do crises catalyze creative destruction? firm level evidence from indonesia

21
Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia Mary Hallward-Driemeier and Bob Rijkers ACES/AEA, January 8, 2012

Upload: zan

Post on 04-Jan-2016

18 views

Category:

Documents


0 download

DESCRIPTION

Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia. Mary Hallward-Driemeier and Bob Rijkers ACES/AEA, January 8, 2012. Do crises facilitate or hamper a more efficient allocation of resources?. COMPETING THEORETICAL PARADIGMS: The “Cleansing” Hypothesis: - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Do Crises Catalyze Creative Destruction?Firm Level Evidence From Indonesia

Mary Hallward-Driemeier and Bob Rijkers

ACES/AEA, January 8, 2012

Page 2: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Do crises facilitate or hamper a more efficient allocation of resources?

• COMPETING THEORETICAL PARADIGMS:– The “Cleansing” Hypothesis: Crises accelerate the Schumpetarian (1939) process of creative

destruction (Caballero and Hammour 1994).• They weed out unproductive firms.• And free up resources for more productive uses

– The “Scarring” Hypothesis: Crises obstruct the reallocative process (Barlevy, 2002; Ouyang 2009)

• They exacerbate market imperfections (e.g. credit or labor markets)• And destroy potentially productive firms

– Implications for appropriate policy response: is there a tradeoff between minimizing short-term impacts and maximizing long-run growth prospects?

Page 3: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Related Literature

• Empirical evidence Is ambiguous– Not clear whether firm- and job turnover are pro- or countercyclical

(Davis and Haltiwanger, 1990, 1992; Caballero and Davis, 1995; Boeri, 1996)

– Weak evidence that plant productivity rises during downturns (Griliches and Regev, 1995, Bailey et al. 1998; Davis et al, 1996)

– Market imperfections may be particularly pernicious during crises (Bergoeing et al., 2005, Blalock and Gertler, 2006, Gallego and Tessadda, 2009)

• Few micro-level studies of the impact of crises on resource allocation– Exceptions, again with ambiguous findings:

• Evidence for scarring: Japan’s Banking Crisis (Nishimura et al, 2005)

• Evidence for cleansing: Chile during debt crisis; Uruguay during the Argentine peso crisis(Tybout and Liu 1995: Casacuberta and Gandelsman: 2009)

Page 4: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

This Paper• Examines the impact of the East Asian Crisis on manufacturing

plant dynamics in IndonesiaComplementary analyses

• Aggregate productivity decompositions • Contributions due to entry, exit, within-plant adjustment and reallocation

between plants• Plant-level regressions of determinants of exit and of employment

growth• Examine plant heterogeneity in adjustment patterns• Disentangle the role of productivity and other plant characteristics• Allow for variation in impact over time• Examine role of regulations that varied over time and location

• Separate effects of economic crisis and political crisis

Page 5: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Testable hypotheses

If crises are cleansing:• Aggregate productivity decompositions should show:

– Increased contribution of reallocation between firms, and positive correlation between productivity and growth

– Relatively larger contributions of entry• Plant-level

– Survival: stronger relationship between firm productivity and survival– Growth: productive firms shed proportionately fewer jobs during the crisis and expand

faster after the crisis

What accounts for the observed patterns?• Plant characteristics: size, sector, need for credit• Policy environment: e.g. labor regulations

Is the political crisis more cleansing?

Suharto fell in the spring of 2008, but time alone is not enough to distinguish the political crisis. Data can identify those where a Suharto family member has an ownership stake, and/or is on the board (Mobarak and Purbasari, 2008).- Does productivity play a stronger role in connected firms?- Are the general results robust to the exclusion of connected firms?

Page 6: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Data BPS’s Indonesian Annual Manufacturing Census

Covers all manufacturing firms ≥20 employees Entry: first entry into the survey Exit: last exit

Sample: 1991-2001 Detailed information on employment, inputs and outputs, industrial

classification and ownership. Productivity measures:

Real value added per worker TFP: OLS, Solow residual, Ackerberg Caves and Frazer (2006)

Non-stationarity (and persistence) of series limits the applicability of some estimation procedures, e.g. Arellano-Bond

Political connectedness: Mobarak and Purbasari (2008) measures of Suharto family ownership

and membership on boards Regulations

Minimum wages, set at the provincial level, World Bank Jakarta Office

Page 7: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Plant Entry and Exit

Exit rates were higher during the crisis.Peak in 2001 reflects the economic slowdown, but is magnified by splits in provinces where plants were seen to exit and reenter (results robust to excluding 2001)

Page 8: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Aggregate Jobs Flows

Excess churning and net job destruction during the crisis, 1997-98

Page 9: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Decomposing Productivity Growth: Within-Adjustment Dominates

)()()(

)(

1111

111

titXi

ittitNi

ittt

Niit

Ciitittit

Ciitit

Ciit

PpPpPP

pPppP

“within” “between” “cross

FHK Decomposition of the growth of real value added per worker

“proportionate entry” “disproportionate entry” “exit”

Do see increasing cross and proportionate entry. But negative contribution of exit.

Page 10: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Firm Survival: Econometric Strategy

itititCrisisitP

itxitxCrisisitxit

vppCrisisp

xxCrisisxtt

PP

iiit

cov*Re*

cov*Re*)(log)(

covRe

covRe*0

)'exp()(

)(')(log)(log

00 iitiit

txi

itxii x

t

txtt

Discrete-time logistic survival model (Cox, 1972)))|(1()|(

1

xtxtS jt

i

ji j

Our strategy: Interact explanatory variables over time to examine how the determinants of survival vary over time:

“Proportional Hazards”

1'exp:2

1'exp:1

1'exp:0

*

*

*

it

it

it

TFPCrisis

TFPCrisis

TFPCrisis

HA

HA

H

Crises have no differential impact on creative destruction (in the short-run)

Crises accelerate creative destruction (in the short-run)

Crises attenuate creative destruction (in the short-run)

Competing Hypotheses

Page 11: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

The crisis attenuated the link between productivity and exit

Logistic Survival model: Baseline Model

Odds Ratios: Relative Probability of Exit

Value-Added TFP -Solow TFP- ACF

coef/ se coef/ se coef/ se

ln(V/ L) 0.834***

(0.011) Crisis*ln(V/ L) 1.187*** (0.025) Recovery*ln(V/ L) 0.990 (0.020) TFP (Solow) 0.755*** (0.033) Crisis*TFP (Solow) 1.556*** (0.090) Recovery*TFP (Solow) 1.111* (0.069) TFP (Solow) 0.724*** (0.034) Crisis*TFP (Solow) 1.144** (0.071) Recovery*TFP (Solow) 1.075 (0.059) Province Dummies Yes Yes Yes Period Dummies Yes Yes Yes Industry Dummies Yes Yes Yes N 153,115 95,966 73,196 Pseudo R2 0.075 0.072 0.058

Page 12: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Results are robust

Robustness tests passed but not presented: – Additional controls (in years for which they are available): self-reported constraints, training,

R&D intensity, foster parent company, education, minimum wages.– Using lagged productivity as a proxy for true productivity– Including anomalous observations– Excluding outliers

VA

TFP (Solow) TFP (Solow)

coef/ se coef/ se coef/ se coef/ se 3 Year Window

Pre-Crisis Crisis Pre-Crisis Crisis Survival Period (from – to)

1993-1996 1996-1999 1993-1996 1996-1999 ln(V/ L) 0.791*** 0.939*** (0.019) (0.019) TFP (Solow) 0.757*** 1.009 (0.052) (0.066)

5 Year Window

Pre-Crisis Crisis Pre-Crisis Crisis Survival Period (from – to)

1991-1996 1996-2001 1991-1996 1996-2001 ln(V/ L) 0.811*** 0.885*** (0.017) (0.016) TFP (Solow) 0.723*** 1.055 (0.040) (0.068)

Page 13: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Are Credit Constraints Driving the Attenuation Effect?

• Hypothesis: Firms that are credit constrained or face greater needs for financing should be disproportionately hurt (exit more, face greater attenuation effect)

• Cannot measure financial constraints directly– Across sectors: use measure of financial dependence

• Use Rajan-Zingales (1998) measure of dependence on external credit, Braun (2003) measure of asset tangibility

– Within sectors: exploit information on the composition of investment financing.• Compare: firms that financed investment with loans versus financed by other means• With depreciation, those with loans could be disproportionately hurt rather than protected

by having credit. If productive firms received loans in dollars and then exited during the crisis, this could account for the attenuation. But results hold excluding them.

• Result: Changing credit conditions amplified exit, those in sectors with greater need for financing experienced higher rates of exit.

• But do not account for the observed attenuation– Attenuation result remains even including financial dependence– And protective role of productivity is actually higher during the crisis in sectors with greater

financial dependence – consistent with the cleansing hypothesis.– Attenuation was strongest for firms that lacked access to credit to start with.

Page 14: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Firms in sectors more dependent on external finance were more likely to exit

… yet the attenuation effect was weakest in these sectorsA Financial dependence (RZ)

Baseline Extended Baseline Extended RZ_ 0.988 0.313*** 1.184** 0.914 fin_dependence (0.070) (0.108) (0.092) (0.349) Crisis* 1.329** 6.727*** 1.167 1.049 RZ_fin_dependence (0.148) (3.818) (0.156) (0.696) Recovery* 0.745*** 1.198 0.446*** 0.169** RZ_fin_dependence (0.077) (0.567) (0.067) (0.133) ln(V/ L) 0.836*** 0.819*** (0.011) (0.012) Crisis*ln(V/ L) 1.160*** 1.194*** (0.026) (0.029) Recovery*ln(V/ L) 0.991 0.999 (0.021) (0.023) RZ*ln(V/ L) 1.168*** (0.053) Crisis*RZln(V/ L) 0.805*** (0.059) Recovery*RZln(V/ L) 0.939 (0.059) TFP (Solow) 0.780*** 0.759*** (0.029) (0.040) Crisis*TFP (Solow) 1.592*** 1.578*** (0.098) (0.138) Recovery*TFP (Solow) 1.173** 1.052 (0.081) (0.112) RZ*TFP (Solow) 1.126 (0.196) Crisis*RZ*TFP (Solow) 1.028 (0.270) Recovery*RZ*TFP (Solow) 1.549 (0.531) Controls Yes Yes Yes Yes N 153,115 153,115 95,966 95,966 Pseudo R2 0.074 0.074 0.071 0.071

Page 15: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Labor market rigidities?

• Labor markets were relatively flexible during Suharto• Minimum wages set at provincial level• MW were rising steadily during the 1990s • Results:

– Interaction terms of plant productivity and provincial minimum wage is >1; more productive firms are more likely to exit when MW are high – and this is more pronounced during the crisis

– Including MW also reduces the overall attenuation results.

Page 16: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Labor market regulations are partial explanation

VA TFP (Solow)

TFP (Solow)

Levels Interacted Levels Interacted coef/ se coef/ se coef/ se coef/ se ln(V/ L) 0.820*** 0.847*** (0.011) (0.014) Crisis*ln(V/ L) 1.189*** 1.049 (0.026) (0.036) Recovery*ln(V/ L) 1.040* 1.005 (0.021) (0.023) MW*ln(V/ L) 1.194*** (0.062) MW*Crisis*ln(V/ L) 1.176*** (0.064) MW*Recovery*ln(V/ L) 0.819*** (0.049) TFP (Solow) 0.753*** 0.746*** (0.031) (0.034) Crisis*TFP (Solow) 1.555*** 1.417*** (0.094) (0.144) Recovery*TFP (Solow) 1.111 1.137* (0.074) (0.081) MW*TFP (Solow) 0.950 (0.137) MW*Crisis*TFP (Solow)

1.489** (0.296) MW*Recovery*TFP (Solow)

0.492** (0.154) Minimum Wage (log) 0.821 0.215*** 0.725* 0.794 (0.111) (0.088) (0.124) (0.280) Crisis*Minimum Wage (log)

0.958 0.795 1.247 1.060 (0.144) (0.134) (0.262) (0.252) Recovery*Minimum Wage (log)

0.168*** 0.189*** 0.117*** 0.130*** (0.028) (0.033) (0.035) (0.041) Controls Yes Yes Yes Yes N 153,115 153,115 95,966 95,966 Pseudo R2 0.076 0.076 0.074 0.074

Page 17: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Political crisis was more cleansingC Suharto Family Member on the Board and/ or JSX connection (A&B)

Baseline Extended Baseline Extended ln(V/ L)’96 0.798*** 0.804*** (0.028) (0.028) TFP (Solow) 1.113 1.160 (0.107) (0.160) Connected (JSX/ Suharto) 1.390 227.982** 1.622 1,871.488* (0.482) (576.428) (0.594) (7,284.427) Connected, (JSX/ Suharto) 0.565**

* ln(V/ L)’96 (0.164) Connected (JSX/ Suharto)

0.046 * TFP (Solow)

(0.093) Controls Yes Yes Yes Yes N 11,877 11,877 6,957 6,957 Pseudo R2 Yes Yes 0.078 0.081

Connected firms were far more likely to exit post-Suharto, but less so for productive firms.

Page 18: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Employment Growth

Estimating Equation

i

iitiitiit

iit

tPostCrisis

txPosttxtxCrisis

tx

PostrePostrecCrisis

xPostrecxxCrisis

xLd

cov'covRe''

*'cov*Re'*'

'ln

covRecovRe

covRecovRe* 111111

1

NB: testing strategy is analogous to that for the survival model

0':2

0':1

0':0

*

*

*

it

it

it

TFPCrisis

TFPCrisis

TFPCrisis

HA

HA

H

Competing Hypotheses

Crises has no differential impact on the reallocative efficiency of employment growthCrises improves the reallocative efficiency of employment growthCrises weakens the reallocative efficiency of employment growth

Page 19: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

The association between employment growth and productivity weakened over time and did not recover post-crisis

Employment Growth Dependent Variable: ∆lnL

Ln (V/ L) empvafe

TFP (Solow)

OLS FE OLS FE coef/ se coef/ se coef/ se coef/ se ln(V/ L) 0.023*** 0.009*** (0.001) (0.002) Crisis*ln(V/ L) -0.005** 0.002 (0.002) (0.002) Recovery*ln(V/ L) -0.001 0.004** (0.002) (0.002) TFP (Solow) 0.004 0.025*** (0.003) (0.005) Crisis*TFP (Solow) -0.004 -0.018*** (0.005) (0.006) Recovery*TFP (Solow) -0.004 -0.021*** (0.004) (0.006) Plant controls Yes Yes Yes Yes

Province Dummies Yes Yes Yes Yes Period Dummies Yes Yes Yes Yes Industry Dummies Yes Yes Yes Yes N 138,997 138,997 88,530 88,530 R2 0.029 0.219 0.025 0.213 Adjusted R2 0.029 0.219 0.024 -0.048

Page 20: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Conclusions• The crisis was excessively punishing in the short-run

– Excess job losses– Attenuation of the link between productivity, survival and growth– Although some improvement in relative productivity of entrants

• Changing credit market conditions amplified exit, but do not fully account for the observed attenuation.– Attenuation was strongest amongst firms that lack access to credit to start with

• Political crisis exhibited more cleansing– Post-Suharto, previously connected firms were more likely to exit, but more

productive ones were more likely to survive• Reallocation dynamics were not permanently scarred

– The link between productivity and survival is restored post-crisis– Yet, the link between productivity and employment growth remains weaker than it

had been pre-crisis• Perhaps because of more stringent labor market regulation?

Page 21: Do Crises Catalyze Creative Destruction? Firm Level Evidence From Indonesia

Broader Lessons from Indonesia Indonesia of interest as a large ‘tiger cub’. But are these likely to be

relevant elsewhere? Crisis was marked by large fall in demand and tightening of credit

markets The extent of the depreciation is significant.

Crisis are more prevalent in developing countries, although current situation shows lessons can be relevant for a larger set of countries

Shock was particularly large But not unprecedented – or unmatched.

Cleansing hypothesis may be more easily rejected if there are large market frictions Labor markets were fairly flexible under Suharto, although regulations

were tightened after his fall