellular
subscriptions (Panel A), fixed broadband subscriptions (Panel B), and Internet users (Panel C), for countries of Latin America,
China, India, Africa, the United States and the OECD countries. All countries and regions selected are above the line that
reflects an expansion of use of ICT. The further the points are from this line the larger the growth of use of ICT between 2004
and 2014.
The use of cellular phones has greatly expanded in the Latin American countries as devices’ prices and access fees have
fallen given the pace of technological progress worldwide. As can be seen in Fig. 1A, this fast improvement has pushed
subscriptions of cellular phones up in that region faster than in the U.S. and OECD countries in that period.
In spite of the vast advance in mobile cellular subscriptions, access to fixed broadband and use of Internet have remained
below levels reached by developed regions as OECD and the United States.
Looking at fixed broadband subscriptions (Panel B), Latin American figures for 2014 remained well below 15 subscriptions per
100 people, while OECD attains around 28 subscriptions per 100 people and the United States reaches 30 subscriptions per 100
people. Only Uruguay shows closer levels to developed regions. Two aspects have favored the Uruguayan performance. First, the
large public investment in fiber optic network, which also impacts on good quality of service. Second , the density of population
and geography of the country that facilitate such investment and penetration. It can be argued that, as happens to mobile cellular
subscriptions, access to mobile broadband can change the figures. ECLAC (2015b) indicates that even the diffusion of mobile
broadband is greater than the fixed one, in 2013 the average penetration of mobile broadband of OECD countries attained 79%,
while in Latin America only 30%.
LATIN AMERICA AND THE CARIBBEAN: ICT INDICATO RS OF ACCESS AND USE RELATIVE TO CHINA,
UNITED STATES, AND TO THE OECD COUNTRIES
Fig. 1. Latin America and the Caribbean: ICT indicators of access and use relative to China, U.S., and to the OECD countries. Source: Prepared by the authors
in the base of WDI Worldbank infrastructure indicators.
The number of users of Internet has increased markedly in Latin American, and some countries are lowering the gap with
the OECD and the United States, but gaps remain large. Part of the explanation can be that unlike in developed countries –
where the mass uptake of the Internet is largely contingent on individual preferences or generational constraints – in Latin
America and the Caribbean, coverage among households is determined by the availability of infrastructure and other
economic, social and demographic variables, such as housing location and income level (ECLAC, 2015b).
The amount of Internet users with fixed broadband subscriptions reflects better the delay in the level of access and use of
ICT in Latin American compared to the U.S. and the OECD countries.
Finally, McKinsey & Company (2012) found that Internet contributes an average of 1.9% to the GDP in a group of
developing countries compared with 3.4% in developed countries. In absolute terms, economic value generated annually by
Internet is $119 per capita in developing countries compared with $ 1488 per capita in developed countries. This difference
is attributed to the composition of the Internet's GDP contribution. In developing countries, Internet-related consumption
forms the vast majority of the contribution to GDP, but this component sums little to GDP. In contrast, the internet's
enterprise benefits – which contribute more to GDP – are more prevalent in the mature internet ecosystems of the
developed countries. Internet-related private investment contributes less to GDP in developing countries (13%) than in
developed countries (29%).
In the same line Colecchia and Schreyer (2001) show that, despite differences between countries, the U.S. has not been
alone in benefiting from the positive effects of ICT capital investment on economic growth nor was the U.S. the sole country
to experience an acceleration of these effects. ICT diffusion plays a key role and depends on the right framework conditions,
not necessarily on the existence of an ICT producing sector.
One of the achievements of this paper is to obtain gross fixed capital formation in ICT in Latin America, which con-
solidates two assets: office machinery and computer equipment and telecommunications equipment. The breakdown of
gross fixed capital formation (GFCF) by type of asset ICT and non-ICT was possible for the 18 Latin American countries
indicated before. A striking result is the large size of the participation of ICT assets in gross fixed capital formation in Brazil,
Fig. 2. Latin America: ICT investment as a percentage of total fixed capital formation, 1996–2009 (Percentage).
Source: Prepared by the authors.
which approximately doubles the rest of the Latin American countries, which in majority start with much lower levels of ICT
investment and increase their participation moderately towards 2009.
In order to put these results in perspective, the Latin American participation levels are compared with the results
provided by EU KLEMS. To do this Spain and Italy are taken into consideration, developed countries with a GDP per
capita closer to Latin America and on the other hand, the results are compared with the USA economy. Fig. 2 shows the
GFCF of ICT assets as a percentage of total fixed capital formation and it is clear that Brazil has a very high level of investment
with respect to the countries of the region and even about the same level as the U.S. and higher than the European
countries.
Latin American countries have low levels of ICT investment, but with upward trends over time, and similar behavior to
Spain and Italy until 2000. Some countries converge to similar levels of Spain and Italy. This is not because of their well
performance but of the decline experienced by Spain and Italy's ICT investment rates since 2000. The GFCF in ICT is led by
Brazil, which shows a higher proportion of investment than other Latin American countries, but with a slight downward
trend in recent years.
3. Methodology
The evolution of per capita GDP allows measuring the growth of the economy and, at the same time, taking into account
its relative dimension by adjusting total GDP with population growth. In order to analyze the factors that might explain the
divergence in 18 Latin American countries with respect to the U.S., the level of GDP per capita is decomposed in the fol-
lowing way (see Eq. (1)) and subsequently compared between countries i and j (see Eq. (2)):
Y Y N
¼ h ð1Þ
P Nh P
, ,
Y Y N N
¼ Y=Nh i = Y=Nh j Þ h =h ð2Þ
P i P j P i P j i j
where Y represents the gross domestic product, P population, h the average worked hours per people employed, and N
the number of persons employed. Per capita GDP depends on productivity per worked hour, Y/(Nh); on the employment rate
in total population, N/P; and on the average number of worked hours by people employed, h.
Additionally, to determine the factors that explain the difference in productivity per hour worked, Y/(Nh), in com-
parison with U.S. Eq. (3) is used. This equation describes, through a methodology known as growth accounting,5 effi-
cient combinations of production and inputs for the economy as a whole. Then the production function takes the
following form:
Y ¼ AK α L1 α ð3Þ
5
Methodology originated in the work of Solow (1956, 1957) and Jorgenson and Griliches (1967), Jorgenson and Yip (2001).
Where K and L are the capital and labor inputs, respectively, A is a neutral factor, in Hicks sense, that increases the
aggregated volume of inputs. As a contribution of this study, the analytical framework – and its corresponding measurement
– expands the usual one in order to highlight the composition of capital and labor inputs as follows:
Y ¼ AKðK tic ; K notic Þα ðNhcÞ1 α ð4Þ
where Ktic and Knotic are the ICT capital assets and non-ICT capital assets, respectively; Nh is the total number of hours
worked, and c an indicator of human capital. The estimation of the contribution of ICT capital is important because of the
surge of ICT investments, caused by the acceleration in the decline in prices of hardware and software, which began around
1995. This is consistent with the observation of Abramovitz (1993), that to adjust growth estimates and its sources, it is
important to take into account the specific characteristics of capital in each historical moment. These features of capital also
helps to explain what otherwise appears in Total Factor Productivity, the residual factor corresponding to A in the equations.
Dividing by Y on both sides, taking Y/Nh to the left side of Eq. (4), and regrouping terms gives
1 α α
1 1 α Y
Y 1 α ¼ A1 α ðK tic ; K notic Þ1 α ðNhcÞ
Y
1 α α
Y 1 ðK tic ; K notic Þ
¼ A1 α c ð5Þ
Nh Y
Using Eq. (5), the ratio of GDP per hour (Y/Nh) between countries i and j is given by
0
, !1 α α
1 1 α
Y
=
Y
¼ Ai =Aj @ KðK tic ; K notic Þ KðK tic ; K notic Þ
ci =cj ð6Þ
Nh i Nh j Y i Y j
where differences in GDP per hour may be the result of three factors: differences in TFP, differences in physical capital as
share of GDP, and differences in human capital per worker.
Additionally, in five of the countries examined it is possible to quantify the contribution to economic growth of the
various inputs in nine sectors of economic activity. However, it is not possible to replicate the above equations for analysis,
since the series are in local currency and no purchasing power parity (PPP) for economic activity are available that would
allow sector comparisons.6 However, based on Eq. (4) it is possible to compare the sector contributions, developed as
follows:
Y ¼ AKðK tic ; K notic Þα ðNhcÞ1 α ð40 Þ
Δ ln Y ¼ αΔ ln K þ ð1 αÞΔ ln Nhc þ Δ ln A
Δ ln Y ¼ αΔ ln K þ Δ ln Nhc αΔ ln Nhc þ Δ ln A
Δ ln Y ¼ αΔ ln K þ Δ ln Nhc αΔ ln Nhc þ Δ ln A
K
Δ ln Y ¼ αΔ ln þ Δ ln Nhþ Δ ln c þ Δ ln A
Nhc
Y K K
Δ ln ¼ αKtic Δ ln tic þ αKnotic Δ ln notic þ Δ ln c þ Δ ln A ð7Þ
Nh Nhc Nhc
Each α represents the share of total income of the input expressed in the subscript. That is, the growth of labor pro-
ductivity in every sector of economic activity may be decomposed into the contribution of ICT capital and non-ICT capital
per hour of labor, human capital and a residual factor (A), total factor productivity.7
4. Data
4.1. Capital stock
The net capital stock is defined as the weighted sum of past investments of different types of assets in the economy.8 The
weighting is defined by the relative efficiency of the various assets of different ages. The net capital stock for the asset j is
6
See for example Jorgenson, Nomura, and Samuels (2015) for a detailed industry level analysis with price level indices for outputs and inputs of 36
industries.
7
It is important to stress that the growth accounting methodology applied here has its shortcomings. Especially in the Latin American case the
interpretation of the residual component is complicated. For example, the often observed pro-cyclical nature of the residual in Latin American countries is
often due to volatile growth which is not completely accounted for in calculating factor inputs. In spite of the great improvements in measurement of TFP at
the industry level and with respect to factor services the adagio of Abramovitz (1956), residual as the measure of our ignorance, still counts.
8
Data coming from official sources and elaborated in the LA-KLEMS project.
estimated using the following formula:
X
Tj
K pt;j ¼ I j;t τ Rj;τ Ej;τ ð8Þ
τ¼0
where I j;t τ represents investment of age τ expressed in constant prices; Rj,τ is the retirement function which determines
the proportion of the investment in τ periods that survives; and Ej,τ represents the age-efficiency profile, which char-
acterizes the efficiency loss of productive assets as they age.
Taking into account the above, in this paper a geometric function is used to jointly model the asset retirement and loss of
efficiency thereof.
The functional form used is
!τ
Rj
Rj;τ Ej;τ ¼ 1 ð9Þ
Tj
where Rj is the parameter that defines the rate of loss of efficiency9 and T j the average service life of the asset.
Once the net capital stock for each type of asset is estimated, the next step is to add the assets. The assumption of perfect
competition in the factor market implies that a profit-maximizing firm will use capital to the point where the rent paid
equals the marginal benefit of good. Therefore, aggregation of capital services of different types of assets is performed using
as weights the user cost of capital.
4.1.1. User cost of capital

The capital stock represents quantities. This quantity concept is associated with the corresponding price concept. This
price is called the user cost of capital (Jorgenson, 1963). In its standard version, the user cost is given by the following
expression:
μj;t;0 pj;t;0 ðit þdj;t;0 qj;t Þ ð10Þ
where it is the nominal rate of return, which is assumed to be equal for all types of assets; the depreciation rate as dj;t;τ and
the rate of return of capital as qj;t and μj;t;τ is the amount received by renting the asset with age τ during period t, or cost of
use, which under the assumptions made equals the marginal product of the asset.
In this article, user cost is estimated using an exogenous rate of return obtained from observable interest rates in the
market. The problem is that the relevant rate depends on the financing profile of each company, so often an average of active
an passive rates are used. The use of an exogenous rate of return implicitly assumes (Harchaoui & Tarkhani, 2002) complete
information, which implies that no agency problem exists between the owners of the factors of production and those who
administer them. There is a complete and efficient market for second-hand assets, which means that investment decisions
are reversible, capital assets are divisible and that different types of assets are substitutes in the production process.
One of the consequences of adopting an exogeneous rate of return is that the total value of capital services will not equal
the gross operating surplus obtained from the national accounts. This discrepancy can be explained as the difference
between the expected and actual costs, what indicates that the production process does not exhibit constant returns to scale
or the existence of non-competitive markets. In particular, assume that the nominal rate of return is given by Fisher's
formula: 1 þ it ¼ ð1 þr Þð1 þ π t Þ where as in Mas, Pérez, and Uriel (2005) it is assumed that r ¼4%, which is about the historical
average interest rate free of risk in the OECD. Once the costs of use of capital for each type of asset are estimated, the
variations in the value index of capital services can be obtained using a Tornqvist index10:
!ν j
Kp
Δζ t;K p ¼ Π j p j;t ð11Þ
K j;t 1
where the weights are defined as

μj;t K p
νj ¼ 0:5ðνj;t þ νj;t 1 Þ νj;t ¼ P j;tp
μj;t K j;t
j
4.1.2. Gross fixed capital formation

The gross formation of fixed capital is an indispensable input for the calculation of capital stock. Since it consists of goods
used to produce or create value in a production process, its disaggregation by type of asset is of utmost importance for the
correct estimation of flows of capital services, as also to the contribution to growth of each type of asset. The differentiation
by type of asset allows to distinguish the role that TIC assets play versus the role of non-ICT assets in economic growth.
9
Estimates for the United States by Hulten and Wykoff (1981a, 1981b, 1981c), with values of 1.65 for machinery and equipment and 0.91 for con-
struction, are generally used.
10
For more detail of the methodology see Mas, Pérez, and Uriel (2005).
Given that ICT assets have been integrated strongly in recent decades, and in an dissimilar way into the economies in
study, national accounts now have official series of investment in computer equipment and telecommunications, but only
short time series. In cases where these series are not available, it is necessary to implement a methodology that allows its
estimation. Particularly, in the case of software assets, which does not count with measurements in many countries of the
region.
For countries and periods of time where there are no official series, computer and telecommunications assets are esti-
mated through the "commodity flow method". This method follows this products from domestic or imported production to
its final destination, consumption or investment. First, industrial surveys and trade (BADECEL database) statistics are used to
obtain the apparent expenditure on office equipment and computer and telecommunications equipment, namely domestic
production plus imports less exports. Second, the proportion of investment in the apparent expenditure of such goods in the
input-product matrices of each country is estimated. Finally, to get series of investments in office equipment and computer
and telecommunications equipment, this proportion is applied to the apparent expenditure obtained in the first step, as
shown below:
I IO
i;t
I i;t ¼ IO
ðQ i;t þM i;t Ei;t Þ ð12Þ
ðQ i;t þ ðM IO IO
i;t Ei;t Þ
where Ii,j represents the investment in goods i, office equipment and computer and telecommunications equipment in year t,
Qit represents domestic production, Mit are imports, and Eit are exports of these goods. The above IO index denotes the use
input–output matrix of each country.
To obtain the real series of ICT assets the current prices series should be deflated, an estimate for specific assets was used,
corresponding to a harmonized process described in Vries de et al. (2010), which allows to deflate the series of ICT assets
using the U.S. hedonic deflators adjusted for domestic inflation, as follows:
P GDP country
P ict country ¼ P ð13Þ
P GDP USA ict USA
4.2. Labor
The calculation of labor services reflect more adequately the contribution of each type of worker to growth, recognizing
their curves of different efficiencies depending on their level of education etc. Hours worked provide a starting point for an
economic measure of labor. However, the results may be affected by the individual attributes of each employee, generating
differences in productivity of these hours, so to consider consistent quality could be a biased estimate that ignores the
heterogeneity of the workforce.
A better estimate of the “quality” of labor would distinguish between a measure that reflects the replacement and quality
versus another simple measure that fails to incorporate the heterogeneity of workers and human capital.
The labor factor is defined as the total number of hours worked in a given (H) period, which equals the total number of
workers (N) multiplied by the average number of hours worked per employee (h) and multiplied by a factor of quality (LQ)
incorporating productivity differences between workers.
L ¼ NhLQ ð14Þ
X Pj
LQ ¼ φj ð15Þ
P
The component of job quality (LQ), according to growth accounting reflects the labor composition based on the level of
study. This component is a weighted average of the level of education (P) with the rates of return to schooling obtained
using Mincer wage equations.
The estimation of labor in the 5 countries of LAKLEMS along with incorporating the economic sector they belong to sum
additional educational level as sex and age group characteristics as detailed in the Appendix table, the rate of employment
growth Lt, is expressed as a transcendental logarithmic (translog) function for the types of features i defined as:
X
Δ ln Lt ¼ vt Δ ln H it ð16Þ
where the weights vt are given by the average of the shares of each category in the value of sector earnings. This breakdown
allows to identify the relationship between the contribution to growth and is achieved when observing the heterogeneity of
the labor market.
5. Results
The gap of average per capita GDP of the countries of Latin America with respect to the U.S. corresponds to the ratio of
both per capita GDPs expressed in the form of Eq. (1). In 1990 the per capita GDP of the U.S. was 36,906 (2011 PPP) while the
countries of Latin America averaged 7370 (PPP of 2011), one-fifth of the U.S. In 2013, the difference in per capita GDP of the
Table 2
Convergence of per capita GDP of Latin America and U.S. gap, and its determinants.
PPP US$ of 2011 U.S.¼1
LA U.S. Y/P Y/(N*h) N/P h
1990–2013 9408 44,988 0.212 0.21 0.79 1.31

1990 7370 36,906 0.187 0.23 0.63 1.31
2000 8845 45,862 0.199 0.21 0.75 1.30
2010 11,273 49,602 0.236 0.21 0.92 1.30
2013 12,499 51,258 0.254 0.22 0.91 1.29
Table 3
Convergence of GDP per capita in Latin America and the determinants of divergence with the U.S. (U.S. ¼1).
PIB/(N*h) (%) N/P (%) h (%)
1990–2013 101.7 15.6 17.2

1990 88.3 27.7 16.0
2000 98.8 17.6 16.4
2010 112.6 5.7 18.3
2013 111.9 6.9 18.8
Fig. 3. Impacts on the gap of the determinants of the per capita GDP of Latin America and U.S.
countries of Latin America with respect to the U.S. decreased, representing a quarter of the U.S. This was due to higher
average annual growth of per capita GDP in Latin America of 2.4% versus 1.5% recorded by the U.S. in 1990–2013. On average,
Latin America reduces its gap in per capita GDP with regard to U.S. by a half percentage point per year (the Latin American
population grew faster than in the U.S.).
Eq. (1) allows disaggregating the determinants of per capita GDP and its gap, into the gaps in productivity, employment
and the number of worked hours. It can be seen that the improvement of the employment rate allowed decreasing the gap.
Calculations of these gaps are presented in Table 2.
The employment rate increases strongly in Latin America, from 30% in 1990 30% it reached 42% in 2013, achieving rates
closer to U.S. which went from 48% to 46% in the same period. A great part of the rising level of people employed with
respect to the total population is a result of the growth in the rate of participation of women from 42.9% to 52.9% in 2012.
(ECLAC, 2015a).
The annual average number of hours worked in Latin America is far superior to the U.S. with around 2200 h in Latin
America and 1700 in the U.S. In contrast to the evolution of per capita GDP, where Latin America grows faster than the U.S.,
labor productivity grows at lower rates than the U.S. The United States grows on average 1.7% per year and Latin America
1.5% per year. This is the only indicator where there is no improvement with respect to the U.S.
The results of the determinants of the gap of per capita GDP in Latin America and U.S. show that while indicators of
employment and the average number of hours worked per employee improved – and thereby contributed to reducing the
per capita GDP gap – in the case of labor productivity the difference between Latin America and U.S. increased.
On the basis of the results of Table 2, the impacts of the determinants were estimated and are shown in Table 3 and Fig. 3.
The average worked hours reduced by 20% the gap in per capita GDP with respect to U.S. This drop has been steadily over
Fig. 4. Percentage of the per capita gap between Latin America and the U.S. explained by: Labor productivity (Y/N*h).
time, due to the continuous higher level of worked hours in Latin America as compared with the U.S. While in 1990,
employment had a negative impact on the gap of 27.7%, in the 2010–2013 period it only contributed negatively to the gap by
6%. Unlike the other determinants, labor productivity is the determinant that mostly explains the gap in per capita GDP,
being increasingly responsible of it in 88% in 1990 and 112% in 2013.
These results show that labor productivity is the factor that explains increasingly the gap of per capita GDP between Latin
American countries and the U.S.. On the other hand, the effect of the employment as percentage of total population
decreasingly explains the gap of per capita GDP. The same occurs with the effect of employment, but in the opposite
direction (see Fig. 3).
Fig. 4 shows that the results for individual countries are equivalent to the regional analysis. The degree of expla-
nation of per capita GDP in the Latin American countries with respect to the U.S. by labor productivity increases from
1990 to 2000, and again up to 2010. However, towards 2013 several countries experienced a slight decrease, but almost
always remaining above the values of the previous years of analysis. Based on these results, we conclude that labor
Fig. 5. Determinants of the gap between Latin America and U.S. of GDP per worked hour.
Table 4
ICT capital/GDP contribution to labor productivity relative to U.S.
1990 1995 2000 2005 2010 2013
Argentina 0.96 0.92 0.77 0.70 0.51 0.53

Bolivia 0.95 0.92 0.80 0.73 0.51 0.54
Brazil 0.95 0.92 0.81 0.76 0.71 0.70
Chile 0.97 0.95 0.85 0.82 0.70 0.79
Colombia 0.97 0.95 0.85 0.81 0.68 0.79
Costa Rica 0.97 0.93 0.79 0.79 0.66 0.78
Ecuador 0.93 0.85 0.76 0.76 0.62 0.65
Guatemala 0.96 0.93 0.81 0.80 0.67 0.71
Honduras 0.95 0.92 0.81 0.77 0.68 0.72
Mexico 0.98 0.93 0.74 0.67 0.44 0.46
Nicaragua 0.96 0.92 0.76 0.73 0.65 0.72
Panama 0.92 0.85 0.78 0.75 0.60 0.61
Peru 0.96 0.91 0.80 0.75 0.58 0.64
Paraguay 0.92 0.88 0.70 0.62 0.52 0.54
Dominican Republic 0.92 0.92 0.81 0.81 0.58 0.58
Uruguay 0.96 0.93 0.78 0.71 0.52 0.56
Venezuela 0.96 0.91 0.71 0.62 0.56 0.62
productivity is the main factor to explain along the period of study the gap of the per capita GDP between the Latin
American countries and the United States.
The factors that explain the gap in GDP per hour worked between Latin America and the U.S. (see Fig. 5 and Appendix
Table 2) such as non-ICT asset as share of product, human capital, and TFP and they remain relatively constant or decrease
with respect to the U.S. From 1990 to 2013, all observed countries have reduced their non-ICT capital. U.S. reduced its capital
by 26% in this period. On average, non-ICT capital as share of GDP in the Latin American region do not present great
differences with the U.S. At the country level, in 2013 all converged at a level between 90% and 105% of non-ICT capital of U.
S. However, developments since 1990 have been disparate. Countries like Venezuela, Ecuador, Nicaragua and Guatemala,
reduced this proportion from more than 120% to levels close to 100%. At the other side, El Salvador and Panama are
characterized by reducing this gap with U.S. from around 60% and 70%, to more than 90% of the non-ICT capital experienced
by the U.S. in 2013.
The quality of employment improved steadily over the years from levels around 60 percent of the U.S. at the
beginning of 1990 to a little more than 70% in 2013. It is related to the significant increase in the level of schooling, the
rate of completion of primary and secondary education, as well as to increased access to post-secondary education of
the population in the region during the last thirty years (ECLAC, 2015a). The upward trend of the regional average of
this factor of quality of employment between the region and the U.S. is a reflection of what happens in all countries. In
all the countries of the region, the levels of this factor of job quality improved in greater proportion than in the U.S.
between 1990 and 2013. The increases of quality of employment levels of El Salvador and Brazil, of 43% and 38%,
respectively, are remarkable.
Fig. 6. Latin America: Sources of growth of GDP.

The total factor productivity gap decreased from 40% to 50%, thanks firstly to improvements since 2003 onwards in Latin
America, as well as to the U.S. slowdown following the crisis of 2009. In a context where all countries surveyed increased
their TFP, those who increased the gap with the U.S. are those that did not exceed the increase of 130%, from 1990 to 2013.
The gap in ICT capital strongly increased since the boom in 1995, see Table 4. It increased by 30 percentage points,
from 90% 1995 to 60% in 2013. The effect of ICT capital for all analyzed countries, including the U.S., is decreasing in the
period 1990–2013, with a greater slope for Latin American countries. This greater slope explains the downward trend of
the indicators of the gap with the U.S., which suffers a break in 2010, when the regional indicators change their trends
with an ICT capital that recovers and increases in the following years. This does not occur with the ICT capital in the U.S.
For the average of the region, the main factor that maintains the labor productivity gap is the increasing gap of ICT
capital, which counteracts the improvement of human capital and somehow of the TFP. Therefore, the delay in the incor-
poration of ICT capital has not allowed higher levels of growth.
5.1. Impact of ICT assets on growth
Given the results which determined ICT capital and TFP as the main determinant of the gap in labor productivity with
respect to the United States, and where its incidence – that in most of the 1990s does not exceed 10%-reached in 2013 30%, it
is necessary to quantify its contribution to GDP growth. First, it is necessary to establish whether the small contribution is
generalized in all countries; and second, it is important to examine if this negative effect only affects specific economic
activities of the economy or the entire economy.
The GDP of Latin America for the full period of analysis grew 3.8% per annum, with labor playing a key role and contributing 56%
to its average growth over the period. Capital services provide 43% of the region's GDP growth. The contribution of total factor
productivity to the growth of the economies was slightly positive, due to its performance during the last ten years.
The difference in the dynamics of growth of countries is originated in productivity. An analysis on the basis of different
economic phases experienced by the region in the periods 1991–1998; 1999–2003; 2004–2008 and 2009–2013, shows that
capital services only exceed the contribution of employment in the last period of analysis, 2009–2013.11 This occurs thanks
to both the decline in the contribution of labor with respect to prior periods, as by the increase of the contribution of capital
services, being labor the factor that contributes most in the rest of the periods. Both the contribution of labor and TFP are
pro-cyclical, while the contribution of capital increases from 1999, see Fig. 6. The contribution of ICT capital has been very
low, representing less than one-sixth of the contribution of total capital.
Given the results which determined ICT capital and TFP as the main determinant of the gap in labor productivity with
respect to the U.S., and where its incidence – that in most of the 1990s does not exceed 10% – reached in 2013 30%, it is
necessary to quantify their contribution to GDP growth. This in order to, first, establish if the small contribution is
11
The first sub-period corresponds to the economic recovery of Latin America after the "lost decade" of the 1980s characterized by a strong mac-
roeconomic adjustment. This period ended with the financial crisis of Russia and Asia that pushed the world economy into a recession from which it only
recovered from the year 2003 onwards. Therefore, the second sub-period covers 1998–2003 towards the end of which a new boom started that sub-
stantially raised the prices of raw materials. The following period covers the years of the boom in raw materials 2004–2008 which ends with the great
recession of 2008–2009 generated by the “sub-prime“ crisis in the United States. Finally, the period 2009–2013 corresponds to the recovery of the world
economy.
Table 5
Determinants of labor productivity by industry, 1991–2010 (growth rate of value added per hours worked).Source: Prepared by the authors.
Value added/hours Total capital/labor Non-ICT capital/ ICT capital/labor Labor composition TFP
worked service labor service service
Argentina (1991–2010)
TOTAL 2.2 2.3 1.7 0.6 0.6 0.7
Agriculture, hunting, forestry and 2.8 1.9 1.9 0.0 2.5 1.6
fishing
Mining and quarrying 7.2 4.3 4.2 0.0 1.9 9.6
Total Manufacturing 3.0 2.9 2.5 0.5 0.7 0.6
Electricity, gas and water supply 7.3 6.1 5.1 1.0 0.4 1.6
Construction 1.3 1.4 1.4 0.1 1.1 1.6
Wholesale, retail trade and hotels and 0.8 2.0 1.0 0.9 0.2 1.4
restaurants
Transport and storage and 5.5 2.4 2.1 0.3 1.1 2.0
communication
Finance, insurance, real estate and 1.8 3.9 2.6 1.4 0.1 2.3
business services
Community social and personal 1.7 0.6 0.3 0.3 0.9 0.2
services
Brazil (1996–2009)
TOTAL 1.5 0.1 0.6 0.5 2.1 3.5
fishing
Construction 2.9 0.7 0.6 0.0 1.3 3.6
restaurants
communication
business services
services
Chile (1991–2010)
TOTAL 0.9 1.4 1.1 0.3 1.8 2.3
fishing
Construction 0.4 0.1 0.0 0.1 1.2 1.7
restaurants
communication
business services
services
Colombia (1991–2010)
TOTAL 0.9 2.2 2.0 0.1 0.6 1.9
fishing
Construction 0.3 0.8 0.8 0.0 2.8 3.8
restaurants
communication
business services
services
México (1991–2009)
Table 5 (continued )
Value added/hours Total capital/labor Non-ICT capital/ ICT capital/labor Labor composition TFP
worked service labor service service
TOTAL 0.8 0.5 0.2 0.3 1.2 2.6

fishing
Construction 2.6 0.8 0.5 0.3 1.2 4.6
restaurants
communication
business services
services
generalized in all countries, and second, to examine whether if this negative effect only affects specific economic activities of
the economy or the entire economy.
Unlike Latin America, for the period 1990–2013, in the US the contribution of ICT capital is similar to that of non ICT
capital and above that of labor and TFP explaining 29% of GDP growth, while in Latin America ICT capital only account for 6%.
5.2. Impact of ICT by sectors
Once having determined that the main cause of the increase in the GDP gap of hours worked in Latin America regarding
the US is ICT capital, the complementary database LAKLEMS is used to disaggregate the data by nine economic sectors and
analyze the contribution to labor productivity by each sector to detect if this lack of ICT capital is widespread to all sectors of
economic activity in Latin American countries. These data cover the period 1990–2010, and have a limited coverage of
countries, available only for Argentina, Brazil, Chile, Colombia and Mexico.
In analyzing the contributions to labor productivity by sector, Table 5 highlights some common features. First, in Brazil,
Colombia and Mexico variations in labor productivity differ significantly by sector, while in Argentina and Chile increases
productivity predominate in most sectors. In general, the sectors that have higher labor productivity increases coincide with
the biggest rise in the contribution of capital services. In Mexico, productivity increases in the tradable sectors and decreases
in non-tradable, and Brazil shows greater heterogeneity: while labor productivity experienced a sharp increase in agri-
culture it reduced in industry, construction and most service sectors.
Second, while the contribution of ICT capital is positive in all sectors of economic activity it is very low compared to the
heterogeneous non-ICT capital contribution. Agriculture and construction are those with the lowest contributions and services
have the largest contribution. This result reinforces the finding that ICT capital is the factor that makes the least contribution to
the increase in labor productivity in the economies of the region, both in terms of the total economy and by activity sector.
Third, agriculture tends to be one of the sectors where labor productivity increases. In the case of Brazil, Chile and Mexico
it is associated with a significant increase in the estimated contribution of TFP, which in fact reflects a decline in
employment of unskilled labor in the sector.
Fourth, the estimation of the contribution of TFP tends to be negative in all sectors. Overall, the sectoral pattern of the
contribution of TFP matches the pattern that characterizes all the countries of the region, with major contributions in
sectors with strongest increases or decreases in labor productivity. A case that illustrates the negative contribution of TFP,
reflecting capacity utilization, occurs in the electricity, gas and water sector in Chile, where the change of the energy matrix
resulted in a significant increase in investment but was not subsequently used (because of the disruption of gas deliveries
from Argentina) or with limited efficiency (little added value generated as a result of using costly imported oil).
Fifth, in the countries the contribution of the quality of labor is positive, with major positive contributions in the case of
Brazil and Chile compared to other countries, due to a sharp increase in the share of employment with high and medium
education.
6. Conclusions
Latin America has reduced the gap in per capita GDP compared to the US. In 1990 the per capita GDP of US was 36,906
(PPP of 2011) while for the average Latin American countries it was only 7370 (PPP of 2011), one-fifth of the US. In 2013, this
gap decreases to a quarter of the US. However, this improvement is not shared by all countries, seven countries – Brazil,
Ecuador, Guatemala, Honduras, Mexico, Paraguay and Venezuela – increased their GDP per capita gap relative to the US, and
the other eleven countries reduced their distance from the observed data twenty-three years earlier.
The literature has established that the acceleration of economic growth in the US since 1995 is explained significantly by
increased ICT investment. Since the OECD index of ICT development in Latin America 2011 was lower than the index for 2002,
reflecting a lag of more than ten years, it seems closer to the truth that that the lack of ICT has prevented the closure of the gap.
The decomposition of the factors that determine the gap in GDP per capita with the US shows that improvements in both
the participation rate of the population and the average hours worked in Latin America are those that have helped to reduce
the gap in GDP per capita relative to the US. Poor labor productivity has contributed negatively.
Once concluded that labor productivity is the main cause of the divergence of per capita GDP compared to the US, the
natural question is what determined low labor productivity in the region: lack of capital, low human capital and/or effi-
ciency problems.
When analyzing the determinants of labor productivity – ICT and non-ICT capital stocks, human capital and total factor
productivity – unexpectedly there is no difference in the levels of stock of traditional capital and the main cause that the
labor productivity gap remains is the widening gap in ICT capital that counteracts improvements in human capital and TFP
in Latin America. The role of ICT has been very low, representing less than one-sixth of the total capital contribution.
The analysis by economic activity, using the LA-KLEMS data base for Argentina, Brazil, Chile, Colombia and Mexico, shows
that the contribution of capital is the main source of growth in the fastest growing industry, the transportation industry and
communications, and went hand in hand with high investment especially in ICT.
Disclaimer
The opinions expressed in this article are those of the authors and do not necessarily reflect the views of ECLAC.
Appendix A
See Tables A1 and A2.
Table A1
Characteristics of classification.
Industries Agriculture, hunting, forestry and fishing

Mining and quarrying
Total manufacturing
Electricity, gas and water supply
Construction
Wholesale, retail trade and hotels and restaurants
Transport and storage and communication
Finance, insurance, real estate and business services
Community social and personal services
Gender Female
Male
Age range Aged 15–29

Aged 30–49
Aged 50 and over
Skills range Low skilled

Medium skilled
High skilled
16
Table A2
Explanatory factors of GDP per hour gap between Latin American countries and the U.S.Source: Elaborated by the authors.
Argentina Bolivia Brazil Chile Colombia Costa Rica Ecuador Guatemala Honduras Mexico Nicaragua Panama Peru Paraguay Dominican Republic Uruguay Venezuela
Non ICT Capital_i/GDP_i/(Non ICT_usa/GDP_usa)

1990 1.44 0.64 0.82 0.94 1.07 0.66 1.41 0.80 0.78 0.91 1.44 0.62 1.18 0.67 0.90 1.23 1.93
1995 1.13 0.62 0.80 0.77 1.01 0.63 1.31 0.73 0.84 1.00 1.33 0.56 0.87 0.72 0.87 1.07 1.68
A. Hofman et al. / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

2000 1.18 0.73 0.85 0.89 1.21 0.67 1.39 0.77 0.99 0.96 1.21 0.66 0.97 0.89 0.92 1.16 1.77
2005 1.08 0.72 0.79 0.87 1.04 0.66 1.17 0.79 0.93 1.05 1.15 0.62 0.85 0.84 1.10 1.10 1.54
2010 0.88 0.65 0.70 0.93 0.97 0.67 1.14 0.74 0.96 1.14 1.08 0.59 0.74 0.71 1.00 0.86 1.41
2013 0.90 0.67 0.81 1.00 1.02 0.71 1.16 0.74 0.97 1.20 1.05 0.61 0.83 0.71 1.06 0.84 1.42
ICT Capital_i/GDP_i/(ICT Capital_usa/GDP_usa)

1990 0.18 0.14 0.13 0.32 0.31 0.23 0.04 0.21 0.14 0.50 0.16 0.04 0.19 0.03 0.03 0.18 0.17
1995 0.15 0.15 0.16 0.32 0.31 0.19 0.03 0.20 0.17 0.21 0.14 0.03 0.13 0.06 0.14 0.18 0.11
2000 0.11 0.15 0.17 0.27 0.26 0.14 0.10 0.17 0.17 0.08 0.10 0.12 0.15 0.05 0.17 0.12 0.06
2005 0.09 0.13 0.17 0.28 0.25 0.22 0.17 0.24 0.18 0.07 0.13 0.16 0.16 0.04 0.25 0.11 0.04
2010 0.12 0.12 0.34 0.32 0.30 0.28 0.23 0.29 0.30 0.08 0.27 0.20 0.19 0.13 0.19 0.13 0.17
2013 0.14 0.15 0.34 0.48 0.49 0.47 0.27 0.35 0.37 0.09 0.36 0.21 0.25 0.15 0.19 0.17 0.23
human capital_i/human capital_usa

1990 0.77 0.70 0.53 0.77 0.61 0.69 0.69 0.47 0.55 0.64 0.47 0.72 0.69 0.62 0.60 0.71 0.56
1995 0.76 0.71 0.57 0.77 0.63 0.70 0.68 0.48 0.57 0.67 0.48 0.75 0.72 0.61 0.61 0.70 0.56
2000 0.76 0.74 0.62 0.78 0.65 0.72 0.67 0.49 0.60 0.70 0.49 0.76 0.74 0.61 0.63 0.73 0.58
2005 0.77 0.79 0.66 0.80 0.65 0.72 0.69 0.48 0.63 0.74 0.48 0.78 0.74 0.69 0.64 0.72 0.61
2010 0.78 0.80 0.68 0.82 0.69 0.74 0.72 0.52 0.66 0.76 0.52 0.79 0.75 0.73 0.66 0.73 0.65
2013 0.78 0.81 0.69 0.83 0.71 0.76 0.73 0.54 0.68 0.78 0.54 0.79 0.76 0.75 0.67 0.75 0.67
TFP contribution_i/TFP contribution_usa

1990 0.29 0.26 0.64 0.36 0.32 0.36 0.32 0.41 0.30 0.61 0.20 0.51 0.33 0.34 0.30 0.35 0.66
1995 0.42 0.23 0.56 0.51 0.32 0.39 0.32 0.44 0.22 0.57 0.21 0.55 0.37 0.32 0.33 0.46 0.69
2000 0.46 0.15 0.49 0.53 0.29 0.39 0.30 0.36 0.17 0.68 0.20 0.52 0.29 0.26 0.34 0.58 0.71
2005 0.48 0.14 0.43 0.54 0.31 0.36 0.30 0.31 0.18 0.66 0.20 0.47 0.32 0.23 0.31 0.58 0.71
2010 0.79 0.19 0.46 0.60 0.36 0.43 0.39 0.35 0.17 1.00 0.18 0.68 0.48 0.30 0.53 0.70 0.76
2013 0.82 0.16 0.48 0.60 0.32 0.39 0.40 0.34 0.17 0.90 0.17 0.81 0.50 0.28 0.55 0.72 0.72
References
Abramovitz, M. (1956). Resource and output trends in the U.S. since 1870. American Economic Review
Abramovitz, M. (1993). The search for the sources of growth: Areas of ignorance, old and new. The Journal of Economic History, 53(2).
Balboni, M., Rovira, S., & Vergara, S. (2011). ICT in Latin America. A microdata analysis. Santiago: ECLAC.
Colecchia, A., & Schreyer, P. (2001). ICT investment and economic growth in the 1990s: Is the United States a unique case? A comparative study of nine OECD
countries. OECD Publishinghttp://dx.doi.org/10.1787/078325004705.
ECLAC. (2013). The digital economy for structural change and equality. Santiago: ECLAC and @LIS2. 〈〈http://repositorio.cepal.org/bitstream/handle/11362/
35954/S2013350_en.pdf?sequence ¼1〉〉.
ECLAC (2015a), Social Panorama 2014, Santiago, Chile.
ECLAC (2015b), The new digital revolution – From the consumer internet to the industrial internet, Santiago, Chile.
Cimoli, M., Hofman, A. A., & Mulder, N. (Eds.). (2010). Innovation and economic development: The impact of information and communication technologies in
Latin America. Cheltenham: Edward Elgar.
Fuentes, J., & García, G. (2014). Una Mirada Desagregada al Deterioro de la Productividad en Chile: Existe un Cambio Estructural? Economía Chilena, 17.
Harchaoui, T. M. & Tarkhani, F. (2002). A comprehensive revision of Statistic's Canada estimates of capital input for productivity accounts. Technical report.
Statistics Canada, 〈www.statcan.gc.ca〉.
Hausmann, Hidalgo, C. A., Bustos, S., Coscia, M., Simoes, A., & Yildirim, M. A. (2014). The atlas of economic complexity: Mapping paths to prosperity. Cam-
bridge: The MIT Press.
Hidalgo, & Cesar. (2015). Why information grows: The evolution of order, from atoms to economies. New York: Basic Books.
Hofman, A. A., Mas, M., Aravena, C., & Fernandez de Guevara, J. (2016). LA KLEMS: Economic growth and productivity in Latin America. In D. W. Jorgenson,
K. Fukao, M. P. Timmer (Eds.), The world economy: Growth or stagnation? Cambridge: Cambridge University Press, in press.
Hulten, C. R., & Wykoff, F. C. (1981a). Economic depreciation and the taxation of structures in United States manufacturing industries: An empirical analysis.
In D. Usher (Ed.), The measurement of capital (pp. 83–120). Chicago: University of Chicago Press.
Hulten, C. R., & Wykoff, F. C. (1981b). The estimation of economic depreciation using vintage assets prices: An aplication of the Box–Cox power trans-
formation. Journal of Econometrics, 15, 367–396.
Hulten, C. R., & Wykoff, F. C. (1981c). The measurement of economic depreciation. In C. R. Hulten (Ed.), Depreciation, inflation and the taxation of income from
capital (pp. 81–125). Washington: The Urban Institute Press.
Inklaar, R., Timmer, M. P., & van Ark, B. (2008). Market services productivity across Europe and the US. Economic Policy, 25(33), 139–194.
ITU (International Telecommunications Union) (2015), World Telecommunications/ICT Indicators database, 19th edition.
Jorgenson, D. (1963). Capital theory and investment behavior. American Economic Review, 53(2), 247–259.
Jorgenson, D. W., & Griliches, Z. (1967). The explanation of productivity change. Review of Economic Studies, 34, 249–283.
Jorgenson, D. W., & Yip, E. (2001). Whatever happened to productivity growth? (pp. 509–540)National Bureau of Economic Research, Inc.509–540.
Jorgenson, D. W., & Vu, K. (2005). Information technology and the world economy. Scandinavian Journal of Economics
Jorgenson, D. W., Nomura, K., & Samuels, J. D. (2015). A half-century of trans-pacific competition: Price level indices and productivity gaps for Japanese and
US Industries. 1955-2012. RIETI Discussion Paper No. 15-E-054, Tokio.
Mas, M., Pérez, F., & Uriel, E. (2005). El Stock y los Servicios del Capital en España, (1964–2002). Nueva Metodología, Fundación BBVA.
McKinsey & Company. (2012). Online and upcoming: The Internet's impact on aspiring countries. High Tech Practice
Restuccia, D. (2013). The latin american development problem, an interpretation. Economía, 13(2), 69–100.
Restuccia, D., & Rogerson, R. (2013). Special issue: Misallocation and productivity. Review of Economic Dynamics, 1, 16.
Solow, R. M. (1956). A contribution to the theory of economic growth. The Quarterly Journal of Economics, 70(1).
Solow, R. M. (1957). Technical change and the aggregate production function. Review of Economics and Statistics, 39.
Szirmai, A., Naudé, W. A., & Alcorta, L. (2013). Pathways to industrialization in the 21st century. Oxford: Oxford University Press.
Szirmai, A. & Verspagen, B. (2010). Is manufacturing still an engine of growth in developing countries? Paper presented at the 31st General Conference of the
International Association for Research in Income and Wealth, St. Gallen, Switzerland.
Timmer, M. P., & de Vries, G. J. (2009). Structural change and growth accelerations in Asia and Latin America: A new sectoral data set. Cliometrica, 3(2),
165–190.
Timmer, M. P., Inklaar, R., O'Mahony, M., & van Ark, B. (2010). Economic growth in Europe. A comparative industry perspective. Cambridge University Press.
van Ark, B., Inklaar, R., & McGuckin, R. H. (2003). Changing gear: Productivity, ICT and service industries in Europe and the U.S.. In J. F. Christensen, &
P. Maskell (Eds.), The industrial dynamics of the new digital economy. London: Edward Elgar.
van Ark, B., Gupta, A., & Erumban, A. A. (2011). Measuring the contribution of ICT to economic growth. In B. van Ark (Ed.), The linked world: How ICT is
transforming societies, cultures, and economies. Madrid: Ariel and Fundación Telefónica.
Vergara, R., & Rivero, R. (2006). Productividad Sectorial en Chile: 1986–2001. Cuadernos de Economía, 43(May).
Vries de, G. J., Mulder, N., Borgo, M. D., & Hofman, A. A. (2010). ICT investment in Latin America: Does it matter for economic growth?. In M. Cimoli, A.
A. Hofman, & N. Mulder (Eds.), Innovation and economic development: The impact of information and communication technologies in Latin America.
Cheltenham: Edward Elgar.

Hofman 2016

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Hofman 2016

Transféré par

Droits d'auteur :

Formats disponibles

Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Information and communication technologies and their

This article examines the contribution of information and communications technologies

2. Evolution of per cápita GDP and ICT investment in Latin America

1990 2000 2010 2013

Latin America 0.20 0.19 0.23 0.24

Y ¼ AKðK tic ; K notic Þα ðNhcÞ1 α ð4Þ

Y ¼ AKðK tic ; K notic Þα ðNhcÞ1 α ð40 Þ

4.1. Capital stock

estimated using the following formula:

4.1.1. User cost of capital

where the weights are deﬁned as

4.1.2. Gross ﬁxed capital formation

PPP US$ of 2011 U.S.¼1

LA U.S. Y/P Y/(N*h) N/P h

1990–2013 9408 44,988 0.212 0.21 0.79 1.31

PIB/(N*h) (%) N/P (%) h (%)

1990–2013 101.7 15.6 17.2

1990 1995 2000 2005 2010 2013

Argentina 0.96 0.92 0.77 0.70 0.51 0.53

Fig. 6. Latin America: Sources of growth of GDP.

5.1. Impact of ICT assets on growth

TOTAL 0.8 0.5 0.2 0.3 1.2 2.6

5.2. Impact of ICT by sectors

See Tables A1 and A2.

Industries Agriculture, hunting, forestry and ﬁshing

Age range Aged 15–29

Skills range Low skilled

Non ICT Capital_i/GDP_i/(Non ICT_usa/GDP_usa)

A. Hofman et al. / Telecommunications Policy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

ICT Capital_i/GDP_i/(ICT Capital_usa/GDP_usa)

human capital_i/human capital_usa

TFP contribution_i/TFP contribution_usa

Vous aimerez peut-être aussi