Learning from others: Benchmarking

Outline

• Concept
• Practices
• Performance and practices
• Implementation

Frederick Winslow Taylor

• “there are many different ways in common use for doing the same thing … a variety in the implements used for each class of work … there is always one method and one implement which is quicker and better than the rest” (The Principles of Scientific Management, 1911, 24-25) [emphasis added]

Benchmarking

• “[M]ost car-owners believe they are above-average drivers, most companies might well believe they have above-average levels of productivity” (Andrew Haldane, Chief Economist, Bank of England, 2017)
• Benchmarking: Report on relative performance and best practices
• Specific way of learning from others -- not necessarily competitors

Benchmarking

• Google: Global, Business & industrial category, past 5 years

​

​

Benchmarking

• Scope
• Internal -- different plants/facilities against each other
• External – against other organizations
• Dimension
• Performance – sales per square metre, inventory turn, customer satisfaction, profit
• Practices
• Performance and practices
• Distinguish from contest – exceed the others

Outline

• Concept
• Practices [student presentation]
• Performance and practices
• Implementation

​

Practice benchmarking (Dalton et al. 2019)

• Research question: How does best practice information affect business practices and performance?
• Practices
• Record-keeping
• Financial planning
• Inventory policy
• Marketing
• Joint decision-making (with co-owners/managers of shop)

Dalton et al. (2019): Context

• Retail shops (warung/toko) in 29 of 112 districts in South, East, and West Jakarta

Dalton et al. (2019): Experimental design

• Identify best practices
• Qualitative study of shops in other areas
• Regression of profit on practices within (baseline) sample
• Sample: 1,301 shop owners
• Treatment – handbook of best practices: 1,040 shops
• Additional treatment
• Watch 25-minute video of model store owners (role modeling)
• 2 x 30-minute consultation
• Video + consultation
• Control: 261 shops

​

​

Dalton et al. (2019): Treatment effect

• Read from vertical axis
• Further to right => higher profit

Dalton et al. (2019): Mechanism

• Best practices which significantly affected profit
• Itemize business revenues and costs
• Budget
• Monitor sales relative to target
• Stock according to profit margin
• Solicit client feedback
• Joint decisions with business partners

Dalton et al. (2019): Alternative explanation

• Better recorded performance due to better record-keeping
• Increase in variance of profit [seems to support alternative explanation: benchmarking should cause performance to converge]
• [Limit sample to those who already kept business records at baseline]

Dalton et al. (2019): Mechanism

• Handbook alone had no effect: Rules out Hawthorne Effect
• No effect on other practices (not in handbook): Rules out Hawthorne Effect
• No self-reporting bias: Surveyors recorded whether shop appeared clean, well-stocked, and if prices clearly marked
• [Check whether objectively recorded implementation of front-of-house practices correlated with back-of-house practices]
• No effect on business aspirations: Increase in performance due to changes in knowledge and practices

​

​

Dalton et al. (2019): Critique

1. The researchers randomized shop owners into four treatment groups and one control group. What purpose did the control group serve?
2. The researchers concluded that the handbook alone did not affect shop owners’ behaviour. Reflecting on the design of handbook, how would you explain the null effect?
3. How did the researchers arrange for shop owners to see the movie? How might the arrangements affect the managerial and policy implications of the empirical results?
4. How would the findings on the effect of practice benchmarking apply among Singapore law firms?

​

Outline

• Concept
• Practices
• Performance and practices [student presentation]
• Implementation

Benchmarking (Song et al. 2018)

1. Why are the following important in the study? (a) Physicians were paid a fixed salary, with no additional compensation for attending to more or working longer hours; and (b) Patients were assigned on a round-robin basis independent of physician work speed or idle time.
2. Song et al. (2018) found that publication of relative performance only improved performance if the identities of top performers were also published. Please explain this result in other ways besides the authors’ theory.
3. Would you want your future employer to publish relative performance? Yes or no? Please explain.

​

Outline

• Concept
• Practices
• Performance and practices
• Implementation

Performance benchmarking:�Limitations (Chandra 2013)

• To benchmark costs effectively, must know the underlying model of costs
• Benchmark the model, not the outcome
• Very difficult in practice
• Limitations
• Economies of scale
• Production inefficiencies

​

Economies of scale

• Economies of scale due to large fixed cost: Cost, C = F + cQ
• F = fixed cost
• c = marginal cost
• Q = quantity
• Two factories serving different markets
• Factory A: average cost = C/Q_A = F/Q_A + c
• Factory B: average cost = C/Q_B = F/Q_B + c
• Suppose that Q_A > Q_B, then C/Q_A < C/Q_B
• Benchmark average cost => produce more at factory A and ship to market B => actually raise total cost by cost of shipping

Production inefficiency

• Cost, C_A = [Q_A + x] c
• Cost, C_B = [Q_B + x] c
• If Q_A different from Q_B, then different average cost but benchmarking does not reveal production inefficiency – because common to both producers

Mechanism

• Private vis-à-vis public performance benchmarking
• Learning about potential performance
• Social comparison
• Self-esteem

Not used

Black and Lynch (Economic Journal, 2004)

• U.S. Census Bureau: Educational Quality of the Workforce National Employers Survey (EQW-NES), 1994 and 1997: Representative sample
• Focus: labour productivity in manufacturing

​

Black and Lynch (2004)

• Management practices themselves may be endogenous: more productive establishments may decide to introduce new practices.
• Difficult, especially at plant level, to find valid instruments

Benchmarking of performance and practices: �Hou and Png (2021)

• Hawkers in Singapore
• 194 hawkers, 60% to treatment and 40% to control
• How do we sample?

Larger hawker centres

Collaborations from the representatives

​

• How do we randomize?

Within hawker centre, by sections separated by physical divisions

​

• How do we collect data?

Face to face interviews

​

Do hawkers know about their relative performance?

Benchmarking and exit

Benchmarking and exit

Why benchmarking affects behavior?

Informational: Social learning

Ability evaluation – Ignorant of own performance

Information about better practices

Psychological

For public benchmarking information among familiar peers

Self-esteem

Social pressure

For private benchmarking information

Competitive preference

Ego utility

See Villeval (2020) for a review on relative performance feedback

Contingency: Incentives? Benchmarked peers?

(C) I.P.L. Png

32