Low labour force participation, high worker turnover (Blattman and Dercon 2017) and absenteeism (Adhvaryu et al. 2020) are common in many low-income countries, hampering productivity and growth. While many studies have focused on the role of weak labour demand (Bloom et al. 2013, McKenzie and Woodruff 2017), we hypothesize labour supply may also play an important role in this gap. With sporadic school attendance and flexible timing of agricultural and casual work, many workers may not develop the habit of regular labour supply, limiting their ability to fully participate in more organized urban labor markets.
To test this hypothesis, we conduct a randomized control trial (RCT) with casual labourers in urban India, harnessing the predictions of a habit formation model whereby workers’ current disutility of supplying labor is a function of accumulated labour supply stock -- i.e., there are intertemporal complementarities in labour (dis-)utility. Thus, temporary financial incentives to work can permanently raise labor supply. This habituation hypothesis is discussed extensively in historical theories suggesting that the transition to an industrial society requires not only technical skills, but also a change in work organization and culture among labourers (Pollard 1963; Clark 1994). We focus on daily casual construction workers whose main source of income is work found at labour stands – public spaces where workers and employers meet and form short-term labour contracts. These stands are common throughout low-income countries - especially in rapidly growing cities - and employ a large fraction of the world’s poor either full time or during lean seasons (Khandker, Khalily, and Samad 2010; ILO 2018). In addition to the broad applicability of the setting, these labour stands provide a directly observable measure of willingness to labour supply that does not rely on self-report: attendance at the stand.
The RCT incentivises a subset of labourers to increase their attendance at the stand, and then removes the incentives and tracks attendance, labour supply, and days worked of all participants in order to examine the persistence of labour supply changes among treated participants. An important feature of the study is the simplicity of the program. This simplicity would allow governments to utilise similar incentive schemes in established labor training centers which already operate at scale—for example, by focusing more heavily on requirements for regular on time attendance.
The RCT has been completed with a final sample of 225 participants across 11 stands. During the incentive phase, we find that incentives increase timely stand attendance. Participants in the Treatment group attend the stand before 8am on average 1.8 days more often, which corresponds to a 122% increase (p-value 0.000). They also attend the stand 0.753 more days per week — a 23% increase (p-value 0.000). In the 7 weeks after incentives are removed, we find that Treatment participants continue to attend the stand more often (0.44 days, 16% increase, p-value 0.043) as well as attend the stand before 8am (0.51 days, 38% increase, p-value 0.010). We find that incentives also increase the number of days that the participants find jobs at the stand – Treatment participants find jobs at the labour stand on 0.32 more days per week relative to the Control participants (22% increase, p-value 0.029).
These results are promising as they suggest that even temporary incentives can have durable impacts on labor supply and job-finding rates. However, the attenuation of these effects over time is also of direct policy interest. We are conducting additional data analyses which we hope will shed light on the causes of attenuation over time. We hypothesise that the many shocks (illness, weather, etc) that individuals in this environment face may lead to reduced habit stock. We will examine this hypothesis in the experimental data as well as in secondary data to shed light on what reinforcement patterns might be most useful in producing high rates of persistence. As the full analysis is ongoing, we have not engaged in policy outreach to date.