2
I.
Introduction
Economists have long been interested in firms’ management and organization as key drivers
of productivity and growth. Adam Smith’s pin factory focused on the impact of
organizational specialization on productivity, while Walker (1887) claimed the single largest
driver of business performance was management quality.
1
While this interest in management
faded after the 1930s, there has been a resurgence of interest in the last decade, driven by the
recent availability of
microdata on management (Roberts, 2018). However, this work has
tended to focus mainly on manufacturing firms in developed countries. This paper presents
the first large scale and nationally representative analysis of management practices in a major
developing country, Mexico, covering both manufacturing and service sectors.
Our analysis is based on a novel firm-level survey implemented by the National Institute of
Statistics and Geography (INEGI). To ensure survey quality and international comparability,
the management part of the survey followed as closely as possible (given the English to
Spanish translation) the US MOPS 2010 and 2015 surveys (Buffington et al., 2018). The
survey was then matched to the Census administrative data, providing additional and rich
measures of performance, demographics, and other firms and area characteristics.
We find more structured management – that is, more systematic collection and use of data
through monitoring, goal-setting, and stronger use of incentives in Human Resource practices
(e.g., over hiring, firing, pay, and promotions) - is associated with superior firm performance
(in terms of size, productivity,
profitability, and innovation) in both manufacturing and
services. This result is robust to a wide variety of checks and controls. While this does not
imply any causal relationship, it does suggest these management practices are tightly
associated with firm-level performance.
2
We first document that aggregate Mexican management scores, like productivity, are well
below those in the US. We then focus on whether this is driven by greater market frictions
1
Walker was the founding president of the American Economic Association and ran the 1870 and 1880 Census
that formed the basis of his views.
2
Below, we discuss randomized control trials around introducing more structured management practices that
also seem to find significant performance improvements (e.g., Bruhn, Karlan and Schoar, 2018, on Mexican
firms).
3
reflected in the fact that well-managed Mexican firms are unable to reap the returns of their
managerial practices in terms of greater scale because of these distortions.
To examine this, we perform a number of tests. First, we show that although firms with higher
management scores attain larger size in Mexican manufacturing,
the size-management
relationship is half as strong as it is in US manufacturing. Moreover, the size-management
relationship is much weaker among firms in the services sector than among those in the more
competitive and open manufacturing sector.
3
Mexican services are more shielded from
international competition (especially since Mexico’s accession to GATT in 1986, and
NAFTA in 1994) and have greater idiosyncratic regulations than manufacturing.
4
This comparison across countries and broad sectors is suggestive but coarse. Our second test
focuses on market size. A wide class of trade and IO models suggest that reallocation should
increase with integration and market size (e.g., Melitz, 2003). We measure market access
(trade costs) by drive times to the US border for manufacturing firms, and by income-
weighted population density for services firms. We find that proximity to the US border
increases the size-management relationship for manufacturing (which
relies heavily on
exports to US markets), but has no effect on services. We also find that this reallocation effect
is driven by the manufacturing industries that are more export-intensive. For services, we
find that local market size strengthens the size-management relationship, whereas this local
market size indicator has no effect on manufacturing firms. Again, this is consistent with our
priors, as reallocation among service firms should be affected by local market size and not
international market access.
3
In particular, the regression coefficient of firm log (employment) on the management score is 1.62 in Mexican
services, is 2.75 in Mexican manufacturing, and is 3.36 in US manufacturing. Hence, for every 0.1-point change
in the management score, the associated change in firm employment size over twice as large in US
manufacturing as it is in Mexican services.
4
See Goodwin et al. (2018) and Levy (2018). A recent World Bank report assessing the regulations at sub-
national level in various Mexican States (e.g., Mexico State, Tabasco, Oaxaca, etc.) found for instance that
firms in the retail sector face various types of barriers at entry (e.g., municipal licenses for which there are not
clearly established time terms and for which the granting criteria are not fully transparent – for instance the
declaration of “urban impact” can take between 1 month and 6 years), and during operations (limitations to use
of English placards and promotional boards, renovation of licenses, limitations to the hours in which retailers
are allowed to be open, etc.). In the transport sector, the report found that the state law in Tabasco does not
allow a private company to propose starting a service to serve a new route but only the Government can propose
a new route and open a market. Similarly, the state law not only prohibits foreign companies from concessions
for local transport services, but even favors local companies against those from other states of Mexico – see
Goodwin et al. (2018).
4
Our final test focuses on institutional frictions. We measure different measures of
institutional frictions relying on data that measures contract enforcement, crime, and
government corruption at the local (municipality) level. We
find that reallocation is
significantly greater where these institutional frictions are lower.
All three designs: across countries and sectors; across market sizes; and across local
institutions suggest that frictions are a major factor in depressing the performance of the
Mexican economy.
The paper is structured with a detailed description of the data in section II, including some
basic results on management practices and firm performance. We discuss the distribution of
Mexican
management
practices
and
how
they
compare
with
the
US in section III. Section IV focus on misallocation, wedges, and the importance of market
frictions, international exposure for manufacturing, and market density for services, while
section V concludes.
5
Dostları ilə paylaş: