Want Fewer Employees to Quit? Listen to Them

mistreated, or without purpose. And those feelings are strongly reflected in the holistic value we derive from our jobs and probability that we stay on versus quit. But what, if anything, can firms do to address workers’ dissatisfaction and prevent turnover? We decided to test whether a simple intervention — asking workers for feedback — might be successful at reducing quit rates.

This idea is rooted in a prominent economic theory, first proposed by Albert Hirschman in 1970 in a book titled Exit, Voice, and Loyalty, which reasons that workers can do one of two things when faced with an unsatisfactory employment situation: exit the relationship (quit), or voice their concerns and try to work through their problems. This theory, despite its huge influence on both academic thought and firm policies, has never been experimentally tested in a large-scale real-world setting.

We picked large manufacturing and services firms in low-income countries because they have grown their workforces rapidly in recent years, bolstered by the rising global demand for goods and services and the increased mobility of workers from rural to urban settings. Yet despite this growth, worker absenteeism and turnover, especially for frontline workers, remain persistently high — as do horror stories of poor working conditions.

Motivated by the idea of a tradeoff between exit and voice, we decided to test whether providing workers with voice, in the form of an employee satisfaction survey, could reduce quitting for frontline workers. To conduct this test, we partnered with Shahi Exports, the largest manufacturer of ready-made garments in India. Wages for frontline workers in this firm are largely influenced by minimum wage policy: the government revises its minimum wage schedule every year, and the firm increases wages to comply. We discovered that the period around these annual wage hikes is particularly turbulent: worker dissatisfaction tends to be high following the wage hike, despite the fact that workers end up receiving larger paychecks.

This puzzle motivated us to test our voice intervention in the month following the 2016 minimum wage hike. We hypothesized that workers might be dissatisfied after the wage hike because they were expecting larger wage increases than they actually received. This seems reasonable in the Indian context, in which there tends to be a great deal of uncertainty and lack of reliable information on the expected size of wage increases.

Our study consisted of two parts: a baseline survey and a field experiment. Before workers learned of the size of the wage increase, we conducted a baseline survey to learn what they thought the increase would be. Then, after the wage hike took place, we provided “voice” to a randomly selected half of the sample. These treatment workers were asked to participate in an anonymous survey, which asked for their feedback on various aspects of the job. We then tracked attendance and quits for five months using the firm’s administrative data.

The surveys revealed some interesting findings. First, workers were, on average, disappointed by the wage increase. The vast majority of workers expected a larger increase than they actually received. This might be because the wage increase in 2014 was over four times larger than the 2016 wage increase. Second, workers did use the satisfaction survey to express dissatisfaction with their job. Over half of respondents responded negatively to one of the six specific statements about the work environment. Importantly, workers also reported low satisfaction with wages (much lower than satisfaction related to supervisors, the workplace environment, and overall).

Most of us, at some time in our lives, have had jobs in which we felt unfulfilled, Importantly, finding an outlet to voice workers’ frustrations improved retention. When we compared the quitting behavior of workers in the treatment and control groups, we found that quitting was 20 percent (approximately 4 percentage points) lower in the group that received the voice intervention in the months following the wage hike. Strikingly, this effect was most pronounced among the most disappointed workers —those who were expecting much higher wage hikes than they actually received. This suggests that the effectiveness of the voice intervention stemmed from its ability to mitigate disappointment. We find similar results when we look at attendance, which we used as a proxy for how hard employees are working in their job. For disappointed workers, the intervention significantly reduced absenteeism in the months following the wage hike.

Read More