The Limited Impact of ICTs on Microenterprise Growth: A Study of Businesses Owned by Women in Urban India

The Limited Impact of ICTs on Microenterprise Growth CHEW, LEVY, ILAVARASAN Research Article The Limited Impact of ICTs on Microenterprise Growth: A Study of Businesses Owned by Women in Urban India1 Han Ei Chew chewhan@msu.edu Doctoral Student Department of Telecommunication Information Studies & Media Michigan State University 409 Communication Arts & Sciences East Lansing, MI 48824-1212 USA Mark Levy mlevy@msu.edu Professor Department of Telecommunication Information Studies & Media Michigan State University 402 Comm Arts & Sciences East Lansing, MI 48824-1212 USA Vigneswara Ilavarasan vignesh@hss.iitd.ac.in Assistant Professor Dept. of Humanities & Social Sciences Indian Institute of Technology Delhi Hauz Khas, New Delhi 110 016 India Abstract This article is presented as a response to the increasing need for rigorous impact assessment in ICT4D. The research reported here empirically examines whether ICTs enable microenterprise growth, to what degree, and under what conditions. We created two theoretical models that predict relationships between selected antecedents of ICT access, ICT use, and business growth. Using data collected through a multistage probability survey of women microentrepreneurs in Mumbai, India, we tested the models by structural equation modeling (SEM). Both models predicted a statistically signiªcant, but limited causal relationship between access to ICTs (as the independent variable) and business growth (as the dependent variable). The theoretical model and the analytical techniques suggest that future research should pay greater attention to the speciªc factors that mediate the impact of ICTs on the growth of very small businesses. Introduction Challenging the long-running claim that information and communication technologies (ICTs) enable development is now a central premise of ICT for development (ICT4D) scholarship. Indeed, numerous researchers have called for rigorous studies that examine the impact of increasingly diffused ICTs (Donner, 2008; Donner & Escobari, 2010; Duncombe, 2009; Heeks, 2010a; Heeks & Molla, 2009). The research reported here is a partial response to the need for studies that investigate whether, to what degree, and under what conditions ICTs enable microenterprises to grow.2 We chose a research site that we believed would be especially fruitful for probing the role of ICTs—namely, microenterprises owned by women in Mumbai, India. With data collected through a multistage probability design, we tested two models of ICT impact on business growth using structural equation modeling (SEM). Both models predicted a statistically signiªcant, causal relationship between access to ICTs (as an independent variable) and business growth (as the dependent variable). However, the set of independent variables investigated could explain only a small amount of the variance in business growth. Thus, these ªndings offer one 1. This research was carried out with the aid of a grant from the International Development Research Centre, Ottawa. 2. There is no consensus in the research literature about the deªnition of a “microenterprise” (Donner & Escobari, 2010). Indeed, in India the National Survey Sample Organization (NSSO, 2000) categorizes very small businesses by the number of workers, while the Indian Ministry of Micro, Small, & Medium Enterprises (2007) bases its deªnition on the value of the physical plant and machinery. © 2011 USC Annenberg School for Communication & Journalism. Published under Creative Commons Attribution-Non Commercial-Share Alike 3.0 Unported license. All rights not granted thereunder to the public are reserved to the publisher and may not be exercised without its express written permission. Volume 7, Number 4, Winter 2011, 1–16 1 THE LIMITED IMPACT OF ICTS ON MICROENTERPRISE GROWTH context-speciªc example of the limited impact of ICTs on microenterprise growth and thereby provide a more nuanced picture of ICTs as enablers of development. Literature Review ICTs and Economic Development From a macroeconomic perspective, ICTs have been shown to have positive consequences for the economies of developing nations. In the countries of Africa, for example, a 10% increase in the availability of mobile phones resulted in a 0.59% increase in per capita GDP (Waverman, Mesci, & Fuss, 2005), while across the Global South, a 10% increase in broadband penetration was projected to produce a 1.38% increase in per capita GDP (Qiang & Rossoto, 2009). At the micro level, there is little evidence about the impact of computers on microenterprises, largely because rates of computer and Internet diffusion are low, and the costs of computer use make it an economically unattractive option for most microentrepreneurs (Chinn & Fairlie, 2010). Mobile phones, on the other hand, do appear to effect microenterprises in ways that might be economically beneªcial. Two wide-ranging reviews of the literature (Donner, 2008; Donner & Escobari, 2010) conclude that mobile phones may facilitate the search for price information, reduce business-associated travel, and aid in communication with existing suppliers and customers, but evidence that mobile phones also make it possible for microentrepreneurs to expand their base of suppliers and customers—key components of economic growth—is mixed. The strongest evidence that mobile phone use by microentrepreneurs helps correct information asymmetries and creates more efªcient markets comes from Abraham (2006), Jensen (2007), and Aker (2008). These studies demonstrate how information acquired by participants in the ªshing industry in Kerala, India (Abraham and Jensen), or by grain traders in Niger (Aker), reduces producer risk, price variability, and cost to consumers. However, while Jensen and Aker report increased producer proªts, Aker found that economic beneªts did not accrue equally across the ªshing industry, with most gains going to the better-off in the supply chain. Urban Women Microentrepreneurs Because of their ubiquity and potential as a path toward alleviating poverty, microenterprises have 2 long been of interest to the development community (Mead & Liedholm, 1998; Nichter & Goldmark, 2005). Moreover, various estimates, both countryspeciªc and regional, suggest that women own upward of half of all microenterprises in the developing world (Chen, 2001; National Sample Survey Organization [NSSO], 2000; Peebles, 2006; Wasihun & Paul, 2010). However, most enterprises owned by women are classiªed as sole proprietorships and/or home-based businesses (Mead & Liedholm, 1998; Nichter & Goldmark, 2005). In India, for example, at one time 69% of urban microenterprises owned by women were home-based and had no hired workers (NSSO, 2000). Recent research has questioned the potential of sole proprietorships to increase revenues or create jobs (Gelb, Mengistae, Ramachandran, & Shah, 2009; International Labor Organization [ILO], 2009; LaPorta & Shleifer, 2008). Therefore, sole proprietorships and home-work microenterprises are excluded from this study. This approach, it is hoped, enriches our sample with greater numbers of microenterprises that are not “livelihood” or “survivalist” enterprises and that might exhibit ICTenabled growth (Duncombe & Heeks, 2005). Several factors make urban settings a strategic research site. First, much previous research on microenterprises has focused solely on rural development. While not ignoring the need for rural development, we contend that the increased urbanization of the developing world provides a largely overlooked locus in which ICTs might be successfully leveraged to support economic growth. Second, with regard to the availability of ICTs, it is clear that mobile phones and even the Internet have diffused earlier and with higher subsequent penetration in the urban areas of developing nations than they have in the countryside (Castells, Fernandez-Ardevol, Qiu, & Sey, 2007; Mariscal, 2005). Thus, access to ICTs, which is a necessary but not sufªcient condition for demonstrating ICT impact, is by and large met in urban settings. Third, compared to very small businesses in the countryside, urban microenterprises have an almost 25% greater chance of surviving beyond their ªrst year (Mead & Liedholm, 1998). Whether urban microenterprises remain in business after a year as a function of ICT use remains a largely unexplored phenomenon, as does the related, possibly recursive relationship between business growth and ICT use. Fourth, cities are generally home to large numbers of microenterprises—the target population of this Information Technologies & International Development CHEW, LEVY, ILAVARASAN research. According to the NSSO (2000), for example, the greatest number of “establishments” (in practice, microenterprises)—some 4.2 million—are found in the cities of India, a ªgure that is three times that of the number located in India’s vast rural areas. In Mumbai, for instance, the informal sector accounts for two-thirds of total employment (Srivastava, 2005). Assessing the Impact of ICTs A variety of theoretical and methodological strategies have been used to examine the effects of ICTs on development (for critical compilations, see Duncombe, 2009; Heeks, 2007; Heeks & Molla, 2009). Duncombe acknowledges the strengths and weaknesses of both qualitative and quantitative methods, but suggests that quantitative analysis is more likely to provide the more rigorous tests of causality required to demonstrate ICT impact. One such strategy is the “LISREL-facilitated approach” (Heeks, 2007). LISREL is a software program for structural equation modeling that is widely used (as is a closely related software known as SPSS AMOS) used in communication research (Kotz, Krishnan, & Wickersham, 2007) and by information systems researchers, especially for testing models of technological adoption (Lee, Kozar, & Larsen, 2003).3 Structural equation software allows researchers to model relationships between variables pictorially and thereby understand the relationships of model variables more clearly. In SEM, all variables in the proposed model are simultaneously tested to assess the degree to which they are collectively consistent with the data. Moreover, unlike regression models, SEM simultaneously takes into account collinear relationships between predictor variables. Finally, as Bentler (1980) observes, “If the model cannot be rejected statistically, it is a plausible representation of the causal structure” (p. 420). Method Sample Design Research for this study was carried out in Mumbai, India, in May and June 2009. As India’s largest city, Mumbai had the greatest number of Internetconnected computers in India at approximately 3.3 million (India Broadband Forum, 2008), the most mobile phones nationwide at 20 million (Telecom Regulatory Authority of India, 2009), and approximately 2,000 Internet cafés (“Cyber Cafés,” 2009). GNN Market Research Pvt. Ltd. of Mumbai conducted in-person interviews with women who owned microenterprises with between one and 19 hired workers. We set the upper boundary at 19 hired workers to try to capture “growth” enterprises—those microenterprises with potentially greater business communication needs that might “show a greater business focus and which deliver broader/longer term beneªts of competitiveness, innovation and exports” (Duncombe & Heeks, 2005, p. 5). Data were collected using a three-stage random cluster probability design. The ªrst stage of the sample consisted of 34 “investigative units” (IVs) chosen by circular systematic random sampling from the 900 IVs in the NSSO Urban Survey Frame for Mumbai city. The NSSO further subdivides every IV unit into a variable number of “blocks,” made up of one or two roughly square city blocks or similar areas circumscribed by other geographic features. All blocks (usually between 15 and 20) in each IV selected in stage 1 were numbered, and for stage 2 of the sample, we selected one block in each IV by simple random sampling. Each corner of every block chosen in stage 2 was numbered, moving clockwise around the block from a starting point at its uppermost left corner. In the ªnal stage of the sample, one corner of each block was selected by simple random sampling and became the starting point for interviews. Interviewers were instructed to seek eligible respondents at each starting point. A snowball approach was used to locate subsequent respondents. Within each of the sampled IV units, we established the following quotas: one business with 1–6 hired workers; four businesses with 6–9 hired workers each; one with 10–19 hired workers; and one from any of the three preceding categories. These quotas reºect the approximate distribution of 3. A search of the three top-ranked ICT4D journals (Heeks, 2010b)—Information Technologies & International Development, The Electronic Journal of Information Systems in Developing Countries, and Information Technology for Development—found only two papers that tested a model using SEM. We believe this is unfortunate. Short of difªcult-to-fund experimental designs or longitudinal studies, structural equation modeling offers ICT4D scholars an important method to probe causal impacts. Volume 7, Number 4, Winter 2011 3 THE LIMITED IMPACT OF ICTS ON MICROENTERPRISE GROWTH urban microenterprises by number of hired employees (NSSO, 2000). In six investigative units, interviewers were unable to locate eligible respondents within the IV boundaries and were instructed to move slightly outside the IV to complete the quotas. Since contiguous IVs shared similar social and economic characteristics, this procedure, we believe, had no serious implications for the integrity of the sample. Interviews generally required 30–40 minutes to complete. A cash incentive of Rs. 600 (approximately US$12 at the time and equal to four days’ wages for a semi-skilled urban worker in India) was offered to respondents. We obtained 231 completed interviews. Dividing the number of completed interviews by the number of encounters with eligible respondents yielded a completion rate of 67%. For the sake of convenience, a draft questionnaire was pilot tested on seven respondents in the Chandni Chowk area of New Delhi, a neighborhood that approximates Mumbai both in density and type of microenterprises. The revised questionnaire was prepared in English and translated into Hindi. The original English text and a back-translated version of the Hindi questionnaire were found to be in substantial agreement, indicating that a culturally and linguistically acceptable text in Hindi had been created. On the advice of the survey research ªrm, it was decided to use the Hindi version of the questionnaire, with the English version as a backup. Sample Characteristics Taking the microenterprise as the unit of analysis, the modal number of hired workers was six (range ⫽ 1–15). Consistent with the sample design, about one-third of the microenterprises (31.6%) had between one and ªve hired workers, 51.9% employed between six and nine paid workers, and 14.4% had 10 or more. More than three-quarters (77.5%) of microenterprises sampled had a business landline; 87.9% of the women who owned those very small businesses had at least one mobile phone. Only 10.4% had a computer in the workplace, and only half of those computers had Internet access. A majority of female-owned microenterprises (55.7%) were in the trading sector. As reported by respondents, typical businesses in trade included sellers of clothing, groceries, small electronics, dress materials, or jewelry. Some 42.6% of the micro- 4 enterprises were service sector businesses. Examples of services offered by respondents included medical care by general physicians, taxicabs and tours, academic tutoring, beauty parlors, and data entry. Only 1.7% of the microenterprises were in the manufacturing sector, including tailoring, dressmaking, or low-end leather work. The average microenterprise had been in business for a decade (mode ⫽ 10 years, range ⫽ 2 months–60 years), and only 5.6% had been started within a year of the survey. While 61.5% of microenterprise customers came from the “neighborhood,” more than one-third (38.5%) came from other parts of Mumbai, with 14.3% of sales resulting from phone calls. Of the 231 women microentrepreneurs sampled, 86.5% were married and had an average of two children (mode ⫽ 2.0, range ⫽ 0–9). Indeed, given the patriarchal nature of Indian society, which prescribes home-based roles for women, ownership of a microenterprise must be understood in the context of dual home-work challenges. Women who own and run a microenterprise are not free from domestic work. Nevertheless, 42.4% had part-time domestic help, and a surprising 19.9% said their husband shared in the work of maintaining the home. More than 8 out of 10 (83.5%) businesswomen had a personal bank account from which they could make withdrawals or payments on their own. Education levels of respondents were high, with 51.9% having a high school or higher secondary school education and 41.1% having earned a bachelor’s degree or higher. All respondents said they speak, read, and write Hindi. More than half (56.3%) spoke English, 70.1% could read English, and 67.5% could write it. More than half (56.3%) of respondents said they could calculate taxes or interest, another 39.8% said they could do simple arithmetic, while only 3.9% said they could recognize or write numbers but could not do calculations. In sum, the probabilitybased survey generated a sample with generally high levels of human capital—a cluster of skills, experiences, and social contexts that are acquired prior to (and are necessary for) ICT access and use (van Dijk, 2005). Testing the Models We examined the question of ICT impact on microenterprise growth by using SEM (Byrne, 2010). The ªrst step in the process is to specify all inde- Information Technologies & International Development CHEW, LEVY, ILAVARASAN model is economic growth, with total ICT access, microenterprise, and owner characteristics as antecedent, independent variables. The wording of most questionnaire items and the internal consistency coefªcients (Cronbach’s alpha) for the multi-item additive indices can be found in the appendix.4 The dependent variable, business growth, was operationalized as the sum of (a) the owner-estimated percentage change in the income of the Figure 1. Model of Microenterprise Growth (Ilavarasan & Levy, 2010, p. 61). microenterprise “compared to a year ago,” and (b) the responpendent (predictor) and dependent variables in the dent’s assessment of the effects on her business of model and to depict hypothesized relationships the Mumbai terrorist attacks in November 2008 between the variables in the form of a path dia(0 ⫽ less business; 1 ⫽ no change in amount of gram. A correlation matrix of all variables is then business; 2 ⫽ more business). None of the microcreated to determine the nature of the relationships enterprises in the sample were targets of those between pairs of variables and to eliminate hypothe- attacks. We did not ask respondents about actual sized variables that are not signiªcantly correlated income, but inquired instead about “percentage with other variables. Regression analyses are then change,” because we had found in the pilot study carried out to determine whether the hypothesized that businesspeople were unwilling to share that paths are statistically signiªcant and to determine information. Coupled with similar experiences how much variance in the dependent variables is reported by the survey interviewers, we decided that predicted by the antecedent variables. Finally, the an indirect approach to this sensitive subject would entire structure of variables is tested using a path yield more valid data (de Mel, McKenzie, & Woodanalysis to evaluate the degree to which the data ruff, 2009). are consistent with the model, an assessment Total ICT access was constructed from 10 items known as “testing goodness-of-ªt.” The use of path that measured ownership of mobiles, phones, peranalysis—a speciªc type of SEM—allows researchers sonal computers, laptop computers, and availability to analyze all the variables in the model simultaof the Internet, both in the respondent’s home and neously instead of separately in different regression at the workplace.5 A logarithmic transformation was models (Chin & Newsted, 1999; Fornell, 1984). performed on the ªnal total ICT scores. If ownership A model that has a good ªt with the data supports of ICTs improves the economic well-being of microthe plausibility of the postulated relationships in the enterprises owned by women, we would expect to model. ªnd a positive relationship between total ICT access and business growth. Initial Model While microenterprises are generally considered The initial model tested is called the “Model of part of the informal economy, it is also true that Microenterprise Growth” (Ilavarasan & Levy, 2010, microenterprises may vary considerably across a p. 61; see Figure 1). The dependent variable in the dimension of informality–formality (Becker, 2004; 4. The appendix contains descriptions of the variables used in the ªnal structural model. Most of these variables were retained from Ilavarasan and Levy, 2010. Revisions to the variables are described in the Methods section. 5. Access to a workplace landline was excluded in the ªnal index of ICT access, because, with its inclusion, the Cronbach’s alpha for the index fell to an unacceptable level (␣⫽0.53). Based on open-ended questions in the survey, we found that many respondents believed that a microenterprise needed to have a landline to remain visible in the business space, even though the landline was rarely used. Volume 7, Number 4, Winter 2011 5 THE LIMITED IMPACT OF ICTS ON MICROENTERPRISE GROWTH Chen, 2007). To capture this characteristic, the business formality variable was constructed from items that measured whether the business was registered with the government, whether it had a business bank account, and what bookkeeping practices it followed. Some development economists contend that informal enterprises can be productive, but are often hampered by unfair taxes, cumbersome government regulations, and limited access to capital (de Soto, 1989). By staying in the unregulated, unlicensed, and untaxed sector of developing economies, microenterprises can remain economically viable and may even thrive (Portes & Haller, 2005). Therefore, we hypothesized that there would be a negative relationship between microenterprise formality and microenterprise growth. Two psychological variables were also created. The ªrst was an index of motivation to use mobile phones for business with eight operational measures (e.g., “My mobile phone helps me keep informed about prices and other business news” or “Using my mobile helps me stay in touch with current customers”) that are roughly similar to indicators of “relative advantage” found in the diffusion of innovation literature (Rogers, 2003). We assumed that female microentrepreneurs who saw potential economic beneªts in applying mobile phones to their business practices would be more inclined to use them, which could, in turn, lead to business growth. The second psychological variable was labeled “social status/power” in the initial model and was constructed from self-reported measures of perceived increases in status and power. We hypothesized that to the extent female microentrepreneurs acquire and use ICTs as a means to achieve business growth and, in turn, to the degree that others perceive the microenterprise to be a “success,” then the women who own such microenterprises would garner increased respect from family and the community, as well as a greater say in making family decisions. Correlation coefªcients for all variables in the “Model of Microenterprise Growth” and results of a standard multiple regression analysis of all hypothesized paths in the model were previously reported in Chew, Ilavarasan, & Levy (2010). As a next step, therefore, the AMOS software package was used to examine the goodness-of-ªt of the initial model. To reiterate, when a model is tested for goodness-of- 6 ªt, all the hypothesized relationships in the theoretical model are simultaneously examined for the extent to which they are supported by the data. A model that has good ªt with the data supports the plausibility of the postulated relationships in the model. Four measures of goodness-of-ªt are commonly used: the overall chi-square test of ªt; a mini$ (CMIN) as Byrne cited mum discrepancy statistic C (1989); the root mean square error of approximation (RMSEA) as Browne and Cudeck noted (1993); and a comparative ªt index (CFI/df) as Hu and Bentler discussed (1999). Results from the initial model showed that the minimum chi-square value was achieved, meaning that the software succeeded in estimating the model parameters and produced a convergent solution for the model. However, the model failed the test for goodness-of-ªt on three of its four mea$ ⬎ 2.00); the sures: minimum discrepancy statistic (C root mean square error of approximation (RMSEA ⫽ 0.08), and the comparative ªx index (CFI/df ⬍ 0.90). Given that the chi-square test of ªt only indicates that the model was correctly speciªed in terms of the relationships between variables (Byrne, 2010, p. 102) and that the data were judged to be a poor ªt on three other more robust measures, we concluded that the data did not provide plausible support for the relationships speciªed in the initial model. Revising the Model To ªnd a structure that offered a better ªt for the data and that also might overcome some of the shortcomings of the initial model, we created a revised model that retained only two unchanged variables (business formality and total ICT access) from the initial model. 1. To create a more generalizable measure of business growth, we deleted the measure referring to the terrorist attacks in Mumbai as an operational indicator and restructured the business growth variable as the sum of (a) the owner-estimated percentage change in the annual income of her microenterprise compared to a year earlier and (b) the change in the number of hired workers over the same period. The two component items of the revised business growth variable were positively correlated (r⫽0.38, p⬍0.01) and the revised variable itself was now based on Information Technologies & International Development CHEW, LEVY, ILAVARASAN theories of ICTs and development (Gill, Brooks, McDougall, Patel, & Kes, 2010; Gurumurthy, 2004, 2006). More speciªcally, the items that make up the perceived empowerment variable can be thought of as tapping into “transformative forms and . . . achievements [that] suggest a greater ability on the part of poor women to question, analyse and act on the structures of patriarchal constraint in their lives” (Kabeer, 2003, p. 174). We predict that the more empowered a female microentrepreneur feels in running her business, the more she will be motivated to own and use ICTs in support of that enterprise. two of the most common indicators of business performance (Delmar, 2006). 2. 3. 4. The variable motivation to use ICTs was renamed “perceived usefulness of mobile phones” and “perceived usefulness of work computers.” The items comprising the perceived usefulness of mobile phones variable point to information-seeking behaviors regarding prices, business news more generally, and possible feedback on the viability of the microenterprise. One additional item points to the dual business-personal dimension of mobile use and recognizes that conversations with family and friends might have business content (Donner, 2009). For the perceived usefulness of work computers, the items comprising this variable suggest information-seeking behaviors also regarding prices and include the affordances of work computers for improving work efªciency and for keeping the microentrepreneurs up to speed with technological advancements. An education variable was reintroduced to the revised model, as more formal education has been consistently associated with higher income for women in developing countries. Additionally, by enhancing cognitive skills and knowledge, education can be seen as promoting an individual’s “power to” make and act on signiªcant social and economic life-choices (Kabeer, 2003, p. 175). The new education variable combined a respondent’s level of formal education—already shown to be correlated with ICT access—with a dichotomous measure of whether the female microentrepreneur had done any formal computer training. The two components of the expanded education variable were correlated, r⫽0.41, p⬍0.001. With this new indicator, we would expect to see an even stronger positive relationship between education and total ICT access than we found in the initial model. The social status/power variable from the initial model continued to be constructed as in the initial model, but was renamed “perceived empowerment” to clarify its meaning and highlight its association with gender Volume 7, Number 4, Winter 2011 5. Finally, a measure of mobile phone use for business purposes was added to the model. We included this new variable to possibly strengthen the causal link between mobiles and microenterprise growth, since it is logically necessary to show that female microentrepreneurs indeed use their mobiles to conduct business. A composite index— total mobile use—was formed from three items measuring whether respondents used a mobile phone to call customers, suppliers, or employees. Index scores ranged from 1 to 3 and had a mean of M⫽1.14, SD⫽0.36. The revised model is pictured in Figure 2. The signs on the paths indicate whether the variables are hypothesized to be positively or negatively correlated. Pearson product-moment correlations between variables in the revised model were ªrst calculated using the Predictive Analytics Software (PASW) Statistics 18 software package. Regression analyses were then run to determine the statistically signiªcant paths in the model. Two versions of this analysis were carried out. In the ªrst, business growth was run as the dependent variable, and in the second, total ICT access was considered to be the dependent variable. This procedure allowed us to test the hypothesis that increased business growth is the cause of increased ICT access as opposed to the predicted direction of the relationship in which ICT access predicts business growth. Then, with AMOS 18.0, using maximum likelihood estimation, a path analysis was performed to test the proposed model. 7 THE LIMITED IMPACT OF ICTS ON MICROENTERPRISE GROWTH rusefulnessCOMP⫽0.21, p⬍0.01), with respondent education (reducation⫽0.36, p⬍0.01), and with perceived empowerment (rempower⫽0.20, p⬍0.01). The nature of all relationships is consistent with those predicted in Figure 2. To verify the paths hypothesized in the model, three standard multiple regressions were conducted. The ªrst regression focused on total ICT access, since it was hypothesized to be a major predictor of business growth. Total ICT access was set as the dependent variable and all other variables, except business growth, were entered as independent Figure 2. Revised Model for Microenterprise Growth. variables. The results of the Note: In developing this revised model, we examined a number of ICTs that respondents might wish to use in various business processes. We found, however, such limited regression (Table 2) show that the ownership of workplace computers and such limited use of Internet cafés that operaset of independent variables tional measures of those two technologies were not statistically related to the other account for a moderate amount variables in the model. Therefore, the model only included the variable “Business use of variance in total ICT access of mobile phones.” (R2⫽0.29, F(6,192)⫽14.39, p⬍0.001). Only business formality is unrelated to total ICT access. This regression analResults ysis suggests that about 29% of the variance in Table 1 shows the Pearson product-moment correlatotal ICT access among female microentrepreneurs tions among the all variables in the model. Business can be accounted for by a combination of business growth is positively associated only with total ICT use of mobile phones, perceived usefulness of access (rgrowth⫽0.12, p⬍0.05, one-tailed) and negamobile phones, education, and perceived empowertively associated with business formality (rformality⫽ ment.6 Moreover, both the correlations in Table 1 ⫺0.16, p⬍0.05). The negative correlation between and the beta weights here point to the possibility microenterprise growth and business formality adds that female microentrepreneurs who feel a greater support to the notion that microenterprises might sense of empowerment from owning businesses are have greater economic success by remaining in the also somewhat more likely to have a greater reperinformal sector (Portes & Haller, 2005). Business toire of ICTs. growth is not related to education (r⫽0.09, n.s.), The second regression tested business use of raising the possibility that the quality of education mobile phones as the dependent variable, and total received might be more important than is the numICT access, business formality, and the characteristics ber of years spent in school (Hanushek & of the entrepreneurs were entered as the independWoessmann, 2007). ent variables. As shown in Table 3, the signiªcant Total ICT access is positively associated predictors of business use of mobile phones are with business formality (rformality2⫽0.22, p⬍0.01), total ICT access (␤⫽0.155, p⬍0.05, one-tailed) and positively associated with both measures of perperceived empowerment (␤⫽0.246, p⬍0.01). The ceived usefulness (rusefulnessMP⫽0.18, p⬍0.01, 6. The t-tests for the standardized beta weights for “business use of mobile phones” and “perceived usefulness of mobile phones” were signiªcant at the 0.05-level of signiªcance, one-tailed, and considered as contributing to the regressions being tested. 8 Information Technologies & International Development CHEW, LEVY, ILAVARASAN Table 1. Pearson Product-Moment Correlation Matrix. 1 2 1. Business growth ⫺1 2. Total ICT access ⫺0.123* 1 3. Business use of mobile phones ⫺0.078 0.179** 3 4 5 6 7 8 1 4. Formality ⫺0.163* 0.218** ⫺0.058 ⫺1 5. Perceived usefulness of mobile phones ⫺0.149* 0.182** ⫺0.120 ⫺0.178** 1 6. Perceived usefulness of work computers ⫺0.106 0.206** ⫺0.078 ⫺0.001 7. Education ⫺0.091 0.356** ⫺0.062 ⫺0.198** 0.015 8. Perceived empowerment ⫺0.030 0.195** ⫺0.279** ⫺0.061 0.066 0.333** ⫺1 ⫺0.060 ⫺1 ⫺0.068 ⫺0.018 1 *Correlation is signiªcant at the 0.05 level ** Correlation is signiªcant at the 0.01 level Table 2. Regression of Business and Entrepreneur Characteristics on Total ICT Access. Model Standardized Beta (Constant) t Sig. ⫺3.120 0.002 Characteristics of businesses Formality 0.104 ⫺1.672 0.096 Business use of mobile phones 0.120 ⫺1.909 0.058 Characteristics of entrepreneurs Perceived usefulness of mobile phones 0.123 ⫺1.858 0.065 Perceived usefulness of work computers 0.236 ⫺3.906 0.000 Education 0.377 ⫺6.173 0.000 Perceived empowerment 0.183 ⫺2.732 0.007 F(6, 192)⫽14.39; p⬍0.001; Adjusted R 2 ⫽0.289. positive relationship between perceived empowerment and business use of mobile phones was not previously posited in the revised model. Further, the beta weights here reinforce the relationship in Table 2 between perceived empowerment and total ICT access, and they further suggest that for female microentrepreneurs a greater sense of self-efªcacy and self-esteem from owning a microenterprise is also associated with using mobile phones for business purposes.7 The third regression tested business growth as the dependent variable and the characteristics of the businesses and of the women who own them as independent variables. As illustrated in Table 4, the results of the regression show that the set of independent variables we examined account for a small, but statistically signiªcant amount of variance in business growth, R2⫽0.075, F(6,191)⫽3.267, p⬍0.01. Consistent with previous analysis and the correlation matrix, the only statistically signiªcant predictors of business growth are business formality (␤⫽ ⫺179, p⬍0.05) and total ICT access (␤⫽0.155, p⬍0.05, one-tailed). Surprisingly, the business use of mobile phones is not a signiªcant predictor of business growth. We had conducted the previous regression with the view that the business use of mobile phones is an important mediating variable between total ICT access and business growth. As it turns out in the path analysis that follows, the very limited effect of ICTs on business growth appears not to require a direct relationship between business use of mobile phones and the dependent variable. 7. We are more fully investigating the antecedents of perceived empowerment and its causal relationship with total ICT access and microenterprise growth in the study of female microentrepreneurs, ICT, and empowerment currently under way in Chennai, India. Volume 7, Number 4, Winter 2011 9 THE LIMITED IMPACT OF ICTS ON MICROENTERPRISE GROWTH Table 3. Regression of Business Characteristics, Entrepreneur Characteristics, and Total ICT Access on Business Use of Mobile Phones. Model Standardized Beta Sig. ⫺3.120 0.002 Formality ⫺0.047 ⫺0.656 0.513 Perceived usefulness of mobile phones ⫺0.010 ⫺0.137 0.892 Perceived usefulness of work computers ⫺0.039 ⫺0.542 0.589 Education ⫺0.115 ⫺1.514 0.132 Perceived empowerment ⫺0.246 ⫺3.258 0.001 Total ICT access ⫺0.155 ⫺1.909 0.058 (Constant) Characteristics of entrepreneurs t F(6, 192)⫽3.894; p⫽0.001; Adjusted R 2 ⫽0.081. Table 4. Regression of Business Characteristics, Entrepreneur Characteristics, and Total ICT Access on Business Growth. Model Standardized Beta (Constant) t Sig. ⫺3.120 0.002 Characteristics of businesses Business use of mobile phones ⫺0.127 ⫺1.753 0.081 Formality ⫺0.179 ⫺2.501 0.013 Characteristics of entrepreneurs Perceived usefulness of mobile phones ⫺0.045 ⫺0.588 0.557 Perceived usefulness of work computers ⫺0.158 ⫺2.196 0.029 Education ⫺0.067 ⫺0.880 0.380 Perceived empowerment ⫺0.003 ⫺0.032 0.974 Total ICT access ⫺0.154 ⫺1.861 0.064 F(6, 191) ⫽ 3.267; p⬍0.01; Adjusted R2 ⫽0.075. That is, the use of the mobile phone for business purposes did not lead to an increase in business growth directly. Instead, mobile phone use for business contributes to an increased repertoire of ICTs, which, in turn, leads to business growth (see Figure 3). Figure 3 displays the standardized path coefªcients for the ªnal structural model. Inter-item correlations between the independent variables are not shown in the model. All path coefªcients are in line with predicted relationships except for the path between business use of mobile phones and business growth, which is not statistically signiªcant and, indeed, produces a model of poor ªt when it is included in the path analysis (␹2⫽32.30, df⫽17, p⫽0.014, CMIN⫽1.90, CFI⫽0.880, RMSEA⫽0.063). However, when business use of mobiles is excluded from the analysis and business growth is treated as the dependent variable, the ªnal model has a good ªt with the data (␹2⫽26.727, df⫽17, p⫽0.062, CMIN⫽1.572, CFI⫽0.924, RMSEA⫽0.050). Finally, 10 and most important in terms of the goals of this article, the model predicts that the antecedent variables have a direct but small impact on business growth based on the amount of variance explained (R2⫽0.075, p⬍0.01). It is possible, of course, that relatively larger microenterprises (ªve or more hired workers) might have greater information and communication needs and perhaps seek to satisfy some of those needs by increased access to and use of ICTs. If that were the case, then perhaps the impact of ICTs on microenterprise growth would also be greater. We tested for this possibility by a regression analysis with the enterprise size entered as a dummy variable (larger microenterprises with ªve or more hired employees equal to 1; microenterprises with fewer than ªve hired workers equal to 0). The beta coefªcient for enterprise size in this regression analysis is not statistically signiªcant and the total variance explained differs from the variance explained in the ªnal Information Technologies & International Development CHEW, LEVY, ILAVARASAN access have a direct impact on business growth. Moreover, an examination of the regression analyses disclosed that business formality contributes about 4% of the variance in business growth, while ICT access accounts for only 2%. In other words, access to ICTs predicts only half the already meager variance in business growth explained by the model. Still, the small, albeit statistically signiªcant, relationship between total ICT access and business growth demonstrates there is a plausible causal link between access to a greater repertoire of ICTs and the economic well-being of a microenterprise. To summarize, adding mobile phones, computFigure 3. Final Structural Model for Microenterprise Growth. ers, and Internet access to a microenterprise might increase model by a minute amount. This suggests, therethe proªts and the number of hired workers, but fore, that the size of the microenterprise in this data only minimally. set does not inºuence business growth, and the staAttempting to demonstrate a causal link tistical signiªcance of the other predictor variables between business use of mobile phones and remain unchanged. microenterprise growth produces a poor-ªtting We also investigated whether the relationship model, suggesting that mobile phone access alone between total ICT access and business growth might does not necessarily produce business growth. be recursive—that is, whether more successful Indeed, fewer than 10% of women who owned microentrepreneurs might have the capital to microenterprises consistently used their mobiles to acquire more ICTs, which, in turn, might promote conduct business. For the impact of ICTs (and business growth. To test for this possibility of endomobile phones speciªcally) to show up on the busigeneity, we reversed the direction of causality ness growth variable, it might at the very least be between total ICT access and business growth in the necessary for female microentrepreneurs to use ICTs ªnal structural model, making business growth the much more frequently and extensively in their busipredictor variable and ICT access the dependent nesses. Succinctly put, our data support a notion of variable. However, the path coefªcient of the limited impact from limited use. reversed causal link turned out to be nonsigniªcant The ªnal model clearly indicates the directionality (␤⫽0.056, n.s.), suggesting that the relationship of the path between the antecedent variables, ICT between ICT access and microenterprise growth, at access, and microenterprise growth. Education, perleast in this data set, is not recursive. ceived usefulness of mobile phones, perceived empowerment, and business use of mobile phones Conclusions all exercise indirect impact on business growth. The advantage of the path analysis is that we now have The statistical analyses yield several potentially a more complete understanding of the relationship signiªcant insights regarding business growth in between those antecedents and the dependent varifemale-owned urban microenterprises. First, we able of business growth. In short, we found that found that only business formality and total ICT these four antecedents predict total ICT access, Volume 7, Number 4, Winter 2011 11 THE LIMITED IMPACT OF ICTS ON MICROENTERPRISE GROWTH which, in turn, increases business growth directly. The directionality of the causation is another important ªnding from the analysis. We tested the plausibility of a recursive relationship between ICT access and business growth and found that a causal path from business growth to total ICT access is improbable. Thus, the data support the notion of a positive impact of total ICT on business growth rather than the other way around. Having greater ICT access appears to be the driver of microenterprise growth. The ªndings of this study are, of course, limited by the nature of the respondents (women who own microenterprises) and by geography (Mumbai, India). While our probability-based sample closely reºects the very small business economy of Mumbai, it is composed almost entirely of microenterprises in the trade and service sectors. The results of our analysis might have been different if the sample had included microenterprises owned by women in the manufacturing sector, especially microenterprises at or near the bottom of global value chains. To the extent that participation in global value chains might be facilitated by ICTs, it might be possible to discern a greater impact of mobile phones and computers on microenterprise growth (Goldmark & Barber, 2005; Nichter & Goldmark, 2009). Moreover, many of the gains from the use of ICTs, be it the telegraph or the mobile phone, have come from the decoupling of communication from transportation. In an urban setting, productivity losses from transportation-related issues are often much lower than they are in far-ºung rural areas where distance imposes signiªcant transaction costs.8 Still, since more than one-third of customers in our sample came from other parts of Mumbai, we might have expected, but did not ªnd, mobiles playing such a role, depending on the products and services rendered. The primary aim of this article was to test a causal model of ICTs and economic development and, by using the powerful statistical technique of SEM, to examine the unique contribution of ICTs to the economic growth of microenterprises. Given the substantial social, economic, and cultural constraints on women who own microenterprises (Dejene, 2007; Mitra, 2005; Wasihun & Paul, 2010), it is scarcely surprising that ICTs, even the ubiquitous mobile phone, might have scant impact. These ªndings might lead some to erroneously conclude that ICT-based development strategies no longer have much of a role as facilitators of microenterprise growth. We would caution against such a hasty judgment. The relationship we found among perceived empowerment, ICT access, and microenterprise growth holds some promise, we believe, for programs with goals of poverty reduction and greater empowerment of women. Moreover, most respondents in this study held generally positive attitudes about how mobiles could help meet business communication needs, even if actual use for business purposes in the microenterprise context was limited. Regulators and providers of mobile services might be able to build on these positive perceptions. Keeping the cost of mobiles affordable, creating relevant mobile applications, and encouraging female microentrepreneurs to ªnd their own business uses for mobiles might lead to greater mobile use and, in turn, to more marked microenterprise growth. In the long run, it might well be true that the productivity gains of ICTs, especially computing resources, are likely to be greater in the small and medium enterprise sector than they would be for microenterprises (Esselaar, Stork, Ndiwalana, & Deen-Swarray, 2007; Legatum Institute, 2011). Nevertheless, our research did ªnd some positive consequences of ICTs on the economic growth of microenterprises, and that ªnding should challenge policymakers, practitioners, and ICT4D researchers to continue exploring how ICTs might improve the lives of the women who own some of the smallest businesses in developing economies. ■ References Abraham, R. (2006). Mobile phones and economic development: Evidence from the ªshing industry in India. The International Conference on Information and Communications Technologies and Development (ICTD 2006) Conference Proceedings. Berkeley, CA: IEEE. Aker, J. (2008). Does digital divide or provide? The impact of mobile phones on grain markets in Niger (Bureau for Research and Economic Analysis of Development [BREAD] Working Paper 8. We are grateful to one of the anonymous referees for suggesting this point. 12 Information Technologies & International Development CHEW, LEVY, ILAVARASAN No. 177). Retrieved from http://ipl.econ.duke .edu/bread/abstract.php?paper⫽177 Becker, K. (2004). The informal economy. (Swedish International Development Cooperation Agency). Retrieved from http://rru.worldbank.org/ Documents/PapersLinks/Sida.pdf Bentler, P. (1980). Multivariate analysis with latent variables: Causal modeling. Annual Review of Psychology, 31, 419–456. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model ªt. In K. A. Bollen & J. Scott Lang (Eds.), Testing structural models (pp. 136–162). Newbury Park, CA: SAGE Publications. Byrne, B. M. (1989). A primer of LISREL: Basic applications and programming for conªrmatory factor analytic models. New York: Springer-Verlag. Byrne, B. M. (2010). Structural equation modeling using AMOS. Basic concepts, applications, and programming (2nd ed.). New York: Routledge. Castells, M., Fernandez-Ardevol, M., Qiu, J., & Sey, A. (2007). Mobile communication and society: A global perspective. Cambridge, MA: MIT Press. Chen, M. (2001). Women in the informal sector: A global picture, the global movement. SAIS Review, 21(1), 71–82. Chen, M. (2007). Rethinking the informal economy: Linkages with the formal economy and the formal regulatory environment. New York: United Nations Department of Social and Economic Affairs. Chew, H. E., Ilavarasan, P. V., & Levy, M. (2010). The economic impact of information and communication technologies (ICTs) on microenterprises in the context of development. Electronic Journal of Information Systems in Developing Countries, 44. Chin, W. W., & Newsted, P. R. (1999). Structural equation modeling analysis with small samples using partial least squares. In R. H. Hoyle (Ed.), Statistical strategies for small sample research (pp. 307–341). Thousand Oaks, CA: SAGE Publications. Chinn, M., & Fairlie, R. (2010). ICT use in the developing world: An analysis of differences in com- Volume 7, Number 4, Winter 2011 puter and Internet penetration. Review of International Economics, 18(1), 153–167. Cyber cafés log out of ªngerprinting plan. (2009, June 29). Times of India. Retrieved from http:// timesoªndia.indiatimes.com/news/city/mumbai/ Cyber-cafes-log-out-ofªngerprinting-plan/ articleshow/4713516.cms Delmar, F. (2006). Measuring growth: Methodological considerations and empirical results. In P. Davidsson, F. Delmar, & J. Wiklund (Eds.), Entrepreneurship and the growth of ªrms (pp. 62–86). Cheltenham, UK: Edward Elgar Publishing, Inc. Dejene, Y. (2007, November 15–17). Promoting women’s economic opportunity in Africa. United Nations Economic Commission for Africa; African Development Bank; African Economic Conference 2007: Opportunities and Challenges of Development for Africa in the Global Arena. Addis Ababa, Ethiopia. Retrieved from http:// www.uneca.org/aec/documents/Yeshiareg %20Dejene.pdf de Mel, S., McKenzie, D., & Woodruff, C. (2009). Measuring microenterprise proªts: Must we ask how the sausage is made? Journal of Development Economics, 88, 19–31. de Soto, H. (1989). The other path: The invisible revolution in the third worlds. New York: Harper and Row Publishers. Donner, J. (2008). Research approaches to mobile use in the developing world: A review of the literature. The Information Society, 24(3), 140–159. Donner, J. (2009). Blurring livelihoods and lives: The social uses of mobile phones and socioeconomic development. Innovations: Technology, Governance, Globalization, 4(1), 91–101. Donner, J., & Escobari, M. X. (2010). A review of evidence on mobile use by micro and small enterprises in developing countries. Journal of International Development, 22, 641–658. Duncombe, R. (2009). Impact assessment of mobile phones on development: Concepts, methods and lessons for practice. (Development Informatics Working Paper Series No. 39). Institute for Development Policy and Management, University of 13 THE LIMITED IMPACT OF ICTS ON MICROENTERPRISE GROWTH Manchester, UK. Retrieved from http://www.sed .manchester.ac.uk/idpm/research/publications/wp/ di/documents/di_wp39.pdf Duncombe, R., & Heeks, R. (2005). Information & communication technologies (ICTs), poverty reduction and micro, small & medium-scale enterprises (MSMEs): A framework for understanding ICT applications for MSMEs in developing countries. Vienna: United Nations Industrial Development Organization. Esselaar, S., Stork, C., Ndiwalana, A., & DeenSwarray, M. (2007). ICT usage and its impact on proªtability of SMEs in 13 African countries. Information Technologies & International Development, 4(1), 87–100. Fornell, C. (1984). A second generation of multivariate analysis: Classiªcation of methods and implications for marketing research. Working paper, Graduate School of Business Administration, the University of Michigan, Ann Arbor. Retrieved from http://deepblue.lib.umich.edu/ bitstream/2027.42/35621/1/b1408124.0001 .001.txt Gelb, A., Mengistae, T., Ramachandran, V., & Shah, M. (2009, July 20). To formalize or not to formalize? Comparisons of microenterprise data from Southern and East Africa. Available at SSRN: http://ssrn.com/abstract⫽1473273 Gill, K., Brooks, K., McDougall, J., Patel, P., & Kes, A. (2010). Bridging the gender divide: How technology can advance women economically. Washington, DC: International Center for Research on Women. Goldmark, L., & Barber, T. (2005). Trade, micro and small enterprises, and global value chains. (AMAP report number 25). Washington, DC: USAID. Gurumurthy, A. (2004). Gender and ICTs: A BRIDGE Overview Report. Brighton, UK: Institute for Development Studies. Gurumurthy, A. (2006). Promoting gender equality? Some development-related uses of ICTs by women. Development in Practice, 16(6), 611– 616. Hanushek, E. A., & Woessmann, L. (2007, February 1). The role of education quality for economic 14 growth. (World Bank Policy Research Working Paper No. 4122). Available at SSRN: http://ssrn .com/abstract⫽960379 Heeks, R. (2007). Impact assessment of ICT4D projects: A partial review of frameworks. Institute for Development Policy and Management, University of Manchester, UK. Heeks, R. (2010a, December 28). The ICT4D value chain. ICT4D blog. Retrieved from http:// ict4dblog.wordpress.com/author/richardheeks Heeks, R. (2010b, April 14). ICT4D journal ranking table. ICT4D blog. Retrieved from http:// ict4dblog.wordpress.com/2010/04/14/ict4djournal-ranking-table/ Heeks, R., & Molla, A. (2009). Impact assessment of ICT-for-development projects: A compendium of approaches (Development Informatics Working Paper No. 36). University of Manchester, UK. Retrieved from http://www.sed.manchester.ac.uk/ idpm/research/publications/wp/di/di_wp36.htm Hu, L., & Bentler, P. M. (1999). Cutoff criteria for ªt indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. Ilavarasan, P. V., & Levy, M. (2010). ICTs and urban microenterprises: Identifying and maximizing opportunities for economic development: Final report. Ottawa: International Development Research Centre. Retrieved from http://www.idrc.ca/ uploads/user-S/12802403661ICTs_and_Urban_ Microenterprises_104170-001.pdf International Labor Organization (ILO). (2009). Globalization and informal jobs in developing countries. (A joint study of the International Labor Ofªce and the Secretariat of the World Trade Organization). Geneva: International Labor Organization. Jensen, R. (2007). The digital provide: Information (technology), market performance, and welfare in the south Indian ªsheries sector. The Quarterly Journal of Economics, CXXII(3), 879–924. Kabeer, N. (2003). Gender mainstreaming in poverty eradication and the Millennium Development Information Technologies & International Development CHEW, LEVY, ILAVARASAN Goals. Ottawa: Commonwealth Secretariat/IDRC/ CIDA. Kotz, J., Krishnan, A., & Wickersham, J. (2007, November 15). Structural equation modeling: The state of the art in communication science. Paper presented at the annual meeting of the NCA 93rd Annual Convention, Chicago. Retrieved from http://www.allacademic.com/ meta/p192055_index.html LaPorta, R., & Shleifer, A. (2008). The unofªcial economy and economic development (NBER Working Paper No. 14520). Cambridge, MA: National Bureau of Economic Affairs. Retrieved from http://www.nber.org/papers/w14520 Lee, Y., Kozar, K. A., & Larsen, K. R. T. (2003). The technology acceptance model: Past, present, and future. Communications of the Association for Information Systems, 12, 752–780. Legatum Institute. (2011). The prosperity index in Africa: The role of entrepreneurship and opportunity in sub-Saharan Africa. Retrieved from http:// www.legatum.com/attachments/The_Prosperity_ Index_in_Africa-May_2011.pdf Mariscal, J. (2005). Digital divide in a developing country. Telecommunications Policy, 29(5–6), 409–428. Mead, D., & Liedholm, C. (1998). The dynamics of micro and small enterprises in developing countries. World Development, 26(1), 61–74. Ministry of Micro, Small, & Medium Enterprises. (2007). Micro, small, and medium enterprises in India: An overview. Annual Report 2006–2007. New Delhi: Government of India. Retrieved from http://msme.gov.in/ssiar-eng-2006-07.pdf .microlinks.org/ev_en.php?ID⫽10015_201&ID2 ⫽DO_TOPIC Nichter, S., & Goldmark, L. (2009). Small ªrm growth in developing countries. World Development, 37(9), 1453–1464. Peebles, D. (2006). Women’s microenterprises growth strategies project: The NGO business subcontracting model. Prepared for Consultative Cycle 2006, Innovations in Export Strategy. International Trade Centre Executive Forum. Retrieved from http://www.docstoc.com/docs/49637954/ WOMENS-MICROENTERPRISES-GROWTHSTRATEGIES-PROJECT-THE-NGO-BUSINESS Portes, A., & Haller, W. (2005). The informal economy and its paradoxes. In N. Smelser & R. Swedberg (Eds.), Handbook of economic sociology (2nd ed.), (pp. 403–425). New York: Russell Sage Foundation. Qiang, C., & Rossotto, C. M. (with Kimura, K.). (2009). Economic impacts of broadband. In World Bank (Ed.), Information and Communications for Development 2009: Extending reach and increasing impact (pp. 35–50). Washington, DC: The World Bank. Rogers, E. M. (2003). Diffusion of innovations (5th ed.). New York: Free Press. Srivastava, R. (2005, January 5). The informal sector and urban poverty. Infochange Urban India. Retrieved from http://www.infochangeindia.org/ urban_india_04.jsp Telecom Regulatory Authority of India (TRAI). (2009). Telecom subscription data as on 30th June 2009. New Delhi: TRAI. Retrieved from http://www .trai.gov.in/PressReleases_content.asp Mitra, A. (2005). Women in the urban informal sector: Perpetuation of meager earnings. Development and Change, 36(2), 291–316. van Dijk, J. (2005). The deepening divide: Inequality in the information society. Thousand Oaks, CA: SAGE Publications. National Sample Survey Organization (NSSO). (2000). Non-agricultural enterprises in the informal sector in India, 1999–2000—Key Results (No. 456). New Delhi: Ministry of Statistics and Programme Implementation. Wasihun, R., & Paul, I. (2010). Growth determinants of women-operated micro and small enterprises in Addis Ababa. Journal of Sustainable Development in Africa, 12(6), 233–246. Nichter, S., & Goldmark, L. (2005). Understanding micro and small enterprise growth (Micro Report #36). Washington, DC: U.S. Agency for International Development. Retrieved from http://www Volume 7, Number 4, Winter 2011 Waverman, L., Meschi, M., & Fuss, M. (2005). The impact of telecoms on economic growth in developing countries. The Vodafone Policy Paper Series, 2(March), 10–23. 15 THE LIMITED IMPACT OF ICTS ON MICROENTERPRISE GROWTH Appendix Survey Items and Reliabilities Index Business growth Questionnaire Items Scoring Alpha Compared to a year ago, has the annual income of this business: Change in percentage r⫽0.38 Change in number Compared to a year ago, has the number of hired workers: Total ICT access Do you have/own a mobile phone; personal computer; laptop; Internet connection; Do you ever use external PCOs/STD booths for this business; Have you ever given mobile phones to your employees to use for business purposes; Do they ever use their personal mobile phones for business purposes; Does this business have computers in the workplace; Does the business have an Internet connection; Have you provided computers for the employees to use at work? Formality Is this business registered with the government; Does this business have a bank account; For maintaining the ªnancial and business records of this business, which of the following statements is most applicable? 1⫽Yes 0.61 1⫽yes; 1⫽yes; 0.64 0⫽no records, .5⫽written records, 1⫽accountant Perceived usefulness of mobile phones Mobile phones help keep me informed of prices; help me be more conªdent that the business will survive; help me stay in touch with friends and relatives; help me stay connected to businesses in other parts of Mumbai. 1⫽strongly disagree 0.73 Perceived usefulness of work computers My work computer helps me keep informed about prices and other business news; helps me come and go without worrying about missing important business phone calls; Having a work computer makes me feel up-to-date. 1⫽strongly disagree 0.67 Formal education 0⫽never been to school; 5⫽master’s degree r⫽0.41 Education Have you undertaken any formal computer education training? 1⫽yes Perceived empowerment Because of my business, I am feeling more conªdent; Because of my business, I have gained respect among my friends and in my neighborhood; Despite my business, my parents do not feel especially proud of me (reversed); Because of my business, my parents-in-law are proud of me; Despite my business, my parents do not feel especially proud of me (reversed); Despite my business, my opinions are not considered to be important in family decisions (reversed). 1⫽strongly disagree 0.70 Business use of mobile phones How often do you use the mobile to call customers; How often do you use the mobile to call employees? How often do you use the mobile to call suppliers? 1⫽Rarely, 4⫽Very Often 0.64 16 Information Technologies & International Development