WEEK 1 LT ‘Exporting and Firm Performance: Evidence from a randomized
Experiment’ by Atkin, Khandelwal and Osman (2017)
Motivation Conducted a randomised experiment trying to identify the impact of
exporting on firm performance, using rug producers as an example.
Theory A movement along the PPF can be defined as firms already knowing how
to produce higher quality products, firms just change their product
specification - they use the same technology, materials and know how.
Outward shift in PPF comes from the increase in efficiency and
productivity through learning-by-exporting
ITT (intent to treat) giving them the option to export, some take it
up and some don’t - always lower than TOT
TOT (treatment on treated) always higher than ITT, 100% pick up
of the treatment
Theoretically, those who export have increased market access, match
sellers with high paying buyers, increase demand for higher quality
products, could lead to increased margins, profits, quality and increased
communication with higher number of industry agents.
Empirical design A randomised control trial is done to limit selection bias that arises as
using current exporters may skew results as they may be inherently better
at making goods or have access to connections or are generally richer.
They create an exogenous opportunity (variation) to export (give access to
certain firms), we can then compare like for like.
They do an IV in which they use allocation to the treatment as an IV for
the take-up; they need exogenous variation in take-up (we want to take out
the possible of inherent differences)
Data (frequency, Set in Fowa, Egypt and used Hamis Carpets as a local intermediary who
questions) brought these export orders to local firms (with very few employees).
Sample of 303 firms.
Rug producers are matched with customers with a higher willingness to
pay for higher quality products.
Surveys may not be enough as you can identify whether firms are more
productive but not the mechanism through which that occurs.
Key Findings Learning by exporting was a key finding in this paper; this led to export-
induced improvements in technical efficiency, with foreign buyers may
demand more detailed/sophisticated/better quality products. By repetition
and ‘doing’, there is an increase in productivity (shift out in the PPF)
Whilst not everyone took up the chance to be an exporter, firms in the
treatment group saw their revenues, and thus, their profits increase as a
result (consistent with a move along and out on the PPF).
They saw quality also increase - as measured by experts using 11 different
metrics.
Even though total production has gone down, the price charged has
increased (43%). Table VIII shows us that the quality metrics have all
improved in the exporting groups case as you have been pushed to do so
by being an exporter and competing with other foreign producers etc – that
increase in quality warrants charging a higher price which then results in
increased profitability and access to larger markets.
Decide to get exporters and non-exporters to make the same ‘old style’ rug
to check quality differences. They found that all exporters made this rug
with much better quality and precision (as judged by artisan experts etc),
while labour and technology is kept constant between the two groups.
Analysis was done on email logs and correspondence and found evidence
of learning through knowledge transfer- showed they focussed on areas
that were pointed out in emails and improved those.
,WEEK 1 LT ‘The Cost of Remoteness: Evidence from German Division and
Reunification’ by Redding, Stephen and Sturm (2008)
Motivation Uses the reunification of East and West Germany as a natural experiment
to provide evidence for the importance of market access for development.
Theory If you are a smaller city, you are more likely to rely on trade more-so
than bigger cities who can produce things within.
Empirical design Natural experiment: Difference in difference (compare growth
performance of WG cities close to the border to those that were not)
They look at basic data on city populations when you have the wall
coming down between East and west Germany - the cities are all similar in
geography, size, weather etc, but clear difference in the sense that some
cities are much closer to the border - not randomly assigned - make some
argument that where that border appeared was not known before
Diff-in-Diff:
- Using the 1919 consensus as your numeraire
- At the beginning populations are growing at very similar rates, but
then after WW2, the ones that are closer to the border plateaus and
declines a little. Closer people are more dependent on trade with
EG markets
- Control for macro-economic effects that affect pop. Growth
Potentially cities closer to the border are inherently different to others in
structure but correct this selection bias by matching to find similar cities.
Check other things like effects of war/bombing etc, if these affected those
at the border more than anywhere else, but that did not seem to make a big
difference. They look at the Western border to see if there has been
economic integration which accounts for an increased population in those
areas and control for this, but the affect still persists.
Three effects on the decline in cities: 1) cost of living effect: increase in
CoL and reduction in real wages; 2) home market effect: reduction in
market access for all firms reduces nominal wages; 3) market crowding
effect: reduction in the number of competing varieties available in WG
cities leading to higher real wages (1 & 2 triggers pop. outflow)
Data (frequency, Balanced panel of WG cities covering the period from 1919-2002
questions) including populations of all WG cities which had more than 20,000 people
in 1919 - total number of cities is 119.
Treatment group 20 WG cities that lie within 75km of the E-W border.
Control group reduces in size when matching happens.
Key Findings The imposition of the border led to a sharp decline in pop. growth in West
German cities, diminishing rapidly with distance from the border and this
decline was more pronounced for small cities vs large cities
No evidence of differences in population growth between treatment and
control cities prior to division
Higher transport costs. Building up both within country trade and
international trade gives cities plentiful options and limits this reliance.
Disproportionate loss of market access for firms far from the border.
Cities closer to the border may have been potentially trading with EG and
selling to WG so will be more affected if trade between the two stops
(0.75% decline).
Close to 25km of the border - you have a large negative change in
population and over 150km you have a large increase in population.
Technology such as the internet makes it so much easier to trade over
large distances and reduces the reliability of location.
There is not only an association but also a causal relationship between
market access and the spatial distribution of economic activity.
, WEEK 2 LT ‘Does Management Matter? Evidence from India’ by Bloom, Eifert,
Mahajan et al (2013)
Motivation Conducted a randomised experiment on large multi-plant textile firms in order to
test the importance of management practices in large firms.
Theory Areas of poor management practices: factory operations (maintenance), quality
control, inventory, HR management. These practices disincentivise workers, slow
production and faster depreciation of machinery
Informational frictions lack of awareness of good practices
Imperfect competition badly run firms not forced to exit market: voters may
want these small firms as they carry a lot of the employment, access, difference in
varieties/amenities offered, lack of information etc.
Empirical design Randomly allocated plants into treatment and control groups.
Treatment Groups received 5 months of extensive management consulting
from a large international consulting firm
- First month diagnosed improvement opportunities in a set of 38
management practices
- Next 4 months received intensive support
Control Groups only received the one month of diagnostics
- this is an ITT equation
Outcome1 of key performance metrics: quality, inventory, output, TFP
TREATtakes value of 1 for treatment plants one month after the end
DURINGtakes value of 1 for treatment plants for 6-month window from the
start of the diagnostic period (b shows short-term impact)
c dummies to control for seasonality
d dummies to control for differences between plants
Data (frequency, Very small cross-sectional data sample of 28 plants across 17 large textile firms in
questions) India (6 control firms and 11 treatment plants) - costs $1.3m
Project firms - data collected by direct observation at factory.
NON-project firms - collected during interview with direct factory observation
and discussion with managers
Key Findings The treatment intervention led to significant improvements in quality, inventory,
and output. We estimate that within the first-year productivity increased by 17%;
based on these changes we impute that annual profitability increased by over
$300,000 and reduced defects!
Fig. V: improvement of management practices in treatment plants, control plants
also improve slightly after the diagnostic phase; even nonexperimental plants in
treatment firms also benefit, suggesting there is some cross-plants learning.
- Better managed firmed seemed to grow faster with evidence that better
management allowed them to delegate more and open more production
plants in the three years following the start of the experiment. Helped
firms expand (only significant at 10% level)
- Table III: col. 1: in the 2011 cross-section the number of plants per firm
is higher for better managed firms and for firms with more male adult
family members.
- Spread these management improvements to other plants owned
- Informational constraints are the most important factor for uncommon
practices and for common practices, they knew of them, but thought they
would not be profitable to adopt
- Badly run firms were not exiting the market due to heavily restricted
competitive pressu res: high tariffs, lack of external finance, limited
managerial time (non-family members were not trusted by firm owners
with any decision-making power - firms did not expand to a size that
couldn’t be managed by family)
Policy Implications Provide firm-specific consulting as in this paper - but this is very costly. May not
scale up if firms have to bear costs also question strong external validity and
whether this can be applied to other countries/industries!
Reduce barriers to competition (allow MNCs in), improve legal/regulatory system
to reduce reliance on social capital for trust in management, run informational
campaigns, government-sponsored training
