Formal vs Informal Finance:

1

• Formal vs Informal Finance
• Evidence from China
• by
• Meghana Ayyagari, Asli Demirgus-Kunt, Vojislav
  Maksimovic
• Discussant
• Bernard Yeung
• Stern
• Mar 2007

2
Interesting and important paper

• First point
• Chinas use of informal financing is just as
  extensive as other developing countries, e.g.,
  Brazil, Africa, East European countries, Africa,
  South Asian

3
Interesting and important paper

• Second point among all firms (2002 data), or
  just private firms (not SOE, listed firms)
• Having bank financing is
• positively and significantly associated with
  sales growth, percent profits reinvested, and,
• not related with productivity
• Self-financing (informal, family, other) is
• Not related with sales growth,
• Possibly negative related to reinvestment
• Positively and significantly associated with
  productivity growth
• Productivity is (sales total material costs)/L
• Control for
• Endogeneity in treatment effects
• City, size, firm type, age, competition
• Contribution Use large scale data to show that
  formal (bank) financing helps growth and
  reinvestment

4
Comment on endogeneity

• Correlation between getting loans and future
  profitability, a classic problem
• Use a treatment effects model, Heckman correction
  (Inverse Mills Ratio)

5
Endogeneity

• Need an instrument to predict getting bank loans
  but not correlated with future performance
• Collateral dummy as
• 1 if has a loan and put in a collateral
• 0 if has a loan, no collateral requirement,
• no loan because cant put in a
  collateral,
• do not apply for a loan b/c no
  collateral.
• The dummy captures the intersection of
  requirement for and ability to put in a
  collateral
• Logic
• Collateral serves as a proxy for the firms
  ability to post collateral. (so, can get a bank
  loan).
• Whether there was a collateral requirement or not
  is less likely to be correlated with firms
  growth and hence need not enter the 2nd stage.
  Thus collateral serves as the identifying
  variable in our selection equation.
• Curiosity
• That some are granted a bank loan but not
  required a collateral are possibly due to their
  expected future performance
• But would that not make a part of the dummy
  correlated with future performance?
• Feel weird to use ex post variable to as an ex
  ante predictor, but not very successful to
  pinning down what is wrong
• Why not use tangibility as an instrument?

6
Why the research is important?

• Does finance matter?
• Certainly
• With finance, growth, no doubt
• Why interested? Allocation efficiency is the key
• Reinterpret
• With bank loan, experiences sales growth,
  reinvest more, but, VA/L does not increase!
• Meaning?
• Self-financing, experiences little sales growth,
  little increase in reinvestment, but VA/L grows!
• Meaning?
• Efficient/Inefficient formal external financing?!

7
Why banks are inefficient?

• Corporate governance
• Who own banks?
• Managers are by political appointment
• Lack of prudent judiciary duties
• Connected lending (short termism in the G)
• Opaque system
• ? corruption, NPL
• Poor internal corporate governance
• Credit officers want to lend!
• My promotion vs. protecting other peoples money
• Bank monitoring of borrower?
• Go hard face against connections

8
Push

• Examine the efficiency of the system
• Dig into the regions and cities and ask where
  bank lending leads to better results?
• Link the efficiency of bank financing with a
  region/citys institutional environment?
• Quality of the government
• Quality of the regulations and supervision?
• Quality of bank executives?
• Banks credit policies equity requirement,
  credit reports, etc
• Loan officers incentive/agency problem?

9
Still puzzled

• Recall the first point
• Chinas use of informal financing is just as
  extensive as other developing countries, e.g.,
  Brazil, Africa, East European countries, Africa,
  South Asian
• But, efficiency?
• Raise the puzzle why is China growing so fast
  even with its lousy capital markets?
• Higher savings?
• Chinas capital allocation is a bit more
  efficient, at least in some major cities?
• Using location level variables to shed light?

10
Conclude

• Interesting paper leading to interesting questions