Millions of women in developing countries would prefer paid work but remain outside the labour force. In India, social norms that place a woman’s role in the home conflict with the jobs available, few of which offer remote working. The median married woman spends only half an hour a day outside her home. New research asks whether reshaping jobs to fit these norms can bring women into paid work, whether such jobs can lead to less flexible forms of employment, and what flexibility costs employers in terms of performance.

Experimental Design
The study took place in eight areas in and around Kolkata between April 2022 and January 2023. It involved 1,670 lower-middle-income households. Eligible women were literate in Hindi or Bangla, had access to an Android smartphone, and were outside the labour force. Most were married homemakers aged about 30 years old; 69% had never worked for pay, and only 36% said they had the final say over their own labor supply. Recruiters did not mention that the baseline survey might lead to a job, meaning that selection into the sample was not based on an expressed interest in working.

Women were randomly assigned to a control group with no job offer or to one of five month-long, part-time, piece-rate jobs on a smartphone speech-data platform, paying up to ₹4,000 (about 34% of average monthly household income). The jobs varied along three dimensions: time flexibility, the ability to multitask with childcare, and working from home versus working in a nearby women-only office. The most flexible job (Home: Fully Flexible) had all three features. After women accepted or declined their initial offers, half of those in less flexible arrangements were randomly given a surprise upgrade to the fully flexible job. This design allowed researchers to distinguish selection effects from treatment effects.

Two to three months later, a second round of jobs was offered, including non-digital work such as sewing masks and making jewelry. The three key results of the study are as follows.

Result 1: Flexibility Dramatically Raises Take-Up
Job take-up rose from 15% for the office job to 48% for the fully flexible job. The researchers note that this effect is larger than those found in other interventions. It also dwarfs the response to wages. In a companion Mumbai experiment, a nearly fivefold wage increase raised take-up by only seven percentage points.

Decomposing the effect shows two roughly equal sources. The ability to work from home raised take-up from 14% to 29% (+15 points), while the ability to multitask with childcare added 17–19 points. Time flexibility had no significant effect. Women could and did set aside consistent hours. Effects were largest for women from more traditional households, and gender attitudes emerged as the most important moderator in the analysis.

Result 2: Flexible Work as a Gateway
Women without previous work experience who had been assigned a more flexible job in Round 1 than in Round 2 were 9 percentage points more likely to start the Round 2 job. There was no significant effect for women with prior experience, and the full-sample effect (4 points) was not statistically significant. The effect appeared whether the Round 2 job was digital or non-digital, suggesting that it was not simply a result of digital skill-building.

The researchers point to gender attitudes as the primary mechanism. Receiving a job offer made women’s attitudes 0.06 standard deviations less traditional, and 0.11 standard deviations for women with more traditional baseline views. These were the same women whose labour supply responded most strongly to flexibility. The researchers contend that this points to a mutually reinforcing cycle between employment and attitudes. 

They consider and largely discount alternate explanations. However, they acknowledge that trust in the employer and perceived safety cannot be ruled out, because all the jobs came from the same firm. This suggests that home-based work offered by third parties may not produce the same stepping-stone effect. 

Work experience did not change psychological well-being, so there is no evidence of a ‘second shift’ burden. It did, however, make women more likely to feel that their talents were being put to good use.

Result 3: Performance Costs of Working from Home
Using the surprise upgrades, the researchers compared women who accepted the office job and were randomly upgraded to working from home with those who remained in the office. Office work reduced task completion time by about 18% and raised accuracy by about 3%. The effects were larger for difficult tasks. The mechanism appears to be interruptions. Work sessions at home were 40% shorter, and tasks undertaken at the start of each session were completed more slowly and less accurately.

Importantly, there was little evidence of selection. Women whose participation depended on flexibility did not perform significantly different from those willing to work in the office. Expanding the labour pool therefore does not appear to lower the average quality of workers. Because pay was piece-rate, workers bore most of the productivity cost. The researchers suggest that the penalty could be larger for salaried jobs, and fragmented work patterns could be especially problematic for roles requiring synchronous collaboration.

Implications and Caveats
For employers, the gains from a much larger pool of willing workers may outweigh the performance penalty, especially since home-based work saves office costs and any productivity loss could potentially be offset by lower wages. The researchers suggest that work from home may be underprovided because firms lack information about the workers they are missing and do not internalise the gateway benefit to women’s future labour supply.

They also raise cautions. Stepping-stone jobs may need to be temporary so that women are not trapped in lower-paid, precarious home-based work, and flexible work could contribute to a more gender-segregated labour market. The sample is not fully representative, since participants were more likely than non-participants to have a say in their own labour supply decisions.

Overall, this research demonstrates how adapting jobs to existing norms can expand women’s employment while potentially helping to change those norms over time.