This is the fifth in our series of posts by students on the job market this year.
Do people perform better when working with friends or do their friends distract them from doing their job well? Does the effect depend on their personality traits? I investigate these questions in the context of a seafood-processing plant in Vietnam in which several workers perform the identical task – cleaning and filleting fish -- at 4-person tables in a processing room. I collaborated with the management to design and implement a field experiment in which employees were randomly assigned to positions within the room each day. I use random variation in a worker’s proximity to friends to estimate the effect of working with friends on job performance. Before the experiment, I administered a baseline survey to collect information on employees’ friendship ties and personality characteristics. I find that employees are less productive when working with friends but only when friends are close enough to socialize with each other. I also find that personality traits matter and explain a significant portion of individual differences in socializing behaviors at work. Conversely, socializing with friends explains a large portion of why workers with certain personality traits – notably, conscientiousness – are more productive workers.
Human resource allocation policies can have substantial impacts on employee productivity (Mas and Moretti, 2009; Bandiera et al., 2010; Hjort, 2014). How friends affect job performance is potentially relevant to developing countries. Evidence shows that friendships are channels of job referrals (Beaman and Magruder, 2012) and informal credit (Fafchamps and Lund, 2003) for individuals in less-developed regions. In my sample of female employees in Binh Thuan, Vietnam, 75% reported to have learned about their current job from their friends and 80% reported to have engaged in lending and borrowing practices with their friends from work.
Working nearby a friend may help employees cope with boredom or create a sense of competition, leading to greater motivation. Alternatively, the presence of a friend could lead to goofing off during work and becoming less productive. However, not everyone may feel the same degree of competition nor indulge in idle chats with friends during work. To take individual differences into account, I draw on the literature on non-cognitive skills, also known as personality traits or socio-emotional skills (Almlund et al., 2011). The literature provides substantial evidence of strong correlations between job performance and conscientiousness, one of the Big Five personality factors that measures a person’s tendency to exert self-control. Accordingly, I examine how the effect of working with friends differs between high- and low-conscientiousness workers.
The plant I study hires female processing workers who produce mackerel fillets. Compensation is a combination of a fixed daily wage plus a piece rate based on each worker’s individual output, measured by kilograms of fish processed. Before the experiment, workers were able to choose work stations on their own. During the five-month experiment period (August 2014 to December 2014), processing workers were randomly assigned to work at different stations each day. This created random variation in the presence of friends at various levels of spatial proximity to the worker.
I collected daily data on each worker’s productivity and work station from the firm’s employee records database. The data set spans both the pre-experiment period and experiment period and consists of 7,570 worker-day observations of 104 workers. Two weeks prior to the start of the experiment, I administered a baseline survey to collect data on each worker’s friendship ties at work, non-cognitive skills, and other background data. Non-cognitive skills are measured using the Big Five personality factors: extraversion, agreeableness, conscientiousness, neuroticism, and openness.
1. How does working with friends affect employee productivity?
I find that an employee is on average 6 percent less productive when a friend is working right next to her. But, I find no effect when friends are working at positions further away from her such as across the table or at a neighboring table. These findings suggest that employees are socializing with their friends when they are at close enough distances to each other. As a hygienic requirement, employees are required to put on face masks during the work time. This makes it difficult for an employee to talk with her friends unless she is right next to them. If just being able to see a friend affected productivity, we would also expect to find effects of having a friend across the table, but we do not.
Since the compensation is partially based on a piece rate, an employee earns less when a friend is present alongside (on average 4% less). Yet, employees will choose to work alongside friends if the benefit they enjoy from socializing is greater than the cost in terms of forgone potential wage.
2. But, then, how much do employees value socializing with friends at work?
Quantifying how much one enjoys or values a non-monetary attribute is challenging when the price of that attribute cannot be randomly offered. Instead, I use observational data on employee positions from the pre-experiment period during which employees were able to choose their own work stations. From this I measure how often each worker chooses to work alongside her friends. Next, to estimate each employee’s value of socializing, I use a probabilistic choice model in which the probability of working alongside a friend is increasing in an employee’s value of socializing and decreasing in the size of the wage loss she incurs from having a friend alongside. The estimates show that, conditional on having a positive valuation (which holds for 90 percent of the worker sample), an employee is willing to forgo approximately 6 percent of her wage to socialize with friends at work.
3. Do employees with high-conscientiousness behave differently?
I find that high-conscientiousness employees have smaller productivity losses and place lower valuation of socializing at work. An employee who is one standard deviation above the average level of conscientiousness has a 3 percent productivity drop when working alongside a friend, whereas an employee one standard deviation below the average shows a 9 percent loss in productivity. Moreover, high-conscientiousness employees place lower valuation on working alongside friends not simply because they have smaller productivity declines—they also choose to work alongside friends less often than low-conscientiousness workers. Looking at it the other way, the reasons that conscientious employees are more productive on this job is socializing behavior at work: high-conscientiousness employees have smaller productivity declines when working alongside friends and also less likely to choose to work alongside friends. (See the Figure below.)
The results suggest that working with friends can hurt job performance and employees are willing to do so even at their own cost. Since employees are partly paid fixed wages, this plausibly implies that the firm can profit from randomizing worker positions; if compensation was completely performance pay the firm should be indifferent. However, this is less clear in the long run. Random positions may lower job satisfactions and heighten job turnovers. Also, friendships at work might be important channels of job-skill transfers. Because the study period was five months, this paper is unable to capture how a change in social dynamics may influence other dimensions of human resources in the workplace.
Finally, in terms of external validity, in my context, conscientiousness is an especially important non-cognitive skill affecting interactions on the job, but we might expect other non-cognitive skills, such as prosociality or extraversion, to have greater relevance in occupations that require team work or effective communication skills.
Sangyoon Park is a PhD Candidate at Northwestern University.