Mobile telephony in Ghana: the effect of the introduction of
the mobile phone on economic performance
Abstract This thesis tries to assess the impact of mobile telephony on economic growth in Ghana. Motivation for this is the exponential growth of mobile phone penetration in the country and the existing evidence of the positive impact of
telecommunications infrastructure on economic performance in both developed and developing economies. Using a macroeconomic growth model similar to those used in previous studies, the impact of mobile phones on per capita GDP growth is estimated. The overall results show a significant positive economic impact. Also, the positive effect of mobile telephony on economic growth seems to be more
pronounced when the level of main telephone lines is low.
Nathalie Duijvesteijn
Student number/ UvA NetID: 6030297 Bachelor’s Thesis Economics
1. Introduction
An article in The Economist (2008) reported: “A device that was a yuppie toy not so long ago has now become a potent force for economic development in the world’s poorest countries.” In many places in rural Africa, the mobile phone was the first kind of modern telecommunications infrastructure that the inhabitants of these places were introduced to. In the last decade, the mobile telephony industry has expanded enormously both in the industrialized nations as in the developing world. The telecommunication infrastructure gap between developed and developing countries decreased rapidly with the introduction of mobile phone technologies in Africa. As this gap was always a great component in explaining why poor countries remained poor (Beuermann et al., 2012, p. 1617), this convergence is expected to have an impact on the economic welfare of the developing world.
In the industrialized nations such as OECD countries, it has been shown that there is a causal relationship between telecommunications infrastructure and economic output (Röller and Waverman, 2001). For the developing world there isn’t much evidence for this relationship. Several papers have been written about how mobile phone coverage helped with the reduction of price dispersion for several commodities (Jensen, 2007 and Aker, 2010). However, these results don’t give general evidence of increased aggregate economic development.
According to the World Bank classifications of economies (World Bank, 2013, pp. 4-5), Ghana is one of the lower middle-income economies of sub-Saharan Africa. In this country, only around 15 percent of the roads are paved, slightly more than half of the population has access to electricity, and there is only one main telephone line per 100 people (World Bank). Therefore, Ghana is classified as one of the developing economies and this makes it an interesting country for research about the effect of the introduction of mobile telephony on economic performance.
In this thesis I will attempt to answer the following research question: Did the introduction of mobile telephony result in a better economic development in Ghana?
To answer this question, I will run a regression between the growth rate of per capita GDP and the penetration rate of mobile phones in Ghana, thus the number of mobile phone subscriptions per 100 people. I closely follow the
macroeconomic growth model used by Datta and Agarwal (2004) and Lee et al. (2012). While correcting for different economic factors, such as population growth, openness to trade, and government expenditures, the main results of the estimation show a clear positive effect of mobile phones on economic performance.
The rest of my thesis is organized as follows. In Section 2, I will give an overview of the literature that is related to my research, and I will end it with my hypothesis. In Section 3, I will give a description of the data and introduce and explain the macroeconomic growth model, on which I will run some regressions. I will show and discuss the estimation results in Section 4. Section 5 concludes this thesis with a brief summary and final conclusion.
2. Related literature
2.1 Rapid growth of mobile phone expansion
With the introduction of the mobile telephone, the continent of Africa was introduced to various new possibilities. Mobile phones connect family members across different regions and countries, they make it easier to find out about job opportunities without having to travel, and they make it possible to obtain information about crop prices and selling prices of products sold in different markets. In some countries even cash transfers can be made using mobile phones, with m-money systems. In many places in rural Africa, the mobile telephone was the first kind of modern telecommunications infrastructure the inhabitants of these places were introduced to. The introduction of this device has greatly reduced communication costs, thereby allowing individuals and firms to send and to obtain information quickly and cheaply on a variety of economic, social, and political topics (Aker and Mbiti, 2010).
In the developed world, the telecommunications industry invested first in main telephone lines before moving to mobile networks. In Africa, however, main telephone line networks were too expensive to invest in, especially in the countries with poor roads, vast distances and low population densities. Mobile telephone networks were much easier and cheaper to provide, via a network of specialized base stations, which can provide service to a 5-10 kilometer radius. Due to unreliable
electricity supplies across Africa, diesel generators primarily power these base stations.
The rapid adoption of mobile phones has generated a great deal of speculation and optimism regarding its effect on economic development in Africa. This rapid adoption has far exceeded expectations. In 1999, for example, the Kenyan-based service provider Safaricom projected that the mobile phone market in Kenya would reach three million subscribers in 2020. Now, not even close to 2020, Safaricom alone has already over fourteen million subscribers. This is even more surprising considering the prevalence of poverty in sub-Saharan Africa, where 300 million Africans are classified as ‘poor’, living on less than US$1 per day, and 120 million as ‘ultra-poor’, living on less than US$0,50 per day (Ahmed et al., 2007).
2.2 Micro- and macroeconomic evidence
In this sub-section I will give some examples of evidence of economic benefits resulting from mobile phones on the micro- and macroeconomic level. Several studies show that a reduction in communication costs associated with mobile phones has economic benefits, by improving agricultural and labor market efficiency and producer and consumer welfare in specific circumstances and countries.
The first example is given by the study of Jensen (2007), about the market performance in the South Indian fisheries sector. In the Indian state Kerala, the adoption and usage of mobile phones by fishermen and wholesalers was associated with a dramatic reduction in price dispersion. Another result was the complete elimination of waste, and the near-perfect adherence to the Law of One Price. Both consumer and producer welfare increased significantly.
Another example is the impact of cell phones on grain markets in Niger, examined by Aker (2008). The estimation results in this study provide evidence that cell phones reduce grain price dispersion across markets by a minimum of 6.4 percent and reduce intra-annual price variation by 10 percent. Cell phones have a greater impact on price dispersion for market pairs that are farther away, and for those with lower road quality. This effect becomes larger as a higher percentage of markets have mobile phone coverage. Aker (2010) conducted another, yet similar,
study about the impact of mobile phones on price dispersion across grain markets in Niger. The results were also similar and show that the introduction of mobile phone service between 2001 and 2006 explains a 10 to 16 percent reduction in grain price dispersion. This effect is larger for markets that are more remote and those connected by unpaved roads.
Klonner and Nolen (2008) wrote a paper about their study on the economic effects of the rollout of mobile phone network coverage in rural South Africa. The results show substantial impacts of cell phone network rollout on labor market outcomes with remarkable gender-specific differences. A gender-differentiated analysis shows that most of this effect is due to the increased employment by women. Another result of this study shows that household income increases in a pro-poor way when cellular infrastructure is provided.
Muto and Yamano (2009) study the impact of mobile phone expansion on market participation in Uganda. They estimate the impact of mobile phones on agricultural markets in Uganda, focusing on farmers’ market participation rather than market efficiency. Their findings show that mobile phone coverage is associated with a 10 percent increase in farmers’ probability of market participation for bananas, although not maize, thereby suggesting that mobile phones are more useful for perishable crops.
2.3 Mobile phones and their potential economic benefits
This sub-section explains how mobile phones could be beneficial for economic development in poor countries. Aker and Mbiti (2010) constructed a model indicating the five potential mechanisms through which mobile phones can provide economic benefits to consumers and producers in sub-Saharan Africa:
1. Mobile phones can improve access to and use of information, thereby reducing search costs, improving coordination across agents, and increasing market efficiency;
2. This increased communication should improve firms’ productive efficiency by allowing them to better manage their supply chains;
3. Mobile phones create new jobs to address demand for mobile-related services, thereby providing income-generating opportunities in rural and urban areas;
4. Mobile phones can facilitate communication among social networks in response to shocks, thereby reducing households’ exposure to risk;
5. Mobile phone-based applications and development projects have the potential to facilitate the delivery of financial, agricultural, health, and educational services.
The fifth mechanism needs some more explanation and examples. There are various ways in which mobile phones can provide services and innovative development projects. The underlying belief of this is the idea that mobile phones can offer a useful platform for providing information and services.
The first, and possibly most useful, mobile service is mobile money, also know as m-money or m-banking. M-money systems allow the user to store value in an account accessible by the handset, convert cash in and out of the stored value account, and transfer value between users by using a set of text messages, menu commands, and personal identification numbers (PINs). The best-known example of such an m-money system is M-Pesa, founded in Kenya and now the world leader in mobile money (the Economist, 2013).
To address the importance of m-money as a mobile development project, the following study shows evidence from an Indian social banking experiment (Burgess and Pande, 2005). Lack of access to finance is often cited as a key reason why poor people remain poor. The study of Burgess and Pande (2005) identifies the impact of opening a rural bank on poverty and output. Findings suggest that the Indian rural branch expansion program significantly lowered poverty, and increased non-agricultural output. M-money systems have effectively expanded the breadth and reach of money transfer systems for the rural and urban poor, but their effect on the welfare of poor users in developing countries is still questionable, since these systems are technically not banking from either a financial or legal perspective.
Another example of m-development is the working of m-health projects. Some of these projects include: monitoring measles outbreaks in Zambia, supporting diagnosis and treatment by health workers in Mozambique, sending health
education messages in Benin, Malawi and Uganda, and sending reminders about anti-retroviral therapy schedules to HIV-positive patients.
M-development also includes the facilitated access to agricultural market information: farmers can type in a code, send a text message, and receive the price of a good or product immediately. Farmers can also call or text hotlines to ask for technical agricultural advice.
The political part of mobile services takes places in election campaigns around the world. Transmitting voting results via text messages to a central system facilitates counting the votes in countries with limited infrastructure and communication systems.
Finally, mobile phones are also used to promote literacy. In a country without local language newspapers and village-level libraries, text messaging makes literacy functional. Aker et al. (2010) conducted a randomized evaluation of a mobile phone literacy and numeracy program in Niger, called project ABC. Preliminary results suggest that the mobile phone-based literacy students have higher test scores than students in normal literacy classes, and these results are maintained six months after the end of classes.
2.4 Mobile phones and opportunities in Ghana
Continuing on the potential of mobile phones in creating economic benefits, this sub-section addresses some attention on mobile phone-based applications and development projects in my country of interest, Ghana.
Although Ghana is a developing country with economic conditions similar to Kenya, there is a lack of awareness of mobile money among Ghanaians and the adoption rate of the product is low. Because of this low adoption rates, the great success of M-Pesa in Kenya, and the therefore assumed potential benefits mobile money has to offer the poor and consumers at large, researchers and mobile money service providers organized a 2-day Mobile Money Conference with the theme: ‘Reaching the Unreached: Mobile Money Uptake in Ghana.’ According to the results of the following article, this intervention will increase the awareness and usage of the m-money systems.
Aker and Wilson (2013) attempt to address some of the barriers to adoption of mobile money systems in Ghana. They chose northern Ghana for some primary reasons. Ghana has a relatively stable economy and a number of formal financial institutions, but the access to formal credit and savings accounts is still limited, especially in rural areas. There is also a long history of informal savings in Ghana, for example through rotating village savings clubs.
Classified as a lower-middle-income country, 30 percent of the population of Ghana lives on less than US$1 per day and 54 percent on less than US$2 per day. Only 29 percent of the population has access to a formal financial institution (World Bank, 2012). To cope with shocks, diversify income and invest, rural households seek other ways than formal financial services, since these services are often unreachable for them. Migration is a very common strategy to do this. More than half of the households have at least one migrant member, typically searching for employment in larger urban centers. They send their earned money back to their family via the postal service MoneyGram, or by bus or via (traveling) friends. Using these informal money transfers bears enormous risks, since money can easily be stolen or get lost in the mail.
For these poor households, who mostly are not included in financial mechanisms, mobile money systems are a potential solution to provide some form of financial system. The monetary and security costs associated with money transfers can be reduced by m-money, which allows households to send and receive money at any time and thereby improves households’ ability to share risk. Additional to these money transfers, users of m-money can create pseudo-savings accounts, where they can deposit smaller saving amounts for immediate needs. The m-money savings channel offers greater security than, for example, savings under a mattress, since it is protected with a password, and it increases access to the saved money, compared to the annual ‘share out’ of savings clubs. M-money could also stimulate individuals to save for particular objectives.
With the rapid growth of mobile phone penetration, mobile financial services have been introduced into Ghana. Five mobile phone operators, including MTN, Vodafone, Tigo, Glo, and Zain (now Airtel), all started offering some type of m-money services. In addition, the Central Bank of Ghana introduced a biometric
bankcard – e-zwich – that can be used at ATMs, in an effort to increase financial inclusion by enabling access to illiterate individuals.
Figure 1 - Five mobile phone operators in Ghana
Despite the supply of m-money services, the adoption and usage of these services has been relatively low. Estimates in the research of Aker and Wilson (2013) show that less than 2 percent of the population has used m-money, which is in sharp contrast to the success of m-money systems in other lower-middle-income countries like Kenya, South Africa, Brazil and the Philippines. They found that the interest in adopting m-money services by the rural population was very high. The main reasons for the low adoption rates were the significant delays in the activation of the m-money service, the insufficient time frame between the initial registration and the follow-up visits to start using the product, and the limited demand for using the service after registration, due to either difficulties in using the service or the lack of trust in its validity.
Boadi et al. (2007) give some insights into m-commerce adoption in Ghana. M-commerce is relatively redefining the business processes of rural businesses of, for example, farmers and fishermen. With the rapid growth of mobile phone usage in Ghana, they gained opportunities of avoiding intermediaries or bypassing of ‘middlemen’, and trading directly with buyers in the marketplace. Not only could they avoid the intermediaries, they were also able to make themselves more visible to the market and target buyers, strengthen their business relationships and create new ones, and get closer ties with their buyers. These possibilities lead to increased
efficiency in rural markets in Ghana, and m-commerce is therefore likely to contribute to economic growth.
One m-health project in Ghana is the mPedigree Network, which uses mobile phones to secure products against faking, counterfeiting and diversion. It partners the principal telecom operators, Fortune 500 technology companies, and leading pharmaceutical industry associations. By doing this, mPedigree offers service to the world’s leading pharmaceutical and consumable companies with the shared goal of protecting consumers from the fatal effects of pharmaceutical counterfeiting, which kills nearly a million people every year in vulnerable parts of the world (mpedigree.net, 2013).
The paper of Frempong (2009) gives an example of mobile phone opportunities for micro- and small enterprises (MSEs) in Ghana. MSEs are the main economic actors, besides those working in agriculture, in the less urban and rural areas in Ghana. They contribute pronouncedly to employment, wealth creation, and poverty reduction. It is estimated that nearly 40 percent of Ghana’s gross national income is attributed to MSEs. Therefore, these enterprises can easily trigger growth in the economy due to their numbers and the niches they occupy in the national economy.
Figure 2 - Mobile phone impact model (source: Frempong, Godfred (2009), Mobile telephone opportunities: the
Mobile phones have positive impacts on the critical pillars of doing business. Using mobile phones reduces costs of doing business and it facilitates access to the market, as well as access to business information. Mobile phones can also be used for financial services such as m-money systems. All this together contributes to the competitiveness of the MSEs in the market. This higher competitiveness leads to higher efficiencies of markets in the economy, and therefore contributes to the overall economic development.
2.5 Telecommunications infrastructure: from state monopoly to liberalization
Given the evidence in the previous sub-sections it is clear that telecommunications infrastructure and mobile phones in particular create potential benefits for poor economies. This section provides some information about how these economies recognized this potential and changed their policies to promote it.
Telecommunication activities have become increasingly integrated into the operations of companies, government departments and many organizations, as well as in the economic and social behavior of individuals and firms. This led to a growing recognition that productivity of the entire economy in a country depends upon the efficiency of the telecommunications system. Therefore, by the late 1980s, most African countries began to reform their telecommunications’ sectors, through structural adjustment programs. With the pressure of the World Bank and other international organizations, these developments led to the privatization of state-owned telecommunications institutions in Africa.
Djiofack-Zebaze and Keck (2008) examined the impact of the telecommunications liberalization in Africa on sectoral performance and economic growth. Regulatory quality plays a major role in bringing down prices and improving access to telecommunications services in Africa. Competition, notably in the mobile segment, also improves sector performance. They first tested for a liberalization indicator, the number of operators in the mobile phone market in Africa, which proved to be insignificant. If instead an indicator of actual performance, the number of mobile phone subscribers, is substituted for it, the results show a positive impact on real GDP per capita: a 1% increase in the number of mobile phone subscribers
translates into a 0.6% higher level of real GDP per capita. The conclusion they make from this is that regulation/liberalization policies have an indirect effect on real GDP developments by contributing to an increase in the number of subscribers. Starting in 1995, the liberalization of the telecommunications sector in Africa led to a drop of the share of African countries maintaining a state monopoly in the mobile segment from 70% to less than 10%. Liberalization in the fixed-line segment has been somewhat slower, but while all fixed operators were state monopolies in 1995, by 2004 this was only the case in 44% of African countries.
This liberalization of telecommunications also took place in Ghana (Overå, 2006). In 1950, Ghana had one of the highest teledensities in Africa of 0.3 fixed line telephones per 100 habitants, but stagnated at this level for the next 40 years due to inefficiency and poor services of the incumbent telephone provider Ghana Post and Telecommunications Corporation (PTC). In 1995, PTC was privatized and became Ghana Telecom (GT). This led to an increase in the number of fixed line telephones to 1.4 per 100 habitants in 2003. More importantly, it led to a phenomenal growth in the number of mobile phone subscriptions. In 2003, four cell phone companies together had 600.000 subscribers and the total mobile phone penetration was 3.9 per 100 habitants, three times as much as the landline penetration (World Bank).
With the privatization of the telecommunications sector, competition was introduced and this contributed significantly to the growth of the telecommunications sector as well as to the structure of the African economies (Djiofack-Zebaze and Keck, 2008).
2.6 Investment
This sub-section gives some information how telecommunications infrastructure attracts foreign direct investment and how investment in telecommunication creates more potential for economic growth.
There is a growing body of research suggesting that telecommunications infrastructure is an essential tool for economic regeneration as they have a significant impact on economic growth and lead to an increase in foreign direct investment. Investment in telecommunications is regarded as one of the most
strategic industries with a strong potential to improve overall productivity (Datta and Agarwal, 2004). Investments in telecommunications have economic returns that are much higher than the return on investment alone, since there are direct and indirect effects on production (Batuo, 2008). The results of Batuo (2008) also indicate that telecommunications investment is subject to diminishing returns, suggesting that countries at an earlier stage of development are more likely to gain from investing in telecommunications infrastructure. Moreover, telecommunications investment contributes strongly to the increase of overall investment in a country and to productivity growth as these investments lead to cost minimization in the use of the country’s factor endowments, improving organizational functioning, and enhancing the spread of the development of factor and product markets (Röller and Waverman, 2001). A study by Becchetti et al. (2003) also confirms the hypothesis that investment in telecommunications positively affects the introduction of new processes and products.
Additional to this increased productivity and efficiency, investment in telecommunications enhances demand for differentiated products and services for firms, for example, through the use of call centers to intensify contacts between producers and consumers. It also allows the reduction of production lags and information asymmetries among sub-contractors and component producers at different levels of the value chain (Overå, 2006). As this also reduces lags between knowledge of consumer tastes and final production, the development of the telecommunications industry creates an opportunity for firms to change their supply to meet market demand (Becchetti et al., 2003).
2.7 From mobile phone to economic growth
In the previous sub-sections it became clear that mobile phones have the potential to improve consumer and producer welfare in developing countries. This sub-section gives a broader perspective of how mobile telecommunication can contribute to economic growth as a whole. The effect of mobile phones on changes in GDP and growth, especially in sub-Saharan Africa, is still relatively unexamined.
Many empirical studies in developed countries have confirmed the existence of a clear positive correlation between telecommunications and economic growth, and most of these studies have, in general, shown that investment in telecommunications infrastructure is one of the significant factors in the economic growth of a country. Empirical literature involving developing countries is scarce, and results from the literature that does exist are ambiguous.
Röller and Waverman (2001) assessed the impact of telecommunications infrastructure on economic development in 21 OECD countries. They found that a 10 percent increase in the telecommunications penetration rate increased economic growth by 1.5 percent.
Waverman et al. (2005) conducted a similar analysis in developing countries, finding that a 10 percent increase in mobile phone penetration levels was associated with a 0.6 percent increase in growth rates. Additional to this result, they found that mobile phones in less-developed economies are playing the same crucial role that fixed-line telephones played in the richer economies in the 1970s and 1980s. Mobile phones substitute for fixed lines in poor countries, but complement fixed lines in rich countries, implying a strong growth impact in less-developed countries. They also attempt to address concerns about finding credible exogenous instruments for mobile phone penetration by using lagged landline penetration as an instrument for current mobile penetration. Yet the same (unobservable) factors that caused certain countries to have high lagged landline penetration could also drive current mobile adoption and growth. This raises questions about the validity of the instrument and hence the direction of causality.
Shiu and Lam (2008) investigated the causal relationship between telecommunications development and economic growth in both developed and developing regions of China. The results show that there is a unidirectional relationship from real GDP to telecommunications development at the national level. Bi-directional causality, so causality also running from telecommunications to real GDP growth, is found only in the more developed provinces in the eastern region, but not in the low-income central and western provinces. These results imply that an improvement in telecommunications infrastructure alone is not sufficient for stimulating growth in these provinces. It is equally important for the Chinese
government to develop and enhance other complementary factors like business environments, transportation networks, education and manpower training in order to make the best use of telecommunications systems in the western and central regions.
Shirdhar and Shirdhar (2007) investigated for both developed and developing countries the relationship between telephone penetration and economic growth. They found that both mobile and fixed-line telephones have a positive impact on national output. They also found that this impact on total output is positive and significantly higher for developing countries than in the OECD countries.
Virta et al. (2010) examined the role of mobile telephone penetration in economic growth in a sample of developing economies and found that mobile telephony facilitates economic growth. This is suggesting that promoting mobile diffusion in developing countries is beneficial for their economies.
Chavula (2013) conducted a study that tries to assess the impact of telecommunications penetration on people’s living standards in Africa through its impact on per capita economic growth. First, a regression was conducted between telecommunications penetration and economic growth among African countries without classifying the countries in income groups. It was found that both fixed-line and mobile phone penetration have a positive impact at the 1% significance level on economic growth. A 1% increase in fixed-line telephone subscribers per 100 people increases per capita GDP growth with 0.15%, and a 1% increase in mobile telephone subscribers per 100 people increases per capita GDP growth with 0.22%.
After that, the sample was divided into three income groups, into upper-middle-income, lower-middle-income and low-income countries, and another regression was conducted. The results of this regression show that fixed-line telephone penetration has a highly significant impact on economic growth in the upper-middle-income countries, but insignificant in the lower-middle-income and low-income countries. Mobile telephony, however, has a highly significant impact on economic growth in all the three groups. A 1% increase in mobile telephone subscribers per 100 people increases per capita GDP by 0.39%, 0.26% and 0.15% for the upper-middle-, lower-middle-, and low-income countries, respectively.
Chavula (2013) also stated that in order for economic development to occur, complementarity between mobile phones and other forms of capital, such as electricity, paved roads and water, is needed. The challenge is now to ensure complementary access to public goods and the development of appropriate policies to evaluate and propagate the benefits of mobile phones throughout the continent.
Datta and Agarwal (2004) examine the long run relationship between telecommunications infrastructure and economic growth, using data from 22 OECD countries. There are two major issues to focus on when analyzing cross-country economic growth. The first is about whether per capita GDP growth rates across countries converge in the long run. The convergence hypothesis of neoclassical growth theory suggests that due to diminishing returns to capital, the growth rate of a country is inversely proportional to its initial level of income. This implies a tendency on the part of countries to converge to a common steady state rate of growth. The second issue concerns the major sources of economic growth, which has been tried to explain with various macroeconomic indicators. One of these indicators is investment in public infrastructure, from which telecommunications infrastructure investment is identified as one with a strong potential to improve overall productivity.
Indirect social returns of an expanding telecommunications sector are far outweighing the more obvious direct positive effects, such as increased employment and demand. Both direct and indirect effects are included in the five potential mechanisms through which mobile phones can provide economic benefits, as stated by Aker and Mbiti (2010). One of these indirect effects is the possibility for firms to adopt flexible structures and locations, contributing to the evolution of complex and large organizations that are geographically dispersed. This allows firms to fully exploit comparative advantages and therefore creates productivity gains.
There is much empirical evidence suggesting and confirming the relationship between telecommunications investment and growth, but only a few studies include the role of telecommunications infrastructure within a macroeconomic growth model. Madden and Savage (1998) analyzed economic growth in transition economies and found that telecommunication infrastructure plays a significant role in the development of these economies. Datta and Agarwal (2004) examine the
determinants of economic growth, including telecommunications infrastructure, using a panel data approach. In the study of Röller and Waverman (2001) it was indicated that the importance of telecommunications in explaining productivity is too large to be true when fixed effects are ignored. Therefore, Datta and Agarwal specify a dynamic fixed effects panel data model, in which the correlation between previous and subsequent values of growth is corrected for. They do this by using the lagged value of the dependent variable – growth – as an instrument in the regression.
They test for causality of the teledensity variable by including lagged values of this variable. The coefficient of the lagged value of the teledensity variable is expected to be insignificant if causation runs only from GDP growth to telecommunications, and not the other way around. When a significant coefficient for the lagged value of the teledensity variable is found, this would confirm a bi-directional causality between telecommunications infrastructure and GDP growth. Their results of including the lagged values of the teledensity variable significantly show that the positive relationship between telecommunications infrastructure and GDP growth is not merely due to reverse causality, meaning that bi-directionality is confirmed.
Lee et al. (2012) investigate the different effects that mobile and landline phones may have on economic development, accounting for the possibility that causation may run in both ways. Previous studies show evidence that telecommunications infrastructure has a significant positive causal effect on economic growth, but most of them used data on telecommunications infrastructure as a whole. However, not so much research is done about what role mobile telecommunications specifically have played in economic growth. This could be particularly interesting in a region where the penetration of mobile telephony far exceeds the existence of fixed-line telephones, like in sub-Saharan Africa. They differ in research method from existing literature in several ways. First, they focus on sub-Saharan countries, where the mobile phone penetration is much higher than the landline telephone penetration, and they search for any causal links between these two types of phones and economic growth. Second, they control separately for mobile phones and fixed-line telephones, and therefore try to determine the extent
to which mobile phones have a more pronounced effect on economic growth when main telephone line penetration is low.
They closely follow the cross-country macroeconomic growth model of Datta and Agarwal (2004). The lagged value of the dependent variable – growth rate of per capita GDP – is controlled for, assuming a dynamic process in which the current value of the dependent variable may be influenced by past ones. They also include the lagged value of GDP per capita, as a standard measure to test for convergence. Besides including variables for both mobile phone penetration and fixed-line telephone penetration, they also include an interaction term of these two, in order to allow the marginal impact of cellular phones to vary with the level of landlines that are already in place. Since the mobile phone penetration far exceeds the penetration of landlines, they expect the coefficient of this interaction term to be negative.
They also run a regression including a combined term of the two telecommunications variables, but the results of this coefficient were much smaller and statistically insignificant. This could be explained by the idea that cellular phones may not be just substitutes for landline phones, since mobile phones provide more services than simply the voice calling service of landline telephones. Therefore, controlling separately for mobile phone and landline phone penetration may help assess their economic impacts more accurately than combining the two variables as one.
They used a panel data for the period of 1975 to 2006, but because the mobile phone service expansion became mostly relevant in the 2000s, they also estimated the growth model for the period 2000-2006. For the period of 1975 to 2006 they found a positive and significant coefficient for the landline penetration variable. The coefficient for the mobile phone penetration variable was also positive, but not significant. This small estimated impact on the economic growth of cellular phones and its statistical insignificance are very likely to result from the fact that the regression was fitted over the entire sample period, while a substantial increase in the penetration of mobile phones only occurred towards the end of the observation period. When this regression was estimated for the smaller time period of 2000-2006, the estimated coefficient for the landline penetration is smaller than in the
larger time period, and no longer significant. The estimated coefficient for the mobile phone penetration variable is higher now, and statistically significant. This finding is consistent with a trend that has reshaped telecommunications in sub-Saharan Africa.
The estimated coefficient of the interaction term is negative in both regressions, and statistical significant in the shorter time period. This implies that the marginal impact of cellphone penetration on the growth rate of GDP per capita is relatively large when the level of landline telephone penetration is relatively low. In the regression of the shorter time period, the marginal effect of mobile phone penetration on GDP per capita growth is:
𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕
𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕100 = 0.0191 − 0.0001 ∙ 𝜕𝜕𝐿𝐿𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕100
It can be inferred that an increase in the spread of mobile phones yields a higher growth rate for countries where landline phones are rare.
The estimated coefficient of the lagged value of GDP per capita growth is negative, and therefore supports the convergence hypothesis that GDP per capita tends to growth at a slower rate in countries with a higher level of per capita GDP.
Following the results of Chavula (2013), for Ghana, being classified as a lower-middle-income country, it could be expected that there is a positive causal relationship between mobile phone subscriptions per 100 habitants and per capita GDP growth of a similar gratitude, so around 0.26%. This expectation would be in line with the results from the study of Lee et al. (2012), which also show a clear positive correlation between mobile phone penetration and per capita GDP growth. Therefore, my hypothesis is as follows: The introduction of mobile telephony and the rapid growth of the mobile phone penetration increased the economic performance of Ghana.
3. Data and empirical model
In order to determine whether mobile phones have any impact on economic growth in Ghana, some information is needed. As seen in the previous section, the telecommunications industry in Ghana started with liberalization around 1995. To cover this change in the industry and the expansion of the networks of both fixed-line telephones as mobile phones that followed from this liberalization, I will use data from the period 1992-2012.
During this period, population has grown from 15 million to 25 million, GDP per capita increased from US$415 to US$1600, and the completion rate of primary school increased from two-thirds to over 90 per cent. The number of fixed telephone lines per 100 people is now 1.1 instead of 0.3 in 1992, with a higher level of around 1.4 between 2003 and 2007. Moreover, the number of mobile phones per 100 people changed from 0 in 1992 to 100 in 2012 (World Bank, 2013).
All the data I use is from the World Development Indicators databank, which is the primary World Bank collection of development indicators. This databank includes the most current and accurate global development indicators and is therefore perfectly useful for my research.
As I will investigate the effect of mobile phone introduction on economic development in Ghana, I will only use data available for Ghana. The data is on a yearly basis, which means that I will have 21 observations for every variable.
The dependent variable of interest is the growth rate of per capita Gross Domestic Product, labeled as GDPPCGR. The control variables are:
- GDP per capita (PPP, in current international $), labeled as GDPPC - Population annual growth rate, labeled as POPGR
- General government final consumption expenditure as percentage of GDP, labeled as GOVCON
- Trade as percentage of GDP, which is a proxy for the degree of openness in the global economy, labeled as TRADE/GDP
- Number of mobile cellular subscriptions per 100 people, labeled as MOBPHONES
Table 1 - Descriptive statistics
Variable Mean St. Dev. Minimum Maximum
GDPPCGR 3.078 2.60 0.569 12.422 GDPPC 1159.279 373.79 728.452 2013.974 POPGR 2.488 0.17 2.175 2.810 GOVCON 12.053 1.81 9.722 16.643 TRADE/GDP 81.161 19.26 45.994 116.048 MOBPHONES 21.740 32.29 0.003 100.284 LANDPHONES 0.972 0.46 0.302 1.658
To determine the effect of mobile phones on economic performance in Ghana, I follow the macroeconomic growth models of Datta and Agarwal (2004) and Lee et al. (2012), leading to the following framework:
𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡 = 𝛼𝛼𝑡𝑡+ 𝛽𝛽1𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡−1+ 𝛽𝛽2𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡−1+ 𝛽𝛽3𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡
+ 𝛽𝛽4𝜕𝜕𝜕𝜕𝐺𝐺𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝛽𝛽5𝑇𝑇𝜕𝜕𝐿𝐿𝜕𝜕𝜕𝜕/𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝛽𝛽6𝑀𝑀𝜕𝜕𝑀𝑀𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡
+ 𝛽𝛽7𝜕𝜕𝐿𝐿𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝛽𝛽8𝑀𝑀𝜕𝜕𝑀𝑀𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡× 𝜕𝜕𝐿𝐿𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝜀𝜀𝑡𝑡
The lagged value of GDPPCGR is included to control for the dynamic process in which the current value of the dependent variable may be influenced by past ones. The lagged value of GDPPC is included as a standard test for convergence. It measures the effect of past levels of per capita GDP on subsequent growth. As already seen in the related literature, the convergence hypothesis of neoclassical growth theory suggests that due to diminishing returns to capital, the growth rate of a country is inversely proportional to its initial level of income. The POPGR variable is included and expected to have a negative sign, as a higher population growth relates to lower GDP per capita. To estimate the effect of government consumption expenditure as a share of GDP on economic growth, GOVCON is included in this equation. TRADE/GDP measures the extent to which the country is integrated into the global economy and is expected to have a positive impact on economic growth. MOBPHONES and LANDPHONES are the numbers of subscriptions per 100 people and are expected to
be positively correlated with economic growth. An interaction term is included between MOBPHONES and LANDPHONES to allow the marginal impact of mobile phones to vary with the level of landline phones. Since the number of mobile cellular subscriptions far exceeded the number of telephone lines in Ghana, the coefficient of this interaction term is expected to be negative. A negative coefficient would imply that the impact of mobile telecommunications on economic growth is more pronounced when the penetration of main telephone lines is relatively low.
Since the mobile phone penetration started to growth exponentially from the year 2001, I will also run a regression over a shorter time period, from 2001 to 2012. Following the results from Lee et al. (2012), this would give a more accurate estimate of the impact of mobile phones.
4. Results en discussion
4.1 Main results
In this section I will show the results from the regressions performed on the macroeconomic growth model stated in the previous section. The regression results are presented in Table 2.
The coefficient of the lagged value of the dependent variable, GDPPCGR, is negative but statistically insignificant. Since this variable was only included to control for the dynamic process in which past values of the dependent variable may influence the current ones, no further discussion about these estimates is needed.
For both Regression (1) and Regression (2), the estimated coefficient for the lagged value of GDPPC is negative, meaning that with a higher level of initial per capita GDP, the growth rate of current per capita GDP is lower. This confirms the convergence hypothesis of neoclassical growth theory, stating that a country tends to grow at a higher rate when the initial level of national income is low.
As expected, the estimated coefficients of POPGR are also negative, and significant at the 1% level in the first regression. This confirms the theory that a higher population growth relates to lower GDP per capita, since with more people the total level of GDP has to be shared by more. In my regression however, the estimation results indicate the effects on the growth rate of per capita GDP, instead
of on per capita GDP itself. This explains the higher absolute values of the estimated coefficients (compared to the other coefficients): in Regression (1), the absolute value of this estimated coefficient is |𝛽𝛽3|=9.8508, and in Regression (2) this is
|𝛽𝛽3|=36.1997. With these higher absolute values of the estimated coefficient of
POPGR, the effect of population growth is not only to slow growth down, but it also might make it negative. This way, the direct effect of population growth on per capita GDP indirectly relates to a negative growth rate when population growth is high.
Table 2 - Estimation results
VARIABLE Regression (1) 1992 to 2012 Regression (2) 2001 to 2012 GDPPCGRt-1 -.2001 (-0.54) (-0.55) -.4672 GDPPCt-1 -.0304 (-1.25) (-0.19) -.0124 POPGR -9.8508*** (-2.90) -36.1997 (-1.18) GOVCON .9131*** (4.12) 1.2529* (1.99) TRADE/GDP -.0672 (-1.66) (-1.34) -.2126 MOBPHONES .6126* (1.77) (0.58) .7965 LANDPHONES 11.5823* (1.76) 23.6046 (0.68) MOBPHONES*LANDPHONES -.2912* (-1.95) (-0.82) -.6326 Number of observations 20 11 R2 0.8611 0.8885 Adjusted R2 0.7602 0.4425 F-statistic 8.53 1.99
Notes: t-values in parentheses; * indicates significance at the 10% level; ** indicates significance at the 5% level; *** indicates significance at the 1% level.
The coefficients of GOVCON are positive, and statistically significant at the 1% level in Regression (1) and at the 10% level in Regression (2), meaning that when
the general government increases final consumption expenditures as a percentage of GDP, this has a positive effect on economic development. In Regression (1), with a 1% increase in governmental consumption expenditures the growth rate of per capita GDP increases with 0.91%. In Regression (2) this is 1.25%.
TRADE/GDP was included as a proxy for the openness of the country in the global economy and the estimated coefficients were expected to be positive, since a more open economy almost always grows faster. In both Regression (1) and (2), the estimation results are negative, which would suggest that the more open Ghana is to the global economy, the lower its economic growth is. However, the estimation results have low absolute values and are not statistically significant at all, so there is not much to conclude from them.
In both the first and the second regression, the estimation results for the mobile phone penetration rate are positive, and in the first regression this is significant at the 10% level. In order to discuss the impact of mobile phone penetration on economic growth, I also have to look at the estimation coefficients of the included interaction term. By combining these estimates, the marginal impact of mobile phones on per capita GDP growth can be determined, as seen in the study of Lee et al. (2012). For Regression (1), the marginal impact of mobile phones on per capita GDP growth is:
𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕
𝜕𝜕𝑀𝑀𝜕𝜕𝑀𝑀𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕 = 0.6126 − 0.2912 ∙ 𝜕𝜕𝐿𝐿𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕 For Regression (2) this is:
𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕
𝜕𝜕𝑀𝑀𝜕𝜕𝑀𝑀𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕 = 0.7965 − 0.6326 ∙ 𝜕𝜕𝐿𝐿𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕
As expected, the marginal impact of mobile phones is higher in the period 2001-2012, in which the number of mobile phones started to grow exponentially and where the mobile phone penetration far exceeded the main telephone line penetration. Also, the coefficient of the interaction term is bigger in Regression (2), suggesting that the impact of mobile telecommunications is indeed more pronounced when the penetration of main telephone lines is relatively low.
The estimated coefficients for landline phone penetration are also positive, and significant at the 10% level in Regression (1). These estimates are higher than
those of mobile phone penetration, which suggests that the impact of main telephone lines on economic growth would be higher than the impact of mobile phones. However, not much investment is made in main telephone line networks due to the high costs. Therefore, despite this potential great positive impact, the number of main telephone lines is very low. Following the results of Regression (1), a 1% increase in the number of main telephone lines is related to a 11.58% increase in GDPPCGR, and in Regression (2) this is 23.6%. Compared to the results of the study of Lee et al. (2012), this is the other way around. They found that in the estimation for the cropped data period, the impact of landline phone penetration is lower than in the estimation for the entire observation period. They did, however, find significant results in the first regression, which were no longer significant in the cropped data regression, just like in my growth model.
The estimated coefficient for the interaction term is negative, as expected, and significant at the 10% level in Regression (1). The estimated coefficient is greater and has, therefore, more economic impact in the estimation for the shorter time period, but it is no longer statistically significant.
4.2 Robustness-checks
In order to test my estimation results on robustness, I performed more regressions on the growth model with some changes in it.
The first change I made was leaving out the lagged value of per capita GDP. This leads to the following growth model:
𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡 = 𝛼𝛼𝑡𝑡+ 𝛽𝛽1𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡−1+ 𝛽𝛽2𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝛽𝛽3𝜕𝜕𝜕𝜕𝐺𝐺𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡
+ 𝛽𝛽4𝑇𝑇𝜕𝜕𝐿𝐿𝜕𝜕𝜕𝜕/𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝛽𝛽5𝑀𝑀𝜕𝜕𝑀𝑀𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝛽𝛽6𝑀𝑀𝜕𝜕𝑀𝑀𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡
× 𝜕𝜕𝐿𝐿𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝜀𝜀𝑡𝑡
Performing a regression on this model gives some interesting results (see Table 3 in Appendix). In Regression (1), both the R-squared and the adjusted R-squared drop only a little, from 0.8611 to 0.8416 and from 0.7602 to 0.7491 respectively. This suggests that the overall regression still fits the model. Besides that, the estimated coefficient of mobile phone penetration is lower than in the main growth model estimation, but this time it is even more significant (1% level). The estimated
coefficient for landline phone penetration is also lower, but also more significant (5% level). In Regression (2), the R-squared drops even less, from 0.8885 to 0.8864, but the adjusted R-squared is much higher than in the main growth model estimation. Since the adjusted R-squared is a modified version of R-squared that has been adjusted for the number of predictors in the model, it increases only if the extra predictor improves the model more than would be expected by chance. According to this and the higher value of adjusted R-squared in the estimation of the growth model where the lagged GDPPC is not included, it seems that the model is better without the lagged value of per capita GDP. However, as this was not the aim of my thesis, I will leave this result here in the discussion instead of changing my initial growth model, and continue with robustness checks on the initial model.
To test whether a drop of one variable changes the estimation results much, I leave out the proxy for openness of the country, the TRADE/GDP variable. This leaves the following adjusted growth model:
𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡 = 𝛼𝛼𝑡𝑡+ 𝛽𝛽1𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡−1+ 𝛽𝛽2𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡−1+ 𝛽𝛽3𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡
+ 𝛽𝛽4𝜕𝜕𝜕𝜕𝐺𝐺𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝛽𝛽5𝑀𝑀𝜕𝜕𝑀𝑀𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝛽𝛽6𝑀𝑀𝜕𝜕𝑀𝑀𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡
× 𝜕𝜕𝐿𝐿𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝜕𝑡𝑡+ 𝜀𝜀𝑡𝑡
The estimation results of this regression can be found in Table 4 in the Appendix. For Regression (1), the overall fit of the model doesn’t differ much from the initial model with the TRADE/GDP term included, since the R-squared and adjusted R-squared dropped only a little, from 0.8611 to 0.8264 and from 0.7602 to 0.7252 respectively. The higher value of the adjusted R-squared in the initial model suggests that that model is better than this adjusted model. All the values of the estimates are somewhat lower, but most of these changes are statistically insignificant. However, the estimation results for mobile phone and landline phone penetration and the interaction term are no longer significant. The same logic could be addressed for Regression (2), where the R-squared and adjusted R-squared are both lower than in the first model. Altogether, leaving out the TRADE/GDP variable does not change the overall estimation of the model much, indicating that my estimation results are robust.
Another check on robustness is performing a regression on the model with leaving out or adding the data of randomly chosen years. I did two more regressions
for the entire observation period, one leaving out the years 1997 and 2006, and another one leaving out the years 2000 and 2003. I also did one more regression for the cropped data period, adding the year 2000 to it. The estimation results are shown in Table 5 in the Appendix. According to these results, leaving out random years doesn’t change the estimates much, so robustness may be assumed. Adding data from the year 2000 to the cropped data regression lowers the economic impact of both mobile phone and main telephone penetration on economic growth, but these impacts are still positive and still not significant.
5. Conclusion
This thesis investigates the impact of mobile telephony on economic growth in Ghana. Related literature and existing empirical evidence suggest that telecommunications infrastructure and mobile phones in particular can contribute to economic development in developing countries. Therefore, a positive correlation between mobile telephony and economic growth is expected for Ghana, as this is a lower-middle-income developing economy. A macroeconomic growth model is used to investigate the relationship between mobile phone penetration rates and the growth rate of per capita GDP. Other variables are included, such as the population growth rate and a proxy for the degree of openness to trade, to control for fixed effects. The results show that mobile phone penetration is positively correlated with per capita GDP growth, as is the penetration rate of main telephone lines. These estimation results are significant at the 10% level in the estimation for the entire observation period. However, they are not significant in the estimation for the cropped data period. Moreover, the estimates of the interaction term show that the marginal impact of mobile phones is greater when main telephone line penetration is low.
Although the results are only statistically significant at the 10% level in Regression (1), the estimation results show clear positive numbers of the mobile phone variable on the dependent variable, growth. Therefore, it can be concluded that the introduction of mobile telephony in Ghana results in a better economic development of the country.
Reference list
Aker, Jenny C. (2008), Does digital divide or provide? The impact of mobile phones on
grain markets in Niger, BREAD Working paper 177
Aker, Jenny C. (2010), Information from Markets Near and Far: Mobile Phones and
Agricultural markets in Niger, American Economic Journal: Applied Economics
2, pp. 46-59
Aker, Jenny C., Christopher Ksoll and Travis J. Lybert (2010), ABC 123: Can you text
me now? The impact of a mobile phone literacy program on educational outcomes, unpublished paper, Tufts University
Aker, Jenny C. and Isaac M. Mbiti (2010), Mobile Phones and Economic Development
in Africa, Journal of Economic Perspectives, 24(3), pp. 207–232
Aker, Jenny C., and Kim Wilson (2013), Can mobile money be used to promote
savings? Evidence from preliminary research Northern Ghana, Swift Institute
Working Paper, 2012(003)
Batuo, Michael Enowbi (2008), The role of telecommunications infrastructure in the
regional economic growth of Africa, MPRA Paper nr. 12431
Becchetti, L., David Andres Londono Bedoya, and L. Paganetto (2003), ICT
investment, productivity and efficiency: Evidence at firm level using a stochastic frontier approach, Journal of Productivity Analysis, 20, pp. 143-167
Beuermann, D.W., C. McKelvey, and R. Vakis (2012), Mobile Phones and Economic
Development in Rural Peru, The Journal of Development Studies, 48(11), pp.
1617-1628
Boadi, Raymond A., Richard Boateng, Robert Hinson and Robert A. Opuku (2007),
Preliminary insights into M-commerce adoption in Ghana, Information
Development, 23(4), pp. 253-265
Burgess, Robin, and Rohini Pande (2005), Do rural banks matter? Evidence from the Indian Social Banking Experiment, The American Economic Review, 95(3), pp. 780-795
Chavula, Hopestone Kayiska (2013), Telecommunications development and economic
Datta, Anusua and Sumit Agarwal (2004), Telecommunications and economic growth:
a panel data approach, Applied economics, 36(15), pp. 1649-1654
Djiofack-Zebaze, C. and A. Keck (2008), Telecommunications Services in Africa: The
impact of WTO commitments and unilateral reform on sector performance and economic growth, World Development, 37(5), pp. 919-940
the Economist (2008), Halfway there: How to promote the spread of mobile phones
among the world’s poorest, http://www.economist.com/node/11465558, May
29th 2008
the Economist (2013), Why does Kenya lead the world in mobile money?, http://www.economist.com/blogs/economist-explains/2013/05/economist-explains-18, May 27th 2013
Frempong, Godfred (2009), Mobile telephone opportunities: the case of micro- and
small enterprises in Ghana, info, 11(2), pp 79-94
Jensen, R. (2007), The Digital Provide: Information (Technology), Market
Performance, and Welfare in the South Indian fisheries sector, Quarterly
Journal of Economics, 122(3), pp. 879–924
Klonner, Stefan, and Patrick Nolen (2008), Does ICT benefit the poor? Evidence from
South Africa, unpublished article, Essex University
Lee, Sang H., John Levendis, and Luis Gutierrez (2012), Telecommunications and
economic growth: an empirical analysis of sub-Saharan Africa, Applied
Economics, 44(4), pp. 461-469
Madden, Gary, and Scott J. Savage (2000), Telecommunications and economic
growth, International Journal of Social Economics, 27, pp. 893-906
Muto, M. and T. Yamano (2009), The Impact of Mobile Phone Expansion on Market
Participation: Panel Data Evidence from Uganda, World Development, 37(12),
pp.1887-1896
Overå, R. (2006), Networks, Distance, and Trust: Telecommunications Development
and Changing Trading Practices in Ghana, World Development, 34(7), pp.
1301-1315
Röller, L.H. and L. Waverman (2001), Telecommunications Infrastructure and
Economic Development: A Simultaneous Approach, American Economic
Shiu, Alice, and Pun-Lee Lam (2008), Causal relationship between
telecommunications and economic growth in China and its regions, Regional
Studies, 42(5), pp. 705-718
Shirdhar, K.S., and V. Shridhar (2007), Telecommunications and growth: Causal
model, quantitative and qualitative evidence, Economic and Political Weekly,
41(25), pp. 2611-2619
Waverman, Leonard, Meloria Meschi and Melvyn Fuss (2005), The impact of
telecoms on economic growth in developing countries, Vodafone Policy Paper
3, pp. 10-23
World Bank (2013), Africa Development Indicators 2012/13, Washington, DC: World Bank
Appendix
Table 3 - Estimation results without lagged GDPPC
VARIABLE Regression (1) 1992 to 2012 Regression (2) 2001 to 2012 GDPPCGRt-1 -.5864** (-2.82) (-1.58) -.6044 POPGR -8.1884** (-2.56) -39.1462 (-1.78) GOVCON .9190*** (4.06) 1.3029** (2.75) TRADE/GDP -.0353 (-1.10) (-1.69) -.2177 MOBPHONES .1891*** (2.97) (0.88) .5835 LANDPHONES 3.616** (2.33) 19.3338 (0.88) MOBPHONES*LANDPHONES -.1231* (-1.91) (-1.05) -.5459 Number of observations 20 11 R2 0.8416 0.8864 Adjusted R2 0.7491 0.6215 F-statistic 9.11 3.35
Notes: t-values in parentheses; * indicates significance at the 10% level; ** indicates significance at the 5% level; *** indicates significance at the 1% level.
Table 4 - Estimation results without TRADE/GDP
VARIABLE Regression (1) 1992 to 2012 Regression (2) 2001 to 2012 GDPPCGRt-1 -.5680* (-1.78) (-0.44) -.4235 GDPPCt-1 -.0049 (-0.24) (-0.38) -.0269 POPGR -5.6473** (-2.33) (-0.04) -.6774 GOVCON .7986*** (3.55) (1.31) .6735 MOBPHONES .2162 (0.81) (0.61) .9369
LANDPHONES 3.333 (0.72) 22.2679 (0.57) MOBPHONES*LANDPHONES -.0999 (-0.99) (-0.60) -.5208 Number of observations 20 11 R2 0.8264 0.7885 Adjusted R2 0.7252 0.2951 F-statistic 8.16 1.60
Notes: t-values in parentheses; * indicates significance at the 10% level; ** indicates significance at the 5% level; *** indicates significance at the 1% level.
Table 5 - Estimation results with gaps in years
VARIABLE Regression (1) 1992 to 2012, but gaps 1997 and 2006 Regression (1) 1992 to 2012, but gaps 2000 and 2003 Regression (2) 2000 to 2012 GDPPCGRt-1 -.4020 (-1.27) (-0.46) -.2103 (-0.66) -.4725 GDPPCt-1 -.0300 (-1.45) (-0.91) -.0275 (-0.10) -.0054 POPGR -10.2058*** (-3.10) -11.5423** (-2.34) -39.6163 (-1.58) GOVCON 1.0657*** (5.82) 1.0352*** (3.03) 1.2875** (2.43) TRADE/GDP -.0018 (-0.04) (-1.41) -.0960 (-1.56) -.2104 MOBPHONES .6790* (2.30) (1.39) .5900 (0.50) .4288 LANDPHONES 9.0393 (1.58) 11.4758 (1.42) 13.1757 (0.68) MOBPHONES*LAND PHONES -.3441** (-2.66) (-1.61) -.2958 (-0.93) -.3980 Number of observations 20 20 12 R2 0.9437 0.8599 0.8891 Adjusted R2 0.8794 0.6997 0.5935 F-statistic 14.67 5.37 3.01
Notes: t-values in parentheses; * indicates significance at the 10% level; ** indicates significance at the 5% level; *** indicates significance at the 1% level.
Figure 3 - Mobile and fixed-line telephone penetration growth for Ghana (source: World Bank, 2013)
Figure 4 - Correlation between GDPPCGR and Mobile phone penetration (source: Stata12)
0 5 10 15 0 20 40 60 80 100 MOBPHONES GDPPCGR Fitted values
