The effect of time-saving household appliance ownership on outcomes for children and married women: evidence from India

Appliance ownership in India has increased during recent decades, owing, in large measure, to rising disposable incomes, urbanization, increasing organized retail, easier financing options, growing demand, and increasing electrification (Ernst & Young, 2015). The proportion of Indian households owning a television increased from 47% in 2011 to 65% in 2018. Refrigerators are the second most prevalent household asset, followed by washing machines.

https://www.livemint.com/Specials/bhWpWqj3AFuETVdsC05fdM/In-India-washing-machines-top-computers-in-popularity.html Source: Livemint, “In India, Washing Machines Top Computers in Popularity.”

In this analysis, we follow Bowden and Offer (1994) by distinguishing between two types of household appliances: time-saving and time-using. Time-saving technologies embody appliances that reduce time required to perform household work. In the IHDS data, available time-saving appliances include washing machines, refrigerators, microwave ovens, electric cooking pots, and pressure cookers. Time-using goods, on the other hand, enhance the quality of discretionary time. Examples available in the IHDS data include air conditioners, air coolers, cameras, color televisions, and electric fans.

In this study, we focus on two large time-saving appliances: washing machines and refrigerators. Ownership of these appliances provides households the opportunity to substitute between capital and labor in home production. When a household purchases a time-saving appliance, household members have the opportunity to reallocate time to leisure, schooling, or market work. Previous literature on durable goods ownership has found that the presence of time-saving durable goods increases adult female LFP in the United States (Greenwood et al., 2005; Coen-Pirani et al., 2010), China (Tewari and Wang, 2016), Latin American countries (Cubas, 2016), Nigeria (Omotoso and Obembe, 2016, 2017), and OECD countries (Cavalcanti and Tavares, 2008). These same appliances are found to increase school enrollment and decrease child LFP in China (Kerr, 2019). Time-conserving appliances provide women and children with more efficient ways to perform domestic tasks, making them more productive in the household, and thus increasing their intrafamily bargaining power.

This paper adds to the literature by studying the effect of household appliances on Indian adult females and children. With the drastic changes in appliance ownership occurring in India, as described above, and the cultural differences between India and the countries studied in previous research, exploring the relationship between appliance ownership and female and child outcomes within the Indian context is an important contribution to current literature.

As described above, appliance ownership has been found to be associated with increases in female LFP in multiple developed and developing countries. While prevalence in these same appliances has increased in India, India's female LFP has been declining while remaining visibly low. The ILO ranked India as the country with the 28th lowest female LFP in 2019, as female LFP declined from 31% in 1990 to 23.4% in 2019.

These are estimates of the participation rate of the female labor force (percentage of the female population aged ≥15 years).

One hypothesis as to why female LFP in India has declined even with extreme economic growth is that the Hindu culture and the caste system create a need for differential preferences for women. While Chinese women have increased their LFP with increased technological advancements, Indian women face different obstacles. India's average household size is 1.2 persons larger than China's, suggesting that family resources may be more diluted and require a larger time commitment at home for women.

https://population.un.org/Household/#/countries/840, accessed April 2020.

Understanding the underlying trends in female LFP is important to those interested in women's well-being. Beyond women's contribution to growth, being economically active affects women's progression toward economic independence, bargaining power, and their children's overall well-being (Klasen and Pieters, 2015; Mammen and Paxson, 2000). With India's decrease in female LFP, it is important to explore whether the appliances found to improve female outcomes in other countries also improve female outcomes in a country that is currently experiencing a consistent decline in female LFP. This study thus helps understand the relationship between appliance ownership and female LFP in the Indian context.

The education system in India, also known as the “10 plus 2” system, consists of 4–5 years of primary education, 3–4 years of middle education, 2–3 years of secondary education, and 2 years of senior secondary education. Children begin primary education by the age of 5 years.

According to Government of India's Ministry of Human Resource Development, the total gross enrollment ratio (GER) at the primary level increased from 109.4 in 2005 to 116 in 2012.

Gross Enrollment Ratio (GER) is defined as the total student enrollment in a given level of education, regardless of age, expressed as percentage of the corresponding eligible official age-group population in a given school year. This allows the GER to be >100%.

Kerr (2019) indicates that the presence of time-saving household appliances alleviates children from household work and provides them the opportunity to reallocate time to schooling in China. With the low secondary and senior secondary enrollment rates present in India, it is useful to explore whether these same appliances can help initiate similar changes in school enrollment rates.

The data used in this analysis come from the IHDS. This survey was conducted by the University of Maryland and the National Council of Applied Economic Research, New Delhi. The IHDS is a nationally representative, two-wave longitudinal data set consisting of a first round of interviews completed in 2004–2005 and a second round of interviews, mostly reinterviews of households from the first wave, completed in 2011–2012. The first wave contains data on 41,554 households in 1,503 villages and 971 urban neighborhoods across India, and the second wave contains data on 42,152 households in the same villages and urban neighborhoods as the first wave.

For a full description of the survey design, see: https://ihds.umd.edu/IHDS.

Our first sample consists of 52,046 married women of working age, with 73,916 observations over the 2 years of data. Table 1 presents the descriptive statistics for this sample, partitioned by household ownership of a washing machine and a refrigerator. Less than 10% of the sample lives in a household that owns a washing machine. Refrigerator ownership is more prevalent, with nearly 30% of the sample living in a household that owns a refrigerator.

The summary statistics are similar regardless of whether we partition the sample by washing machine or refrigerator ownership. Thus, we discuss general differences between appliance ownership rather than washing machine and refrigerator ownership separately.

Our variable of interest in the sample of married women is employment status. Since the IHDS does not provide data on LFP, we construct the variable using information on different work categories. The work categories include the following: work on family farm, work tending to the household's animals, family's nonfarm business work, agricultural wage labor, nonagricultural wage labor, and salaried position. Participation in each category is given a value of one if the individual spent ≥240 hours in that activity in the past year and zero otherwise. Employment status is indicated using a dummy variable, equaling one if the woman participates in any of the work categories listed above and zero otherwise. It appears that married women who live in households that own time-saving appliances are less likely to be employed than women who do not live in households that own a time-saving appliance.

In addition to studying the effect of time-saving appliance ownership on the overall employment rate, we analyze the effect of time-saving appliance ownership on different types of employment, as some employment types are viewed as more valuable than others. For example, Desai et al. (2010) suggest that salaried work (typically found outside the home) is at the top of the job ladder, while manual labor positions (typically found inside the home) are valued much lower. We use IHDS data on each individual's work category, listed above, to classify whether an individual's employment is in work inside or outside the household. Using the definitions provided by the IHDS, we construct the variable work inside the household by combining the three work categories of animal care, farm work, and nonfarm family business work. These are mainly manual labor roles and are typically unpaid or low paid. Similarly, we construct the variable work outside the household for household members who work for payment (wage or salary) by combining the three work categories of agricultural wage labor, nonagricultural wage labor, or salaried position worker. These positions are viewed as more valuable, as they typically come with higher pay than for work inside the household. Our categories follow the categories used in previous literature (such as by Desai et al., 2010).

Table A1 in the Appendix presents the trends in employment rate separately for married and single women in the working-age population. The overall employment rate for married women declined from 52% in 2005 to 48% in 2012. At the same time, the overall employment rate for single women declined from 36% in 2005 to 34% in 2012. However, the table indicates that the decline in employment rate observed for married and single women occurs in all work inside the household job types. Married and single women experience an increase in two of the work roles outside the household (nonagricultural labor and salaried position) between 2005 and 2012, suggesting that the breakdown by employment type in the empirical analysis is warranted.

It is beneficial to examine the differences between married and single women beyond their employment decisions as we will use average appliance ownership rate among single women to estimate a married woman's ownership of an appliance, described in the following section. Table A1 in the Appendix presents additional summary statistics for married and single women separately. The proportion of women living in urban areas has increased over our two-wave sample for both married and single women, with a higher percentage of single women living in urban areas in each wave. Married women are older than single women, are more likely to be Hindu, and are more likely to have more kids living in the household than single women. When looking at education-related variables, we see that single women have more years of schooling, on average, and are more likely to be literate than married women. Wealth indices for both groups have increased over the two waves of data and are similar in size. While there are many differences between the married- and single-women samples, both groups have seen an increase in washing machine ownership from roughly 6% to 12% and an increase in refrigerator ownership from 24%–25% to 34%–36% (not shown in the table), suggesting that household appliance ownership decisions may be due to something other than marital status.

Given the higher prevalence of refrigerator ownership, we additionally present differences among women across waves partitioned by refrigerator ownership in Table A2 in the Appendix.

To understand how an appliance affects a household as a whole, we additionally perform our empirical analysis, described in the following section, on a child sample consisting of 50,160 children over the 2 years of data. Descriptive statistics for the child sample are presented in Table 2, which follows the same format as Table 1. We analyze children between the ages of 12 years and 18 years, as the enrollment rate begins to decline from 96% to 92% and the employment rate doubles from 3% to 6% at the age of 12 years in the IHDS data. We observe that primary school completion rate is 10 percentage points higher for children living in households that own an appliance than children living in households that do not own an appliance. In addition, these appliance-owning children are more likely to be enrolled in school and less likely to be employed. Children who live in a household that does not own an appliance are more likely to be from a SC or ST background and from a rural area. Similar to married women, most children who live in a household that owns a washing machine also own a refrigerator and color television. Parents’ ages appear to be similar across children who live in households with and without appliances, but both parents possess a higher number of years of schooling for those children who live in a household that owns an appliance. Not surprisingly, the table suggests a higher index of wealth among appliance owners.

One concern with estimating the effect of household appliances on individual outcomes is that the appliances considered in this analysis are a large purchase and are typically available to wealthier households. These appliances might pick up a wealth effect, whereby wealthier households simultaneously purchase appliances and alter their daily activities. To address this concern, we control for wealth using a wealth index, which uses detailed data and places households in the analysis on a continuous scale of relative wealth. This control assists with distinguishing between the wealth effect and the true effect of the time-saving appliance on LFP and schooling. Wealth indices are commonly used in research in developing countries. The Demographic and Health Surveys (DHS) regularly create wealth indices when studying health services, as “the distribution of health services to the poor can be determined by a wealth index as well as or better than an income or expenditure index. This is because of the lower volatility of wealth as compared with that of income and expenditures” (https://dhsprogram.com/pubs/pdf/cr6/cr6.pdf). Their logic follows when studying other services available in developing countries.

We use information on each household's ownership of time-using appliances and household infrastructure to construct a wealth index. A detailed description of the construction of this index is available in Kerr (2019). Briefly, we use a factor analysis approach and assign factor scores to household assets and infrastructure (see Kennedy [2008, p. 200] for an explanation on factor analysis). We then combine the scores by household to rank each household's overall wealth. The time-using assets in our wealth index include the following: air cooler, black- and-white television, bicycle, cable, car, cellphone, chair, computer, cot, generator, laptop, mixer, motorcycle, sewing machine, shoes, telephone, and two sets of clothes.

Air conditioner and color television are not included in our wealth index, as these two appliances are included in our instrument, which is described in the following section.

Household infrastructure variables include discrete indicators for access to drinking water and toilet facilities; principal sources for water; and the types of chulha,

A chulha is a traditional Indian stove.

We use these infrastructure variables and asset ownership indicator variables to estimate the factor scores for each household in each year of the survey. Again, following Kerr (2019), we sum the factor scores within each household each year to create an index between the values 0 and 3.18, with 0 representing the least wealthy households and 3.18 representing the wealthiest households. We compare the index with available household income to verify that the households that report higher incomes also show higher wealth indices.

Summary statistics presented above suggest differences in married women's employment and children's schooling and employment, which appear to be identified on the basis of ownership of durable appliances. The purpose of the empirical analysis is to ascertain the extent to which this difference is sustained in the multivariate context of Eqs (1) and (3), described in the following section.

As a point of departure, Table 3 presents the OLS estimates of Eq. (1), with appliance ownership defined as the presence of a washing machine in Columns 1 and 2 and the presence of a refrigerator in Columns 3 and 4. Living in a household that owns a washing machine increases married women's probability of employment by 3 percentage points (Column 1 of Table 3). Refrigerator ownership is found to be insignificant in married women's employment decisions (Column 3 of Table 3). Estimates obtained using PSU-level controls (which estimate the effect of appliance ownership on married women living in villages only) indicate insignificant changes in employment with ownership of either appliance (Columns 2 and 4 of Table 3).

Some of the control variables provide further insight into the Indian setting, where women appear to be reducing their LFP when a concomitant increase in other wealth-indicating characteristics occurs, which is inconsistent with general intuition. Our results suggest that women decrease employment when years of schooling increase and when the household becomes wealthier (as measured through the wealth index). This finding is consistent with previous literature (Bhargava, 2020): Indian women exhibit negative income elasticity of labor supply, which dominates over their positive own-wage elasticity of labor supply. Among other controls, age, rural status, and Hindu status are significant in all cases.

The directions of these effects are plausible and consistent with previous literature on female labor supply in India. See Bhargava (2020), Klasen and Pieters (2015), Afridi et al. (2018), and Neff et al. (2012), among others.

Tables 4 and 5 present the OLS estimates of Eq. (1) for the child sample for washing machines and refrigerators, respectively. Columns 1–6 and 7–12 indicate the results for enrollment and employment, respectively. We find an insignificant increase in the probability of being enrolled and an insignificant decrease in the probability of being employed in nearly all specifications for both appliances. The wealth index is highly significant in all models and for both genders. Unlike the married-female sample, the direction of the effect follows intuition and is plausible: households characterized by greater wealth tend to (1) more actively enroll children in school and (2) engage children to a lesser extent in market work. Among the controls, parents’ years of schooling are significant in all cases, with signs that largely match those of the wealth index. The remaining controls are only sporadically significant, yet their inclusion is useful as a general principle in isolating the key independent variable, appliance ownership, after controlling for household and PSU characteristics.

The estimates presented in Tables 4 and 5 suggest that time-saving household appliances do not consistently exert significant effects on employment or enrollment for children. However, the important caveat is the possible endogeneity of appliance ownership. We address this by using a 2SLS estimation of Eq. (1), the results of which are discussed in the following sections.

To address the concern of appliance ownership being endogenous, we begin our 2SLS estimation by estimating Eq. (2), which regresses household ownership of time-saving appliances for married women against average PSU ownership of time-saving appliances for single women, ownership of an air conditioner, ownership of a color television, and the set of controls included in our OLS estimation. The model is estimated separately for ownership of washing machines and refrigerators.

Estimates of Eq. (2) are compressed to the first four rows of Table 6 and suggest that the likelihood of owning either appliance increases for married women when the average PSU ownership of the same appliance increases for single women living in the same PSU.

Tables including first-stage results that present all controls are available upon request.

The remaining rows of Table 6 present our 2SLS results using Eq. (3) for married women. Columns 1 and 2 suggest insignificant effects of washing machine ownership on married women's employment. In contrast, we find that married women living in households that own a refrigerator decrease their probability of employment by 10 percentage points (Columns 3 and 4 of Table 6). This finding is significant in both the full sample (Column 3) and the village sample (Column 4) and does not follow most intuition or previous studies. The results suggest that time-saving appliances allow women to reduce their employment rather than increase it. We will explore this finding later in the paper. Similar to the OLS results, we find statistically significant effects of age, years of schooling, rural, Hindu, and the wealth index.

For robustness, we generate another index by regressing appliance ownership on all household assets/infrastructure variables from the original wealth index, and then, we use the residual in our final specification as an alternative index to control for household wealth. The coefficients on washing machine and refrigerator ownership estimated using this strategy are larger than our main results and are highly significant. These results further validate our results and are available upon request. We thank an anonymous reviewer for suggesting this robustness check.

Before moving on to the child sample, we split our sample of married women on the basis of their education to further distinguish the human capital effect. Table A3 in the Appendix presents these results, following the same empirical strategy as above, separately for women who did not complete primary education and for women with an education level of primary and above. The results presented in Table A3 in the Appendix are consistent with those presented in Table 6, suggesting insignificant effects of washing machine ownership for both education levels and significant effects of refrigerator ownership on the probability of married women's employment for both education levels. The coefficient is larger for the sample of married women who did not complete primary (−18.4% in Column 5 compared to −13.1% in Column 7 of Table A3 in the Appendix), suggesting that these time-saving appliances have a larger impact on the probability of being employed for less-educated married women.

We thank an anonymous reviewer for suggesting this robustness check.

First-stage estimates using Eq. (2) for the child sample are presented in the first three rows of Tables 7 and 8. Similar to the married women sample, the results suggest that ownership of time-saving appliances among households with children increases as the average PSU ownership among households with no children in the same PSU increases.

The remaining rows of Tables 7 and 8 present the estimates of Eq. (3) for ownership of a washing machine and a refrigerator, respectively. Table 7 is consistent with the findings presented in Table 6, suggesting that washing machine ownership does not significantly alter enrollment or employment decisions for children in the sample. One notable difference is found in the village sample of female children. These individuals experience a 16.2 percentage point decrease in the probability of being employed when living in a household that owns a washing machine, suggesting that washing machines alleviate some of the employment opportunities or needs for older female children. An analysis using time-use data would be beneficial to further understand why female children experience this significant change. However, time-use data are not available in this data set.

Observing refrigerator ownership in Table 8, we find statistically significant effects on both enrollment and employment decisions for children. Beginning with enrollment, we observe a 15.3 percentage point increase in the probability of being enrolled when a child lives in a household that owns a refrigerator (Column 1 of Table 8). These significant results hold when the sample is partitioned by gender (Columns 3 and 5). We find that the effect of refrigerator ownership on enrollment is larger for males than for females. Males (females) living in a household that owns a refrigerator experience an increase of 17.5 (12.7) percentage points in the probability of being enrolled. These refrigerator-owning children also experience a 9.1 percentage point decrease in the probability of participating in the labor force (Column 7 of Table 8). Again, the results hold when the sample is partitioned by gender (Columns 9 and 11). Females, however, experience a larger decrease in the probability of employment than males.

It is interesting to note that, while the full child sample presents significant findings on refrigerator ownership, performing the analysis on villages provides suggestive evidence that purchasing an appliance is not sufficient in altering labor force decisions for children (Columns 8, 10, and 12 of Table 8). One reason for this finding may be that refrigerator ownership is less prevalent in rural villages (17% ownership in villages compared to 26% ownership in the full sample). We do find a statistically significant increase in the effect of refrigerator ownership on enrollment for the village sample (Columns 2, 4, and 6 of Table 8). These results hold when partitioning the sample by gender. Males and females from the village sample experience a 16.1 percentage point and 13.5 percentage point increase, respectively, in the probability of being enrolled if they live in a household that owns a refrigerator.

Overall, results using the child sample indicate that refrigerator ownership significantly alters both education and employment decisions for older children. When focusing on village-level data, refrigerators increase the probability of enrollment but do not initiate a significant change in employment decisions for older children. Washing machines do not cause similar, significant changes in enrollment or employment decisions.

The results from the married women sample seem perplexing at first, as most research finds appliance ownership to be liberating, allowing women to reduce household work and enter the labor force.

For example, see Coen-Pirani et al. (2010) and Greenwood et al. (2005).

Table 9 presents the effect of washing machine ownership on married women's employment segmented by work inside and outside the household (Columns 1–2 and 3–4, respectively). We find that married women living in households that own a washing machine experience a decrease in the probability of being employed in work inside the household by 6.8 percentage points in the full sample (Column 1). Employment in work inside the household is reduced by 14.7 percentage points in the village sample (column 2). We find that married women's probability of employment in work outside the household increases by 15.4 and 8 percentage points in the full and village samples, respectively (columns 3 and 4).

Table 10 presents similar results of owning a refrigerator on married women's employment by different work types. The results suggest that ownership of a refrigerator decreases the probability of married women's participation in work inside the household by 14.5 percentage points in the full sample (Column 1). These results hold when we extend our analysis to the village sample, with married women's probability of being employed in work inside the household being reduced by 10.6 percentage points in the village sample (Column 2). Similar to the washing machine-related findings, refrigerator ownership is found to increase the probability of married women's employment in work outside the household by 8.2 percentage points in the full sample.

The results presented in Tables 9 and 10 are suggestive of appliances being used as a tool to assist women in altering their time commitments between employment in different work types. Appliances allow married women to reallocate their time to jobs that are of more value to them as both appliances are found to significantly alter work force decisions for married women. The results suggest that time-saving appliances allow women more freedom of choice, allowing them to select employment that is less home-based and potentially more lucrative and liberating.