I. Introduction

Workers care about more than just pay, valuing job security, supportive relationships, work-life balance, among many other non-monetary aspects of employment (OECD, 2013). These diverse work rewards can be aggregated using two approaches designed to capture overall non-monetary rewards (Clark and Kozak, 2023). The first approach combines multiple job attributes into a composite index, using methods like principal component analysis (Lee and Green, 2025). A limitation of this approach is that it likely fails to capture the full complexity of work rewards and ignores that workers assign different weights to specific job attributes. The second approach uses single-index measures like overall job satisfaction, offering a more holistic view and addressing weighting issues, but it prevents comparisons between specific work-related reward.

In a recent paper, Clark et al. (2024) introduce a simple yet innovative method to estimate the non-monetary rewards at the occupational level. This method uses an OLS regression to estimate the correlations between occupations and life satisfaction, and then scales the results into wage-equivalent units. By doing so, it enables the combination of monetary and non-monetary rewards into a single measure of overall rewards. The authors apply this framework to compare inequality in monetary, non-monetary, and overall rewards in the UK. This method is well-suited to analyze the inequality of labor market rewards in South Korea, where job quality has received growing attention in recent years. High rates of temporary employment, long work hours and educational mismatches are among the many problems faced by workers. This study applies the framework to examine how such factors contribute to the inequality of overall work rewards in South Korea. The theory of compensating wage differentials offers a useful framework for understanding the relationship between monetary and non-monetary rewards. It posits that workers are willing to accept negative job attributes—like irregular work hours—in exchange for higher wages, or conversely, accept lower wages in exchange for positive job attributes—like flexible work hours. Thus, the theory predicts a negative relationship between the monetary and non-monetary rewards. This implies that overall inequality should decrease when non-monetary rewards are included, since wage differences are offset by non-monetary differences.

An alternative theory, gift exchange, posits that employers provide workers with higher monetary and non-monetary rewards—such as higher wages and autonomy—in return for enhanced effort and lower turnover. In contrast to the theory of compensating wage differentials, gift exchange predicts a positive relationship between monetary and non-monetary rewards. Of course, it is possible that both theories are valid but apply to different job attributes and labor market contexts. One aim of this study is to assess the applicability of these theories in explaining the relationship between monetary and non-monetary rewards in South Korea. Several studies have examined non-monetary aspects of jobs in South Korea. Lee and Green (2025) develop a composite index to identify “bad jobs,” finding a U-shaped relationship between job quality and its marginal effects on well-being. In a comparative study, Jung and Cho (2016) find that Korean women are less likely than their Australian counterparts to be employed in “good” jobs, defined by hourly wages, job security, and working hours. Aum et al. (2025) estimate the willingness to pay for a horizontal work culture, absence of overtime requirements, career development opportunities, and flexible commuting times, showing that these job attributes contribute to greater inequality in overall rewards. Park et al. (2020a) show that although unskilled manual workers report lower well-being, work stressors and stress relievers matter more than occupational categories. Finally, several studies have documented the harmful effects of long working hours on well-being (Park et al., 2020b; Song and Lee, 2021). While previous studies tend to emphasize specific non-monetary aspects of jobs in South Korea, our research focuses on estimating overall non-monetary rewards and examining how these contribute to inequality in overall labor market rewards.

Our main findings are as follows. First, labor market inequality is understated when considering wages alone, as including non-monetary rewards increases measures of dispersion like the standard deviation, 90-50 gap, and 50-10 gap. While wage inequality is more pronounced in the top half of the distribution, inequality of non-monetary rewards is more pronounced in the bottom half. We also find that incorporating non-monetary rewards narrows gender gaps but widens education gaps. The positive correlation between average wages and non-monetary rewards across occupations suggests that any compensating wage differentials are offset in the aggregate. However, specific factors like atypical work schedules and public sector jobs show evidence of compensating wage differentials. These results underscore the importance of considering both monetary and non-monetary rewards when assessing the inequality of labor market rewards.

II. Data and Method

To estimate the rewards of work, we use data from the Korean Labor and Income Panel Study (KLIPS), a longitudinal survey that began in 1998. The initial wave surveyed 13,319 individuals from 5,000 urban households regarding their labor market activities. Every year, the sample includes new individuals who form family ties with existing participants and replaces nonrespondents with randomly selected households. To mitigate attrition, 1,415 households were added in 2009, and another 710 households were added in 2018. As a result, the most recent wave includes data on 23,971 individuals. Our main analysis uses data from 2023 (Wave 26), but we also use data from 2009-2023 (Waves 12-26) to conduct robustness tests.

We design our sample following Clark et al. (2024) to ensure comparability. Specifically, we include respondents aged 18–64 who are in full-time employment (working at least 35 hours per week), while excluding daily workers, the self-employed,1 and unpaid family workers. To reduce the influence of outliers, we exclude 14 workers reporting more than 72 work hours per week, and those with wages in the bottom 1st percentile. Additionally, we limit the sample to occupations with at least 10 workers. The final sample comprises 6,788 individuals, with sample weights applied in all analyses to ensure population representativeness.

Our main variables for estimating the non-monetary rewards of work are life and work satisfaction. Although economists have been skeptical of subjective measures, Freeman (1978) shows that work satisfaction predicts quitting, while Krueger and Schkade (2008) demonstrate the stability of life satisfaction over time. The KLIPS has included a question on overall life satisfaction since 2017 (Wave 20), with responses recorded on an 11-point scale ranging from 0 (not at all satisfied) to 10 (completely satisfied). Work satisfaction has been surveyed since 2006 (Wave 9), using a 5-point scale from 1 (very satisfied) to 5 (very dissatisfied). For consistency, we reverse the order of the work satisfaction responses and rescale them to an 11-point scale.

In line with Clark et al. (2024), we begin by estimating the following equation:

where S_ij represents life or work satisfaction for individual i in occupation j. On the right-hand side, X_i is a vector of control variables, including gender, age, and age squared (divided by 100). Gender is controlled for using a dummy variable, female, which equals 1 for women and 0 for men. Wages are measured as the logarithm of the hourly wage, which we calculate by dividing monthly earnings by monthly work hours. The sparse set of controls is used to ensure the inclusion of only truly exogenous variables.

We include 2-digit occupation codes based on the 7^th revision of the Korean Standard Classification of Occupations (KSCO). A total of 44 occupations are included in the estimation equation, with one category omitted as the reference group. The coefficients capture the relationship between occupations and life satisfaction or work satisfaction, conditional on wages. These coefficients can thus be interpreted as representing non-monetary rewards at the occupational level . Following Clark et al. (2024), we transform these coefficients into deviations from the employment-share-weighted mean to enhance interpretability. We then scale the resulting deviations by the coefficient on log wages, expressing the non-monetary rewards of occupations in units of log wages. Finally, we sum log wages and non-monetary rewards to obtain an estimate of overall job rewards.

This approach’s main strengths are its simplicity and its ability to capture overall non-monetary rewards. While some studies, like Aum et al. (2025), focus on specific non-monetary rewards, such as corporate culture and overtime work, they likely miss the full range of work-related benefits. The overall approach used here provides a broader view but less detail on specific reward types. Nonetheless, we can still examine how specific attributes relate to overall non-monetary rewards. Another limitation is that non-monetary rewards are measured at the occupational level, assuming workers in the same occupation receive similar rewards, essentially reflecting occupational averages. Finally, the analysis is descriptive rather than causal. Nevertheless, our findings align with previous empirical work and labor economic theory, lending credibility. In the final section, we examine how specific attributes relate to monetary, non-monetary, and overall work rewards. These include the control variables from Equation (1), along with dummies for education levels (high school or less, college, university, and graduate school). Educational mismatch is indicated by a dummy equal to 1 if the job requires less education than attained. Additional dummy variables capture regular employment, public sector status, union membership, overtime work (any overtime), night work (working at least two hours between 10 p.m. and 5 a.m.), and weekend work (working on Saturday and/or Sunday). Commute length is categorized into short, medium, and long, corresponding to the lower tertile (< 40 minutes), middle tertile (40–60 minutes), and upper tertile (> 60 minutes). Summary statistics for all variables used throughout the paper are presented in Table 1 below.

1)The self-employed are excluded because both monetary and non-monetary rewards are largely under their own control.

III. Results

1. Inequality of Overall Rewards

Table 2 shows the estimation results for Equation (1), using life and work satisfaction as dependent variables. The R² values of 0.09 and 0.12 indicate that differences in wages, gender, age, and occupations explain 9% and 12% of the variation in life and work satisfaction, respectively, among full-time workers in 2023.

The coefficients on log wages—0.886 for life satisfaction and 1.136 for work satisfaction—are much larger than those reported by Clark et al. (2024) for the UK, suggesting that wages are more strongly associated with subjective well-being in Korea. These coefficients suggest that a 10% higher wage across individuals is associated with a 0.09-point higher level of life satisfaction and a 0.11-point high level of work satisfaction, both measured on an 11-point scale, which correspond to 0.07 standard deviations for both outcomes.

Women report higher levels of life and work satisfaction than men in Korea, with estimated coefficients of 0.090 and 0.132, equivalent to 0.07 and 0.09 standard deviations, and both exceeding the corresponding estimate of 0.057 for the UK. The estimated coefficients on age imply that life satisfaction and work satisfaction decline from 7.23 to 6.97 and 7.43 to 7.20, respectively between the ages 30 and 50, highlighting a U-shaped relationship which is commonly documented in the literature.

Finally, the estimated coefficients on 44 occupation dummies (with one omitted) reflect the non-monetary rewards of work. Following Clark et al. (2024), we transform these coefficients into deviations from the employment-share-weighted mean to improve interpretability. We then divide the resulting deviations by the coefficients on log wages in order to express the non-monetary rewards in units of log wages.

Table 3 reports estimates of various measures of inequality in monetary rewards, non-monetary rewards, and overall rewards. The standard deviation of log wages is 0.42, while that of non-monetary rewards is 0.20, whether estimated using life or work satisfaction. These values suggest that monetary rewards are more dispersed than non-monetary ones.2 Moreover, the standard deviations of overall rewards and 0.48—are higher than those of monetary rewards alone, implying that measures of in- equality based solely on wages likely understate the true dispersion in work-related rewards in Korea. These patterns align with the findings of Aum et al. (2025) for Korea and Clark et al. (2024) for the UK.

Extending on Clark et al. (2024), we estimate 90–50 and 50–10 percentile gaps to identify distributional asymmetries in rewards. For wages, the 90–50 and 50–10 gaps are 0.61 and 0.43, respectively, indicating that wage dispersion is more pronounced in the upper half of the distribution. One explanation is labor market institutions such as minimum wages compress wage differences in the low-wage jobs. In contrast, for non-monetary rewards, the 90–50 and 50–10 gaps are 0.11 and 0.40, based on life satisfaction, and 0.19 and 0.27, based on work satisfaction. This pattern suggests that inequality in non-monetary rewards is more pronounced in the lower half of the distribution. Together, this finding implies inequality in wages is greater among high-wage jobs, whereas inequality in non-monetary rewards is greater among low-quality jobs.

We assess the robustness of these findings using alternative estimation methods. First, we compare the OLS regression results with those from an ordered probit regression that accounts for the ordinal nature of life satisfaction. The estimates, presented in Appendix Tables A1 and A2, are consistent across both models—supporting Ferrer-i-Carbonell and Frijters (2004), who find minimal differences between ordinal and cardinal approaches. Next, we use panel data from 2009–2023 to estimate work satisfaction with individual fixed effects,3 thereby addressing potential omitted variable bias. These estimates, presented in Appendix Tables A3 and A4 also support our main findings. For the remainder of the paper, we follow Clark et al. (2024) by relying on the OLS estimates of the relationship between occupations and life satisfaction, rather than work satisfaction, as the former likely captures a broader set of rewards. Moreover, individual fixed effects estimates capture only within-individual occupational changes, and since only 32% of individuals switch occupations, they may not represent the non-monetary rewards across the larger population.

2. Overall Rewards across Occupations

In this section, we investigate how the ranking of 43 occupations varies according to their monetary, non-monetary, and overall rewards.4 Figure 1 plots the average log wage and overall rewards for each occupation. The difference between these two measures reflects the non-monetary rewards. Since non-monetary rewards are expressed as deviations from the employment-share-weighted mean, negative values indicate below-average non-monetary rewards, while positive values indicate aboveaverage non-monetary rewards.

Figure 1 shows that occupations with higher average wages tend to offer greater non-monetary rewards, with a Spearman correlation coefficient of 0.28.5 This relationship helps explain the greater inequality observed when both monetary rewards and non-monetary rewards are considered together. In the aggregate, this pattern is at odds with the theory of compensating wage differentials, which posits that less desirable job attributes are offset by higher wages. While compensating wage differentials may exist, they appear to be outweighed by the positive correlation between monetary and non-monetary rewards. Therefore, this pattern is more consistent with gift exchange theory, which posits that employers offer higher monetary rewards and non-monetary rewards in order to elicit greater effort and reduce turnover.

The three occupations with the highest non-monetary rewards are professional service managers (OCC13), education professionals (OCC25), and science professionals and related occupations (OCC21). In contrast, the three occupations with the lowest non-monetary rewards are food processing machine operators (OCC81), other craft workers (OCC79), and agriculture, forestry, fishery, and other elementary service workers (OCC99). While a detailed analysis of the underlying factors is beyond the scope of this paper, the latter group typically experiences less autonomy and more hazardous working conditions. Notably, their lower non-monetary rewards are not compensated by higher wages but are instead paired with lower wages.

However, the correlation of 0.28 shows that the positive association between monetary and non-monetary rewards is far from perfect. Some relatively high-wage occupations—such as construction, electrical, and production managers (OCC14) and business and financial professionals (OCC27)—offer relatively low non-monetary rewards. Conversely, some relatively low-wage occupations—such as caregiving, health, and personal service workers (OCC42) and cooking and food service workers (OCC44)—offer relatively high non- monetary rewards. While the former group may involve demanding work environments, the latter may offer intrinsic rewards. Even without pinpointing the exact determinants, the negative link between monetary and non-monetary rewards in these occupations lend support to the theory of compensating wage differentials.

3. Specific Attributes and Overall Rewards

In this final section, we examine how specific attributes correlate with monetary, non-monetary, and overall rewards using OLS regressions. Table 4 presents the results in columns 1–3, respectively. Note that, due to the absence of a clear identification strategy, these estimates reflect correlations rather than causal effects.

The coefficients on the female dummy variable suggest that although women earn lower wages than men, they experience greater non-monetary rewards. This pattern has been documented across numerous countries, and Redmond and McGuinness (2020) attribute it to women’s stronger preferences for jobs with better work-life balance and intrinsic satisfaction. Aum et al. (2025) find that Korean women are more willing to accept lower wages in exchange for no overtime, particularly those with children. These findings imply compensating wage differentials can explain some gender differences in labor market outcomes.

Although age exhibits strong relationships with wages and life satisfaction (Table 2), its association with the non-monetary rewards of work appears weak. Aum et al. (2025) find that older workers place greater value on a horizontal corporate culture, whereas younger workers place greater value on future career prospects and commuting time flexibility. These specific job attributes, among others, may offset one another, resulting in little difference in overall non-monetary rewards across ages.

In contrast, educational attainment exhibits a strong positive relationship with non-monetary rewards. One possible explanation is that education both increases access to high-wage occupations and shapes preferences for desirable job attributes, while strict educational requirements in many high-reward occupations further reinforce this pattern. Unsurprisingly, workers who report being educationally mismatched tend to experience both lower wages and lower non-monetary rewards. As Hamermesh (2001) argues, dissatisfaction with the returns to one’s educational investment may reduce overall job satisfaction.

One salient feature of the Korean labor market is the clear duality between regular and irregular workers. Our findings show that regular workers not only earn higher wages but also enjoy greater non-monetary rewards compared to their irregular counterparts. This is not surprising, given that irregular workers typically face lower job security and limited benefits. In contrast, workers in public organizations tend to receive high non-monetary rewards but low wages. This likely reflects the greater job stability and reduced work intensity in the public sector. Such a trade-off is consistent with the theory of compensating wage differentials.

Unionized workers receive higher wages but lower non-monetary rewards from work. This pattern may, however, reflect reverse causation: lower non-monetary rewards in occupations may increase the likelihood that workers join unions. Freeman (1978) similarly finds a negative association between unionization and job satisfaction, suggesting that unions serve as a “voice” mechanism—allowing workers to express dissatisfaction without resorting to exit. In this sense, unions may function as an institutional channel for securing compensating wage differentials in occupations with less desirable non-monetary attributes.

In terms of work schedules, employees who work overtime earn higher wages but lower non-monetary rewards, consistent with a compensating wage differential. In contrast, those who work on weekends tend to receive lower wages, with no significant difference in non-monetary rewards. Night work is associated with both higher wages and, somewhat unexpectedly, higher non-monetary rewards, suggesting that the downside of atypical hours may be offset by other favorable job characteristics. Finally, longer commute times are correlated with higher wages but show no significant relationship with non-monetary rewards—possibly because workers accept longer commutes in exchange for other desirable aspects of their jobs.

2)Clark (2024) notes that this pattern can be easily reversed if the coefficient on log wages is much smaller.

3)Life satisfaction data is only available from 2018 onward.

4)We exclude skilled crop and livestock workers (OCC61) and construction and mining elementary workers (OCC91) due to their low employment shares of 0.17% and 0.11% which make their estimates more prone to measurement error.

5)The Spearman correlation coefficients are 0.33 when using the OLS regression on work satisfaction, 0.26 when using the ordered probit regression on life satisfaction, and 0.52 when using the individual fixed effects regression on work satisfaction.

IV. Conclusion

This study analyzes the inequality of monetary, non-monetary, and overall rewards among full-time workers in South Korea. We estimate non-monetary rewards by applying an OLS regression to quantify the relationship between occupations and life satisfaction. The resulting estimates are then transformed into wage-equivalent units, allowing us to combine monetary and non-monetary rewards into a single measure of overall rewards. Our analysis reveals that while monetary rewards are more dispersed than non-monetary ones, combining the two results in even greater overall inequality. One important implication is that policymakers and researchers should consider both monetary and non-monetary factors when assessing worker welfare and designing labor market interventions. Distributional patterns show that inequality in non-monetary rewards is greater among low-quality jobs. This implies that policies aimed at improving non-monetary aspects of low-quality jobs may achieve significant results in reducing overall labor market inequality.

The increased inequality observed when combining monetary and non-monetary rewards suggests a positive association between the two types of rewards. In other words, occupations with higher wages also tend to offer greater non-monetary rewards. However, exceptions exist that support the theory of compensating wage differentials, where some occupations with lower non-monetary rewards offer higher wages, and conversely, some occupations with higher non-monetary rewards offer lower wages. This analysis also underscores the importance of including non-monetary attributes when assessing differences in occupational rewards.

A closer look at specific traits of the Korean labor market and their relation to both monetary and non-monetary rewards reveals several notable patterns. The duality between regular and irregular workers in Korea is particularly evident, as irregular workers not only earn lower wages but also receive lower non-monetary rewards, suggesting that policymakers should address both aspects to reduce labor market segmentation. Similarly, workers experiencing educational mismatches earn lower monetary and non-monetary rewards. This may call for improvements in education and job placement policies with a focus on overall rewards. Finally, employees who work overtime earn higher wages but lower non-monetary rewards. When combined, the overall rewards of overtime work appear positive. Although these relationships are not necessarily causal, they suggest that policymakers should exercise caution in regulating work hours.

Finally, our analysis has several key limitations. Due to sample size constraints, we rely on 2-digit occupational codes rather than 3-digit classifications, which may hide important within-occupation heterogeneity. Similarly, while our estimates capture average non-monetary rewards by occupation, they cannot capture within-occupation variation across workers. Lastly, our findings are based on correlations rather than causal relationships, and should therefore be interpreted as descriptive. These limitations suggest avenues for future research using richer data and causal identification strategies.