How Much Does Good Management Matter? Evidence from India.

1,726 views
1,601 views

Published on

The 19th Barcelona GSE Lecture presented by Prof. John Roberts (Stanford GSB) on May 6, 2010. Prof. Roberts conducts a field experiment to discover the economic impact of suboptimal management techniques in the Indian textile industry.

More info: http://www.barcelonagse.eu/Lecture19_JohnRoberts.html

Published in: Business
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,726
On SlideShare
0
From Embeds
0
Number of Embeds
14
Actions
Shares
0
Downloads
0
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide
  • Above-left and right: mending department, Riviera Fashions. Below-left: mending department, Mahajan silk mills.
  • How Much Does Good Management Matter? Evidence from India.

    1. 1. Does management matter: evidence from India Nick Bloom (Stanford) Benn Eifert (Berkeley) Aprajit Mahajan (Stanford) David McKenzie (World Bank) John Roberts (Stanford GSB) Barcelona Lecture May 2010
    2. 2. 2.6 2.8 3 3.2 3.4 mean of management US Germany Sweden Japan Canada France Italy Great Britain Australia Northern Ireland Poland Republic of Ireland Portugal Brazil India China Greece 2 Management appears worse in developing countries Average country management score, manufacturing firms 100 to 5000 employees (monitoring, targets and incentives management scored on a 1 to 5 scale methodology: Bloom & Van Reenen (2007, QJE), data: Bloom, Sadun & Van Reenen (2010, AR)) 695 336 270 122 344 312 188 762 382 92 231 102 140 524 171 620 559 # firms
    3. 3. 3 Firm-Level Management Scores 0 .2.4.6.8 1 2 3 4 5 management 0 .2.4.6.8 1 2 3 4 5 management US manufacturing, mean=3.33 (N=695) Indian manufacturing, mean=2.69 (N=620) India’s low score is due to a tail of badly managed firms DensityDensity Firm level histograms underlying the country averages
    4. 4. 4 This raises three obvious questions 1. Does “bad” management really exist? • Syversson (2010) notes: “no potential driving factor of productivity has seen a higher ratio of speculation to empirical study”. • Literature goes back to beginnings of social science, e.g. Francis Walker’s (1887) “On the sources of Business Profits” 2. Does “bad” management matter for development? • “Prominent development textbooks such as the four-volume Handbook of Development Economics do not contain a single chapter or substantial section on entrepreneurship” (Naude, 2009) 3. Why are so many Indian firms “badly” managed?
    5. 5. 5 Summary of the experiment Experiment on 28 plants owned by 17 large firms (≈ 300 employees each) outside Mumbai making cotton fabric Randomized treatment plants get heavy management consulting, controls plants get very light consulting (just enough to get data) Collect weekly performance, management, organizational and IT data on all plants from 2008 to 2010
    6. 6. 6 Summary of the results Impact of better management: 1. Raised TFP by about 10% and profits by about 17% 2. Owners decentralized more decisions to plant managers 3. Computer use increased for data collection and processing Why did firms not improve management earlier? • Information – firms not aware of existence or impact of many modern management practices (e.g. quality control) • Ability - family ownership limits pool of potential directors • Competition - high entry costs, and little reallocation (inability to decentralize limits ability of well managed firms to grow)
    7. 7. Exhibit 1: Plants are large compounds, often containing several buildings. Plant surrounded by grounds Front entrance to the main building Plant buildings with gates and guard post Plant entrance with gates and a guard post
    8. 8. Exhibit 2: The plants operate 24 hours a day for 7 days a week producing fabric from yarn, with 4 main stages of production (1) Winding the yarn thread onto the warp beam (2) Drawing the warp beam ready for weaving (3) Weaving the fabric on the weaving loom (4) Quality checking and repair
    9. 9. Exhibit 3: Many parts of these Indian plants were dirty and unsafe Garbage outside the plant Garbage inside a plant Chemicals without any coveringFlammable garbage in a plant
    10. 10. Exhibit 4: The plant floors were often disorganized and aisles blocked Instrument not removed after use, blocking hallway. Tools left on the floor after use Dirty and poorly maintained machines Old warp beam, chairs and a desk obstructing the plant floor
    11. 11. Yarn piled up so high and deep that access to back sacks is almost impossible Exhibit 5: The inventory rooms had months of excess yarn, often without any formal storage system or protection from damp or crushing Different types and colors of yarn lying mixed Yarn without labeling, order or damp protection A crushed yarn cone, which is unusable as it leads to irregular yarn tension
    12. 12. No protection to prevent damage and rustSpares without any labeling or order Exhibit 6: The spare parts stores were also disorganized and dirty Shelves overfilled and disorganizedSpares without any labeling or order
    13. 13. 0 .2.4.6.8 1 2 3 4 5 management 13 Management scores (using Bloom and Van Reenen (2007) methodology) Brazil and China Manufacturing, mean=2.67 0 .2.4.6.8 1 2 3 4 5 management 0 .2.4.6.8 1 1 2 3 4 5 management 0 .5 1 1.5 1 3 5 management Indian Manufacturing, mean=2.69 Indian Textiles, mean=2.60 Experimental Firms, mean=2.60 These firms appear typical of large manufacturers in Brazil, China, India and Mexico
    14. 14. 14 Management practices before and after treatment Performance of the plants before and after treatment Decentralization and IT Why were these practices not introduced before?
    15. 15. 15 The experiment used consulting to randomly change management practices • Obtained details of the population of 529 woven cotton fabric firms (SIC 2211) near Mumbai with 100 to 5000 employees. • Selected 66 firms in the largest cluster (Tarapur) • Contacted every firm: 17 participated containing 28 plants • A team of 6 consultants from Accenture, Mumbai was hired to help improve the practices in some of these firms
    16. 16. 16 Treatment and control, on-site and off-site splits • Intervention and performance data collection is time consuming, so focused on 20 plants – the “on-site” plants: • Treatment: 1 month diagnostic + 4 months implementation • Control: 1 month diagnostic • Collecting management, organization and IT data is easier as slow moving (bi-monthly) so gathered for all 28 plants • Hence, performance regressions on 20 plants only, all other regressions on all 28 plants.
    17. 17. 17 Intervention aimed to improve 38 core textile management practices in 6 areas Targeted practices in 6 areas: operations, quality, inventory, loom planning, HR and sales & orders
    18. 18. 18 Intervention aimed to improve 38 core textile management practices in 6 areas Targeted practices in 6 areas: operations, quality, inventory, loom planning, HR and sales & orders
    19. 19. .2.3.4.5.6 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 Months after the diagnostic phase Notes: Average adoption rates of the 38 key textile manufacturing management practices listed in Table 2. Shown separately for the 14 on-site treatment plants (round symbol), 6 Control plants (diamond symbol) and the 8 off-site plants (+ symbol). Scores range from 0 (if none of the group of plants have adopted any of the 38 management practices) to 1 (if all of the group of plants have adopted all of the 38 management practices). Initial differences across all the groups are not statistically significant (e.g. the initial difference between treatment and control has a p-value of 0.248). Treatment plants (on-site) Control plants (on-site) Shareofkeytextilemanagementpracticesadopted Off-site plants (treatment and control) Adoption of these 38 management practices did rise, and particularly in the treatment plants
    20. 20. 20 Management practices before and after treatment Performance of the plants before and after treatment • Quality • Inventory • Output Decentralization and IT Why were these practices not introduced before?
    21. 21. Poor quality meant 19% of manpower went on repairs Workers spread cloth over lighted plates to spot defectsLarge room full of repair workers (the day shift) Defects lead to about 5% of cloth being scrappedDefects are repaired by hand or cut out from cloth
    22. 22. 22 Previously mending was recorded only to cross- check against customers’ claims for rebates Defects log with defects not recorded in an standardized format. These defects were recorded solely as a record in case of customer complaints. The data was not aggregated or analyzed
    23. 23. 23 23 Now mending is recorded daily in a standard format, so it can analyzed by loom, shift, design & weaver
    24. 24. 24 The quality data is now collated and analyzed as part of the new daily production meetings Plant managers now meet regularly with heads of quality, inventory, weaving, maintenance, warping etc. to analyze data
    25. 25. 0 20406080 100120140 -20 -10 0 10 20 30 40 weeks since diagnostic phase 2.5th percentile Figure 3: Quality defects index for the treatment and control plants Notes: Displays the average weekly quality defects index, which is a weighted index of quality defects, so a higher score means lower quality. This is plotted for the 14 on-site treatment plants (+ symbols) and the 6 on-site control plants (♦ symbols). Values normalized so both series have an average of 100 prior to the start of the intervention. To obtain confidence intervals we bootstrapped the firms with replacement 250 times. Control plants Treatment plants Weeks after the start of the diagnostic Qualitydefectsindex(higherscore=lowerquality) Start of Diagnostic Start of Implementation Average (+ symbol) 97.5th percentile Average (♦ symbol) 97.5th percentile End of Implementation 2.5th percentile
    26. 26. Estimating management effect in regressions (A) OLS: plant FEs and weekly time dummies Outcomei,t=αi+ λt + βmanagementi,t + vi,t (B) IV: 2nd stage as above, 1st stage instruments management Managementi,t=αi+ λt + β1log(1+intervention weeks)i,t + ei,t (C) ITT: regress on outcome on intervention Outcomei,t=αi+ λt + βinterventioni,t + vi,t All standard errors bootstrapped clustered at firm level
    27. 27. 27 Data is weekly at the plant level. Standard errors are boostrap clustered at the firm level. Impact of management on quality
    28. 28. 28 Management practices before and after treatment Performance of the plants before and after treatment • Quality • Inventory • Output Decentralization and IT Why were these practices not introduced before?
    29. 29. 29 Stock is organized, labeled, and entered into an Electronic Resource Planning (ERP) system which has details of the type, age and location. Bagging and racking yarn reduces waste from rotting (keeps the yarn dry) and crushing Computerized inventory systems help to reduce stock levels. Organizing and racking inventory enables firms to substantially reduce capital stock
    30. 30. 30 Sales are also informed about excess yarn stock so they can incorporate this in new designs. Shade cards now produced for all surplus yarn. These are sent to the design team in Mumbai to use in future products
    31. 31. 708090 100110120 -20 -10 0 10 20 30 40 weeks since diagnostic phase 2.5th percentile Figure 4: Yarn inventory for the treatment and control plants Notes: Displays the weekly average yarn inventory plotted for 12 on-site treatment plants (+ symbols) and the 6 on-site control plants (♦ symbols). Values normalized so both series have an average of 100 prior to the start of the intervention. To obtain confidence intervals we bootstrapped the firms with replacement 250 times. Slow moving fluctuations due to seasonality. 2 treatment plants maintain no on-site yarn inventory so are not included in the figures. Control plants Treatment plants Weeks after the start of the intervention Yarninventory(normalizedto100priortodiagnostic) Start of Diagnostic Start of Implementation Average (+ symbol) 97.5th percentile Average (♦ symbol) 2.5th percentile 97.5th percentile End of Implementation
    32. 32. 32 Data is weekly at the plant level. Standard errors are boostrap clustered at the firm level. Impact of management on inventory
    33. 33. 33 Management practices before and after treatment Performance of the plants before and after treatment • Quality • Inventory • Output Decentralization and IT Why were these practices not introduced before?
    34. 34. 34 Many treated firms have also introduced basic initiatives (called “5S”) to organize the plant floor Worker involved in 5S initiative on the shop floor, marking out the area around the model machine Snag tagging to identify the abnormalities on & around the machines, such as redundant materials, broken equipment, or accident areas. The operator and the maintenance team is responsible for removing these abnormalities.
    35. 35. 35 Spare parts were also organized, reducing downtime (parts can be found quickly) and waste Nuts & bolts sorted as per specifications Tool storage organized Parts like gears, bushes, sorted as per specifications
    36. 36. 36 Production data is now collected in a standardized format, for discussion in the daily meetings Before (not standardized, on loose pieces of paper) After (standardized, so easy to enter daily into a computer)
    37. 37. 37 Daily performance boards have also been put up, with incentive pay for employees based on this
    38. 38. 38 Data is weekly at the plant level. Standard errors are boostrap clustered at the firm level. Impact of management on output
    39. 39. 39 Impact on productivity and profitability looks large Estimate increased profit by about $322,000 per firm (≈16.8%) Productivity increased by about 10.5% Long-run impacts potentially much larger as flexibility on changing inputs & product choice, but also chance of slippage
    40. 40. 40 Management practices before and after treatment Performance of the plants before and after treatment Decentralization and IT Why were these practices not introduced before?
    41. 41. Improved management also led to more decentralization in decision making • Firms in developing countries are typically very centralized (Bloom, Sadun and Van Reenen, 2009) • Owners take decisions to avoid expropriation by their middle managers, which they risk because: – Rule of law is weak, so punishing theft is hard – Management is poor, so detecting theft is hard • When management improves the ability to detect theft increases, so we should see more decentralization • This matters for growth since the inability to decentralize limits the growth of productive firms, impeding reallocation
    42. 42. 0 0 11 1 1 1 1 0 0 1 0 1 1 1 0 1 1 0 1 1 1 0 01 1 1 1 0 .2.4.6.8 0 .2 .4 .6 Change in management practices Note: Decentralization index is the principal component factor of 8 measures of decentralization around weaver hiring, manager hiring, spares purchases, maintenance planning, weaver bonuses, new products, investment, and departmental co-ordination. Changes defined over the period from pre-treatment to March 2010. Changeinthedecentralizationindex Change in management practices correlation 0.594 (p-value 0.001) 1=treatment plant, 0=control plant Figure 6: Decentralization and changes in management practices
    43. 43. Decentralization regression results Data is pre-intervention and March 2010. Standard errors are boostrap clustered at the firm level.
    44. 44. Improved management also led to greater use of computers • Most modern management practices involve extensive data collection, analysis and presentation • Suggests that computer use (hours and number of users) should rise with the adoption of better management • Better management appears to be one type of skill-biased technical change
    45. 45. 0 0 1 1 00 1 0 1 1 0 1 0 1 1 0 0 1 1 1 0 .5 1 1.5 2 0 .2 .4 .6 Change in management practices Note: Computerization index is the principal component factor of 10 measures, which are the use of an ERP system, the number of computers, the number of computers less than 2 years old, the number of employees using computers for at least 10 minutes per day, and the cumulative number of hours of computer use per week, an internet connection at the plant, the firm has a web-site, if the plant-manager uses e-mail, if the directors use of e-mail, and the intensity of computerization in production. Changeinthecomputerizationindex Change in management practices correlation 0.778 (p-value 0.001) 1=treatment plant, 0=control plant Figure 7: Computerization and changes in management practices
    46. 46. Computerization regression results Data is pre-intervention and March 2010. Standard errors are boostrap clustered at the firm level.
    47. 47. 47 Management practices before and after treatment Performance of the plants before and after treatment Decentralization and IT Why were these practices not introduced before?
    48. 48. 48 So why did these firms have bad management? Asked the consultants to investigate the non-adoption of each of the 38 practices, in each plant, every other month They did this by discussion with the owners, managers and workers, observation of the factory, and from their experiences of trying to change management practices. We also spent many months visiting the firms, and spending time with the managers and owners The next slide shows this data over time
    49. 49. 49 1 month before 1 month after 3 months after 5 months after 7 months after 9 months after Lack of information (not aware of the practice) 38.6 12.8 2.2 0.5 0.4 0.3 Incorrect information (wrong cost-benefit analysis) 29.3 33.3 31.9 29.2 28.5 27.5 Owner ability, time and/or procrastination 1.3 9.1 7.2 7.5 7 6.7 Manager incentives and/or authority 0 2.1 2.4 3.0 3 3.2 Not profitable (non-adoption is correct) 0 0.2 0.4 0.5 0.5 0.5 Other (variety of other reasons) 0 0.2 0.4 0.2 0.5 0.5 Total (% practices not adopted) 73 57.7 44.3 40.9 39.8 38.6 Reason for the non-adoption of the practices in the treatment plants (as a % of all 38 practices) Notes: covers 532 practices (38 practices in 14 plants) in the treatment plants. Table 9 (in the paper) also has values for control and non-experimental plants.
    50. 50. 50 Lack of information Incorrect infor- mation Owner ability pro- crastination Plant manager incentives or authority Other Doing Lack of information 33 22.3 13.5 1.8 0.9 28.4 Incorrect information 92 0.4 7.6 Owner ability, time or procrastination 81.8 18.2 Manager incentive and/or authority 100 Not profitable 100 Other 100 Transition matrix for the reasons for non-adoption 2 months ahead (t+1) Current(t) Note: All blank cells are zero. Shows transition of reasons for non adoption to other reasons or implementation (“doing”) over each two month period. Averaged over all treatment firms for months 1 to 11.
    51. 51. 51 Why does competition not fix badly managed firms? Bankruptcy is not (currently) a threat: at weaver wage rates of $5 a day these firms are profitable Reallocation appears limited: Owners take all decisions as they worry about managers stealing. But owners time is constrained – they already work 72.4 hours average a week – limiting growth. One explanation for Hsieh and Klenow (2009) results. As an illustration firm size is more linked to number of male family members (corr=0.689) - who are trusted to be given managerial positions - than management scores (corr=0.223) Entry appears limited: Capital intensive ($13m assets average per firm), and no guarantee new entrants are any better
    52. 52. 52 Summary Firms in developing countries often have poor management practices, which lowers their productivity Reasons include lack of information about modern management practices, and limited CEO ability and procrastination Policy implications A) Competition and FDI: free product markets and encourage foreign multinationals B) Rule of law: improve rule of law to encourage reallocation and ownership and control separation C) Training: improved basic training around management skills
    53. 53. Finally, not to pick on the Indians, one country even exports TV shows about bad managers..... Michael Scott (USA) David Brent (Britain) Basil Fawlty (Britain)
    54. 54. The production technology has not changed much over time Warp beam Krill The warping looms at Lowell Mills in 1854, Massachusetts
    55. 55. 55 “Non adoption flow chart” used to collect data Was the firm previously aware that the practice existed? Lack of information Can the firm adopt the practice with existing staff & equipment? Did the owner believe introducing the practice would be profitable? Low ability of the owner and/or procrastination Does the firm have enough internal financing or access to credit? Do you think the CEO was correct about the cost-benefit tradeoff? Could the firm hire new employees or consultants to adopt the practice? Credit constraints External factors (legal, climate etc) Is the reason for the non adoption of the practice internal to the firm? Could the CEO get his employees to introduce the practice? Did the firm realize this would be profitable? Would this adoption be profitable Not profit maximizing Incorrect information Lack of local skills Other reasons Limited incentives and/or authority for employees Yes No Legend Conclusion Hypothesis No Yes
    56. 56. Are these Hawthorne effects (temporary increases in performance due to monitoring? • Treatment and control plants both had initial 1 month of diagnostics and extended follow-up • Improvements take time to arise, and in areas (quality, inventory and efficiency) where practices are changing • Improvements persisted for several months after the intervention phase (although still collecting data) • Firms themselves also believe the improvements work and have rolled these out to other plants
    57. 57. 57 But before the photos, I want to note that this is not a cost-benefit evaluation of management consulting We hire consultants as a practical mechanism to achieve an improvement in management practices. Our findings suggest a large impact of this in the treatment plants, but much less impact in the control plants. Assessing the cost-benefit for both groups depends on a number of assumptions around long-run output and cost impacts, open market consulting costs, discount rates, and rival firm copying. We have not done this, and it is not the focus of the paper.
    58. 58. Exhibit 7: The path for materials flow was often heavily obstructed Unfinished rough path along which several 0.6 ton warp beams were taken on wheeled trolleys every day to the elevator, which led down to the looms. This steep slope, rough surface and sharp angle meant workers often lost control of the trolleys. They crashed into the iron beam or wall, breaking the trolleys. So now each beam is carried by 6 men. A broken trolley (the wheel snapped off) At another plant both warp beam elevators had broken down due to poor maintenance. As a result teams of 7 men carried several warps beams down the stairs every day. At 0.6 tons each this was slow and dangerous
    59. 59. 59 Our plants are large by Indian and US standards Source: Hsieh and Klenow, 2009 Average size of our plants Employment weighted size distributions, workers per plant
    60. 60. 60 Sample of firms and plants we worked with
    61. 61. 61 Why doesn’t the consulting market fix this? 90% of the reason for non-adoption is informational, so firms not aware they are badly managed But, surely consultants could contact firms telling them about their services? • In India there is an active telesales market selling variety of cost reduction services, so not easy But, why don’t consultants advertise free consulting and get paid through profit sharing? • But, firms not reporting honest profits to the tax authorities so unlikely to do so to consulting firms And, firms are breaking tax, labor and safety laws so are also nervous about outsiders (we had WB and Stanford endorsement)

    ×