<< . .

. 9
( : 53)



. . >>

Internet users and this number is heavily influenced by the large number of Internet users
in Bermuda, Canada and the United States. Collectively, the statistics reported in Table
1 indicate that IT communication tends to be the highest in the prosperous regions of
the world.
Access to affordable technology to improve the flow of information is crucial to the
development of an economy. Currently four-fifths of the world™s cellular subscribers live
in developed economies. The prospects for growth in cellular phone subscribers is
largest in the developing countries. For many large developing countries, like Russia, it
can take up to ten years to get a fixed line telephone (Economist, 1999). The problem, of
course, is that it doesn™t make economic sense to install wire-line services in regions of
the world where there are very few people. Cellular phone services provide an affordable
alternative to expensive wire-line telephone because cellular phone companies only need
to install transmission towers to send and receive signals and do not have to dig holes
in the ground. It is much cheaper and easier to install transmission towers. With lower
costs, cellular phone companies can break even with a small number of subscribers and
also more easily tailor cellular phone packages to regional tastes. Cellular phone
companies also bring much needed competition and foreign investment to the telecom-
munications industry in many parts of the world.
Internet usage is an important determinant in closing the Digital Divide (Jain, 2002).
Internet access brings many benefits to a country and its citizens. These benefits include
information on health and education, finding lower prices for goods and services,
increased business efficiency, the creation of new jobs, and increased trade. Here,
however, the problems (costs) associated with installing wire-lines re-surface. In North
America and Europe, two regions of the world with high Internet usage rates, the vast
majority of Internet traffic travels across wire-line. Wireless Internet is available but at
a higher cost. In developed countries, wireless products are seen as a luxury. In
developing countries, because of the high cost of laying fixed wire-line, wireless
technology is seen as more of a necessity. As a result, the demand for wireless products
in developing countries may just possibly drive technological innovations in wireless
products.
E-commerce is one area where the Digital Divide is extreme. The statistics are startling.
Currently, three-quarters of all e-commerce is done in the United States and 90% of all
commercial Web sites are located in the United States. There are vast opportunities
awaiting companies and countries that participate in the globalization of e-commerce
(Iyer, Taube and Raquet, 2002). Macroeconomic advantages include innovation (through
bigger rewards in a global market place and knowledge diffusion through the use of the
Internet), efficiency (by transparency through intermediaries and global availability of
information) and trade (via global reach to export goods and the opportunity to export
services). Microeconomic advantages of international e-commerce growth include


Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written
permission of Idea Group Inc. is prohibited.
44 Henriques & Sadorsky


service (with customized and integrated services worldwide and global knowledge about
consumer™s preferences), distribution (lower costs by bypassing retailers and global
reach through electronic channels) and costs (create new distribution channels through
electronic supply chain integration and global selection of suppliers). Clearly, consum-
ers and businesses in both developed and developing economies stand to gain from the
globalization of e-commerce.
Closing the Digital Divide could bring many benefits to developing countries. Bringing
the benefits of IT to developing counties is possible but the governments of these
countries need to be aware that the process is going to cost money and require
institutional changes. Bortolotti, D™Souza, Fantini and Megginson (2002) have com-
pleted a study in which they examined the financial and operating performance of 31
national telecommunications companies in 25 countries that were fully or partially
privatized through public share offerings. Firm profitability was measured in several
different ways including return on assets, returns on sales and return on equity. Their
results indicate that the financial and operating performance of telecommunication
companies significantly improved after privatization, but that regulatory changes also
played a major role.




Methodology
This section, which follows the methodology in Sadorsky and Henriques (2003) and
Sadorsky (2003), describes the empirical approach used to calculate the various risk
measures and associated required returns. The cost of equity for a firm is the minimum
rate of return required to induce investors to buy or hold the company™s stock. This
required rate of return consists of a risk-free rate (representing the time value of money)
and a premium for risk. Alternatively, the cost of equity is the rate used to capitalize
corporate cash flows. It can be used to measure the required rate of return of future equity
investments as long as the future investments are very similar to the current projects
being undertaken by the firm.
Any required rate of return on an equity investment consists of a risk-free rate and a risk
premium.


RRi = Rf + (RPM)(RMi) (1)


In equation (1), RRi is the required return on equity i (or alternatively, the cost of equity),
Rf is the risk-free rate, RPM is the market risk premium and RMi is a risk measure for equity i.
Several risk measures are considered and studied. Systematic risk (SR) is measured by
the capital asset pricing model (CAPM) beta (Brealey and Myers, 2003). Systematic risk
has a long history and is one of the most widely used measures of risk (Brealey and Myers,
2003; Campbell, Low and Mackinlay, 1997). Systematic risk calculated from the CAPM
also has its critics and Fama and French (1997) is one well-known example of this. Total
risk (TR) is measured by the standard deviation of stock returns. Total risk is also widely



Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written
permission of Idea Group Inc. is prohibited.
Risk and Investment in the Global Telecommunications Industry 45


used and is particularly appropriate in segmented markets (which describes the stock
markets in developing countries). Total risk, the combination of systematic and unsys-
tematic risk, is important to the value of the firm. Total risk may have a negative impact
on the firm™s expected cash flow because financial distress is most likely to occur for firms
with high total risk (Shapiro, 2003). Companies experiencing financial distress face
business uncertainty and this uncertainty imposes costs on consumers, suppliers and
employees. When a company is in financial distress, suppliers often charge higher prices
than normal. In addition, some customers become concerned about the long-term
commitment of the company and cancel orders. Consequently, high total risk is likely to
adversely affect a firm™s value via lower sales and higher costs.
Value at risk (VAR) is a well-known measure of the expected losses in extreme downturns
(Alexander, 2001; JP Morgan, 1996). VAR is the expected loss to be exceeded during a
particular time period with a specific probability. The greater the value of VAR, the greater
the potential loss within the defined time period for a particular confidence interval. It
is usual to assume a 95% confidence interval and a time period of one month. In this case,
VAR is the expected loss to be exceeded with a 5% probability during the next month.
For example, assume that the United States stock market (measured by the S&P 500) has
a VAR of $9. This means that for every $100 invested in the United States stock market,
there is a 5% chance of losing $9 or more in any given month. Equation (2) specifies the
VAR calculation.


VAR = (1 “ exp(-1.645σ ))*100 (2)


In equation (2), 1.645 is the critical value for a 95% confidence interval, assuming a normal
distribution for stock returns, and σ is the monthly standard deviation of the monthly
continuously compounded total returns.
Another perhaps lesser known measure of risk is downside risk. Downside risk (DR) is
measured using semi-standard deviation of returns.


T
Σ B = (1 / T )‘ ( Rt ’ B) 2 for all Rt < B (3)
t =1



In equation (3), Rt are monthly returns and B is the benchmark. There are a number of
different benchmark returns and consequently several different measures of downside
risk. A popular choice for B is the arithmetic mean (µ) of the continuously compounded
monthly returns. In this case, equation (3) reads Σ µ. Other measures for the benchmark
include the risk free rate (f) and zero (0) (Estrada, 2000; Harvey, 2000). Downside risk was
discussed by Markowitz (1959) who recognized that investors have asymmetric prefer-
ences towards risk. Most investors like upside risk, but dislike downside risk. Investors
are interested in minimizing risk for two reasons. First, only downside risk or safety first
is relevant to an investor. The idea of safety of principal can be traced to Roy (1952) who
proposed that investors prefer safety of principal first and will set some minimum


Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written
permission of Idea Group Inc. is prohibited.
46 Henriques & Sadorsky


acceptable return that will protect this principal. Second, asset returns may be non-
normally distributed. A downside risk measure is an appropriate risk measure to use when
asset returns are non-normal. Nawrocki (1999) and Sortino and Satchell (2001) provide
a good review of downside risk and Bernstein (1998) provides a good historical
perspective of risk.
Recently, Dembo and Freeman (2001) have introduced regret as a measure of downside
risk


Regret = -E(min(0,R-B)) (4)


where E is the mathematical expectations operator, R is return and B is the benchmark.
As in equation (3), the benchmark, B, can be time varying or fixed. Notice that regret has
the same form as the pay-off function to a put option with a strike price equal to the
benchmark return. In this chapter, regret is measured using a risk-free benchmark (f) and
the zero benchmark (0).




Data
The data used in this study consist of monthly data on 26 telecommunications compa-
nies™ total returns (price returns plus dividends) over the period from January of 1997 to
December of 2002. The total return series are expressed in American dollars. The data were
collected from the Center for Research in Security Prices (CRSP) database. Total returns
on the value weighted U.S. market portfolios of NYSE, AMEX and NASDAQ stocks are
also included in the data set. The companies were selected on the basis of their inclusion
in the www.adr.com telecommunications database and a continuous set of stock price
data over the period from January of 1997 to December of 2002.
American depository receipts (ADR) is a name for foreign company shares that trade on
a local stock exchange and have receipts that trade on a U.S. exchange like the New York
Stock Exchange (NYSE). ADRs are an easy way for U.S. investors, or international
investors with a U.S. trading account, to buy the shares of foreign companies. ADRs were
first introduced in 1927 by Morgan Guaranty. Currently approximately 70% of the ADRs
(330 companies) trade on the NYSE. To round out the sample and for comparison
purposes, three U.S. companies and one Canadian company were included in the data
set. All companies in our sample trade on NYSE.




Risk Measures
Monthly summary statistics for the 26 global telecommunications companies are re-
ported in Table 2. Mean monthly continuously compounded returns are very small and



Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written
permission of Idea Group Inc. is prohibited.
Risk and Investment in the Global Telecommunications Industry 47


Table 2. Summary statistics

Company Country mean stdev Sharpe ratio rho beta skewness kurtosis



APT Satellite Holdings Ltd. Hong Kong 0.00 0.26 -0.06 0.39 1.76 2.14 7.78

Asia Satellite Telecom Hong Kong 0.01 0.15 0.11 0.62 1.63 0.19 -0.40

BCE Inc. Canada 0.02 0.10 0.70 0.53 0.94 0.39 0.72

BT Group PLC U.K. 0.00 0.11 -0.20 0.59 1.18 -0.21 -0.59

Cable & Wireless U.K. -0.03 0.16 -0.63 0.56 1.55 -0.35 0.38

China Mobile Ltd China 0.02 0.17 0.34 0.61 1.78 1.06 1.44

Deutsche Telekom AG Germany 0.01 0.15 0.09 0.54 1.40 0.34 0.00

France Telecom France 0.01 0.21 0.09 0.47 1.71 0.59 2.00

Indonesian Satellite Corp Indonesia 0.02 0.21 0.22 0.35 1.28 1.87 6.60

Nippon Telegraph & Telephone Japan -0.01 0.11 -0.40 0.47 0.90 0.65 0.11

Nokia OYJ Finland 0.04 0.17 0.64 0.65 1.94 0.07 -0.29

Philippine Long Distance Philippines -0.02 0.13 -0.54 0.51 1.16 0.55 0.33

Portugal Telecom Portugal 0.00 0.11 -0.03 0.48 0.95 0.10 0.57

SK Telecom Korea 0.04 0.21 0.54 0.47 1.67 1.64 3.58

TDC A/S Denmark 0.01 0.13 0.05 0.49 1.11 0.26 1.99

Telecom Argentina Argentina -0.02 0.22 -0.31 0.44 1.65 0.46 1.02

Telecom Italia Italy 0.01 0.10 0.18 0.46 0.79 0.51 0.52

Telefonica de Argentina Argentina -0.02 0.20 -0.39 0.43 1.47 0.83 2.80

Telefonica del Peru Peru -0.03 0.15 -0.89 0.54 1.38 0.47 1.21

Telefonica SA Spain 0.01 0.12 0.07 0.61 1.24 0.49 0.61

Telefonos de Mexico Mexico 0.02 0.11 0.63 0.66 1.21 -0.04 0.47

Telekomunidasi Indonesia Indonesia 0.02 0.20 0.29 0.46 1.59 0.92 2.05

Vodafone U.K. 0.01 0.11 0.24 0.47 0.91 0.01 -0.65

Nextel U.S. 0.03 0.24 0.32 0.54 2.19 0.50 0.64

AT&T U.S. -0.01 0.12 -0.25 0.45 0.96 0.62 0.85

Verizon U.S. 0.01 0.10 0.13 0.31 0.55 1.10 2.39

Average 0.01 0.16 0.03 0.50 1.34 0.58 1.39

US Market 0.00 0.06 -0.15 1.00 1.00 -0.46 -0.46



Notes: Means (rates not percentages), standard deviations, rho, skewness, and kurtosis
reported for monthly stock return values. Sharpe ratios calculated for annual values.



Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written
permission of Idea Group Inc. is prohibited.
48 Henriques & Sadorsky


range from a low of -3.00% to a high of 4.00%. The sample average for the mean monthly
return was 1.00% and the mean monthly return on the U.S. market index was 0.00%. The
five-year period from January of 1997 to December of 2002 was one characterized by no
capital appreciation in the broad-based stock market. Most of the stock returns had some
evidence of either skewness or kurtosis. Most of the companies in the sample had high
standard deviations and, as a result, the values for the Sharpe ratios are low. Risk-averse
individuals prefer high values of the Sharpe ratio. BCE had the highest value of the Sharpe
ratio in the sample (0.70) and this value was 23 times larger than the Sharpe ratio for the
sample company average. On a risk-adjusted basis, BCE was a better investment than the
value-weighted U.S. market. In comparison, Telefonica del Peru, with a Sharpe ratio of
-0.89, was a particularly risky investment. The variable rho measures the correlation
between monthly company total returns and monthly U.S. market total returns. All of the
rho values are positive indicating that each company™s stock returns are positively
correlated with the broad based U.S. market. U.S. companies do not necessarily have the
highest correlation with the U.S. market. For example, Telefonos de Mexico had the
highest correlation with the U.S market while Verizon had the lowest correlation. This
result is useful to investors interested in building a portfolio of global telecommunica-
tions stocks.
The second column in Table 3 reports company beta values. Company beta values are
calculated from a single factor market model. More specifically,


Rit = ±i + βi Rmt + uit (5)


where Rit is the company return, R mt is the return on the U.S. stock market index, and uit
is the error term. Systematic risk (SR) is measured by the beta, βi, and idiosyncratic risk
^
(IR) is measured by the standard deviation of the residuals uit .
Many of the company betas are larger than unity (Table 3). Company beta values show
considerable variation and range from a low of 0.55 for Verizon to a high of 2.19 for Nextel.
Notice that in this sample, two U.S. companies have the lowest and highest values for
systematic risk. Based on these risk measures, Verizon stock is much less risky than the
U.S. market while Nextel stock is much more risky than the U.S. stock market. The
systematic risk values indicate that there is a wide variation in the risk of these
telecommunications companies. The company average systematic risk measure is 34%
larger than the systematic risk value for the U.S. stock market.
As discussed in Harvey (2000) and Estrada (2002), the single factor model in equation
(5) can also be used to calculate two downside beta measures. The first measure,
downside beta 1 (DB1), is calculated as the coefficient on the market return using
observations when company returns and U.S. market returns are both negative. The
second measure, downside beta 2 (DB2), is calculated as the coefficient on the market
returns when U.S. market returns are negative.
For each company, ten risk variables are calculated (Table 3). The risk variables are
systematic risk (SR) measured by beta, total risk (TR) measured by the standard deviation
of returns, value at risk (VAR), downside risk (DR) measured by the semi-standard


Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written
permission of Idea Group Inc. is prohibited.
Risk and Investment in the Global Telecommunications Industry 49


Table 3. Risk measures

Σf Σ0
Σµ
Company SR TR VAR DB1 DB2 REGf REG0




APT Satellite Holdings Ltd. 1.76 0.26 -35.00 0.13 0.16 0.13 2.33 2.74 0.12 0.08

Asia Satellite Telecom 1.63 0.15 -22.20 0.10 0.12 0.10 2.02 2.35 0.08 0.06

BCE Inc. 0.94 0.10 -15.33 0.07 0.08 0.05 1.04 1.38 0.05 0.03

BT Group PLC 1.18 0.11 -17.19 0.08 0.11 0.08 1.57 1.94 0.07 0.05

Cable & Wireless 1.55 0.16 -22.99 0.12 0.16 0.13 2.14 2.28 0.10 0.07

China Mobile Ltd 1.78 0.17 -24.19 0.10 0.11 0.09 1.59 2.49 0.08 0.05

Deutsche Telekom AG 1.40 0.15 -21.63 0.10 0.12 0.09 1.63 2.03 0.08 0.06

France Telecom 1.71 0.21 -29.24 0.14 0.16 0.13 2.64 2.12 0.09 0.07

Indonesian Satellite Corp 1.28 0.21 -29.41 0.12 0.13 0.11 2.27 1.82 0.09 0.06

Nippon Telegraph & Telephone 0.90 0.11 -16.74 0.07 0.10 0.08 1.05 1.25 0.08 0.05

Nokia OYJ 1.94 0.17 -24.51 0.12 0.12 0.10 1.70 2.51 0.07 0.05

Philippine Long Distance 1.16 0.13 -19.42 0.08 0.12 0.10 1.52 1.60 0.09 0.06

Portugal Telecom 0.95 0.11 -17.17 0.08 0.10 0.08 1.45 1.55 0.07 0.04

SK Telecom 1.67 0.21 -28.64 0.11 0.11 0.09 1.44 2.35 0.08 0.05

TDC A/S 1.11 0.13 -19.19 0.09 0.11 0.08 1.96 1.62 0.07 0.04

Telecom Argentina 1.65 0.22 -30.12 0.15 0.18 0.15 2.12 2.28 0.11 0.09

Telecom Italia 0.79 0.10 -15.06 0.06 0.08 0.06 0.97 1.22 0.06 0.03

Telefonica de Argentina 1.47 0.20 -27.75 0.13 0.16 0.14 1.96 2.07 0.11 0.08

Telefonica del Peru 1.38 0.15 -21.34 0.10 0.14 0.12 2.21 2.41 0.10 0.07

Telefonica SA 1.24 0.12 -17.55 0.08 0.10 0.07 1.41 1.83 0.06 0.04

Telefonos de Mexico 1.21 0.11 -15.99 0.07 0.08 0.06 1.16 1.80 0.05 0.03

Telekomunidasi Indonesia 1.59 0.20 -27.95 0.12 0.13 0.11 1.89 2.37 0.09 0.06

Vodafone 0.91 0.11 -16.76 0.08 0.10 0.07 1.25 1.32 0.06 0.04

Nextel 2.19 0.24 -32.13 0.16 0.17 0.14 2.55 3.07 0.10 0.08

AT&T 0.96 0.12 -18.29 0.08 0.11 0.08 1.09 1.38 0.08 0.05

Verizon 0.55 0.10 -15.68 0.06 0.08 0.06 0.69 0.65 0.06 0.03

Average 1.34 0.16 -22.36 0.10 0.12 0.10 1.68 1.94 0.08 0.06

US Market 1.00 0.06 -9.04 0.04 0.07 0.04 0.98 1.47 0.05 0.02




deviation of returns, using three different benchmarks, two downside beta measures, and
two measures of regret.
The TR values range from a low of 0.10 to a high of 0.26. The sample average for TR is
almost three times larger than the TR for the U.S. market.
The -35.00 VAR value for APT Satellite Holdings means that for every $100 invested in
this company there is a 5% probability of losing $35.00 or more in any given month. Based
on the VAR measure, this company is very risky to invest in. Telecommunications

<< . .

. 9
( : 53)



. . >>