6
Networking for foreign direct investment: the telecommunications industry and its effect on investment Taylor Reynolds * , Charles Kenny, Jia Liu, Christine Zhen-Wei Qiang International Telecommunication Union, Place des Nations, CH-1211 Geneva 20, Switzerland Available online 16 September 2003 Abstract It is increasingly recognized that the level and quality of infrastructure may have an im- portant causal relationship with inflows of foreign direct investment (FDI). In recent litera- ture, information infrastructure has been singled out as a potentially significant source of productivity increases and economic growth, in part through its role in attracting investment and increasing the returns to that investment. This paper examines the empirical relationship between FDI flows and the level of telecommunications infrastructure present in host coun- tries, and finds preliminary evidence of a significant link. Ó 2003 Elsevier B.V. All rights reserved. 1. Introduction Many researchers have attempted to pinpoint factors that firms are interested in when making their decision to invest abroad. Economic stability and the political climate are two determinants of investment found in many previous studies (Agar- wal, 1980; Schneider and Frey, 1985; Nigh, 1985; Culem, 1988; Hein, 1992; Singh and Jun, 1995). Early theoretical work on foreign direct investment also focused on infrastructure as a key determinant. Hymer (1970) develops a model highlighting the interaction between governments and foreign firms that determines the level of * Corresponding author. E-mail address: [email protected] (T. Reynolds). 0167-6245/$ - see front matter Ó 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.infoecopol.2003.07.001 www.elsevier.com/locate/econbase Information Economics and Policy 16 (2004) 159–164 INFORMATION ECONOMICS AND POLICY

Networking for foreign direct investment: the telecommunications industry and its effect on investment

Embed Size (px)

Citation preview

Page 1: Networking for foreign direct investment: the telecommunications industry and its effect on investment

INFORMATION

ECONOMICSAND POLICY

www.elsevier.com/locate/econbase

Information Economics and Policy 16 (2004) 159–164

Networking for foreign direct investment:the telecommunications industry and its effect

on investment

Taylor Reynolds*, Charles Kenny, Jia Liu,Christine Zhen-Wei Qiang

International Telecommunication Union, Place des Nations, CH-1211 Geneva 20, Switzerland

Available online 16 September 2003

Abstract

It is increasingly recognized that the level and quality of infrastructure may have an im-

portant causal relationship with inflows of foreign direct investment (FDI). In recent litera-

ture, information infrastructure has been singled out as a potentially significant source of

productivity increases and economic growth, in part through its role in attracting investment

and increasing the returns to that investment. This paper examines the empirical relationship

between FDI flows and the level of telecommunications infrastructure present in host coun-

tries, and finds preliminary evidence of a significant link.

� 2003 Elsevier B.V. All rights reserved.

1. Introduction

Many researchers have attempted to pinpoint factors that firms are interested in

when making their decision to invest abroad. Economic stability and the political

climate are two determinants of investment found in many previous studies (Agar-

wal, 1980; Schneider and Frey, 1985; Nigh, 1985; Culem, 1988; Hein, 1992; Singh

and Jun, 1995). Early theoretical work on foreign direct investment also focused on

infrastructure as a key determinant. Hymer (1970) develops a model highlighting the

interaction between governments and foreign firms that determines the level of

*Corresponding author.

E-mail address: [email protected] (T. Reynolds).

0167-6245/$ - see front matter � 2003 Elsevier B.V. All rights reserved.

doi:10.1016/j.infoecopol.2003.07.001

Page 2: Networking for foreign direct investment: the telecommunications industry and its effect on investment

160 T. Reynolds et al. / Information Economics and Policy 16 (2004) 159–164

infrastructure investment, and ultimately, the level of economic development. Root

and Ahmed (1979) provide some empirical support for Hymer�s position. They ex-

amine 38 variables that may affect FDI in a sample of 70 developing countries using

discriminant analysis with a stepwise procedure, to pinpoint which variables may

play a role. Their results are the basis for much of the existing foreign direct in-

vestment research. They find per capita GDP to be highly significant, with the GDPgrowth rate, the level of economic integration, extent of urbanization, number of

presidential transfers, and an aggregate of commerce-transportation-communica-

tions also significant. 1

More recently, a number of studies have suggested a potential role for advanced

infrastructure, in particular, in attractingFDI.A recent survey of international firms in

HongKong, Singapore and Taiwan, for example, found that the presence of advanced

infrastructure was the most important consideration in the placement of regional

headquarters, services and sourcing operations. It was the second most importantfactor in determining the production site (Mody, 1997). Many countries seem keenly

aware of the connection between advanced infrastructure andFDI and have developed

special ‘‘high-tech corridors’’ to attract investment and foster technology transfer.

Malaysia�s multimedia city ‘‘Cyberjaya,’’ Ireland�s ‘‘Digital Park’’, and Korea�s‘‘Digital Media City’’ are all examples of governments building specialized infra-

structure to lure high-tech investment. Telecoms can also attract FDI as an investment

sector itself. The process of privatizing state owned telecom companies and liberal-

ization of the regulatory and tax environments in which they operate has increasedFDI into developing countries. Eastern Europe, for example, witnessed an FDI boom

in its cellular, landline, and data-transmission sectors. Western European firms have

taken multi-billion dollar stakes in Hungary, Poland, Croatia and the Czech Republic

(Euromoney, 12/99). In Morocco, a consortium of firms from Spain and Portugal

recently acquired a US$ 1.1 billion license to build a new cellular network.

At the same time, an extensive literature has emerged linking telecommunications

rollout to economic growth (some recent examples include Canning, 1997a,b; R€ollerand Waverman, 2001; Easterly and Levine, 1997). One avenue for the impact ofrollout on growth may be through an impact on FDI, given the close link found

between FDI and growth (see Blomstrum et al., 1994).

This paper is an attempt to combine two branches of research to produce a more

detailed understanding of how telecommunications provision might affect FDI.

First, the literature has suggested a relationship between general infrastructure

rollout and foreign direct investment. More recently, there is anecdotal evidence that

telecommunications may be both an important destination for, as well as an im-

portant catalyst of, FDI. Second, a number of recent authors suggest that tele-communications infrastructure might affect GDP growth, and one potential avenue

is through increasing FDI. This paper analyzes that relationship more closely. As the

1 The social indicator with the highest significance is the extent of urbanization. It is important to note

the urbanization variable might in part, be acting as a proxy for some type of infrastructure not picked up

by the commerce-transportation-communications variable. Infrastructure investment is usually focused in

urban areas where the fixed costs can be spread among the most users.

Page 3: Networking for foreign direct investment: the telecommunications industry and its effect on investment

T. Reynolds et al. / Information Economics and Policy 16 (2004) 159–164 161

level and growth of GDP is a significant factor in determining foreign direct in-

vestment as well as the level of telecommunications provision, it will be necessary to

separate GDP and telecommunication infrastructure into distinct parts to measure

their individual effects on foreign direct investment. The paper uses statistical tech-

niques that allow for such a division.

2. Models, estimation and data

Data sources and descriptions are given in Table 1.

The initial estimation model, Eq. (1), involves components from previous re-

search. It is similar to the model employed by Root and Ahmed (1979) except it

separates out the telecommunications infrastructure variable in the regressions. This

method keeps important components such as GDP per capita and GDP growth ratebut also disaggregates certain infrastructure measures. In particular, it looks at a

measure of infrastructure provision (telephone mainlines/100 people), and two

measures of private involvement in telecommunications (privatization of the in-

cumbent and involvement of a separate regulator)

Table

Data s

FDI

GD

TEL

OPE

PRI

REG

POP

FDIi;t ¼ a1 þ a2ðGDPi;t�1Þ þ a3ðGDPGROWi;t�1Þ þ a4ðTELi;t�1Þþ a5ðOPENi;t�1Þ þ a6ðPOPi;t�1Þ þ a7ðPRIVDUMi;t�1Þþ a8ðREGDUMi;t�1Þ: ð1Þ

1

ources and descriptions

The FDI variable represents the amount of foreign direct investment as a

percentage of GDP in a given year. FDI data are from the World Bank�sWorld Development Indicators (WDI) and run through 1997. The WDI

cover 212 countries for the years 1960–1998

P (GDPGROW) GDP per capita is also from the World Bank�s World Development

Indicators and is real per-capita GDP as given by purchasing power parity.

All GDP terms are listed in US dollars. Again, the WDI cover 212 countries

for the years 1960–1998. GDPGROW is the annual GDP growth rate

The number of mainlines (telephone lines) per 100 inhabitants comes from

the International Telecommunication Union (ITU). The data cover 206

countries and cover the years 1960, 1965, 1970, 1975–1998

N The World Bank�s WDI include data on imports, exports, and GDP that

are used to calculate openness. The WDI openness variable is comprised of

(Exports + Imports) as a percentage of nominal GDP. The data cover 212

countries from 1960–1998

VDUM The privatization data consist of a dummy variable with a value of one if

the telecommunications sector was privatized. The data are a compilation

of privatization data from Wallsten (1999) and the World Bank

DUM The regulation data consist of a dummy variable with value one if the

country has a separate regulatory agency. Regulation data are compiled

from the ITU and World Bank

Population figures are obtained from the World Bank�s World Develop-

ment Indicators and cover 212 countries for the years 1960–1998

Page 4: Networking for foreign direct investment: the telecommunications industry and its effect on investment

Table 2

Correlation tests

Variables Correlation

GDP and mainlines 0.90794

GDP growth and mainline growth rates 0.74225

162 T. Reynolds et al. / Information Economics and Policy 16 (2004) 159–164

However, initial runs using fixed effects for Eq. (1) produce extremely high levels of

collinearity between two sets of variables as shown in Table 2. GDP per capita and

mainlines per 100 are highly collinear as are the growth rates of mainlines and GDP.

As a result, this research will employ a two-step method for examining the tele-

communications data in order to solve multicollinearity problems. The two-step

procedure replaces the variable for mainline levels per 100 with mainline residuals in

order to capture the essence of the variable and, at the same time, separate it fromGDP. These residuals are deviations from predicted values given by the relationship

of GDP, GDP2, and the level of mainlines per 100. This should help solve the col-

linearity problems evident in other types of analysis and will show how having more

or less phones than predicted by GDP/capita will affect FDI decisions. 2

A two-stage procedure separates the effects of GDP and mainlines per 100 in a

rather simple way.

2 It

potent

regress

TELi;t ¼ a1 þ a2ðGDPi;t�1Þ þ a3ðGDP2i;t�1Þ; ð2Þ

TELPREDICTED ¼ a1 þ a2ðGDPi;t�1Þ þ a3ðGDP2i;t�1Þ; ð3Þ

TELRESIDS ¼ TEL� TELPREDICTED: ð4Þ

Eq. (2) first determines how the number of mainlines per 100 is influenced by GDP

and GDP2 using a fixed effects estimation, where a1 is the intercept for each re-

spective country and a2 and a3 are parameter estimates. These estimates are then

used in Eq. (3) to find a predicted number of mainlines. Lastly, Eq. (4) calculates the

mainline residual, which is then put into the subsequent regression in place of

mainlines per 100 (see Eq. (5)):

FDIi;t ¼ a1 þ a2ðGDPi;t�1Þ þ a3ðGDPGROWi;t�1Þ þ a4ðTELRESIDSi;t�1Þþ a5ðOPENi;t�1Þ þ a6ðPOPi;t�1Þ þ a7ðPRIVDUMi;t�1Þþ a8ðREGDUMi;t�1Þ: ð5Þ

3. Results

Regressions 1 and 2 in Table 3 are estimations using the residuals for mainlines inplace of levels. Instead of measuring how overall telecommunication infrastructure

should be noted that, given a potentially high level of error in GDP per capita numbers and a

ially lower error in telecoms per capita, it is possible that some of the residual in the telecoms/GDP

ion captures measurement error in GDP per capita.

Page 5: Networking for foreign direct investment: the telecommunications industry and its effect on investment

Table 3

Dependent variables FDI as a percentage of GDP

1975–1997

1 2

GDP 0.00003 )5.97E) 6

2.28** )0.42GDP Growth )0.05948 )0.09379

)0.64 )1.03Mainlines Resids 0.012592 0.028836

1.07 2.39**

Openness 0.011903 0.010992

6.46** 6.04**

Population 1.887E) 9 1.163E) 9

1.99** 1.24

Privatization 0.626482 0.516406

3.89** 3.22**

Regulation 0.010739 )0.136880.07 )0.87

Method CS CS&TS

Countries 133 133

Years 23 23

Adjusted R2 0.5440 0.5709

All regressions on panel data with fixed effects. Independent variables have been lagged one period. T-

statistics are below parameter estimates. **Significant at 5%. CS, country specific; TS, time-series specific

dummy variables.

T. Reynolds et al. / Information Economics and Policy 16 (2004) 159–164 163

affects FDI, the two-stage method shows how having more or less phones than

predicted by GDP influences FDI.Table 3 reports results for the two-stage results. GDP is a significant determinant

in the absence of time-series specific dummies, along with openness, population and

privatization. When time-series dummies are used in addition, GDP and population

drop out, to be replaced by main line residuals. Privatization and openness remain

significant. The coefficient on mainline residuals is 0.029. This means having one

more phone line per 100 than is predicted in Eq. (3) increases FDI by almost .029

cents per dollar of GDP. This effect highlights the relationship between telecom-

munication infrastructure and FDI. Using mainline residuals, we are able to showhow both privatization of the telecommunication sector and the relative number of

mainlines per 100 influence the decisions of firms investing abroad. 3

3 A number of regressions were run as robustness checks, but not reported above. These included (1)

fixed effects results using mainlines per 100 in place of mainline residuals. In this regression, with GDP and

mainlines entered separately, neither were significant. Openness and privatization were significant in both

cases, regulation and population in neither (2) fixed effects using telecoms residuals but a PWT dataset for

openness which restricted the period under observation to 1975–1992. In this case, results were similar

expect GDP remained significant with time-series effects, but residuals did not enter as significant (one

might take this as evidence that telecoms rollout has become more important to FDI decisions over time)

(3) averaging data over the 1975–1997 period to produce a regression based on one observation for each

country. In this case only openness entered significantly. Full results are available from the authors.

Page 6: Networking for foreign direct investment: the telecommunications industry and its effect on investment

164 T. Reynolds et al. / Information Economics and Policy 16 (2004) 159–164

4. Conclusion

There are several conclusions we can make about the determinants of foreign

direct investment. The analysis shows a relationship between GDP, openness,

privatization of the telecommunications sector, the level of telecommunication in-

frastructure and the level of FDI. First, telecommunication residuals in a two-stageregression are a significant predictor of FDI. The residuals give us a glimpse into

telecommunication�s marginal effects. We are able to see countries with one more

phone per 100 people than predicted by their income level will have 0.03 cents more

FDI per dollar of GDP. We also find a very strong relationship between privati-

zation of the telecommunications sector and the amount of FDI a country receives.

Unreported regressions suggest that privatizing increases the number of phones per

100 by 1.2. Privatization also increases the amount of FDI by 0.52 cents per dollar

of GDP. Although preliminary, the results suggest that telecommunications caninfluence inflows of FDI and further support the importance of the sector to

development.

References

Agarwal, J.P., 1980. Determinants of Foreign Direct Investments: A Survey. Weltwirtschaftliches 116,

739–777.

Blomstrum, M., Lipsey, R., Zejan, M., 1994. What explains the growth of developing countries. In:

Baumol, W., Nelson, R., Wolff, E. (Eds.), Convergence of Productivity: Cross National Studies and

Historical Evidence. Oxford University, Oxford.

Canning, D., 1997a. A Database of World Infrastructure Stocks 1950–1995, mimeo. Harvard Institute for

International Development.

Canning, D., 1997b. Does infrastructure cause economic growth? International Evidence for Infrastruc-

ture Bottlenecks, mimeo. Harvard Institute for International Development.

Culem, C.G., 1988. The locational determinants of direct investments among industrialized countries.

European Economic Review 32, 885–904.

Easterly, W., Levine, R., 1997. Africa�s growth tragedy: policies and ethnic divisions. Quarterly Journal of

Economics 112, 1203–1250.

Hein, S., 1992. Trade strategy and the dependency hypothesis: a comparison of policy, foreign investment,

and economic growth in Latin America and East Asia. Economic Development and Cultural Change

40 (3), 495–521.

Hymer, S., 1970. The efficiency of multinational corporations. American Economic Review 54 (4), 65–94.

Mody, A., 1997. Infrastructure Strategies in East Asia: The Untold Story. World Bank, Washington, DC.

Nigh, D., 1985. The effect of political events on United States direct foreign investment: a pooled time-

series cross-sectional analysis. Journal of International Business Studies 16, 1–15.

R€oller, L.-H., Waverman, L., 2001. Telecommunications infrastructure and economic development: a

simultaneous approach. American Economic Review 91 (4), 909–923.

Root, F.R., Ahmed, A.A., 1979. Empirical determinants of manufacturing direct foreign investment in

developing countries. Economic Development and Cultural Change 27, 751–767.

Schneider, F., Frey, B.S., 1985. Economic and political determinants of direct investment. World

Development 13 (2), 161–175.

Singh, H., Jun, K.W., 1995. Some new evidence on determinants of foreign direct investment in developing

countries. World Bank Working Papers – International Economics, Trade, Capital Flows, pp. 1–41.

Wallsten, S., 1999. An empirical analysis of competition, privatization and regulation in Africa and Latin

America. Stanford Institute for Economic Policy Research 107 (2), 407–437.