white tech pay asian wage workers gap
The State of Wage Inequality in Tech in 2018August 23, 2018The tech creed of “go fast and break things” has helped restructure Americanlife. How we communicate, get around and get our information have alldrastically changed because tech companies have disrupted outdated ways ofdoing things. There is, however, one aspect of American life that tech has notchanged and may have even exacerbated: racial and ethnic wage inequality.According to a 2016 EEOC report, black and Hispanic workers are drasticallyunderrepresented in the fast-growing and lucrative tech sector, while whiteand Asian workers are overrepresented. But tech’s diversity problem goes muchdeeper than questions of representation. For certain minority groups,discrimination extends all the way into their wallets.Using PayScale’s own data, we found that all else being equal, black, Hispanicand Native American workers earn less than their white counterparts in thetech industry. On the other hand, Asian tech workers earn more than theirwhite colleagues. Once we control for relevant factors, Pacific Islanders areroughly on par with their white peers.Looking only at those with the job title of software developer in the techindustry, we see the same pattern: on average, black and brown coders earnless than their white counterparts, while Asian developers earn more. Thepicture is more nuanced when we look at the full distribution of wages and notjust typical earnings. While Asian developers’ earnings skew higher, there isno statistical difference between the pay of white and non-Asian PoC softwaredevelopers. This offers us a glimmer of hope in terms of pay equity for thosewho can get these lucrative jobs.# Pay Inequity is a Major Element of Tech’s Diversity ProblemBetween January 2016 and June 2018, over 52,000 workers in the tech industrytook PayScale’s online salary survey and reported their race in addition toinformation about their occupation, location and other compensable factors. Wedefine the racial wage gap as the cents on the dollar that a racial/ethnicgroup makes compared with white workers.To determine the magnitude of this gap, we looked at two different measures:the uncontrolled wage gap and the controlled wage gap. The uncontrolled racialwage gap simply compares the median total cash compensation for each group.The controlled wage gap compares the salary for similarly qualified workers ofdifferent races. For the controlled wage gap we used PayScale’s proprietarysalary model to take into account compensable factors such as job title, yearsof experience and education.Race/Ethnic Group | Uncontrolled Pay Gap | Controlled Pay Gap —|—|— Asian | $1.23 | $1.02 American Indian & Alaska Native | $0.94 | $0.99 Black or African American | $0.87 | $0.98 Hispanic | $0.91 | $0.98 Native Hawaiian and Other Pacific Islander | $0.98 | $1.00 When we look at the uncontrolled gap, we see that Native American, black andHispanic all make less than white workers. The largest gap is for blackworkers, who earn $0.87 for every dollar white tech workers earns. Even whencontrolling for a long list of compensable factors, the wage gap remains forthese race/ethnic groups. Our controlled wage gap shows that Native Americanworkers earn $0.99 on the dollar in tech, while black and Hispanic workersmake $0.98.The controlled racial wage gap numbers are significantly smaller when we takeinto account various compensable factors. However, we can still see that manyracial and ethnic groups do not get equal pay for equal work in tech.Furthermore, the disparity in the uncontrolled and controlled racial wage gapshighlights that black and brown tech workers are underrepresented in higher-earning positions.Consider this together with another disturbing trend: even as the number ofblack and Hispanic students getting STEM degrees is increasing, theirrepresentation in the ranks of tech firms is not. Until tech closes the gap onhiring and promoting workers of all races, the racial wage gap for black andbrown tech workers will remain stubbornly in place.Earning differentials for Pacific Islanders fare somewhat better. In terms ofthe uncontrolled wage gap, Pacific Islanders make $0.98 on the dollar. Theyare also the only group for whom the racial wage gap completely disappearswhen controlling for compensable factors.Asian workers earn more than white workers in the tech industry. They make$1.23 for every dollar a white work earns. When we look at the controlled wagegap, Asian workers earn $1.02 for every dollar paid to their white peers. Thistightening of the pay gap reflects some degree of occupational segregation,with Asian workers tending to hold higher-paying technical positions.According to the Bureau of Labor Statistics, Asians represent 20.8 percent ofall computer and mathematical occupations even though they represent only 6.2percent of the labor market.While the pay numbers may look good for Asian tech workers, we don’t want tosuggest that they have it easy. First, the “Asian” umbrella is one that coversmany ethnic backgrounds, and how someone from Korean origin is perceived andtreated in the workplace will generally be different than how someone ofFilipino origin is perceived and treated. No doubt that these differences,along with a myriad of other factors, have caused income inequality amongAsians to be greater than for any other racial/ethnic group.Second, while Asians may be overrepresented in tech, they tend to beunderrepresented in middle- and senior-level management positions. Indeed,research suggests that of all demographic groups, Asian women are the leastlikely to reach the top of the ranks in tech.# Software Developer Pay by RaceThe industry-level data that we presented in the previous section includesmany jobs that are not typically considered “tech jobs.” This includes peopleworking in tech companies in non-technical roles such as office managers,content writers and HR specialists.When people talk about tech workers, frequently they mean software developers.To get a deeper look at the racial wage gap in tech, we narrowed ourinvestigation to workers with this highly sought-after title and examined thedistribution of their pay. To get the clearest picture of how pay varies byrace/ethnic group, we limited our sample to early career programmers who metthe following criteria: * Work in the tech industry; * Are individual contributors (i.e. do not supervise others); * Have between 0 and 5 years of experience; and * Have earned a bachelor’s degree and no higher.These restrictions help us get an apples-to-apples comparison while alsohaving a sufficient number of observations. Nevertheless, due to sample sizeconstraints, we used three broad racial/ethnic groups: Asian, non-Asian peopleof color and white. The Asian group includes Native Hawaiian and Other PacificIslanders. Non-Asian PoC includes black, Hispanic and Native American workers.Median Pay for Software DevelopersIndividual Contributors with Bachelors’ Degrees and 0-5 Years of Experience — Race/Ethnic Group | Median Pay | Pay Gap (Compared to white workers) Asian | $84,000 | $1.20 Non-Asian PoC | $68,000 | $0.97 White | $70,000 | N/A When we look at national median pay, white developers make $70,000. Non-AsianPoC average $68,000, which results in a wage gap of $0.97 for every dollar awhite developer makes. Asian developers are the highest paid group, withmedian pay of $84,000 and a positive wage gap of $1.20.Next, we looked at the distribution of pay for each group as a whole, and notjust the medians. This allows us to see nuances in the data that we couldotherwise miss. By looking at the distribution of pay, we can see if, forexample, workers cluster around certain amounts of pay, or if pay skews highor low. The median doesn’t give us as complete of a picture.The pay distribution for Asian developers has more weight to the right (i.e.toward higher pay) than either of the other groups, showing a largerpropensity for high earners. The distribution for non-Asian PoC developers hasmore weight both on the left and the right compared to the distribution forwhite developers, meaning that there are relatively more low- and high-earningnon-Asian PoC developers. The distribution for white developers is moretightly bound around its center. This is partially due to there beingsignificantly more observations for white developers than the other groups,which tends to dampen any noise in the sample.We tested to see whether the underlying distributions for these three groupsare really the same (see Methodology for details). The pay distribution forAsian developers is greater than that of the other two groups, which suggeststhat they make more than their white and non-Asian PoC counterparts. Again, wedo not want to suggest that all ethnic groups that are considered Asian aretreated equally or that they have an easy time in tech. For example, becausewe restrict our sample to non-managers, we cannot address the research wecited above about Asians being less likely to ascend into managementpositions. We also do not control for location, so some of the differences arelikely due to Asian workers being concentrated in high-paying but also high-cost locations like California.When we compared the distributions of pay of white and non-Asian PoC, we foundno evidence that they are different. This is encouraging as it suggests thatblack and brown developers are being paid in-line with their white colleaguesonce they get their foot in the door.# Breaking Inequity in TechThe tech industry prides itself on disrupting old and outdated ways of doingthings, and they have undeniably succeeded in many dimensions. Few peopletoday think of finding a pay phone to make a call. Most of us do not stand ona corner to hail a ride. We have more information at our fingertips thanwhat’s contained in entire libraries. But when it comes to achieving equity inpay, the tech industry as a whole is behind the times.Given the importance of tech in shaping our lives and in driving the globaleconomy, intentional changes that promote pay equity would go a long way tomaking the world a fairer place. The first step tech companies could take isto evaluate their current pay structure and promotion practices and assesswhere there might be gaps. Companies should also communicate transparentlywith employees the criteria for raises and promotions. In an effort to be evenmore transparent, companies could choose to share salary ranges for jobs.Transparent conversations about pay not only lead to more equitable outcomes –they also improve employee engagement.These recommendations are a departure from how many companies normally dobusiness. But tech has never been about doing things the normal way. It’s pasttime to go fast and break racial/ethnic inequities in tech.# Methodology##### Density plotsThese are the kernel density estimation (using R’s default bandwidth) of payby race/ethnicity for over 950 white, 230 Asian and 150 non-Asian PoCdevelopers.To test if the data are pulled from the same underlying distribution, we usedthe Kolmogorov-Smirnov test. The null hypothesis is that the distributions arethe same.Kolmogorov-Smirnov Test Null hypothesis: The underlying distributions are the same — Groups Compared | Alternative Hypothesis | P-Value | Conclusion White & Asian | Distribution for the Asian group is greater than that for thewhite group. | < 0.05 | The distribution of pay for the Asian group is greaterthan that of the white group. White & non-Asian PoC | Distribution for the white group is greater than thatof the non-Asian PoC group. | 0.12 | The distribution of pay for the whitegroup is not greater than that of the non-Asian PoC group. Asian & Non-Asian PoC | Distribution of the Asian group is greater than thatof the non-Asian PoC group. | < 0.05 | The distribution of pay for the Asiangroup is greater than that of the non-Asian PoC group. ##### DefinitionsIndividual Contributor: Employees who do not manage others.Race/Ethnicity: Only respondents who chose exactly one of the below, but not“Prefer Not to Answer,” were included: * American Indian and Alaska Native * Asian * Black or African American * Hispanic * Native Hawaiian and Other Pacific Islander * White * Prefer Not to AnswerSoftware Developers: Includes those who report their job title as “softwaredeveloper” as well as more specific titles such as “Java developer” or “VideoGame Programmer.”Tech Industry: PayScale uses a custom aggregate of the North American IndustryClassification System (NAICS) to define the tech industry.Total Cash Compensation: TCC combines base annual salary or hourly wage,bonuses, profit sharing, tips, commissions, and other forms of cash earnings,as applicable. It does not include equity (stock) compensation, cash value ofretirement benefits, or value of other non-cash benefits (e.g., healthcare).