Transcript Click here to add a title

Formal Credit and Informal jobs:
Micro evidence from Brazil
Luis Catão
Carmen Pagés
Feb 2009
M. Fernanda Rosales
Road Map
•
•
•
•
•
Introduction
Prima facie evidence
Econometric results
Links with Productivity (very preliminary)
Conclusion
Introduction
•
Formality is an optimizing decision of firms
based on benefits and costs of being formal.
•
However informality may have important costs
for aggregate welfare and productivity.
•
Increased supply of credit increases
opportunity cost of being informal
– If given only to formal firms.
Introduction II
• Brazil is an interesting case to look at because improved
macroeconomic conditions have led to a substantial
increase in the aggregate supply of credit.
• Plus has a rich household survey dataset that has not
been used to look at links between formal credit and job
informality.
• Key question we ask: to what extent increased supply of
credit has led to higher formality, controlling for other
factors.
Introduction III
• Increase formality may be driven by:
– Informal firms go formal (“within”)
– Formal firms that hire informal workers stop doing
so and/or formalize existing ones (“within”)
– Formal firms expand faster and crowd out
informal firms (“between”)
Stylized Facts
• Macro:
Rapid Credit Expansion
Lower Interest Rates
Huge REER Appreciation
• Labor Market:
Significant Rise in Formalization Rates
(2 definitions)
Weak (Labor) Productivity Growth
Credit (as % of GDP) has increased substantially since
2004…
Figure 3. Brazil: Bank Credit
(as share of GDP)
40%
Credit to Private Sector
Credit to Private Firms
35%
30%
25%
20%
15%
2002Q1
2002Q3
2003Q1
2003Q3
2004Q1
2004Q3
2005Q1
2005Q3
2006Q1
2006Q3
2007Q1
2007Q3
2008Q1
And interest rates have gone down…
Figure 4. Brazil: Interest Rate Indicators
(% a year)
60
Money Market Interest Rates
Intermediation Spreads
50
40
30
20
10
0
2002Q1
2002Q3
2003Q1
2003Q3
2004Q1
2004Q3
2005Q1
2005Q3
2006Q1
2006Q3
2007Q1
2007Q3
Since 2004 formality rates have increased
Carteira-salaried workers
Figure 1. Brazil: Share of Urban Workers with Formal Employment Contrato ("carteira")
in Economically Active Urban Population
46%
44%
42%
40%
38%
36%
34%
2002.03 2002.08 2003.01 2003.06 2003.11 2004.04 2004.09 2005.02 2005.07 2005.12 2006.05
2006.1
2007.03 2007.08 2008.01
Since 2004 formality rates have increased
Share with social security-all workers
Figure 2. Share of Workers that Contribute to Social Security
in Total Active Labor Force
49%
48%
47%
46%
45%
44%
43%
42%
41%
40%
1995
1996
1997
1998
1999
2001
2002
2003
2004
2005
2006
2007
And this took place amidst strong currency appreciation which shifts
resources to non-tradables where informality is deemed higher…
Figure 5. Brazil: Real Effective Exchange Rate
(2000=100)
130
120
110
100
90
80
70
60
2002Q1
2002Q3
2003Q1
2003Q3
2004Q1
2004Q3
2005Q1
2005Q3
2006Q1
2006Q3
2007Q1
2007Q3
2008Q1
Formality rates can be decomposed in:
F
FL
FS


E
E
E
Yielding:
L
L
 EL  F L  F S  ES
 ES  F S
F F E
    L     L   S     S
E E  E
 E E
E  E
 E E
Within L
Between L
Within S
Where: “L” = >11 employees
“M” = between 2 and 11 employees
“S” = self-employed
Between S
Table 1. Decomposition of Aggregate Labor Formalization between Self-Employed, Smaller and Larger Firms
(percent)
Labor Contract Definition
("carteira assinada")
 (F/E)
==
 (F/E)s (Es/E)  (Es/E) (F/E)s  (F/E)L (EL/E)  (EL/E) (L/E)L
2002-2007
100.00
41.63
-22.02
40.06
40.33
2004-2007
100.00
44.66
-14.14
43.82
25.66
Social Security Definition:
("contribuicao a previdencia)
 (F/E)
==
 (F/E)se(Ese/E) (Ese/E) (F/E)se  (F/E)s (Es/E)  (Es/E) (L/E)s  (F/E)L (EL/E)  (EL/E) (L/E)
2002-2007
100.00
5.38
-8.16
30.45
-7.20
26.86
52.66
2004-2007
100.00
6.74
-5.69
35.08
-3.84
33.10
34.61
Note: "se" denotes self-employment, "s" denotes firms of 2-10 employees, and "L" denote firms with 11 or more
employees. For the formality measure "cartera assinada", employment is divided into workers employed in small (2-10)
and larger firms (>11) only, since the category self-employment is not included because "carteira assinada" only
refers to employees.
L
Can the increase in credit explain increasing
formality rates?
Methodology
• Examine whether sectors “structurally” more
dependent of external financing formalize more
when credit becomes more abundant.
• Following Rajan and Zingales (1998), we
measure external dependency of financing as:
FDj=(K invj-cash flowsj)/K invj
Methodology
 Ec 
   0 C   tC   Cj   C (Credit / GDP ) t * FD j   Cjt
 E  jt
• Where c=size categories: (1); (2-10); (10+)
Table 2. Formal Employment and Financial Dependence: Panel Regressions Using the "Carteira" Definition of Formality
Dependent variable: Formal employment/total employment
Formal employment defined as carteira
Coefficients on FDj*Credit/GDP
Credit to private sector/GDP
Credit to firms/GDP
Observations Sectors
Firms with 2-10 employees
0.2067***
(0.0446)
0.4218***
(0.1146)
156
26
Firms with more than
11 employees
0.1111***
(0.0192)
0.2168***
(0.0575)
240
40
Robust Standard errors in parenthesis. * Significant at 10%; ** significant at 5%; significnat at 1%
Each entry corresponds to the coefficient of the measure of dependence on external finance per sector based on
Rajan and Zingales (1998) methodology interacted to a measure of financial depth (Credit/GDP). In addition, the
Table 3. Formal Employment and Financial Dependence: Panel Regressions Using the Social Security
Definition of Formality
Dependent variable: Formal employment/total employment
Coefficients on FDj*Credit/GDP
Credit to private sector/GDP
Credit to firms/GDP
Observations Sectors
ISIC
150
25
Self- employment
0.1309
(0.1166)
0.3684
(0.2545)
Firms with 2-10 employees
0.1748***
(0.0520)
0.3438***
(0.1294)
156
26
Firms with more than
11 employees
0.1054***
(0.0285)
0.1996**
(0.0780)
240
40
Robust Standard errors in parenthesis. * Significant at 10%; **significant at 5%; significant at 1%
Each entry corresponds to the coefficient of the measure of dependence on external finance per sector based on
Rajan and Zingales (1998) methodology interacted with the measure of financial depth (credit/GDP) In addition,
Table 4. Formal Employment and Financial Dependence: Panel Regressions Using a Semi-log specification
Dependent variable: Ln(Formal employment/ total employment)
All the sample without breaking by size
Coefficients on FDj*Credit/GDP
Formal employment
definition:
Credit to private sector/GDP
Credit to firms/GDP
Observations Sectors
0.3581**
(0.1630)
240
40
Social security
0.1846***
(0.0600)
0.2941***
(0.0977)
240
40
Carteira
0.1392***
(0.0420)
Robust Standard errors in parenthesis. * Significant at 10%; ** significant at 5%; ***significant at 1%
Each entry corresponds to the coefficient of the measure of dependence on external finance per sector based on
Rajan and Zingales (1998) methodology interacted to a measure of financial development. In addition, the
Bottom-line of “within” regressions:
• Expansion of aggregate formal credit significantly fosters
formalization within each size category.
• The results are robust to the alternative definitions of
formalization.
• Effects are much stronger for “middle-sized” and “larger”
firms [consistent with the unconditional decomposition
exercise of Table 1].
• Results also robust without breaking by size (log spec.):
 is higher the lower F/E, so stronger for smaller caps.
Table 5. Formal Employment and Financial Dependence: Panel Regressions of Employment Growth ("between") Effects
(2002-2007)
Dependent variable: Employment by firm size/total employment
Dependence on external finance as index
Credit to private sector/GDP
Credit to firms/GDP
Observations Sectors
Sel- employment
-0.0848*
(0.0475)
-0.1470
(0.1120)
150
25
Firms with 2-10 employees
0.0353
(0.0346)
0.0941
(0.0754)
150
25
Firms with more than
11 employees
0.0448
(0.0592)
0.0526
(0.1337)
150
25
Robust Standard errors in parenthesis. * Significant at 10%; ** significant at 5%; ***significant at 1%
Each entry corresponds to the coefficient of the measure of dependence on external finance per sector based on
Rajan and Zingales (1998) methodology interacted to a measure of financial development. In addition, the
Table 6. Formal Employment and Financial Dependence: Panel Regressions of Employment Growth ("between") Effects
(2004-2007)
Dependent variable: Employment by firm size/total employment
Coefficients on FDj * Credit/GDP
Credit to private sector/GDP
Credit to firms/GDP
Observations Sectors
Self- employment
-0.0818**
(0.0357)
-0.1685**
(0.07)
100
25
Firms with 2-10 employees
0.0259
(0.0217)
0.0532
(0.0423)
100
25
Firms with more than
11 employees
0.0786**
(0.0343)
0.1561**
(0.0678)
100
25
Robust Standard errors in parenthesis. * Significant at 10%; 88 significant at 5%; significnat at 1%
Each entry corresponds to the coefficient of the measure of dependence on external finance per sector based on Rajan and
Zingales (1998) methodology interacted to a measure of financial development. In addition to the variable listed, the
Bottom-Line of Between Regressions
• Greater credit availability tends to shrink the self-employment in the
size distribution
 Since much of the self-employment business is informal, this helps
lower aggregate informality.
• Conversely, the shares of upper sizes are boosted.
• In particular, effect more significant among larger firms.
• This may mean that credit allows firms to expand into higher size
segments but we can’t test that with this dataset with no information
on gross flows.
Link with Productivity [very preliminary]
• Greater availability of formal credit facilitates entry and
growth of smaller firms.
• To the extent that these firms have a lower ratio of
capital to labor, aggregate labor productivity is dragged
down.
• All the analysis here is from PIA which only covers
industry and excludes informal firms.
• And eliminates most firms with <30 employees
Brazil: Number of Firms in Industry
30000
New Firms
2004 break-point
Old Firms
Linear (New Firms)
25000
20000
15000
trend
10000
Firms Alive throughout
5000
0
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
Brazil: Average No. of Workers per Firm
500
All
Larger Firms
450
400
350
300
250
200
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
Brazil: Average Labor Productivity in Industry
(Deflated Value Added per Worker, 1996=100)
120
All Firms
Larger Firms
110
100
90
80
70
60
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
Brazil: Inter-Period Comparisons of Output, Employment and Labor Productivity Growth
(% a year)
1999-2002
2002-2004
2004-2006
Growth of Value Added
-0.1%
6.6%
5.0%
Growth of Employment
3.8%
6.3%
5.6%
Growth of Investment
0.8%
0.2%
9.9%
Growth of Value Added/Worker
-3.8%
0.2%
-0.5%
Growth of Value Added
-0.6%
7.3%
3.9%
Growth of Employment
2.2%
4.8%
2.3%
Growth of Investment
3.1%
3.4%
9.1%
Growth of Value Added/Worker
-2.8%
2.5%
1.5%
1. All Firms
2. Larger/Traditional Firm Segment
Conclusions
• Highly imperfect credit markets seem to be a factor
behind high informality rates.
• The evidence suggests that improved access to credit
has led to higher formalization.
• Much of reduction of informality takes place via “within”
effects: holding relative sector size constant, firms within
each sector have an incentive to formalize their
employees as credit becomes more abundant.
• But also some evidence of a significant between
effect – i.e. crowding out of self-owned firms by
larger formal firms and faster expansion of the
largest size segment during the credit boom
period 2004-2007.
• This is consistent with self-owned firms dying out
and/or simply becoming bigger and moving to
the larger size segment.
• The same applies to the middle segment, which
does not expand relatively much.
• This is not necessarily with inconsistent with a
literature on credit constraints and firm size
(Bernanke et al, 1995), since our larger size
category (>11) still encompasses a number of
small firms.
• And what is typically meant “large” in a LA
context is small on a world scale (Herrera and
Lora, 2003).
• Alas, the PNAD does not allow a less coarse
size breakdown, nor to look at gross entry/exit.
• So, using alternative data sources to gauge the
size composition effects of financial deepening is
left for future research.
• In any event, to the extent that wider credit
access and formalization boost long-run
productivity, our findings suggest that financial
deepening can have far-reaching effects on
long-term economic growth.
• Our findings also highlight the importance of
sound macro policies:
• As you reduce macro volatility and keep inflation
stable, this boosts credit supply, lower risk and
hence interest spreads, thus fostering formality.
• So, there is clear a link between informality,
productivity and macroeconomic policy that is
typically overlooked in the macro literature and
policy debate.
• Yet, such a positive association between credit
and (labor) productivity may be tempered in the
short-run.
• This is because, abundant credit seems to favor
the entry of smaller and lower cap firms which
typically have lower labor productivity to begin
with.
• What happens to TFP needs to be documented.
• At any rate, to the extent that credit is allocated
efficiently, this (-) effect on labor productivity
should be gradually overcome in the longer-run.
• But this again is a matter for further research.