Transcript What is optimal financial structure?

Comments on:
Firm Growth and Finance:
Are Some Financial Institutions Better
Suited to Early Stages of Development
than Others?
by
Robert Cull and L. Colin Xu
World Bank
Stijn Claessens
World Bank Conference on Financial Structure
and Economic Development
June 16, 2011, Washington, D.C.
Question and Answer


Does a financial sector better matched in structure to
country’s optimal (based on endowment, development)
industrial structure deliver more (labor) growth?
Yes: some evidence that more bank-oriented systems
help labor growth in countries at lower development



Labor growth higher when banks more important
Stock markets help more for rich countries
But: not because small firms grow faster


Rather, labor at large, capital intensive firms grows faster
Contradicts the financial structure channel
Praise/Main Comments

Surely a worthwhile topic
Need to go beyond private credit to GDP
 Financial crisis has raised (new) questions on finance

Comments
 Like to see more clearly defined what the optimal
economic/industrial and financial structures are


What is H0? And how does this translate into tests?
Would like to see other RHS and LHS variables
LHS: Not just labor growth
 RHS: More on financial structures and controls


Of course, questions on endogeneity/instruments
What is optimal?


Paper does not claim to test why there are
“deviations” from “optimal” financial structure
But has specific optimal structure view
Banks do better job at lower levels of development
 And makes industrial structure = development


Story about financial structure can go other ways
Large commercial banks in Germany during
industrial revolution mobilized financing → growth
 (State) banks in Japan similar coordinated economy

What is optimal industrial structure?
What is optimal financial structure?

Industrial = level of development (GDP/Capita)?
But story more about the structure of economy.
Implies at similar levels of development can have
large differences in what is optimal
 Suggests to introduce some proxies like agriculture,
services, industry, natural resources, land-locked, etc.
to be used for some interactions


Financial structure

Can one be more precise in terms of what are the
market failures? And then decide test?
Comments on Dependent
Variable

Dependent variable not so obvious
Why labor growth? If development is labor-intensive,
why does financing and financial structure matter (so
much)? Also, is this most reliable indicator?
 Short sample over which labor growth is measured
 Look at labor growth in specific sectors?


Other variables that are candidates?
# of firms, growth of some specific sectors that are
know to vary in financial structure demands (“RZ”)
 Sales, productivity, profitability, etc.

Comments on Independent
Variables

Independent variables not so obvious either


Now size of banks and capital markets
Why not (include):
Size of banks, concentration: more related to H0?
 Besides ownership structures, foreign vs. domestic
banks, and state/private: also type of banks?
 Number of branches/access indicators
 Activity restrictions? BCL?
 Presumably: legal/property rights as control variables?

Other Comments on RHS/LHS

Does one want so much firm controls, like age?


Some are LHS possibly (size? age? foreign ownership)
Can one use more actual financing of firms?
Surveys indicate role of external financing
 Look at sample more constrained?


Financial development
If stock market is not developed, still include “0”
 How about other non-bank finance: insurance, leasing?
 Does international financial integration matter?

Methodology

Does instrumental variables (GMM, LIML)


Find robust results with some instruments
Logical instruments? Many stories
Oil needs external financing? Trust is endogenous?
 Also, expect crops growth is endogenous


Do not do a formal horse race of 3 H0s

Do not check whether political economy matters
Other Comments


Legal variables could be LHS too, e.g., more
relationship lending, more credit information
Are there substitutes out there?
Do business groups correct for the lack of optimal
structure?
 Informal forms of financing?


Interpretation: provide economic magnitudes

Effects at country levels of financial structure
Summary

Paper shows that:
Financial structures matter for labor growth
 But not as expected for specific types of firms


But paper does not yet
Address H0 directly: why does structure matter?
 Can not conclude certain structure preferred


Suggestions
More analytical modeling on financial structure
 More empirical work, some is feasible with this data


Look more at industrial structure explicitly