Finance & PSD Impact FEBRUARY 2018 The Lessons from DECFP Impact Evaluations ISSUE 47 Our latest note returns to Indian textile firms that were the subject of a management intervention in 2008-10, and looks at what happens in the long-term Do management interventions last? Evidence from an experiment in India Nicholas Bloom, Aprajit Mahajan, David McKenzie, and John Roberts practices, but the control plants also Between 2008 and 2010, we worked dropped a lot of the practices they had with a leading global consulting company to adopted after the diagnostic. The result attempt to improve management in Indian is that there is still a large, persistent, textile firms (see impact note 10). We conducted a randomized experiment and statistically significant 19.7 involving 28 plants in 17 firms in the woven percentage point gap in the management cotton fabric industry. These were large practices between treatment and control firms, with an average of 270 employees, plants 8-9 years later. and a median of two plants per firm. 2. Over time, the level of management We divided these plants into three practices converges between the groups: experimental treatment plants, experimental and non-experimental experimental control plants, and non- experimental plants. Both the treatment and plants within firms – we see this control plants received a one-month occurring for both the treated firms, and diagnostic, and the treatment plants then for the control firms. This convergence received four months of implementation also occurs at the individual practice support. Non-experimental plants were other level, with treated firms transferring the plants in the same firms as treated or control practices they kept in the experimental plants. In the first year after this plants across to their non-experimental intervention, treated plants had increased their adoption of these 38 management plants. practices by an average of 38 percentage points compared to a 12 percentage point Figure 1: Management Practices Over Time increase in the control plants, resulting in improvements in quality, an increase in output, and a 17 percent increase in productivity. We were then curious as to whether these good management practices persist, or whether firms would return to their old ways, and so returned to these firms in 2017, eight to nine years after the intervention. Long-run Impact on Management Figure 1 shows what happened to We interviewed the directors and plant management practices over time. We see managers about each of the 38 management two key results: practices to understand why practices were 1. Treated plants did drop about 40% of adopted or dropped over time. The most their original gain in management Do you have a project you want evaluated? DECRG-FP researchers are always looking for opportunities to work with colleagues in the Bank and IFC. If you would like to ask our experts for advice or to collaborate on an evaluation, contact us care of the Impact editor, David McKenzie (dmckenzie@worldbank.org) common reason for adopting a new practice fostered by also improving their marketing was spillover from another plant in the same practices. firm. Within the treated experimental plants, there were three main reasons why practices Policy Implications were dropped. The most important, Returning back to these firms after nine accounting for over half the dropped years offers hope that efforts to improve practices, was a new plant manager coming management can yield lasting improvements in. The second most important reason was a in firms. We see several implications for lack of director time, while some practices policy. were also dropped because firm owners saw 1. Cost-benefit calculations for the return no benefit in them. This highlights the on such interventions depend heavily on importance of key employees in maintaining how long they can be expected to persist. good management. Our results show some depreciation of practices, but that there can be persistent Long-run impact on firm performance impacts, making these interventions We find the treated firms to be using better value-for-money. more looms and fewer workers over time, 2. This makes it even more of a puzzle why although neither is statistically significant. more firms don’t improve management The result is a statistically significant by themselves. Understanding what increase in looms per worker, a classic constrains them from doing so, and measure of productivity in the industry. developing policy interventions that help Fewer quality problems means workers the market for management spend less time fixing machines and defects, improvements function better therefore and more time producing output. Based on seems a useful area for policy work. this improvement, we calculate that labor 3. Finally, our results provide an productivity increased 19 percent since illustration of the usefulness of returning 2011, with the long-run effect 35 percent. to projects long after the intervention has Revealed preference also suggests ended. Typical World Bank and that firms have found the management government projects do not provide improvements useful: treated firms were funding or incentives for doing this, so more likely to hire consultants on their own developing mechanisms for policies to after our intervention, and have look beyond immediate funding cycles accompanied the improvement in when examining impacts is an important operational management practices that we area to develop. For further reading see: Nicholas Bloom, Aprajit Mahajan, David McKenzie, and John Roberts “Do management interventions last? Evidence from India”, World Bank Policy Research Working Paper no. 8339, February 2018. Recent impact notes are available on our website: http://econ.worldbank.org/programs/finance/impact