Women and the Internet: Is There an Economic Payoff? - CiteSeerX

Women and the Internet: Is There an Economic Payoff? Ernest Goss and Uma Gupta As the pervasiveness of the Internet spreads rapidly, Peter Drucker’s prediction about the coming of the new organization, labeled as the “knowledge society” [7] or the “virtual corporation” [9] is now a reality. In an information-based economy, any radical shift in our approach toward information and knowledge influences more than an elite group of individuals or firms—it has a far-reaching effect on national and global economies. The Internet is viewed by many as a paradigm shift forcing individuals, firms, businesses, and even societies to rapidly adapt to a newfound virtual world in which agility and responsiveness are the hallmarks of success [5]. At a 1999 meeting of Congress’s Joint Economic Committee Microsoft Chairman Bill Gates told committee members: “We will absolutely see productivity increases coming out of the use of technology for many years to come.” At same meeting, Federal Reserve Chairman Alan Greenspan told the panel that “something special has happened to the American economy in recent years because of computerization.” That “something special” was the increased productivity generated by the Internet and information technology [10]. Bureau of Labor Statistics data at least superficially support this linkage showing the share of employees using a computer at work rising from 25% in 1984 to 52% in 1997. Computer usage, in this context, refers to the use of computers to perform any work-related function, such as word processing, information retrieval, data analysis, and other related tasks. In addition to enhancing productivity, revolutionary technologies such as the Internet revive hopes of removing barriers to economic prosperity, increasing workplace productivity, and improving quality of life. The much-proclaimed anonymity of the Internet user leads to the hypothesis that the Internet may help close racial and gender-based wage gaps. The significant increase in women using the Web since its early origins [8] suggests a relationship between the Internet, gender, and the wage structure, but little or no research has been done in this area. Could the Internet lead to a full realization of the legal concept of equal pay for equal work?

Ernest Goss ([email protected]) is the MacAllister Chair and a Professor of Economics at Creighton University, Omaha, NE.

Uma G. Gupta ([email protected]) is president of Alfred State College, Alfred, New York. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. © 2003 ACM

160

September 2003/Vol. 46, No. 9ve COMMUNICATIONS OF THE ACM

Women earned 76% what men earned in 2000, according to U.S. Department of Labor data; women’s median weekly earnings were $491, compared to $646 for men. While this gender wage gap appears dramatic, women earned only 63% of the median weekly earnings of men in 1979. Thus, one could argue that the explosion in IT usage has helped narrow the gap between the wages of women and men. While IT and the Internet offer the promise of reducing labor market barriers and outcomes, women and minorities including African Americans, Hispanics, and Native American are currently under-represented in IT. According to the Department of Commerce, only 1.1% of undergraduate women choose to major in IT-related disciplines compared to 3.3% of male undergraduates [3]. Furthermore, Freeman and Aspray found the percentage of women entering the computer science pipeline and earning the bachelor’s degree in these IT fields has been dropping steadily since 1984. These researchers found that while the number of computer and information science degrees awarded decreased every year between 1986 and 1994, the decrease occurred at a faster rate for women. [3]. While past studies have examined the relationship between IT stock and productivity, to our knowledge no study has statistically tested the hypothesis that the Internet has been an important contributor to recent U.S. wage growth. The relatively recent arrival of web browser technology and a lack of data have been major inhibiting factors to this line of research. It was not until 1996 that web browser technology was sufficiently widespread to have an impact on the workplace [1]. Furthermore, the U.S. Census Bureau’s Current Population Survey (CPS) did not include survey questions that addressed Internet usage until December 1997. The Census Bureau repeated their Internet survey in December 1998, August 2000, and September 2001. The goal of the research described here was to examine the impact of the Internet on the gender wage gap using this dataset. If the Internet has indeed contributed to greater worker productivity, wage data should reflect this usage, unless the entire value of the increased productivity went to higher profits or lower prices.

Research Background Several researchers have studied the impact of computers on the wage structure of individuals in the U.S. These studies document an unambiguous and positive relationship between computer usage at work and wage growth. Entorf and Kramarz found that technology users had a wage advantage of 16% over non-users. In another study using Current Population Survey (CPS) data, Krueger concluded that employees with computer skills earned 10% to 15% more than workers without this skill set [6]. Kochhar showed that computer usage was twice as high among employees in higher wage quartiles than those in the lower quartiles. The highest premium for computer usage was in small firms in industries with the highest growth rate, where a premium of 24.8% in wages was observed. In fast-growing firms, the wage premium was 23.8% for computer users. Kochhar concluded that women used computers at a higher rate than men in firms of all sizes, noting that more than 50% of women in small firms use computers on the job, while the rate for men was below 40% [4]. Studying the relationship between technology-driven changes and the wage structure of individuals, as proxied by Internet usage, is a complex task for several reasons. First, separating the impact of the Internet from the rest of IT is fraught with difficulties. In addition to the Internet, IT covers a wide spectrum of devices ranging from COMMUNICATIONS OF THE ACM September 2003/Vol. 46, No. 9ve

161

embedded microprocessors to supercomputers. Second, the idea that the full impact of technological change on wage structures can be effectively represented using a single variable is simplistic [4]. Third, it is difficult to isolate the impact of technological change from other explicit and implicit variables that influence wage structures. Finally, firms more likely to use the Internet may also be more inclined to embrace other technological advances that affect wages, making it difficult to separate impacts. Data used to test our hypotheses comes from the Current Population Survey (CPS), a monthly survey of approximately 50,000 households conducted by the U.S. Bureau of the Census for more than 50 years. The CPS is the primary source of information on the U.S. labor force and includes the most robust measures of labor market activity for the civilian, non-institutional population. Supplemental CPS survey questions provide additional information on a variety of topics, including school enrollment, income, previous work experience, health, employee benefits, and work schedules. The December 1997, December 1998, August 2000, and September 2001 surveys are the only surveys currently available that include questions on Internet usage at work. The person in the household over 15 years of age who was most knowledgeable about the Internet or about computers answered Internet-related questions. This study focuses only on job-related Internet usage. The next section presents the empirical results. Table 1 shows the sample data used in this empirical investigation. Excluding observations for which there was inadequate wage data, and excluding persons under Table 1. Profile of sample. Percent of Sample

College Graduate Some College High School Grad Non-High School Grad

Education 27.1% 29.7% 30.9% 12.3%

Percent Internet User

32.3% 15.1% 6.4% 1.5%

Occupation Professional Teacher Technician Sales Clerical Service Craft Operations & Handler Other Occupations

13.8% 5.5% 3.6% 12.0% 14.9% 16.9% 9.3% 9.9% 14.1%

31.4% 29.7% 22.7% 10.8% 16.2% 4.7% 4.3% 2.7% 24.6%

Manufacturing

16.3%

Construction Trans, & Public Utilities Wholesale & Retail Trade Finance, Insurance, Real Estate Business Services Medical

5.4% 7.6% 21.1% 6.6% 5.6% 9.5%

15.9% 6.6% 16.0% 7.4% 22.4% 19.3% 11.2% 21.8%

Industry

Other Industries White Non-While Male Female Average Age

27.9% Demographics

85.0% 15.0% 50.4% 49.6% 38.1 Source: 1998 CPS

Table 1. Profile of sample. 162


15.9% 15.7% 15.6% 15.3%

16 and over 65 years of age, the sample consisted of 12,415 individuals, of which 1,916, or 15.4%, used the Internet on-the-job. The respondent profile shows that the higher the educational level, the greater the likelihood of using the Internet at work. In terms of occupations, professional workers, clerical workers, teachers and “other” occupations, which are occupations not listed in the survey, are the most predominant users of the Internet on the job. Industry analysis shows that employees in finance, insurance, and real estate (FIRE) used the Internet more than other industries. Furthermore, the average wage for the full sample was $15.23. Individuals who used the Internet for job-related purposes earned an average of $22.20 per hour, while workers who did not use the Internet at work earned $13.97 per hour.

The Impact of Internet Usage on Wages Survey respondents were classified along two dimensions, gender and Internet usage. Along gender lines, the sample size was divided into three groups: the full sample, males, and females. Based on Internet usage, the sample size was classified into three groups: The full sample (ALL), Internet Users (IU), and non-Internet Users (NIU). Table 2 shows the hourly wages for males and females according to Internet usage without controlling for personal, family, and labor market factors. Our findings show that women earn less than men, ranging from 27.1% for NIU to 44.9% for IU, with men enjoying a wage advantage in each case. But the analysis illustrated in Table 2 does not control for other important factors influencing wages. In other words, the wage advantage for men may stem from industry, occupational, or educational differences. We used regression analysis to control for factors such as race, marital status, age, industry, occupation, urban residency, and per capita income of state of residency, in order to untangle the impact of Internet usage on wages in several arenas: Education. Table 3 lists estimated wages after controlling for other factors that influence wages. Results indicate that among all educational groups for both men and women, Internet usage had a positive and statistically significant impact on hourly earnings. Among college graduates, however, the difference in hourly wages between IU and NIU was more prominent for men than women. The wage advantage for Table 2. Hourly wage comparisons of men and women by Internet usage (no control for other factors). Difference between Males and Females All Workers Men Women Wage Advantage for Men All $15.23 $17.39 $13.05 $4.34 (n = 12,415) (n = 6,257) (n = 6,158) 33.3% Internet Users (IU) $22.20 $26.18 $18.07 $8.11 (n = 1,916) (n = 975) (n = 941) 44.9% Non-Internet Users $13.97 $15.77 $12.41 $3.36 (NIU) (n = 10,499) (n = 5,282) (5,217) 27.1% Wage Advantage for $ 8.23 $ 10.41 $ 5.66 Internet usage 58.9% 66.0% 45.6% Source: 1998 CPS (number of observations in parentheses) Table 2. Hourly wage comparisons of men and women by Internet usage (no control for other factors). COMMUNICATIONS OF THE ACM September 2003/Vol. 46, No. 9ve

163

Table 3. Hourly pay by education and Internet usage (controlling for other factors). IU All College Graduate Some College High School Graduate Non-High School Graduate

$15.99 $21.17 $14.13 $12.92 $9.68

MALES NIU

Percent Gain For IU $14.04 13.9% $18.33 17.4% $12.60 12.1% $11.72 10.2% $8.12 19.2% Regression Analysis of 1998 CPS

FEMALES IU NIU $11.62 $17.78 $11.66 $10.29 $8.26

$10.32 $16.02 $9.86 $9.00 $6.53

Percent Gain For IU 12.6% 11.0% 18.3% 14.3% 26.5%

Table 3. Hourly pay by education and Internet usage (controlling for other factors).

Table 4. Hourly wages by Internet usage and industry (controlling for other factors).

Internet usage for both men and women was most significant among those with the least education. For women, the story emerges rather clearly that for the less educated, the Internet tends to close the wage gap between men and women. In fact, results in Table 3 indicate that Internet usage on the job closes the wage gap between men and women for all educational groups, except college graduates. For college graduates, Internet usage on the job widened the wage gap between men and women. Industry. Next we performed an industry-wide comparison of wages for IUs and NIUs by gender. These results are presented in Table 4, controlling for personal, family, and labor market factors. The top three industries that pay the highest premium for men with Internet skills were manufacturing (30.1%), business services (28.8%), and wholesale and retail (20.4%). For women, wholesale and retail (36.5%), business services (34.5%) and FIRE (27.0%) paid the highest premium for Internet skills. Surprisingly, construction and medical services paid women without Internet skills more, 8.5% and 1.2% respectively, than they did those with Internet skills. These two industries also paid the lowest premium, 3.3% and 1.0% respectively, for men with Internet skills. According to data in Table 4, Internet usage on-the-job tended to close the wage gap between men and women for wholesale and retail trade, FIRE, and business services, while it widened the wage gap between men and women for manufacturing, construction, transportation and public utilities, and business services. Occupation. Next we conducted an occupational analysis of wages (see Table 5), which revealed that Internet usage pays the highest premium for men in sales, service 164


Table 5. Wage premium by occupation and Internet usage, men vs. women (controlling for other factors).

IU Professional Technician Teachers Clerical Craft Sales Service Occupations Operators & Handlers Other Occupations

Men NIU

% Gain for IU

$21.31 $18.75 13.7% $16.71 $15.78 5.9% $16.05 $16.83 -4.6% $12.28 $11.25 9.2% $14.43 $11.88 21.5% $14.32 $11.33 26.4% $11.70 $9.44 23.9% $9.58 $9.90 -3.2% $20.55 $17.80 15.5% Regression Analysis of 1998 CPS

IU $17.80 $13.95 $15.81 $10.81 $13.24 $10.88 $8.22 $10.69 $16.85

Women NIU $15.46 $13.46 $14.69 $9.70 $8.82 $8.87 $7.26 $7.58 $15.30

% Gain for IU 15.1% 3.6% 7.6% 11.4% 50.1% 22.7% 13.2% 41.0% 10.1%

Table 5. Wage premium by occupation and Internet usage, men versus women (controlling for other factors).

operations, and craft occupations, and for women in craft, operating and handling, and sales occupations. Interestingly, among males, teachers with Internet skills made less (4.6%) than those without these skills. This did not occur for women in any occupation. By occupation, the Internet closed the wage gap among professionals, teachers, clerical workers, craft workers, and operators and handlers. Race and the Internet. Next we explored the relationship between Internet usage, race, and wages. Among males, Internet usage on-the-job produced a wage gain of 11.7% for whites and 6.4% for nonwhites. Among females, Internet usage on-thejob generated a wage premium of 14.8% for whites and 21.3% for non-whites. Thus, Internet usage closed the wage gap between white men and women and between nonwhite men and women. Among whites with Internet usage, men earned more than women by 24.1%, while among non-whites the wage difference was only 2.6%. Among whites without Internet usage, men earned 27.6% more than women, while among non-whites the difference was 17%.

Conclusion Several interesting results emerged from this study, including the following: •

• • • •

It is highly worthwhile for individuals to acquire jobs that make use of Internet technology. Holding education, age, race, marital status, industry, and occupation constant, Internet usage increases male hourly earnings from $13.95 to $15.52, or 11.3%, and female hourly earnings from $11.07 to $12.80, or 15.6%. The heaviest users of Internet on-the-job are found in finance, insurance, and real estate, and in transportation and public utilities. Regardless of education, age, marital status, industry, occupation, or Internet usage, women make less than men. Internet usage narrowed the gender wage gap for less educated workers, but widened it for college-educated workers. For men, manufacturing, business services, and wholesale and retail trade industries paid the greatest premium for Internet usage, while for women, wholesale and retail trade; business services; and finance, insurance and real estate paid the greatest premium for Internet usage. COMMUNICATIONS OF THE ACM September 2003/Vol. 46, No. 9ve

165

• •

In terms of occupations, men received the highest premium for Internet usage in sales, service, and craft occupations while women experienced the largest gain for Internet usage in craft, sales, and professional occupations. Construction and medical services pay the lowest premium for Internet usage.

References

1. Berners-Lee, T. Weaving the Web. Harper San Francisco, New York, 1999. 2. Blackburn, M.L., and Bloom, D. The effects of technological change of earnings and income inequality in the United States. The Impact of Technological Change on Employment and Economic Growth. Cyert, R. and Mowery, D., eds. Ballinger Publishing, Cambridge, MA., (1988), 223–66. 3. Freeman, P. and Aspray, W. The supply of information technology workers in the United States. Computing Research Association, Washington, D.C. (1997), . 4. Kochhar, R. The effect of computer use on the earnings of workers by firm size. National Technical Information Service, U.S. Department of Commerce, VA, No. SBA8033OA93 (1994), www.sba.gov/advo/research/rs156.html. 5. Kraut, R. Social impact of the Internet: what does it mean? Commun. ACM 41, 12 (Dec. 1998), 21–23. 6. Krueger, A. How computers have changed the wage structure: evidence from microdata, 1984–1989. The Quarterly Journal of Economics (Feb. 1993), 33–60. 7. Nonaka, I. The ‘ART’ of knowledge: systems to capitalize on market knowledge. European Management Journal 16, 6 (Dec. 1998), 673–681. 8. Pitkow, J.E. Summary of WWW characterizations. Computer Networks 30, 7 (April 1998), 551–559. 9. Rosecrance, R. The rise of the virtual state. Foreign Affairs 75, 4 (Jul/Aug 1996), 45–63. 10. U.S. Congress. Testimony before the 106th U.S. Congress. Joint Economic Committee, June 17, 1999.

166
