Transcript Document title - World Management Survey

Key messages of lectures 1 to 4
1. Exists a set of core practices for talent management, target
management and performance management (scoring grid)
2. Associated with better performance across a wide range of
countries and industries, especially in larger firms
3. Not universal truths, but important benchmarks against which
all firms should be evaluated
4. Firms are often unaware that their practices are lacking, so
good management is similar to a new technology
5. Hard to change practices in firms – anecdotal evidence this
takes several years
Nick Bloom and John Van Reenen, Management Practices, 2010
Improving management in
Indian factories
Nick Bloom (Stanford Economics)
John Van Reenen (Stanford GSB/LSE)
Lecture 5
Nick Bloom and John Van Reenen, Management Practices, 2010
2
Management appears worse in developing countries
US
Germany
Sweden
Japan
Canada
France
Italy
Great Britain
Australia
Northern Ireland
Poland
Republic of Ireland
Portugal
Brazil
India
China
Greece
# firms
695
336
270
122
344
312
188
762
382
92
231
102
140
559
620
524
171
2.6
2.8
3
3.2
3.4
management
Average Country Management mean
Score,of
firms
100 to 5000 employees
Nick Bloom and John Van Reenen, Management Practices, 2010
3
(from
Bloom & Van Reenen (2007, QJE), Bloom, Sadun & Van Reenen (2009,
AR))
US manufacturing, mean=3.33 (N=695)
0
.2
Density
.4
.6
.8
India’s low score is mainly due to many badly managed firms
2
3
management
4
5
Indian manufacturing, mean=2.69 (N=620)
0
.2
Density
.4
.6
.8
1
1
2
3
management
Firm-Level Management Scores
Nick Bloom and John Van Reenen, Management Practices, 2010
4
5
4
This raises two obvious questions
1. Does “bad” management reduce productivity, or are these
practices dues to difference circumstances in India (i.e. poor
infrastructure, less capital, weak rule of law)?
2. If it does matter, why are so many Indian firms badly managed?
Nick Bloom and John Van Reenen, Management Practices, 2010
5
Summary and photos
Experiment on plants in large (≈ 300 person) Indian textile firms
Randomized treatment plants get heavy management consulting,
control plants get very light consulting (just enough to get data)
Collect weekly performance data on all plants from 2008 to 2010
• Improved management practices led to large and significant
improvements in productivity and profitability
• Appears informational constraints were a major reason for
lack of prior adoption, but often other constraints also present
Before explaining research and results in detail, I want to show
some slides to provide some background
Nick Bloom and John Van Reenen, Management Practices, 2010
6
Exhibit 1: Plants are large compounds, often containing several buildings.
Plant entrance with gates and a guard post
Plant surrounded by grounds
Nick Bloom and John Van Reenen, Management Practices, 2010
Front entrance to the main building
Plant buildings with gates and guard post
Exhibit 2: These 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
Nick Bloom and John Van Reenen, Management Practices, 2010
(3) Weaving the fabric on the weaving loom
(4) Quality checking and repair
This production technology has not changed much over time:
Lowell Mill warping looms (1854, Lowell, Massachusetts)
Krill
Warp
beam
Nick Bloom and John Van Reenen, Management Practices, 2010
Exhibit 3: Many parts of these plants were dirty and unsafe
Garbage outside the plant
Garbage inside a plant
Nick Bloom and John Van Reenen, Management Practices, 2010
Flammable garbage in a plant
Chemicals without any covering
Exhibit 4: The plant floors were disorganized
Instrument
not
removed
after use,
blocking
hallway.
Dirty and
poorly
maintained
machines
Old warp
beam, chairs
and a desk
obstructing the
plant floor
Tools left on
the floor
after use
Nick Bloom and John Van Reenen, Management Practices, 2010
Exhibit 5: The inventory rooms had months of excess yarn, often without
any formal storage system or protection from damp or crushing
Yarn without
labeling, order or
damp protection
Different types
and colors of
yarn lying mixed
Nick Bloom and John Van Reenen, Management Practices, 2010
Yarn piled up so high and
deep that access to back
sacks is almost impossible
A crushed yarn cone, which
is unusable as it leads to
irregular yarn tension
Exhibit 6: Yet more material was often stored around the plant
Inventory was
also regularly
stored in
corridors,
hallways,
doorways and on
stairs. This is
dangerous and
impedes efficient
movement of
materials around
the plant.
Nick Bloom and John Van Reenen, Management Practices, 2010
Inventory was also
often stored around
machinery.
Exhibit 7: The parts stores were also disorganized and dirty
Spares without any labeling or order
No protection to prevent damage and rust
Nick Bloom and John Van Reenen, Management Practices, 2010
Spares without any labeling or order
Shelves overfilled and disorganized
Exhibit 8: The path for materials flow was often 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 - two serious accidents occurred in
our time at the plant.
Nick Bloom and John Van Reenen, Management Practices, 2010
Exhibit 9: Routine maintenance was usually not carried out, with repairs
only undertaken when breakdowns arose, leading to frequent stoppages.
Broken machine parts being repaired
Parts being cleaned and replaced on jammed loom
Nick Bloom and John Van Reenen, Management Practices, 2010
Workers investigating a broken loom
Loom parts being disassembled for diagnosis
1.5
These firms appear typical of large manufacturers in
India, China and Brazil
01
.5
1
Experimental Firms, mean=2.60
1
3
management
5
.8 .2
0
.4
.6
.8
Indian Textiles, mean=2.60
1
2
3
management
4
5
0
.2
.4
.6
Indian Manufacturing, mean=2.69
.8
1
2
3
management
4
5
.2
.4
.6
Brazil and China Manufacturing, mean=2.67
0
Nick Bloom and John Van Reenen, Management Practices, 2010
1
2
3
management
Management scores
4
17
5
So ran an experiment to evaluate impact of
changing the management of large Indian firms
• 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 & Urmagaon)
• Contacted every firm: 17 willing to participate in straight-away,
so randomly picked 20 plants from these 17 firms
• A team of 6 consultants from Accenture, Mumbai was hired to
help improve the practices in some of these firms
• Control: 1 month of diagnostic
• Treatment: 1 month diagnostic + 4 months implementation
• All: follow-on data collection for next 12+ months
• Collecting data from April 2008 to December 2010
Nick Bloom and John Van Reenen, Management Practices, 2010
18
Sample of firms we worked with
Nick Bloom and John Van Reenen, Management Practices, 2010
19
Our plants and firms are large by Indian & US standards
Average size of our plants
Nick Bloom and John Van Reenen, Management Practices, 2010
Source: Hsieh and 20
Klenow, 2009
Management practices before and after treatment
Performance of the plants before and after treatment
• Quality
• Inventory
• Operational efficiency
Why were these practices not introduced before?
Nick Bloom and John Van Reenen, Management Practices, 2010
21
Intervention aimed to improve 38 core textile
management practices in 6 areas (1/2)
Targeted
practices in 6
areas:
operations,
quality,
inventory,
loom planning,
HR and sales
& orders
Nick Bloom and John Van Reenen, Management Practices, 2010
22
Intervention aimed to improve 38 core textile
management practices in 6 areas (2/2)
Targeted
practices in 6
areas:
operations,
quality,
inventory,
loom planning,
HR and sales
& orders
Nick Bloom and John Van Reenen, Management Practices, 2010
23
Share of the 38 management practices adopted
.2
.3
.4
.6
.5
Adoption of these 38 management practices did
rise, and particularly in the treatment plants
Wave 1 treatment plants:
Diagnostic September
2008, implementation
began October 2008
Wave 2 treatment plants:
Diagnostic April 2009,
implementation began
May 2008
Control plants:
Diagnostic July 2009
Non-experiment plants:
No intervention
2008.25
April 2008
2008.5
July 2008
2008.75
October 2008
2009
ym
January 2009
April 2009
2009.25
2009.5
July 2009
2009.75
October 2009
Nick
Bloom and John
Management
Practices,
24 practices over
Notes:
Non-experiment
plantsVan
areReenen,
other plants
in the treatment
firms2010
not involved in the experiment. They improved
this period because the firm internally copied these over themselves. All initial differences not statistically significant (Table 2)
Take away summary points
1. These firms are not adopting basic management practices,
in large part due to a lack of awareness
2. Changing practices is very slow – we are still introducing
new practices into firms 18 months later, because of:
a. Takes time for firms to advice (Accenture in our case)
b. Changes are complementary – e.g. monitoring & pay
3. Change may not be permanent – need to fix both
processes and incentives to avoid backsliding
Nick Bloom and John Van Reenen, Management Practices, 2010
Management practices before and after treatment
Performance of the plants before and after treatment
• Quality
• Inventory
• Operational efficiency
Why were these practices not introduced before?
Nick Bloom and John Van Reenen, Management Practices, 2010
26
Exhibit 10: Quality was so poor that 19% of manpower was spent on
repairing defects at the end of the production process
Large room full of repair workers (the day shift)
Workers spread cloth over lighted plates to spot defects
Nick Bloom and John Van Reenen, Management Practices, 2010
Defects are repaired by hand or cut out from cloth
Non-fixable defects lead to discounts of up to 75%
Previously mending was recorded only to crosscheck 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
Nick Bloom and John Van Reenen, Management Practices, 2010
28
Now mending is recorded daily in a standard format,
so it can analyzed by loom, shift, design & weaver
Nick Bloom and John Van Reenen, Management Practices, 2010
29
29
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
Nick Bloom and John Van Reenen, Management Practices, 2010
30
Defect rates have rapidly fallen in treatment plants
140
Diagnostic start
Implementation start
Implementation stop
60
Quality defects index
100
120
80
Control
plants
40
Treatment
plants
-10
0
10
20
30
Weeks after the starttiming
of the intervention (diagnostic phase)
Notes: Displays the average quality defects index, which is a weighted index of quality defects, so a higher score means lower
Nick
Bloom
and for
John
Reenen, Management
2010the 6 control plants (+ symbols). Values31
quality.
This
is plotted
theVan
14 treatment
plants (squarePractices,
symbols) and
normalized so
both series have an average of 100 prior to the start of the intervention.
Management impact on quality, regressions
Nick Bloom and John Van Reenen, Management Practices, 2010
32
Management practices before and after treatment
Performance of the firms before and after treatment
• Quality
• Inventory
• Operational efficiency
Why were these practices not introduced before?
Nick Bloom and John Van Reenen, Management Practices, 2010
33
Organizing and racking inventory enables firms to
reduce capital stock and reduces waste
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.
Nick Bloom and John Van Reenen, Management Practices, 2010
34
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 to use
in future designs
Nick Bloom and John Van Reenen, Management Practices, 2010
35
And yarn for products ranges no longer made by
the firm (e.g. suiting fabric) was sold
This firms
used to make
suiting and
shirting yarn,
but stopped
making
suiting yarn 2
years ago
Nick Bloom and John Van Reenen, Management Practices, 2010
36
Inventory is falling in treatment firms
Implementation
120
130
Diagnostic
90
100
110
Control firms
80
Treatment firms
-40
-20
0
timing
Weeks
after the start of the intervention
20
Bloomthe
and
John Van
Practices,
2010
37(+ symbols).
Notes:Nick
Displays
average
rawReenen,
materialsManagement
for the 14 treatment
firms
(square symbols) and the 6 control firms
Values normalized so both series have an average of 100 prior to the start of the intervention.
Management impact on inventory, regressions
Nick Bloom and John Van Reenen, Management Practices, 2010
38
Spare parts were also organized, reducing downtime
(parts can be found quickly), capital stock and waste
Nuts & bolts
sorted as per
specifications
Tool
storage
organized
Nick Bloom and John Van Reenen, Management Practices, 2010
Parts like
gears,
bushes,
sorted as per
specifications
39
Management practices before and after treatment
Performance of the firms before and after treatment
• Quality
• Inventory
• Operational efficiency
Why were these practices not introduced before?
Nick Bloom and John Van Reenen, Management Practices, 2010
40
The treated firms have also started to introduce
basic initiatives (called “5S”) to organize the plant
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.
Nick Bloom and John Van Reenen, Management Practices, 2010
This is all part of the routine maintenance
41
Production data is now collected in a standardized
format, for discussion in the daily meetings
Before
(not standardized, on loose
Nick Bloom and pieces
John Van of
Reenen,
Management Practices, 2010
paper)
After
(standardized, so easy to enter
42
daily into a computer)
Daily performance boards have also been put up,
with incentive pay for employees based on this
Nick Bloom and John Van Reenen, Management Practices, 2010
43
Management impact on efficiency, regressions
Nick Bloom and John Van Reenen, Management Practices, 2010
44
Estimated impacts on productivity and
profitability are large and rising
Estimate the intervention has increase profits by about
$250,00 per firm and productivity by 9% so far from:
- reduced repair manpower costs
- reduced wasted materials (from less defects)
- lower inventory
- higher efficiency levels
Full impacts of better management should be much larger:
- short-run impacts only
- narrow set of management practices (almost no HR)
Nick Bloom and John Van Reenen, Management Practices, 2010
45
Nick Bloom and John Van Reenen, Management Practices, 2010
46
Management practices before and after treatment
Performance of the firms before and after treatment
• Quality
• Inventory
• Operational efficiency
Why were these practices not introduced before?
Nick Bloom and John Van Reenen, Management Practices, 2010
47
So why did these firms have bad management?
• Information: management is a technology and India is far
behind the technology frontier, e.g. Lean manufacturing
• Incentives: managers have no incentive pay or within firm
promotion possibilities so have limited motivated to perform
• CEO ability: family firms with directors who struggled to
change practices and sometimes procrastinated
Nick Bloom and John Van Reenen, Management Practices, 2010
48
Why does competition not fix badly managed firms?
Bankruptcy is still avoided : wage of $5 a day means firms are
profitable
Reallocation appears limited: Owners take all decisions as they
worry about managers stealing. But owners time is constrained –
they current work 72.5 hours average a week – limiting growth.
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: Production is very capital intensive ($13m
assets average per firm)
Nick Bloom and John Van Reenen, Management Practices, 2010
49
Summary
Firms in developing countries seem badly managed
Our results suggest this has a material impact on productivity
Also appear to find bad operations management arises from
lack of information and poor HR management
But far from clear….yields as many questions as answers so far
Nick Bloom and John Van Reenen, Management Practices, 2010
50
Back-up
Nick Bloom and John Van Reenen, Management Practices, 2010
51
Figure 3: Quality defects index for the treatment and control plants
Start of Diagnostic
Start of Implementation
Control plants
Data (+ symbol)
Cubic Spline
120
Quality defects index (higher score=lower quality)
140
Spline + 2 SE
60
80
100
Spline - 2 SE
Treatment plants
40
Spline + 2 SE
Cubic Spline
Data (♦ symbol)
Spline - 2 SE
-10
-5
0
5
10
15
20
timing
Weeks after the start of the intervention
Notes: Displays the average quality defects index, which is a weighted index of quality defects, so a higher score means lower
quality. This is plotted for the 14 treatment plants (♦ symbols) and the 6 control plants (+ symbols). Values normalized so both
series have an average of 100 prior to the start of the intervention. “Data” is plotted using a 5 week moving average. To obtain
series (rather than point-wise) confidence intervals we used a cubic-spline with one knot at the start of the implementation period.
The spline estimate is labeled (“Cubic Spine”), the 95% confidence intervals labeled (“Spline + 2SE”) and (“Spline – 2SE”) from
plant-wise block boostrap. Timing based on weeks after the intervention (positive values) or before the intervention (negative
values). For wave 1 treatment plants this is relative to September 1st 2008, for Wave 2 treatment and control firms April 7th 2009.
Nick Bloom and John Van Reenen, Management Practices, 2010
The control group’s rise in weeks 10+ are due to the pre Diwali and Ede production increase, which usually leads to a
deterioration in quality due to increased speeds of production.
We work in Tarapur because textile mills no longer exist in Mumbai
The textile factories in downtown
Mumbai are now all closed as land
prices are too high. The last few
remaining building are now being
demolished and turned into
apartment blocks and shopping malls
Apartment blocks
being built on the
site of an
demolished old
textile mill, on the
opposite side of
the road from the
one being
demolished
picture above
Nick Bloom and John Van Reenen, Management Practices, 2010
53