Digitalization and income inequality: Evidence from households

Tian, Shu; Wu, Yu; Zhou, Wenwen Working Paper Digitalization and income inequality: Evidence from households ADB Economics Working Paper Series, No. 764 Provided in Cooperation with: Asian Development Bank (ADB), Manila Suggested Citation: Tian, Shu; Wu, Yu; Zhou, Wenwen (2025) : Digitalization and income inequality: Evidence from households, ADB Economics Working Paper Series, No. 764, Asian Development Bank (ADB), Manila, https://doi.org/10.22617/WPS250008-2 This Version is available at: https://hdl.handle.net/10419/310439 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/3.0/igo/ ASIAN DEVELOPMENT BANK ASIAN DEVELOPMENT BANK 6 ADB Avenue, Mandaluyong City 1550 Metro Manila, Philippines www.adb.org ADB ECONOMICS WORKING PAPER SERIES NO. 764 January 2025 Digitalization and Income Inequality Evidence from Households This paper examines how digitalization affects income inequality, using household data from the People’s Republic of China. The findings indicate that digitalization significantly reduces income inequality, particularly in less-developed areas and among lower-educated households. The effect remains significant after addressing robustness and endogeneity concerns. Digitalization narrows the income gap by increasing earnings more for lower-income households through improved employment and investment opportunities. It also boosts business income for entrepreneurial households. These results suggest that promoting digitalization can help reduce income inequality in developing economies. About the Asian Development Bank ADB is committed to achieving a prosperous, inclusive, resilient, and sustainable Asia and the Pacific, while sustaining its efforts to eradicate extreme poverty. Established in 1966, it is owned by 69 members —49 from the region. Its main instruments for helping its developing member countries are policy dialogue, loans, equity investments, guarantees, grants, and technical assistance. DIGITALIZATION AND INCOME INEQUALITY EVIDENCE FROM HOUSEHOLDS Shu Tian, Yu Wu, and Wenwen Zhou ASIAN DEVELOPMENT BANK The ADB Economics Working Paper Series presents research in progress to elicit comments and encourage debate on development issues in Asia and the Pacific. The views expressed are those of the authors and do not necessarily reflect the views and policies of ADB or its Board of Governors or the governments they represent. ADB Economics Working Paper Series Shu Tian, Yu Wu, and Wenwen Zhou No. 764 | January 2025 Shu Tian ([email protected]g) is a principal economist at the Economic Research and Development Impact Department, Asian Development Bank. Yu Wu ([email protected]) is a professor at Nanjing Agricultural University. Wenwen Zhou ([email protected]) is a PhD candidate at Southwestern University of Finance and Economics, Chengdu. Digitalization and Income Inequality: Evidence from Households Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) © 2025 Asian Development Bank 6 ADB Avenue, Mandaluyong City, 1550 Metro Manila, Philippines Tel +63 2 8632 4444; Fax +63 2 8636 2444 www.adb.org Some rights reserved. Published in 2025. ISSN 2313-6537 (print), 2313-6545 (PDF) Publication Stock No. WPS250008-2 DOI: http://dx.doi.org/10.22617/WPS250008-2 The views expressed in this publication are those of the authors and do not necessarily reflect the views and policies ofthe Asian Development Bank (ADB) or its Board of Governors or the governments they represent. ADB does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The mention of specific companies or products of manufacturers does not imply that they are endorsed or recommended by ADB in preference to others of a similar nature that are not mentioned. By making any designation of or reference to a particular territory or geographic area inthis document, ADB does not intend to make any judgments as to the legal or other status of any territory or area. This publication is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) https://creativecommons.org/licenses/by/3.0/igo/. By using the content of this publication, you agree to be bound bytheterms of this license. For attribution, translations, adaptations, and permissions, please read the provisions andterms of use at https://www.adb.org/terms-use#openaccess. This CC license does not apply to non-ADB copyright materials in this publication. If the material is attributed toanother source, please contact the copyright owner or publisher of that source for permission to reproduce it. ADB cannot be held liable for any claims that arise as a result of your use of the material. Please contact [email protected] if you have questions or comments with respect to content, or if you wish toobtain copyright permission for your intended use that does not fall within these terms, or for permission to use theADB logo. Corrigenda to ADB publications may be found at http://www.adb.org/publications/corrigenda. Note: ADB recognizes “China” as the People’s Republic of China. ABSTRACT By capturing the adoption of digital services and applications as well as digital industry development, this paper constructs a comprehensive index to measure digitalization at the city level in the People’s Republic of China (PRC) and investigates the impact of digitalization on income inequality using household-level data in the PRC. The findings reveal that a one-unit advancement in the digitalization index significantly reduces the income gap by 1.83% for an average household. This mitigating effect remains statistically and economically significant after addressing endogeneity and robustness concerns. The impact is more pronounced in less-developed areas and among households with lower education levels. Further analysis shows that digitalization narrows the income gap by increasing earnings more for lower-income households, primarily through enhanced employment and investment opportunities. Additionally, digitalization boosts business income for entrepreneurial households. These findings provide valuable policy insights, suggesting that developing economies can reduce income inequality by promoting digitalization, supporting digital-related job creation, and enhancing financial literacy. Keywords: digitalization, inclusiveness, income inequality JEL codes: D30, O10, O30 1. Introduction Income inequality, which is prevalent in both developing and developed countries, is a persistent phenomenon and a fundamental issue of concern. According to the 2022 World Inequality Report released by the World Inequality Lab, the richest 10% of the global population currently accounts for 52.0% of global income, whereas the poorest half of the population earns only 8.5% (Chancel et al. 2023). Rising income inequality is also seen in Asia and the Pacific, though income levels have increased along with rapid economic growth and poverty reduction (Zhuang 2023). With the increasing penetration of digital technology in the lives of residents, digitalization has been embraced by governments because of its potential for promoting economic growth and reducing inequality. Digitalization can transcend geographical limits; offers opportunities to new businesses and easy access to finance; and enhances accessibility to information, goods, and services to reach less-developed regions and vulnerable groups with limited connectivity. However, the lack of digital connectivity and digital literacy can limit people from enjoying the benefits of the digital dividend, which can also widen the income gap between groups. Thus, the influence of digitalization on income distribution is still controversial. On the bright side—digital technologies such as big data, cloud computing, fintech, and online platforms—can promote economic growth and offer opportunities to vulnerable groups, thus increasing household income and narrowing income gaps (Ahmed and Al-Roubaie 2013; Asongu and Odhiambo 2019; Faizah, Yamada, and Pratamo 2021; Demir et al. 2022). Meanwhile, on the negative side, opportunities brought by digital technologies may not be equally accessible to all groups, leading to a significant digital divide that deteriorates income equality (Guellec and Paunov 2017; Daud, Ahmad, and Ngah 2021). However, most of the existing studies are based on macro data at the country level, and it is difficult to deeply discuss heterogeneity and the impact mechanism of digitalization on income inequality. Using household micro survey data, this paper empirically examines the impact of digitalization on income inequality at the household level in the People’s Republic of China (PRC). The evidence from the PRC offers useful policy implications to other developing Asian economies in several ways. First, the PRC has experienced rapid but unequal progress in digitalization.1 The PRC has a variety of digitalization development levels across cities and regions, which offers a good sample to assess how the development of digitalization can affect income gaps. Second, despite an increase in income levels in recent decades, there remains significant inequality in income distribution across the PRC. According to data released by the National Bureau 1 CAICT. 2022. White Papers Research. caict.ac.cn.com/white-papers-research/202220. 2 of Statistics of China, the PRC’s Gini coefficient rose from 0.28 in 1981 to 0.47 in 2022, crossing the international warning line of 0.40 and exceeding the global average. Thus, the PRC offers a good sample to test how digitalization can influence income gaps. Utilizing micro household-level survey data from the China Household Finance Survey (CHFS) covering about 10,000 randomly selected households across approximately 544 cities from 2013 to 2019, this study measures income inequality using the Kakwani index at the household level, following Bossert and D’Ambrosio (2006) and Beck, DemirgüçKunt, and R. Levine (2007). Following Bukht and Heeks (2018), this study constructs a digitalization index at the city level to capture various aspects of digital technology development, including digital industry development, and adoption of digital devices and applications. The analysis results show that a city’s digitalization development can significantly reduce income gaps at household level. In particular, a one unit increase in the digitalization index measure is associated with a 0.0183 reduction in Kakwani index of an average household, which is 1.94% of the sample mean. This means, for example, when the average digitalization development during the sample period increases from 1.06 in Hohhot to 1.68 in Chengdu, the household-level income inequality narrows by 0.011, or 2.04% of an average household in the sample. Moreover, this impact is more pronounced in less-developed regions in the country and households with relatively lower education levels. To address the possible endogeneity issues, we employed the instrument variable approach. This instrument variable is derived by grouping cities with similar economic profiles and calculating the average digitalization index of other cities within the same group. Such an instrument is related to digitalization development as it captures economic development of a typical city in the groups but does not necessarily link to the income inequality of local households of the city that is not part of the instrument. This variable is a strong instrument, and the instrumented digitalization of a city is still significantly and negatively related to income gaps in local households. We also conducted various robustness checks and found robust results on the negative impact of digitalization of a city on its households’ income inequality. To better understand the underlying driver of this effect, this study explores potential mechanisms. There is evidence to show that digitalization is able to increase income levels, especially in less-developed regions and for lower-educated populations. The increase in income is related to employment, investment, and business income. We find evidence that digitalization is related to more job opportunities, especially for lesseducated populations and in less-developed regions, but we did not find significant increases in wages associated with digitalization. Meanwhile, digitalization also fosters participation in financial markets, which offers more investment opportunities to 3 households. And the investment opportunities via access to financial markets apply to all populations, regardless of local development status or education levels. This study contributes to the existing literature in the following way. First, this paper adds to extant literature with a measure that captures more comprehensive information on the level of digitalization development. Existing studies tend to measure digitalization focusing on certain digital technologies, such as digital finance (Daud, Ahmad, and Ngah 2021; Yang and Zhang 2022) and artificial intelligence (Stevenson 2019), or accessibility to digital infrastructure and devices, such as mobile phone and internet penetration rates (Asongu and Nwachukwu 2018, Wang and Xu 2023). Such measures may not have comprehensively gauged the digital transformation in recent years. The measure adopted in this study adds to existing measures with information on development of the digital industry, which presents a more comprehensive picture of digitalization in economic activities. Second, current literature tends to show evidence at the country level, while this paper provides consistent and robust empirical evidence at the household level. Moreover, the empirical estimates obtained from the fixed-effects model are convincing by controlling for heterogeneity across the panels. Lastly, this study lends support to the inclusiveness of digitalization from the perspective of income inequality in less-developed regions and for less-educated populations, and it unveils the possible working channels of this effect. Such evidence has policy implications for peer developing economies in Asia and the Pacific by demonstrating that digitalization may work well as a tool to boost income levels and close income gaps in less-developed areas by fostering job creation. This paper is structured as follows. Section 2 reviews the current literature and outlines the testable hypothesis. Sample construction and research methods are outlined in section 3. Section 4 discusses the empirical results, and section 5 discusses the findings and policy implications. 2. Literature Review The literature on the impact of technological advances on income distribution, as well as income inequality, is inconclusive and offers many different explanations and conclusions. One stream of literature argues that digitalization helps narrow income inequality via the following three channels. First, digitalization enhances connectivity by breaking geographical constraints and making virtual markets and remote workplaces possible 4 (Shaikh and Karjaluoto 2015). Enhanced connectivity opens new business opportunities and enables new jobs for populations facing constraints such as physical location and education levels, and it helps to optimize resource allocation across different regions. For example, emerging digital business models—such as e-commerce, short videos, and online live streaming—provide new job and income opportunities for people from remote areas and low-income groups (Atasoy 2013, Dauth et al. 2017). Second, digitalization enables greater access to services such as finance for vulnerable groups and populations who are domiciling in areas with limited financial resources and services (Jack and Suri 2011; Tchamyou, Erreygers, and D. Cassimon 2019; Demir et al. 2022). By enhancing access to financial services, digitalization enables more investment opportunities and also lowers financing costs, thus narrowing the income gap (Hong, Lu, and Pan 2020; Asongu and Odhiambo 2019). Third, digitalization reduces information barriers for vulnerable groups and facilitates information acquisition and analysis for investment, business, and job opportunities, which could increase their income and narrow the income gap. For example, Hong, Lu, and Pan (2020) find that digitalization allows residents to be exposed to various digital application scenarios, which can effectively alleviate information asymmetry and reduce risk aversion among residents, thereby promoting their participation in financial market investment; this effect is more visible among vulnerable groups in alleviating income inequality through income gains. Meanwhile, other scholars argue that digitalization can widen income inequality (Yeo, Hwang, and Lee 2023). This is because the Schumpeterian type of creative destruction brought by technological advances does not bring the same opportunities to different groups. For example, high-income and better-educated people have better chances to apply the latest digital technology and enjoy digital dividends, thus widening the gap with low-income and less-educated groups (Acemoglu and Restrepo 2020; Daud, Ahmad, and Ngah 2021). Moreover, skilled workers may adapt to new technology and learn new knowledge more quickly compared to unskilled workers, leading to a widening of the income gap (Guellec and Paunov 2017; Acemoglu and Restrepo 2018; Tica, Globan, and Arčabić 2022). In addition, algorithmic mechanisms, based on digital technologies such as big data and machine learning, may strengthen the supply discrimination in credit (Philippon 2016)— that is, higher-income and better-educated people have greater access to low-cost financial products and digital technologies, thereby contributing to a widening of the income gap. There is also evidence that a nonlinear relationship may exist between income inequality and technological progress. On the one hand, according to the technological Kuznets curve hypothesis proposed by Kim (2012), income inequality first rises and then declines with technological progress (i.e., an inverse U-shaped relationship). In the initial stage of technological innovation, the cost of adopting a new technology is relatively large and 11 as a percentage of GDP. Within each group, we further divide cities into three groups by ranking their GDP. This exercise results in nine groups of cities. Within each group, we exclude the city where a household is located and compute the mean value of the digitalization index of other cities in the group to be the instrument variable. To ensure the robustness of the results, we also repeat this exercise using a 5 by 5 grouping and obtain an instrument variable based on the 25 groups (5 x 5) of cities. To further exclude possible similarity of cities from the same province, we reconstructed the instrumental variable by excluding cities from the same province in sorted city groups as an extra robustness test. The instrument that captures digitalization of similar cities may be related to the digitalization level of a city but may not affect the household income distribution of this city directly. Table 4: Digitalization and Income Inequality (1) (2) KKWN Index KKWN Index Digitalization index –0.0183*** –0.0220*** (0.0056) (0.0067) Digitalization index sq. 0.0029 (0.0022) Age –0.0075*** –0.0075*** (0.0013) (0.0013) Age sq. 0.0078*** 0.0077*** (0.0012) (0.0012) Household size 0.0087*** 0.0087*** (0.0018) (0.0018) Children ratio –0.0437*** –0.0432*** (0.0148) (0.0148) Old ratio –0.0195** –0.0196** (0.0096) (0.0096) Unhealthy ratio 0.0416*** 0.0414*** (0.0100) (0.0100) Ln (household total asset) –0.0251*** –0.0251*** (0.0010) (0.0010) Ln (per capita GDP) 0.0333* 0.0364* (0.0196) (0.0198) Year FE Y Y Household FE Y Y N 107671 107671 R 2 0.014 0.014 Adj. R2 0.014 0.014 Continued on the next page 12 FE = fixed effects, GDP = gross domestic product, N = number of observed values. Notes: *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. The standard errors are clustered at the city-year level to ensure heteroscedasticity robustness. Source: Authors’ calculations. The results using the instrumental variable are reported in Table 5. The results of the firststage regression in columns (1) and (3) suggest that the instrumental variable is significantly associated with digitalization of the city. And the second-stage results shown in columns (2) and (8) demonstrate that the instrumented digitalization is still significantly and negatively associated with household income inequalities, confirming the findings from the baseline regression in Table 4. To mitigate the influence of linear trend factors, we also use first-order differences of variables. The results from column (9) of Table 5 indicate that, even after first-order differencing, the digitalization index remains significant and negatively correlated with income inequality. To ensure robustness of the above results, we conduct various additional tests. Column (1) of Table 6 reports the first set of robustness check where we replace the digitalization index from previous period in the baseline with the digitalization in the current period. As shown, the coefficient remains significant and negative for the Kakwani index, which is consistent with the finding stated earlier. Other robustness checks include: 1) computing the Kakwani index using total household income rather than per capita income of the households in Column (2); 2) removing the top and bottom 5% of households by per capita income, then recalculated the Kakwani index for the remaining households in column (3); 3) excluding households that suffered from major events in the recent 5 years from our sample in column (4); and 4) excluding rural samples that may have a large gap with urban households to avoid the results being driven by rural households in column (5). Overall, these robustness checks generate consistent findings that confirm that digitalization is significantly and negatively correlated with income inequality. To further understand the source of the impact of digitalization on income inequality, we run separate regression of individual digitalization indicators on income inequality and report the results in Table 7. As shown, the results are largely driven by development in digital industry, such as the patent licenses in the 5G industry, the number of digital enterprises, and the number of digital enterprises. The literature shows that digitalization may have a heterogeneous impact on households in different economic development regions and with different educations levels. (Knight 2014, Zhang 2021, and Li, Wu, and Xiao 2020) In empirical tests, we divide the sample into different subsamples and explore how digitalization can affect income inequality differently. Columns (1) and (2) of Table 8 report the impacts of digitalization on income inequality in the more advanced regions of the eastern provinces and the relatively less- 13 developed central and western areas provinces. As shown, while the coefficients of digitalization are significantly and negatively associated with the Kakwani index in both subsamples, the impact is more pronounced for less developed central–western area. In columns (3) and (4) of Table 8, we examine how digitalization may affect household income inequality with different education levels. In doing this, we divide the sample into two groups based on the median of average schooling years of the household labor force in 2017. The results show that the effect of digitalization on the Kakwani index is more pronounced in households with less education than in households with more education. Overall, Table 8 shows that the impact of digitalization on income inequality is more pronounced in less-developed areas and for households with lower education levels, further shedding light on the inclusiveness benefit of digitalization. 4.2 How Does Digitalization Narrow Income Inequality? The previous session shows that digitalization can narrow income inequality, especially in less-developed areas and among less-educated households. It is natural to ask how digitalization can reduce income inequality. This section answers the question by looking at how digitalization affects income levels of households. Table 9 reports the estimated results of digitalization on household income, by replacing income inequality with household income in the baseline model of equation (2). It shows that city-level digitalization is significantly and positively associated with household income across different quantile household income levels. The effect is both statistically and economically significant. Moreover, the magnitude of coefficients increases as income level decreases. Overall, Table 9 shows that digitalization positively affects household income, and the impact is more pronounced for lower income households, thus narrowing income inequality. To check further whether the finding is consistent with the previous session, we repeat the test in Table 9 by examining the results in less-developed areas and less-educated households. The results are reported in Table 10. Consistently, we find that a stronger positive impact of digitalization on household income in less developed central–western regions in columns (1) and (2), and for less-educated households in columns (3) and (4). These additional results confirm that the impact of digitalization on household income is more pronounced for disadvantaged groups, thus reducing income inequality. 4.3 Working Mechanisms In this section, we further explore the possible working mechanism in which digitalization may affect income distribution. First, we breakdown household income into three major sources: (i) wage income from employment, (ii) business income from commercial activities, and (iii) investment income from investing in financial markets. Table 11 presents the results on how digitalization is related to different types of household income, 14 showing that digitalization is significantly and positively related to all three types of household incomes. To further investigate how digitalization could boost household wage income, we examine the impact of digitalization on employment opportunities. Table 12 reports the estimated relationship between the digitalization level of a city and household labor force participation using a Probit model. Results show that digitalization is positively and significantly associated with household-level employment status, and the impact is more pronounced for households in the less-developed regions and with lower-education levels. Besides employment status, it is also interesting to know what types of jobs are related to digitalization. We analyze the relationship between digitalization and employment types and report the results in Table 13. As shown, digitalization is significantly related to jobs in private companies and MSME business. This implies that the role of digitalization enables flexible and new jobs in the private sector as well as MSMEs. To understand how digitalization boosts household investment income, we examine the association between digitalization and new financial market participation of households. A household is considered a participant in the financial markets if that household holds any financial assets in the form of stocks, funds, bonds, derivatives, or gold. If a household did not participate in the financial market in the previous year but participated in the current year, it is defined as a new participant of the financial market. As shown in Table 14, digitalization is not significantly positively correlated with overall financial market participation in the full sample. However, among households in the less developed central–western regions and those with lower education levels, digitalization is significantly and positively correlated with new financial market participation. This indicates that digitalization offers new income opportunities by enabling accessibility to financial markets and investment opportunities for less-privileged households. This provides additional evidence of the inclusive characteristics of digitalization. 5. Conclusion This study examined how digitalization is associated with household income inequality. By constructing a comprehensive index to capture digitalization and utilizing household level data from the CHFS in 2013, 2015, 2017, and 2019, this paper documents a positive role of digitalization in reducing income inequality at the household level. This income inequality reduction impact is more pronounced in the less-developed central–western regions and among less-educated households. This impact on income inequality remains statistically and economically significant after addressing endogeneity concerns and a few robustness checks. 15 We further demonstrate how digitalization reduces income inequality. Evidence shows that digitalization has a significant and positive impact on household income. The income promotion impact is larger for lower-income households and households in less developed areas and with lower education levels, thereby reducing income inequality. This further confirms the inclusiveness benefit of digitalization. Furthermore, evidence shows that digitalization boosts household incomes by boosting wage income, business income, and investment income. This study provides micro-level evidence for policymakers to speed up digitalization and promote income opportunities for disadvantaged groups, narrowing income inequality. Our findings imply that digitalization can promote inclusive development, with a larger marginal effect for less-developed areas and households with lower levels of education. Promoting digitalization in less-developed areas will help boost inclusiveness. Policymakers may strengthen digitalization, particularly in remote and underdeveloped areas, to deliver inclusive benefits. 16 Table 5: Addressing Endogeneity in the Estimation (1) (2) (3) (4) (5) (6) (7) (8) (9) First stage Second stage First stage Second stage First stage Second stage First stage Second stage First-order difference estimation Digitalization index KKWN index Digitalization index KKWN index Digitalization index KKWN index Digitalization index KKWN index Difference of KKWN index Digitalization index in 9 subgroups 0.6213*** (0.0040) Digitalization index in 25 subgroups 0.7932*** (0.0028) Digitalization in 9 subgroups excluding cities from the same province 0.5433*** (0.0039) Digitalization index in 25 subgroups excluding cities from the same province 0.6787*** (0.0029) Instrumented Digitalization index –0.0190*** –0.0191*** –0.0202** –0.0323*** (0.0038) (0.0039) (0.0088) (0.0088) Difference of Digitalization index –0.0155* (0.0081) Year FE Y Y Y Y Y Y Y Y Y Household FE Y Y Y Y Y Y Y Y Y Controls Y Y Y Y Y Y Y Y Y N 107,671 107,671 107,671 107,671 107,671 107,671 107,671 107,671 48,957 R2 0.703 0.792 0.692 0.757 0.011 Adj. R2 0.703 0.792 0.692 0.757 0.010 FE = fixed effects, N = number of observed values. Notes: *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. The standard errors are clustered at the city-year level to ensure heteroscedasticity robustness. Source: Authors’ calculations. 17 Table 6: Robustness Check—Digitalization and Income Inequality (1) (2) (3) (4) (5) KKWN Index Based on Income per Capita Based on Total Income Removed 5% Top and Bottom Income of Households Excluded Households that Suffer from Major Events Exclude Rural Sample Digitalization index –0.0119** –0.0189*** –0.0091*** –0.0234*** (0.0052) (0.0056) (0.0033) (0.0062) Digitalization in contemporary period –0.0165** (0.0088) Year FE Y Y Y Y Y Household FE Y Y Y Y Y Controls Y Y Y Y Y N 82,034 107,541 97,894 98,995 72,417 R2 0.008 0.015 0.004 0.005 0.015 Adj. R2 0.008 0.015 0.004 0.004 0.015 FE = fixed effects, N = number of observed values. Notes: ** and *** denote significance at the 5% and 1% levels, respectively. The standard errors are clustered at the city-year level to ensure heteroscedasticity robustness. Source: Authors’ calculations. 18 Table 7: Digitalization indicators and Household Income Inequality (1) (2) (3) (4) (5) (6) KKWN Index Mobile Telephone Subscribers Telecommunicatio n service revenue Digital Payment Coverage Number of digital enterprises Digital industry employment Patent licenses in the 5G industry Subindex of Digitalization index –0.0156 0.0106 –0.0494 –0.0802*** –0.0742*** –0.1427** (0.0297) (0.0276) (0.0696) (0.0239) (0.0230) (0.0615) Year FE Y Y Y Y Y Y Household FE Y Y Y Y Y Y Controls Y Y Y Y Y Y N 107,671 107,671 107,671 107,671 107,671 107,671 R2 0.013 0.013 0.013 0.014 0.014 0.013 Adj. R2 0.013 0.013 0.013 0.014 0.014 0.013 FE = fixed effects, N = number of observed values. Notes: *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. The standard errors are clustered at the cityyear level to ensure heteroscedasticity robustness. Source: Authors’ calculations. Table 8: Digitalization and Income Inequality—Heterogeneity Tests (1) (2) (3) (4) KKWN Index Eastern Region Central and Western Regions High Education Households Low Education Households Digitalization index –0.0108* –0.0335*** –0.0178*** –0.0221*** (0.0064) (0.0116) (0.0062) (0.0077) Year FE Y Y Y Y Household FE Y Y Y Y Controls Y Y Y Y N 53,562 54,109 43,961 33,411 R2 0.015 0.014 0.003 0.002 Adj. R2 0.015 0.013 0.003 0.002 FE = fixed effects, N = number of observed values. Notes: * and *** denote significance at the 10% and 1% levels, respectively. The standard errors are clustered at the city-year level to ensure heteroscedasticity robustness. Source: Authors’ calculations. 19 Table 9: Digitalization and Household Income (1) (2) (2) (3) (4) (5) Ln (household income per capita) Full sample Q10 Q25 Q50 Q75 Q90 Digitalization index 0.0622*** 0.2550*** 0.1304*** 0.0621*** 0.0623*** 0.0567*** (0.0142) (0.0268) (0.0154) (0.0107) (0.0100) (0.0140) Year FE Y Y Y Y Y Y Household FE Y Y Y Y Y Y Controls Y Y Y Y Y Y N 106,226 106,098 106,098 106,098 106,098 106,098 R2 0.032 Adj. R2 0.032 FE = fixed effects, N = number of observed values. Notes: *** denotes the significance at the 1% level. The standard errors are clustered at the city-year level to ensure heteroscedasticity robustness. Source: Authors’ calculations. Table 10: Digitalization and Household Income—Heterogeneity Tests (1) (2) (3) (4) Ln (household income per capita) Eastern Region Central and Western Regions High Education Households Low Education Households Digitalization index 0.0399** 0.0994*** 0.0371*** 0.0899*** (0.0161) (0.0333) (0.0128) (0.0201) Year FE Y Y Y Y Household FE Y Y Y Y Controls Y Y Y Y N 52,888 53,338 43,577 32,764 R2 0.064 0.060 0.064 0.065 Adj. R2 0.064 0.060 0.064 0.064 FE = fixed effects, N = number of observed values. Notes: *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. The standard errors are clustered at the city-year level to ensure heteroscedasticity robustness. Source: Authors’ calculations. 20 Table 11: Digitalization and the Categories of Household Income (1) (2) (3) Ln (wage income per capita) Ln (business income per capita) Ln (investment income per capita) Digitalization index 0.0698*** 0.1052** 0.3197*** (0.0179) (0.0484) (0.0727) Year FE Y Y Y Household FE Y Y Y Controls Y Y Y N 106,098 15,662 106,098 R2 0.041 0.059 0.296 Adj. R2 0.041 0.058 0.296 FE = fixed effects, N = number of observed values. Notes: *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. The standard errors are clustered at the city-year level to ensure heteroscedasticity robustness. Source: Authors’ calculations. Table 12: Digitalization Development and Employment Status (1) (2) (3) (4) (5) Employment Status Whole Sample Eastern Region Central and Western Regions High Education Group Low Education Group Digitalization index 0.0184*** 0.0142* 0.0357** –0.0023 0.0277*** (0.0066) (0.0074) (0.0144) (0.0070) (0.0092) Year FE Y Y Y Y Y Individual FE Y Y Y Y Y Controls Y Y Y Y Y N 149,231 75,917 73,314 74,586 75,848 R2 0.008 0.008 0.008 0.016 0.005 Adj. R2 0.008 0.008 0.008 0.016 0.005 FE = fixed effects, N = number of observed values. Notes: *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively. The standard errors are clustered at the cityyear level to ensure heteroscedasticity robustness. Source: Authors’ calculations.