The fashion retail industry is an important segment of the global fashion supply chain. Retailers perform at the most downstream location within the supply chain, interacting directly with consumers [1
]. The close proximity to consumers also heightens the need to meet today’s consumers’ wants and needs for sustainability. Unfortunately, the fashion retail industry is still criticized for its negative social and environmental impact, often stemming from its high rate of employee turnover and competitive market environment [2
]. To combat these criticisms and improve sustainability performance, many of today’s fashion retail businesses both voluntarily and involuntarily integrate social and environmental practices and policies into their business models [4
]. Organizational sustainability performance is defined as a dynamic process that necessitates achieving short-term performance (meeting current needs) without compromising long-term performance (meeting future needs) of the triple bottom lines—financial, social, and environmental [4
]. However, most of these sustainability initiatives primarily focus on products or services or consumer behavior, and only limited attention has been paid to the role of employees in meeting companies’ sustainability goals [5
]. Employees are important conduits to implementing and achieving a variety of organizational goals toward sustainability. Accordingly, human and social capital theories brought great attention to the development of research related to turnover in many industries and suggested that turnover negatively influences organizational sustainability performance [6
]. To positively influence competitive advantages over the long term, many organizations have recognized the value of ethical climate in enhancing an organization’s image and reputation, as well as its sustainability performance [7
]. However, little is known as to how employees’ attitudes, perceptions of ethical climate, and turnover intentions affect fashion retail businesses’ sustainability performance.
The U.S. Fashion Retail Industry and Employment Issues
Overall, the fashion industry is of great economic importance in terms of trade, employment, investment, and revenue throughout the world. The U.S. fashion industry accounts for US$1.2 trillion of global economic activity, and the United States alone spends more than US$250 billion annually on fashion-related products and services [9
]. More specifically, in the U.S., retail is the largest employer, with nearly 15.7 million employees [10
]. Under the retail industry umbrella, apparel and accessories retail stores are the third largest employers after general merchandise stores and vehicle and parts dealers, accounting for 1.4 million workers [11
]. Employment in retail is projected to grow 7% between 2014 and 2024 [11
As a sector within the supply chain, retail employees, including salespersons, cashiers, stock clerks, order fillers, and supervisors of retail sales workers, account for over 10 million employment positions in the U.S. retail trade industry [12
]. These retail employment positions require working in retail stores, standing for long periods of time, and being assigned varied working hours even during holidays and weekends. Thus, retail employment is characterized by a high degree of casual, part-time, and seasonal employment [9
]. Approximately 70% of retail employees are working full-time, and 68% of part-time employees want to work less than a full-time schedule [13
]. More interestingly, nearly half of part-time retail employees are under the age of 25, while 20% of part-time retail employees are age 55 or older [13
Therefore, retail employees’ turnover rate has increased, with a 73.7% turnover rate in 2013, which is 15.9% above the national average [14
]. This trend toward accelerated turnover is expected to continue [14
]. The average fashion retail sales employee in the U.S. makes an annual income of less than US$30,000 [11
]. However, the cost of replacing an employee making less than US$30,000 per annum is about 16% of that person’s annual wage [15
]. To prevent high turnover among retail employees, some retailers, such as Gap, TJ Maxx, Marshalls, and Uniqlo, have raised their wages [15
]. In short, the fashion retail industry’s working environment and high rate of employee turnover were cited as one of the key reasons for negative organizational performance, including sustainability, in both the short- and long-term [16
Moreover, a work culture that supports and encourages ethical behavior is identified as another key antecedent of employees’ turnover intentions and the sustainability performance of organizations [2
]. The collective effect of organizational employees’ ethical perceptions is found to generate shared ethical work norms within a work environment, in turn positively stimulating individuals’ attitudes toward their jobs and the organization, and influencing ethical decision-making processes within the organization [17
]. Many organizations recognize the value of ethical practices in enhancing their image and reputation, and subsequently the potential to positively increase their long-term competitive advantage [17
]. Yet, few studies are available to on how employees’ job attitude, ethical climate, and turnover intention affect fashion retail businesses’ sustainability performance. Consequently, this study was designed to understand the underlying relationships between fashion retail employees’ turnover intention, its antecedences and its impact on fashion retail businesses’ sustainability performance.
4.1. Data Screening
Prior to testing hypotheses in this study, it was necessary to check the data screening techniques and assess the basic assumptions in IBM Statistical Package for the Social Sciences 22 (SPSS, Armonk, NY, USA), such as missing data, outliers, multicollinearity, and normality. Outliers were evaluated for response patterns that recorded either being all “1” or “5” on all 40 measurement items, and three datasets were excluded from further analysis. Multivariate outliers are observations that are inconsistent with the correlational structure of the dataset, and can be identified with Mahalanobis distance [86
]. Mahalanobis distance measures the distance of a data point from the calculated centroid of the other cases where the centroid is calculated as the intersection of the mean of the variables being assessed. The Squared Mahalanobis distance of 40 items was evaluated, with 73.40 critical value as the cut off, if potential multivariate outliers existed in the SPSS. Twenty-eight responses were detected as multivariate outliers exceeding the critical value (>73.40), and were excluded from further analysis. Multicollinearity was measured in a multiple regression model by observing the correlation matrix for high-correlation coefficients (>0.9) between predictor variables [86
]. The items possessing multicollinearity with values of variance inflation factor (VIF) that exceeded 10 and tolerance that surpassed one were further evaluated, but no multicollinearity was found among the measurement items; therefore, they were included in the data analysis. For normal univariate distribution, values for asymmetry and kurtosis between ±2 were considered acceptable [86
]. Overall, non-normality did not hinder this data analysis, and was determined as the measured variable in the model. After basic assumptions were evaluated, 278 data were analyzed in total. Then, the two-step approach was employed to conduct empirical data analysis using structural equation modeling [87
Sample characteristics were as follows: 145 participants were female (52.2%) and 133 were male (47.8%). Participants were young, in general, with 93 participants in the range of 18 to 24 years of age (33.5%), 98 between 25 and 44 years of age (35.3%), 83 between 45 and 64 years of age (29.9%), and 4 being 65 years of age or older (1.4%). Most participants’ educational backgrounds included at least some higher education; only 68 reported their level of education as high school graduate or less (24.5%). For the remainder, 70 attended some college (25.2%), 82 had received an associate’s or bachelor’s degree (29.5%), 51 held a master’s or professional degree (18.3%), and 7 had earned a doctoral degree (2.5%). As to marital status, 133 participants were married (47.8%), 113 never married (40.6%), 26 were divorced (9.4%), 4 were widowed (1.4%), and 2 were separated (0.7%). With respect to ethnicity, the majority of participants were White/Caucasian with 185 (66.5%), followed by 38 Hispanic/Latinos (13.7%), 29 Black/African Americans (10.4%), 17 Asians (6.1%), and 4 Others (1.4%).
In addition, of the 278 data, the majority of participants, 182, worked in privately-owned retail (65.5%), while 73 participants worked in publicly-owned retail (26.3%). The remaining 23 participants (8.3%) did not know whether their employment was in privately or publicly-owned retail. Participants worked in companies of varying sizes: 41 worked in companies having between 2 and 10 employees (14.7%); 75 between 11 and 50 employees (27%); 82 between 51 and 250 employees (29.5%); 28 between 251 and 500 employees (10.1%); and 52 over 501 employees (18.7%). In terms of job responsibilities, 114 of the participants have work relating to sales (41%), followed by 81 in store management (29.1%), 51 in buying or merchandising (18.3%), 16 in human resources (5.8%), and 16 in various other areas (5.8%), such as creative design, department leadership, marketing, wholesale, quality assurance, and logistics. Relating to income level in 2015, 102 participants (36.7%) reported an earned income in the range of US$25,001–US$50,000, followed by 74 participants (26.6%) earning US$50,001–US$75,000; 39 participants (14%) earning US$75,001–US$100,000; 35 participants (12.6%) earning US$100,001 or more; and 28 participants (10.1%) earning US$25,000 or less.
4.3. Measurement Model
Prior to examining hypothesis relationships, the original a priori measurement model of variables in the hypothesized structural model was condensed and specified over several iterations to reduce standardized residuals and obtain acceptable model fit by using Mplus 7 [87
]. In consideration of modification indices, covariance relations were added, and items were deleted one at a time. Through this process, ten items were removed due to low factor loadings and high correlation.
A cross-sectional study in organizational research can lead to a false internal consistency, and/or cause systematic measurement errors that either inflate or deflate the observed relationships between constructs [88
]. According to Podsakoffe et al. [89
], under the circumstance of single source data collecting in same context, it is important to evaluate common method variance (CMV) conducting “single-factor-approach” (p. 898). According to Harman’s [90
] single-factor method, single-factor analysis has been evaluated and revealed its variance as 33.41%. Moreover, single-factor solution’s fit indices revealed to be a poor fit with a comparative fit index (CFI) of 0.59, a tucker lewis index (TLI) of 0.53, a root mean square residual (RMSEA) of 0.17, and the standardized root mean square residual (SRMR) of 0.12.
On the other hand, the nested model comparison was examined with a χ2
, by comparing a single-factor solution (χ2
= 1704.07, d.f. = 187; p
-value < 0.000) with the research model (272.776, d.f. = 172; p-
value < 0.001) [87
]. Based on the result of
= 1431.27 and
d.f. = 15 (p
-value < 0.000), the two models were different, indicating that the single-factor solution results in a substantial worsening of the overall model fit. Consequently, the measurement model of all latent variables in the hypothesized structural model had an
value of 290.68 (d.f. = 190; p
-value < 0.001), a corresponding CFI of 0.97, a TLI of 0.97, an RMSEA of 0.04, and an SRMR of 0.04 at acceptable levels of model fit. Although the p
was less than 0.10 due to a high sample size, all other fit indices showed an acceptable fit. The loadings for all the factors were high, above 0.60, and corresponding standardized estimate/standard error (est./s.e.) values were statistically significant (see Table 1
Furthermore, the convergent and discriminant validity of the measurement model was checked with the average variance extracted (AVE). First, the overall AVEs of all latent constructs within the measurement model satisfied the criteria of the convergent validity, as the AVE of each latent construct was evaluated with a cut off criterion of 0.50 or higher [91
]. Second, discriminant validity was evaluated with AVE estimates and the correlation matrix (see Table 2
). All values of the square root of the AVE of paired constructs exceeded the correlation estimates between these constructs, which satisfied the discriminant validity of the measurement model [91
]. Overall, the reliability of the measurement model was supported as the scales included in this model exhibited acceptable reliability with Cronbach’s α over 0.70 [91
]. Cronbach’s α value ranged from 0.78 (ethical climate) to 0.93 (job attitude) while composite reliability ranged from 0.81 (social performance) to 0.93 (job attitude). Consequently, by evaluating all model fit indices, reliability and validity, the measurement model was confirmed. Detailed information is listed in Table 1
and Table 2
4.4. Strucutral Model
The structural model reported its goodness-of-model-fit indices to be
= 294.76 (d.f. = 193, p
-value < 0.000), CFI = 0.97; and TLI = 0.97; RMSEA = 0.04, and SRMR = 0.04, suggesting acceptable levels of model fit (see Figure 2
). Hypothesis 1a through 1c examined the impact of ethical climate on perceived organizational sustainability performance. Hypothesis 1a posited a positive effect of ethical climate on perceived financial performance. Analysis results indicated that it was statistically significant (H1a: standardized coefficients
= 0.62; p
-value < 0.001). Hypothesis 1b assessed the positive influence of ethical climate on perceived social performance. This relationship was shown as statistically significant (H1b:
= 0.66; p
-value < 0.001). Hypothesis 1c examined the positive relationship between ethical climate and perceived environmental performance. The result indicated that it was statistically significant (H1c:
= 0.62; p
-value < 0.001).
Hypothesis 2a through 2c examined the relationship between turnover intention and perceived organizational sustainability performance. Hypothesis 2a posited a negative influence of turnover intention on financial performance. Analysis results indicated that it was statistically significant (H2a: = −0.16; p-value < 0.05). Hypothesis 2b assessed the negative influence of turnover intention on perceived social performance. This relationship was shown as statistically significant (H2b: = −0.21; p-value < 0.001). Hypothesis 2c examined the negative relationship between turnover intention and perceived environmental performance. The result indicated that it was statistically significant (H2c: = −0.13; p-value < 0.05).
In addition, hypothesis 3 assessed the negative relationship between ethical climate and turnover intention. Analysis results indicated that it was statistically significant (H3: = −0.20; p-value < 0.05). Hypothesis 4 investigated whether ethical climate could positively influence an employee’s job attitude. This relationship was also shown as statistically significant (H4: = 0.56; p-value < 0.001). Hypothesis 5 predicted a negative relationship between job attitude and turnover intention. The result supporting this hypothesis was statistically significant (H5: = −0.28; p-value < 0.001).
4.5. Further Analysis of Indirect Effects
In addition to hypothesized relationship testing, the results of our analysis also recognized the indirect effects of relationships in this study (see Table 3
). First, turnover intention was influential in the relationship between job attitude and perceived organizational sustainability performance. Turnover intention indeed mediated the relationship between job attitude and perceived financial performance (standardized parameter estimate = 0.044; p
-value < 0.05), and between job attitude and perceived social performance (standardized parameter estimate = 0.058; p
-value < 0.05). However, turnover intention was shown to be statistically insignificant between job attitude and perceived environmental performance.
On the other hand, turnover intention was shown to have mediating effects between ethical climate and perceived organizational sustainability performance. Not only did turnover intention mediate a relationship between ethical climate and perceived financial performance (standardized parameter estimate = 0.03; p-value < 0.05) but job attitude also mediated between ethical climate and perceived financial performance (standardized parameter estimate = 0.03; p-value < 0.05). These mediating effects enhanced the total effect of the relationship between ethical climate and perceived financial performance (standardized parameter estimate = 0.68; p-value < 0.001).
Moreover, turnover intention mediated a relationship between ethical climate and perceived social performance (standardized parameter estimate = 0.04; p-value < 0.05), and job attitude was also shown to mediate between ethical climate and perceived social performance (standardized parameter estimate = 0.03; p-value < 0.05). These mediating effects also enhanced the total effect of the relationship between ethical climate and perceived social performance (standardized parameter estimate = 0.73; p-value < 0.001). Therefore, the indirect effects of turnover intention (standardized parameter estimate = 0.05; p-value = 0.052) and job attitude (standardized parameter estimate = 0.03; p-value = 0.092), and between ethical climate and perceived environmental performance were shown to be statistically insignificant.
5. Discussions and Conclusions
Given that fashion retail businesses often create negative perceptions amongst the public in relation to sustainability, stemming from high turnover rates and competitive market environments, this study investigated what influences companies’ sustainability performance from the employees’ perspective. According to social and human capital theories, employees are the key to success in any organization [38
]; therefore, the study hypothesized that the low turnover rate within organizations with ethical climates would help achieve superior organizational performance, including social and environmental dimensions.
The result of a nationwide survey study using U.S. fashion retail employees as a sample framework found that that employees’ positive attitude toward their jobs not only reduced their turnover intention, but also that their organizations’ environment or climate could reduce employees’ turnover intention. As job satisfaction had been known to be a major contributing factor for employees’ turnover intentions in various domains and disciplines [8
], this study supports the theory that positive job attitude is important to employees’ turnover intention. Moreover, as ethical climate and turnover intention have also been found to be highly influential for organizational performance in service-related industries [8
], the results of this study support the notion that ethical climate is significantly important in lowering employees’ turnover intentions and increasing positive job attitudes. These findings provide a strong support for Blaus’ [68
] social exchange theory. That is, if employees feel satisfied with their jobs and work in a perceived ethical environment, their attitude toward turnover intention tends to decrease [58
Interestingly, this study found that ethical climate not only lowered employees’ turnover intention but also positively affected employees’ job attitude. Similarly, previous studies posited that ethical climate could influence a number of outcomes, such as employees’ job attitude [31
], turnover intentions [8
], and organizational citizenship behavior [54
]. This finding carries an important message: that it is critical to build the perception of an ethical work environment for employees and that this can reduce the negative attitude of turnover intention while encouraging employees’ positive attitudes toward their jobs. This may be because if employees perceive their working environment to be more ethical and perceive their employers as doing right thing voluntarily rather than out of obligation, it may have a significant positive influence on job attitude as it places great importance on the meaningfulness of work. Likewise, for the fashion retail industry where businesses are highly involved with people, it is strongly recommended that the expectations of employees be outlined by developing written standards of ethical workplace conduct, providing training to ensure awareness of these expectations, and equipping managers to reinforce the company’s values through their actions.
The relationship between turnover intention and each of the three dimensions of organizational sustainability performance were found to be positive in this study. Similar to previous study findings, the present study shows that high turnover intention may harm not only organizations’ financial performance but also their social and environmental performance. This is an important new theoretical contribution because few studies focus on organizational sustainability performance from the perspective of low turnover intention. Moreover, because the fashion retail industry has high turnover rates compared to other industries, understanding turnover intention should be extremely important for many retail businesses [14
]. That is, since they are less likely to move, organizations would reap the benefits of training long-term employees. This implies that any fashion retail businesses should consider ways to reduce turnover intention among employees, resulting in positive organizational performance in achieving sustainability.
This study found that ethical climate was positively associated with each of the three dimensions of organizational sustainability performance. These findings support previous studies [27
] which found that ethical climate plays an essential role in stimulating organizational performance. This result emphasizes the importance of creating an ethical climate, not only in a retail environment, but also in any organization, as organizational performance measures success and achievement in any context. According to Paine’s [33
] argument, ethical climate could offer “hand-in-hand” advantages, not only in financial but also in social and environment performance. In other words, ethical climate could bring positive externalities to prevent costs related to ethical relations and policies while strengthening intangible gains in employees’ positive attitudinal engagement and commitment in the long run. Thus, this study’s results suggest that an ethical climate is critically important for encouraging employees to have positive attitudes toward their jobs and to develop organizational performance in the U.S. fashion retail industry. In other words, a working environment that is perceived as ethical is particularly critical to enable fashion retail employees to work well within their teams, departments, or organizations. To stimulate all three aspects of organizational sustainability performance, ethical climate must be embedded within the retail business environment. Accordingly, to provide an improved ethical climate, performance evaluation systems could be introduced to survey the workplace from the employee perspective. This may assist employees in developing more trusting and caring perceptions toward their organizations, which could lead to superior social, financial, and environmental performance through organizational citizenship behavior.
The study showed that turnover intention mediated the relationship between job attitude and all three dimensions of organizational sustainability performance. This relationship suggests that when employees have positive attitude toward their employment, turnover intention may decrease, which will, in turn, improve the triple bottom line for organizational sustainability performance. This implies that employees’ positive attitudes toward their jobs are a central component of improving organizational sustainability performance.
Lastly, three correlations were found between financial and social performance, financial and environmental performance, and social and environmental performance. This study was one of the few to investigate the three dimensions of organizational sustainability performance. Previous studies examining sustainability performance primarily focused on the social and financial aspects of sustainability [7
]. Because sustainability has begun to encompass the triple bottom lines—social, financial, and environmental—this study therefore included environmental performance to measure a more complete framework for the sustainability performance of organizations. This study’s findings show that environmental performance is positively associated with financial and social performance, supporting the triple bottom line theory of sustainability. These correlations imply that all three relationships simultaneously help each other to achieve superior organizational performance [92
]. In the past, fashion businesses have been criticized for causing a negative environmental impact and damage around the world [2
]. To fulfill the profit-driven nature of the industry, many fashion retail businesses have focused solely on the financial aspect of organizational performance, often causing social and environmental problems [2
]. Therefore, the results addressed the importance of including social and environmental aspects of organizational performance. If the three dimensions of organizational sustainability performance are equally or similarly weighted in evaluating fashion retail business performance, it would not only positively affect fashion retail businesses but would certainly influence other related parts of fashion supply chains. Thus, fashion retail businesses should consider evaluating organizational performance based on all three components of sustainability, including financial, social, and environmental factors, as they all seem to go hand in hand.
This study has several limitations, which may lead to future research opportunities. First, the objective of this study focused exclusively on U.S. fashion retail employees. Although the results are valuable, different findings may result in terms of the three dimensions of organizational sustainability performance in other parts of the fashion supply chain operating in other countries or in wholesales that primarily deal with manufacturers in other countries. Therefore, it is suggested that future research compare and contrast the cultural influences on outcomes of the triple bottom line of organizational sustainability performance. Second, the study employed a purposive sampling technique, and it cannot be assumed that this study is representative of the whole population of U.S. fashion retail employees. Also, in order to accommodate data collection in a timely manner, a research firm recruited panels from among their contacts to participate in the survey. Thus, it may be difficult to generally apply this study’s findings to the whole population of U.S. fashion retail employees. Hence, future research could garner a larger number of participants, in addition to utilizing a randomized sampling technique.
Third, this study is one of few studies to examine organizational sustainability performance from employees’ perspectives as to perceived organizational sustainability performance. Although this may provide an interesting view of organizational performance, future research can be considered to include managers’ or other higher-level positions’ perception of perceived organizational sustainability performance. Moreover, this study aimed to investigate influences of turnover intention on organizational sustainability performance. Although the findings of this study led an interesting point for human resources, future research could consider possibilities in the opposite direction to examine the vital role of organizational sustainability performance in the retail industry. Fifth, to investigate employees’ internal attitudes toward their jobs, the items measuring their job attitudes were adopted from Schwepker’s [58
] INDSALES. Therefore, there are other instruments measuring job attitudes which can be used for future study and possible measurement for job attitudes include the JDI or from the Minnesota Satisfaction Questionnaire (MSQ). A multidimensional organizational commitment can also be investigated in future research as this may relate differently to turnover intention. Lastly, the study was designed as a cross-sectional study to investigate U.S. fashion retail employees. Although the common method variance was evaluated to prevent deflating or inflating the results, the work environment may influence the perception of organizational sustainability performance as a halo effect. To avoid such issues, future research could design a two-source method for manager/upper level and employee levels, or develop measurement items with different anchors.