subdizing labor hoarding in recessions employment & welfare...
TRANSCRIPT
Subdizing Labor Hoarding in RecessionsEmployment & Welfare Effects of Short-Time Work
Giulia Giupponi Camille Landais
Mannheim
October 1, 2018
G. Giupponi C. Landais (LSE) Short Time Work 1/28
Introduction
Motivation
Short-time work (STW)
Subsidy for hour reductions to firms experiencing temporary shocks
Big renewal of interest in STW: main policy tool to encourage labourhoarding
Aggressively used during Great Recession
G. Giupponi C. Landais (LSE) Short Time Work 2/28
Introduction
Motivation
Short-time work (STW)
Subsidy for hour reductions to firms experiencing temporary shocks
Big renewal of interest in STW: main policy tool to encourage labourhoarding
Aggressively used during Great Recession
02
46
Perc
enta
ge o
f em
ploy
ees
PRT
USA
CAN
KOR
NOR
AUT
FRA
ESP
CHE
FIN
LUX
DEU
ITA
TUR
BEL
Source: Hijzen and Venn (2010), OECD data
Average monthly take-up rate in 2007 and 2009
2007 2009G. Giupponi C. Landais (LSE) Short Time Work 2/28
Introduction
Motivation
Short-time work (STW)
Subsidy for hour reductions to firms experiencing temporary shocks
Big renewal of interest in STW: main policy tool to encourage labourhoarding
Aggressively used during Great Recession
But WHY?
Remarkably little knowledge about effects of STW on firms, workers &welfare
. Lack of good-quality data
. Lack of credible sources of identification
. Lack of conceptual framework for welfare analysis
G. Giupponi C. Landais (LSE) Short Time Work 2/28
Introduction
This project:
Leverage unique data from INPS records and unique policy setting:
. Universe of administrative data on STW at individual and firm level
. Quasi-experimental variation from Italian STW policy rules
Offer compelling evidence on effects of STW:
1. On firms’ employment, output and balance sheet
2. On short & long-term insurance of workers
3. On reallocation in the labor market
Develop conceptual framework:
1. To rationalize empirical evidence on STW effects
2. To clarify welfare trade-offs for optimal STW design
3. To calibrate model & conduct counterfactual policy analysis
G. Giupponi C. Landais (LSE) Short Time Work 3/28
Introduction
This project:
Leverage unique data from INPS records and unique policy setting:
. Universe of administrative data on STW at individual and firm level
. Quasi-experimental variation from Italian STW policy rules
Offer compelling evidence on effects of STW:
1. On firms’ employment, output and balance sheet
Large effects (-) on hours and on employment (+)
2. On short & long-term insurance of workers
3. On reallocation in the labor market
Develop conceptual framework:
1. To rationalize empirical evidence on STW effects
2. To clarify welfare trade-offs for optimal STW design
3. To calibrate model & conduct counterfactual policy analysis
G. Giupponi C. Landais (LSE) Short Time Work 3/28
Introduction
This project:
Leverage unique data from INPS records and unique policy setting:
. Universe of administrative data on STW at individual and firm level
. Quasi-experimental variation from Italian STW policy rules
Offer compelling evidence on effects of STW:
1. On firms’ employment, output and balance sheet
No significant effect on output / balance sheet
2. On short & long-term insurance of workers
3. On reallocation in the labor market
Develop conceptual framework:
1. To rationalize empirical evidence on STW effects
2. To clarify welfare trade-offs for optimal STW design
3. To calibrate model & conduct counterfactual policy analysis
G. Giupponi C. Landais (LSE) Short Time Work 3/28
Introduction
This project:
Leverage unique data from INPS records and unique policy setting:
. Universe of administrative data on STW at individual and firm level
. Quasi-experimental variation from Italian STW policy rules
Offer compelling evidence on effects of STW:
1. On firms’ employment, output and balance sheet
2. On short & long-term insurance of workers
Almost perfect insurance in short run
3. On reallocation in the labor market
Develop conceptual framework:
1. To rationalize empirical evidence on STW effects
2. To clarify welfare trade-offs for optimal STW design
3. To calibrate model & conduct counterfactual policy analysis
G. Giupponi C. Landais (LSE) Short Time Work 3/28
Introduction
This project:
Leverage unique data from INPS records and unique policy setting:
. Universe of administrative data on STW at individual and firm level
. Quasi-experimental variation from Italian STW policy rules
Offer compelling evidence on effects of STW:
1. On firms’ employment, output and balance sheet
2. On short & long-term insurance of workers
No insurance in long run
3. On reallocation in the labor market
Develop conceptual framework:
1. To rationalize empirical evidence on STW effects
2. To clarify welfare trade-offs for optimal STW design
3. To calibrate model & conduct counterfactual policy analysis
G. Giupponi C. Landais (LSE) Short Time Work 3/28
Introduction
This project:
Leverage unique data from INPS records and unique policy setting:
. Universe of administrative data on STW at individual and firm level
. Quasi-experimental variation from Italian STW policy rules
Offer compelling evidence on effects of STW:
1. On firms’ employment, output and balance sheet
2. On short & long-term insurance of workers
3. On reallocation in the labor market
Significant negative effect on employment of untreated firms
Develop conceptual framework:
1. To rationalize empirical evidence on STW effects
2. To clarify welfare trade-offs for optimal STW design
3. To calibrate model & conduct counterfactual policy analysis
G. Giupponi C. Landais (LSE) Short Time Work 3/28
Introduction
This project:
Leverage unique data from INPS records and unique policy setting:
. Universe of administrative data on STW at individual and firm level
. Quasi-experimental variation from Italian STW policy rules
Offer compelling evidence on effects of STW:
1. On firms’ employment, output and balance sheet
2. On short & long-term insurance of workers
3. On reallocation in the labor market
Develop conceptual framework:
1. To rationalize empirical evidence on STW effects
2. To clarify welfare trade-offs for optimal STW design
3. To calibrate model & conduct counterfactual policy analysis
Small positive welfare gains from STW, but dW /dτ ≈ 0
G. Giupponi C. Landais (LSE) Short Time Work 3/28
Introduction
Outline
1 Institutional Background & Data
2 Effects of STW on Employment & Firm Outcomes
3 Insurance Value to Workers
4 Selection and Reallocation Effects
5 Model & Welfare Implications
G. Giupponi C. Landais (LSE) Short Time Work 4/28
Institutional Background & Data
Cassa Integrazione Guadagni Straordinaria (CIGS)
CIGS: Main pillar of STW during Recession
Targets firm experiencing shocks: demand/revenue shocks, companycrisis, restructuring, reorganization, insolvency etc.
Subsidy for hour reductions, available to workers
Replaces about 80% of foregone earnings due to hours not worked
Weak conditionality requirements:
. Simply provide justification for economic need & recovery plan
. No prohibition of dismissals/layoffs
. Workers: No training provision or search requirement
Minimal cost to firm ≈ 3-4.5% of subsidy
Duration: up to 12 months (with possibility of extension)
G. Giupponi C. Landais (LSE) Short Time Work 5/28
Institutional Background & Data
Data
Administrative data from Italian Social Security Archives
Universe of matched employer-employee data for the private sector
Monthly data 2005-2015 and annual data 1983-2015
Information on workers and firms
DemographicsWorking historiesSocial insurance and social assistance program participationFirm characteristics (size, sector)
Information on CIG eligibility, applications, authorizations, duration andpayment for the years 2005-2015
Matching with firm-level balance-sheet data (approx. 50%)Firms eligibility. Firms take-up. Workers treatment.
G. Giupponi C. Landais (LSE) Short Time Work 6/28
Effects of STW on Employment & Firm Outcomes
Identification
Exploit variation in firm’s eligibility for CIGS based on:
Firms’ industry × contributory codes: Details
Size: more than 15 FTE employees in 6 mths prior to application
Triple Diff. within the 5-digit industry codes. Compare outcomes of firms:
1. In eligible vs non-eligible industry ×contributory codes
2. Just below vs just above 15 FTE-threshold
3. Before vs after the start of the Great Recession
Identifying assumption:No unobservable time shocks that would be, within each 5-digit industrycode, specific to firms that are eligible to CIGS and whose size is just abovethe 15 FTE threshold.
G. Giupponi C. Landais (LSE) Short Time Work 7/28
Effects of STW on Employment & Firm Outcomes
Fraction of Firms Receiving CIGS Treatment: 1st Stage
02
46
8P
erce
ntag
e po
ints
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
G. Giupponi C. Landais (LSE) Short Time Work 8/28
Effects of STW on Employment & Firm Outcomes
Intensive-Margin Employment: Log # Hours
βIV=-.514(.034)
-.06
-.04
-.02
0.0
2Es
timat
ed e
ffect
of I
NPS
cod
e x
FTE
size
inte
ract
ion
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
. STW decreases # of hours worked per employee by ≈ 40%
Weeks Worked
G. Giupponi C. Landais (LSE) Short Time Work 9/28
Effects of STW on Employment & Firm Outcomes
Extensive-Margin Employment: Log Firm Size Headcount
βIV=.380(.036)-.0
20
.02
.04
.06
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
. STW increases headcount employment by ≈ 40%
G. Giupponi C. Landais (LSE) Short Time Work 10/28
Effects of STW on Employment & Firm Outcomes
Wage Rates: Log Earnings Per Week
βIV=-.018(.026)
-.06
-.04
-.02
0.0
2
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
. STW has no significant effect on wage rates
G. Giupponi C. Landais (LSE) Short Time Work 11/28
Effects of STW on Employment & Firm Outcomes
Log Wage Bill Per Employee
βIV=-.585(.044)
-.06
-.04
-.02
0.0
2
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
. STW decreases wage bill per employee by ≈ 45%
G. Giupponi C. Landais (LSE) Short Time Work 12/28
Effects of STW on Employment & Firm Outcomes
Additional Results & Robustness
Dual labor market effects Open-ended vs fixed term
Additional effects on firms’ outcomes Balance-Sheet
Negative effect on output per worker
Slight negative effect on labor productivity & TFP
No significant effect on balance-sheet apart from liquidity (+)
Robustness
No significant size manipulation Size Manip.
No significant eligibility manipulation Eligibility Manip.
No significant differential trend btw eligible & non-eligible Trends
Placebos & permutation-based s.e. Placebos & Permutation
Similar effects for firms without change in dismissal rule at 15FTENo 15FTE dismissal rule
G. Giupponi C. Landais (LSE) Short Time Work 13/28
Effects of STW on Employment & Firm Outcomes
Dynamic effects
IV estimates capture total effects on firms of exposure to STW
Instrument determines both past + current treatment Past Treatment
Develop methodology similar to Cellini & al. [2010] for recursiveidentification of dynamic effects of STW
Intuition:
βDDD2009 = βTOT
0 · dT2009
dZ2009(1)
βDDD2010 = βTOT
0 · dT2010
dZ2009+ βTOT
1 · dT2009
dZ2009(2)
etc...
Graphical Results
G. Giupponi C. Landais (LSE) Short Time Work 14/28
Effects of STW on Employment & Firm Outcomes
Intensive Employment Effects Dissipate After Treatment
STWTreatment
Post STWTreatment
-1-.5
0.5
1TO
T ef
fect
0 1 2 3 4Time since STW treatment (years)
. 1 year after treatment, intensive margin responses disappear
G. Giupponi C. Landais (LSE) Short Time Work 15/28
Effects of STW on Employment & Firm Outcomes
Dynamic Extensive Employment Response
STWTreatment
Post STWTreatment
-1-.5
0.5
1TO
T ef
fect
0 1 2 3 4Time since STW treatment (years)
. No significant long term effects on employment Back
G. Giupponi C. Landais (LSE) Short Time Work 16/28
Effects of STW on Employment & Firm Outcomes
Table: Dynamic TOT Effects of STW on Firm Outcomes
Time since STW Treatment in Years...
0 1 2 3 4(βTOT
0 ) (βTOT1 ) (βTOT
2 ) (βTOT3 ) (βTOT
4 )Log # Hours -.284 .142 -.084 -.033 -.110
(.063) (.081) (.084) (.087) (.091)Log # Weeks Worked per Employee -.276 .160 -.035 .005 -.058
(.058) (.075) (.077) (.080) (.084)Log Employment .210 -.273 -.160 .089 .127
(.152) (.197) (.204) (.211) (.221)Log Wage Bill -.188 .068 -.079 .062 -.066
(.107) (.138) (.143) (.148) (.155)Log Open-Ended Contracts .412 -.317 -.069 -.054 .094
(.125) (.162) (.167) (.173) (.182)
Significant short run employment effects upon treatment
G. Giupponi C. Landais (LSE) Short Time Work 17/28
Effects of STW on Employment & Firm Outcomes
Table: Dynamic TOT Effects of STW on Firm Outcomes
Time since STW Treatment in Years...
0 1 2 3 4(βTOT
0 ) (βTOT1 ) (βTOT
2 ) (βTOT3 ) (βTOT
4 )Log # Hours -.284 .142 -.084 -.033 -.110
(.063) (.081) (.084) (.087) (.091)Log # Weeks Worked per Employee -.276 .160 -.035 .005 -.058
(.058) (.075) (.077) (.080) (.084)Log Employment .210 -.273 -.160 .089 .127
(.152) (.197) (.204) (.211) (.221)Log Wage Bill -.188 .068 -.079 .062 -.066
(.107) (.138) (.143) (.148) (.155)Log Open-Ended Contracts .412 -.317 -.069 -.054 .094
(.125) (.162) (.167) (.173) (.182)
No significant long term effects on employment
G. Giupponi C. Landais (LSE) Short Time Work 17/28
Insurance Value to Workers
STW Event Studies: Workers’ Employment Probability
STWTreatment
-.6-.4
-.20
Rel
ativ
e to
tim
e -1
-4 -3 -2 -1 0 1 2 3 4 5Time since STW treatment (years)
Treated N-E counterfactual N-E Layoff counterfactual
Identification - Details
G. Giupponi C. Landais (LSE) Short Time Work 18/28
Insurance Value to Workers
STW Event Studies: Workers’ Total Hours Worked
STWTreatment
-.6-.4
-.20
Rel
ativ
e to
tim
e -1
-4 -3 -2 -1 0 1 2 3 4 5Time since STW treatment (years)
Treated N-E counterfactual N-E Layoff counterfactual
Identification - Details
G. Giupponi C. Landais (LSE) Short Time Work 19/28
Insurance Value to Workers
STW Event Studies: Earnings + Transfers
STWTreatment
-.6-.4
-.20
Rel
ativ
e to
tim
e -1
-4 -3 -2 -1 0 1 2 3 4 5Time since STW treatment (years)
Treated N-E counterfactual N-E Layoff counterfactual
STW provides high insurance level in the short run
G. Giupponi C. Landais (LSE) Short Time Work 20/28
Insurance Value to Workers
STW Event Studies: Earnings + Transfers:
STWTreatment
-.6-.4
-.20
Rel
ativ
e to
tim
e -1
-4 -3 -2 -1 0 1 2 3 4 5Time since STW treatment (years)
Treated N-E counterfactual N-E Layoff counterfactual
But no insurance in the long run
G. Giupponi C. Landais (LSE) Short Time Work 21/28
Insurance Value to Workers
STW Event Studies: Earnings + Transfers:
STWTreatment
-.6-.4
-.20
Rel
ativ
e to
tim
e -1
-4 -3 -2 -1 0 1 2 3 4 5Time since STW treatment (years)
Treated N-E counterfactual N-E Layoff counterfactual
Limited role of STW in preserving experience / specific human K Bounds vs IV
G. Giupponi C. Landais (LSE) Short Time Work 21/28
Insurance Value to Workers
STW Event Studies: Earnings + Transfers:
STWTreatment
-.6-.4
-.20
Rel
ativ
e to
tim
e -1
-4 -3 -2 -1 0 1 2 3 4 5Time since STW treatment (years)
Treated N-E counterfactual N-E Layoff counterfactual
No dynamic returns to work in low productivity empl. (Card & Hyslop)
G. Giupponi C. Landais (LSE) Short Time Work 21/28
Selection and Reallocation Effects
Reallocation: Equilibrium Effects
Low productivity firms select more into STW Graphs
By increasing employment in low-productivity firms, STW may preventreallocation of workers to more productive firms
Identification of equilibrium effects:
Estimate effect of increase in fraction of workers treated by STW inLLM on employment outcomes of non-eligible / high productivity firms
Instrument fraction of workers treated by STW by fraction of workerseligible in LLM due to size and INPS codes in pre-recession period
Identification - Details
G. Giupponi C. Landais (LSE) Short Time Work 22/28
Selection and Reallocation Effects
Equilibrium Effects: Employment Spillovers
βIV=-.937 (.216)
-.12
-.11
-.1-.0
9-.0
8C
hang
e in
log
firm
siz
e he
adco
unt (
Mar
ch)
.1 .2 .3 .4 .5Fraction of eligible workers 2005-2008
1 ppt ↑ in fraction treated by STW ⇒ ≈ 1% ↓ in empl. of non-eligible firms
G. Giupponi C. Landais (LSE) Short Time Work 23/28
Selection and Reallocation Effects
Equilibrium Effects: Employment Spillovers
βIV=-.937 (.216)
-.12
-.11
-.1-.0
9-.0
8C
hang
e in
log
firm
siz
e he
adco
unt (
Mar
ch)
.1 .2 .3 .4 .5Fraction of eligible workers 2005-2008
For 1 job “saved” by STW, employment in non treated firms ↓ by .03 job
G. Giupponi C. Landais (LSE) Short Time Work 23/28
Selection and Reallocation Effects
Equilibrium Effects: Average Firm TFP in LLM
βIV=-2.093(.606)
-.05
0.0
5.1
.15
Cha
nge
in a
vera
ge T
FP
0 .1 .2 .3 .4 .5Fraction of eligible workers 2005-2008
First-Stage Robustness
G. Giupponi C. Landais (LSE) Short Time Work 24/28
Model & Welfare Implications
Model: Set-Up
Two employment margins: extensive (employment) and intensive (hrs)
Firms exposed to idiosyncratic productivity shocks ε
Can claim STW for workers when ε = εlow
Risk-averse workers, identical ex-ante.
Firms and workers match randomly on labor market with frictions
Concave matching function M(ut , vt)Labor market tightness θt ≡ vt
ut
Wage/Hours negotiated to split the match surplus btw workers and firms
General hours schedule h = h(b, τ, t, ε, θ,w)
G. Giupponi C. Landais (LSE) Short Time Work 25/28
Model & Welfare Implications
Mechanisms
STW effects on employment magnified by wage rigidity
Hours contraint
Workers only accept employment iff W u ≤W ek
Workers Value Function
Implicitly defines lower bound on hours h = h(b, τ)
STW relaxes hrs constraint: dh/dτ < 0 Graph
Firms’ intensive vs extensive margin choices Firms’ problem
When hit by TFP shock, firms may to “hoard” laborEspecially in Recession when hiring is cheap (low θ),h constrains ability to substitute h for n
STW relaxes h & ↑ demand for n for low-productivity firms Graph
Equilibrium & Spillover effects
↑ demand for n for low-productivity firms ⇒ ↑ θ in equilibrium
↑ θ ⇒ ↓ n in firms non-eligible to STW Graph
G. Giupponi C. Landais (LSE) Short Time Work 26/28
Model & Welfare Implications
Mechanisms
STW effects on employment magnified by wage rigidity
Hours contraint
Workers only accept employment iff W u ≤W ek
Workers Value Function
Implicitly defines lower bound on hours h = h(b, τ)
STW relaxes hrs constraint: dh/dτ < 0 Graph
Firms’ intensive vs extensive margin choices Firms’ problem
When hit by TFP shock, firms may to “hoard” laborEspecially in Recession when hiring is cheap (low θ),h constrains ability to substitute h for n
STW relaxes h & ↑ demand for n for low-productivity firms Graph
Equilibrium & Spillover effects
↑ demand for n for low-productivity firms ⇒ ↑ θ in equilibrium
↑ θ ⇒ ↓ n in firms non-eligible to STW Graph
G. Giupponi C. Landais (LSE) Short Time Work 26/28
Model & Welfare Implications
Mechanisms
STW effects on employment magnified by wage rigidity
Hours contraint
Workers only accept employment iff W u ≤W ek
Workers Value Function
Implicitly defines lower bound on hours h = h(b, τ)
STW relaxes hrs constraint: dh/dτ < 0 Graph
Firms’ intensive vs extensive margin choices Firms’ problem
When hit by TFP shock, firms may to “hoard” laborEspecially in Recession when hiring is cheap (low θ),h constrains ability to substitute h for n
STW relaxes h & ↑ demand for n for low-productivity firms Graph
Equilibrium & Spillover effects
↑ demand for n for low-productivity firms ⇒ ↑ θ in equilibrium
↑ θ ⇒ ↓ n in firms non-eligible to STW Graph
G. Giupponi C. Landais (LSE) Short Time Work 26/28
Model & Welfare Implications
Optimal STW: Welfare Trade-Offs
Optimal STW rate τ such that:
Value of Transfer = Fiscal Extern. + Employment Extern. + Hrs Extern.
Fiscal Externality:
1 + εn,ττ − R
τ− εh,τ
h
1− h
Large elasticities εh,τ and εn,τ ⇒ Fiscal extern. >>1
G. Giupponi C. Landais (LSE) Short Time Work 27/28
Model & Welfare Implications
Optimal STW: Welfare Trade-Offs
Optimal STW rate τ such that:
Value of Transfer = Fiscal Extern.︸ ︷︷ ︸Baily-Chetty
+Employment Extern. + Hrs Extern.
Fiscal Externality:
1 + εn,ττ − R
τ− εh,τ
h
1− h
Large elasticities εh,τ and εn,τ ⇒ Fiscal extern. >>1
G. Giupponi C. Landais (LSE) Short Time Work 27/28
Model & Welfare Implications
Optimal STW: Welfare Trade-Offs
Optimal STW rate τ such that:
Value of Transfer = Fiscal Extern. + Employment Extern. + Hrs Extern.
Value of Transfer:∝∼ ESTW [u′(c)]− E[u′(c)]
At τ ≈ .8, marginal insurance value may be small
G. Giupponi C. Landais (LSE) Short Time Work 27/28
Model & Welfare Implications
Optimal STW: Welfare Trade-Offs
Optimal STW rate τ such that:
Value of Transfer = Fiscal Extern. + Employment Extern. + Hrs Extern.
Fiscal Externality:
1 + εn,ττ − R
τ− εh,τ
h
1− h
Large elasticities εh,τ and εn,τ ⇒ Fiscal extern. >>1
G. Giupponi C. Landais (LSE) Short Time Work 27/28
Model & Welfare Implications
Optimal STW: Welfare Trade-Offs
Optimal STW rate τ such that:
Insurance Value = Fiscal Extern. + Employment Extern. + Hrs Extern.
Employment Externalities: Landais & al. [2017], Michaillat & Saez [2015]
∝∼dθ
dτ
{f ′(θ)∆U + q′(θ)C
}︸ ︷︷ ︸
“Hosios” Term
Positive welfare effect of STW if
θ suboptimally low in RecessionAnd dθ/dτ > 0
G. Giupponi C. Landais (LSE) Short Time Work 28/28
Model & Welfare Implications
Optimal STW: Welfare Trade-Offs
Optimal STW rate τ such that:
Insurance Value = Fiscal Extern. + Employment Extern. + Hrs Extern.
Hours Externalities: Missing Market for Hours!
∝∼dh
dτ
{(F ′(h)− w)− (MRSc,h − w(1− τ))︸ ︷︷ ︸
Deviation from FB Hrs
}
First-Best: MRT = MRS = wage ratePositive welfare effect of STW if:
F ′(h) < MRS in Recession
G. Giupponi C. Landais (LSE) Short Time Work 29/28
Model & Welfare Implications
Calibration & Counterfactual Analysis
Calibration
Use reduced-form evidence to identify key parameters of model Calibration
Counterfactual policy I: changes in STW generosity (τ)
STW reduces unemployment by ≈ 3 pptSTW decreases total TFP by ≈ 1.5%STW increases welfare by ≈ 2%
But dW /dτ ≈ 0 Graph
Counterfactual policy II: changes in UI generosity (b)
STW and UI are highly complementary↑ b increases hours constraint on low TFP firms
STW alleviates effects of ↑ b Graph
G. Giupponi C. Landais (LSE) Short Time Work 30/28
Model & Welfare Implications
Calibration & Counterfactual Analysis
Calibration
Use reduced-form evidence to identify key parameters of model Calibration
Counterfactual policy I: changes in STW generosity (τ)
STW reduces unemployment by ≈ 3 pptSTW decreases total TFP by ≈ 1.5%STW increases welfare by ≈ 2%
But dW /dτ ≈ 0 Graph
Counterfactual policy II: changes in UI generosity (b)
STW and UI are highly complementary↑ b increases hours constraint on low TFP firms
STW alleviates effects of ↑ b Graph
G. Giupponi C. Landais (LSE) Short Time Work 30/28
Conclusion
Concluding Remarks & Next Steps
Provide compelling evidence that STW:
Induces sharp ↓ in hrs and a large ↑ in empl. in short termVery small net long-run effectsOffers large insurance to workers but only in the short termMostly targets distressed & low productivity firmsNegative externalities on higher productivity firms in LLM
Provide framework to:
Rationalize evidenceClarify welfare trade-offsCalibrate counterfactual policies
STW have overall small positive welfare effects, can be useful when pairedwith generous UI
G. Giupponi C. Landais (LSE) Short Time Work 31/28
Appendix
Additional slides
G. Giupponi C. Landais (LSE) Short Time Work 32/28
Appendix
Table: Firms by eligibily status
Eligible firms Non eligible firmsMean Sd Mean Sd
Employees (headcount) 9.78 5.55 8.22 4.90Employees (FTE) 9.35 5.38 7.42 4.33Employees on open-ended contracts 8.96 5.35 7.25 4.60Employees on fixed-term contracts 0.81 1.78 0.98 2.25Annual weeks worked 455.41 271.73 432.03 231.68Annual weeks worked per employee 53.55 25.79 55.19 28.66Annual wage bill (000) 218.01 157.17 158.61 120.35Annual wage bill per employee (000) 22.49 13.22 19.80 11.86Value added per week worked (000) 1.22 14.41 1.01 7.42Net revenue per week worked (000) 5.94 52.77 6.48 46.31Profit/loss per week worked (000) -0.20 17.55 -0.03 3.09Liquidity (share of total assets) 0.09 0.13 0.12 0.15Investment in tangible assets (share of total assets) 0.07 0.10 0.07 0.11Investment in intangible assets (share of total assets) 0.01 0.04 0.02 0.06Observations 102757 218823
Notes: Back
G. Giupponi C. Landais (LSE) Short Time Work 33/28
Appendix
Table: Firms by take-up status in eligible sectors
Takers Non takersMean Sd Mean Sd
Employees (headcount) 17.99 5.62 9.57 5.39Employees (FTE) 17.46 5.44 9.15 5.22Employees on open-ended contracts 16.98 5.67 8.77 5.18Employees on fixed-term contracts 1.01 1.80 0.81 1.78Annual weeks worked 950.72 256.38 494.27 262.71Annual weeks worked per employee 55.06 30.92 53.51 25.64Annual wage bill (000) 435.46 170.60 212.64 152.89Annual wage bill per employee (000) 25.18 15.81 22.42 13.14Value added per week worked (000) 0.97 1.71 1.23 14.60Net revenue per week worked (000) 4.47 8.20 5.98 53.48Profit/loss per week worked (000) -0.13 2.07 -0.20 17.79Liquidity (share of total assets) 0.07 0.10 0.10 0.13Investment in tangible assets (share of total assets) 0.08 0.10 0.07 0.10Investment in intangible assets (share of total assets) 0.02 0.05 0.01 0.04Observations 2517 100334
Notes: Back
G. Giupponi C. Landais (LSE) Short Time Work 34/28
Appendix
Table: Workers by treatment status in eligible sectors
Treated Non treatedMean Sd Mean Sd
Proportion female 0.27 0.44 0.22 0.41Age 40.00 9.26 38.41 10.684Experience (years) 19.19 10.35 16.29 10.96Tenure (months) 93.31 87.56 63.66 76.36Annual labor earnings (000) 20.83 9.22 19.34 12.17Annual weeks worked 47.62 9.54 42.08 14.77Proportion on full-time contract 0.94 0.24 0.92 0.27Proportion on open-ended contract 0.94 0.24 0.86 0.35Proportion on fixed-term contract 0.06 0.24 0.14 0.35Proportion on seasonal contract 0.00 0.02 0.00 0.02Proportion blue collar 0.75 0.43 0.66 0.47Proportion white collar 0.21 0.41 0.26 0.44Proportion manager 0.00 0.02 0.01 0.08Observations 36574 73728
Back
G. Giupponi C. Landais (LSE) Short Time Work 35/28
Appendix
Triple Difference Specification
Yigst =∑j
γj1 ·{1[g ∈ E] ∗ 1[Ni,t−1 > 15] ∗ 1[j = t]
}
+∑j
∑k
γjk2 ·
{1[g ∈ E] ∗ 1[k = s] ∗ 1[j = t]
}
+∑j
∑k
γjk3 ·
{1[k = s] ∗ 1[Ni,t−1 > 15] ∗ 1[j = t]
}
+∑j
∑k
γjk4 ·
{1[k = s] ∗ 1[j = t]
}
+∑k
γk5 ·{1[g ∈ E] ∗ 1[k = s] ∗ 1[Ni,t−1 > 15]
}
+∑k
γk6 ·{1[g ∈ E] ∗ 1[k = s]
}+ vigst
(3)
i is firm, s 5-digit industry code, t calendar year
Industry group g . Group of industries eligible to receive CIGS: g ∈ ENi,t−1 max 6 month window FTE size in calendar year t − 1.
Plot estimated coefficients γt1 for all years t
Back
G. Giupponi C. Landais (LSE) Short Time Work 36/28
Appendix
IV First Stage Specification
Tigst = κ1 ·{1[g ∈ E] ∗ 1[Ni,t−1 > 15] ∗ 1[t > 2008]
}+∑j
∑k
κjk2 ·{1[g ∈ E] ∗ 1[k = s] ∗ 1[j = t]
}
+∑j
∑k
κjk3 ·{1[k = s] ∗ 1[Ni,t−1 > 15] ∗ 1[j = t]
}
+∑j
∑k
κjk4 ·{1[k = s] ∗ 1[j = t]
}
+∑k
κk5 ·{1[g ∈ E] ∗ 1[k = s] ∗ 1[Ni,t−1 > 15]
}
+∑k
κk6 ·{1[g ∈ E] ∗ 1[k = s]
}+ νigst
(4)
i is firm, s 5-digit industry code, t calendar year
Industry group g . Group of industries eligible to receive CIGS: g ∈ E
Ni,t−1 max 6 month window FTE size in calendar year t − 1.
Back
G. Giupponi C. Landais (LSE) Short Time Work 37/28
Appendix
Firm Eligibility
Eligibility defined by law supplemented by detailed regulations issued by the Ministry ofLabor and made operational by INPS (essentially in the 1970s)
INPS uses 5-digit INPS industry codes × additional administrative codes (called codiceautorizzazione) to determine eligibility
⇒ within 5-digit INPS industry codes, some firms are deemed eligible, other ineligible
We exploit variation in eligibility within these fine-grained 5-digit industry codes (594industries)
E.g.:
1. 5-digit code 11306, 11307 and 11308 = firms in construction specialized installationof electrical machinery. Only those with administrative code 3N are eligible
2. 5-digit code 10106 produce seeds and beans. Only eligible only if codice
autorizzazione 3A, i.e. if they are cooperatives
Back
G. Giupponi C. Landais (LSE) Short Time Work 38/28
Appendix
Table: Effects of STW on Output, Productivity & Balance-Sheet Outcomes
IV Estimate Std Error N
First StageProba. of CIGS Take-Up .05 (.002) 45336
Balance-Sheet & Productivity OutcomesFirm Survival Probability (in t + 1) -.018 (.024) 45336Firm Value-Added .095 (.159) 10438Value-Added Per Worker -.508 (.120) 10438Labor Productivity -.142 (.104) 10438Total Factor Productivity -.056 (.143) 10438Liquidity .939 (.461) 10438
Notes: The Table reports for each outcome the IV estimate scaled by average baseline outcome: βIV/Y .Value added = total revenues + unsold stocks - cost of goods and services used in production
Labor productivity = value added / total number of week worked in the firm Back
G. Giupponi C. Landais (LSE) Short Time Work 39/28
Appendix
Table: Robustness of Baseline Effects
(1) (2) (3) (4) (5) (6)“Doughnut” Only Only Permutation No DismissalRegression ≤ 15 FTE >15 FTE Test Rule Change
>60FTE 50FTE(Placebo) (Placebo) Across Italy threshold
First StageProba. of .053 .002 .051 .000 .055 .041CIGS Take-Up (.002) (.000) (.002) (.000) (.005) (.004)
OutcomesIV RF IV RF IV IV
Log Hrs per wker -.449 -.011 -.602 .000 -.670 -.156(.037) (.020) (.081) (.010) (.230) (.132)
Log Empl. .284 -.020 .306 -.001 .848 .338(.032) (.030) (.099) (.009) (.297) (.258)
Log Wage Bill -.544 -.026 -.498 .000 -.568 -.390(.049) (.030) (.155) (.013) (.297) (.709)
N 2686140 2608383 429490 2978239 152753 44793
Notes: Permutation P-Values Back
G. Giupponi C. Landais (LSE) Short Time Work 40/28
Appendix
Size Manipulation: FTE size pdf
0.0
5.1
.15
.2D
ensi
ty
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25Firm size
2000-2015
Note: McCrary test -0.008 (s.e. 0.005).Source: INPS.
All sectorsDistribution of firms by firm size
G. Giupponi C. Landais (LSE) Short Time Work 41/28
Appendix
McCrary Test Statistic of Discontinuity in Firms’ SizeDistribution
-.15
-.1-.0
50
.05
.1.1
5
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
G. Giupponi C. Landais (LSE) Short Time Work 42/28
Appendix
McCrary Test Statistic of Discontinuity in Firms’ SizeDistribution
-.15
-.1-.0
50
.05
.1.1
5
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Back
G. Giupponi C. Landais (LSE) Short Time Work 43/28
Appendix
Eligibility Code Manipulation
0.2
.4.6
.81
Perc
ent
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
NE to E E to NESource: INPS.
Change of eligibility status
back
G. Giupponi C. Landais (LSE) Short Time Work 44/28
Appendix
P-Values of Permutation Test: Weeks Worked
0.1
.2.3
.4.5
Pro
b. o
f mor
e ex
trem
e va
lue
than
bet
a D
D
2009 2010 2011 2012 2013 2014
Back
G. Giupponi C. Landais (LSE) Short Time Work 45/28
Appendix
P-Values of Permutation Test: Employment
0.1
.2.3
.4.5
Pro
b. o
f mor
e ex
trem
e va
lue
than
bet
a D
D
2009 2010 2011 2012 2013 2014
Back
G. Giupponi C. Landais (LSE) Short Time Work 46/28
Appendix
Intensive-Margin Employment: Log # Weeks Worked
βIV=-.486(.033)
-.06
-.04
-.02
0.0
2
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
. STW decreases # of weeks worked per employee by ≈ 40%
Back
G. Giupponi C. Landais (LSE) Short Time Work 47/28
Appendix
Probability of CIG Treatment in Previous 5 yrs
05
1015
20Pe
rcen
tage
poi
nts
2006 2007 2008 2009 2010 2011 2012 2013 2014
Back
G. Giupponi C. Landais (LSE) Short Time Work 48/28
Appendix
Employment: Dual labor market effects
βIV=-.616(.043)-.0
20
.02
.04
.06
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
. Open-ended employment largely benefits from STW: ↑ by ≈ 85%
Back
G. Giupponi C. Landais (LSE) Short Time Work 49/28
Appendix
Employment: Dual labor market effects
βIV=-.403(.117)
-.1-.0
50
.05
.1
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
. While fixed-term contracts are substituted away: ↓ by ≈ 35%
Back
G. Giupponi C. Landais (LSE) Short Time Work 50/28
Appendix
Event Studies: Methodology
Panel of all employees of firms active between 2000 and 2015 and with firmsize ∈ (5; 25] in the year prior to the first worker’s STW spell
Treated individuals: workers with a STW event
Control individuals: NN matching based on pre-event characteristics
Selection:
. Focus on control individuals who cannot access STW because ofsize×eligibility
Bounds on selection:
. Counterfactual 1 [upper bound]: average worker in similar firms noneligible to CIGS...
. Counterfactual 2 [lower bound]: laid-off workers in similar firms noneligible to CIGS Back
G. Giupponi C. Landais (LSE) Short Time Work 51/28
Appendix
Event study approachIndividual outcomes
Panel of all employees of firms active between 2000 and 2015 and with firmsize ∈ (5; 25] in the year prior to the first worker’s STW spell
Individual working histories:
Start in the first year the individual appears in the sampleEnd in 2015 unconditional on employment, or before if the individualretires or dies
An event is defined as the first ever STW spell since 2005
Treated individuals (i.e. those with a STW event) are matched with controlindividuals (i.e. those who never experience STW) based onnearest-neighbour matching
Matching on gender, age, job characteristics at event time t-1, employmentstatus, annual weeks worked, earnings and firm size at t-1, t-2, t-3 and t-4,and main industry at t-1
Implementation
G. Giupponi C. Landais (LSE) Short Time Work 52/28
Appendix
Nearest-neighbour matchingIndividual outcomes
For each calendar year j from 2004 to 2014, select:
. All treated that are at event time t = −1 in calendar year j
. All controls employed in calendar year j
Mahalanobis nearest-neighbour matching without replacement
Matching on gender, age, job characteristics at event time t-1, employmentstatus, annual weeks worked, earnings and firm size at t-1, t-2, t-3 and t-4,and main industry at t-1
Matched controls are assigned a placebo event in calendar year j+1 (eventtime t=0)
Panel is balanced over event time for the four years prior to event
Matched sample: 21,273 treated individuals matched to 18,472 controls
Back
G. Giupponi C. Landais (LSE) Short Time Work 53/28
Appendix
Bounds vs IV-based Recursive Dynamic Estimates
STW
Treatment
Post STW
Treatment
-1-.5
0.5
1
0 1 2 3 4 5Time since STW treatment (years)
Dynamic TOT estimateUpper bound (counterfactual)Lower bound (counterfactual)
No dynamic returns to work in low productivity empl. (Card & Hyslop)
BackG. Giupponi C. Landais (LSE) Short Time Work 54/28
Appendix
Selection into Take-Up
02.
55
7.5
1012
.5P
roba
bilit
y of
STW
take
-up
(%)
1 2 3 4 1 2 3 4 1 2 3 4
Quartile Quartile Quartile
Labord Productivity TFP Predicted Layoff Score
Back
G. Giupponi C. Landais (LSE) Short Time Work 55/28
Appendix
Heteregeneous Treatment Effects: Hours
-1-.5
0.5
1Lo
g ho
urs
1 2 3 4 1 2 3 4 1 2 3 4Quartile Quartile Quartile
Labor Productivity TFP Predicted Layoff Score
Back
G. Giupponi C. Landais (LSE) Short Time Work 56/28
Appendix
Heteregeneous Treatment Effects: Employment
-.50
.51
1.5
2Lo
g H
eadc
ount
Em
ploy
men
t
1 2 3 4 1 2 3 4 1 2 3 4Quartile Quartile Quartile
Labor Productivity TFP Predicted Layoff Score
Back
G. Giupponi C. Landais (LSE) Short Time Work 57/28
Appendix
Liquidity constrained firms also take up more strongly
34
56
7P
erce
ntag
e po
ints
1 2 3 4Quantile of liquidity
95% CI Coeff
Rank firms according to pre-recession levels of liquidity Back
G. Giupponi C. Landais (LSE) Short Time Work 58/28
Appendix
Liquidity constrained firms also take up more strongly
34
56
7P
erce
ntag
e po
ints
1 2 3 4Quantile of liquidity
95% CI Coeff
Liquidity constrained firms much more likely to use STW Back
G. Giupponi C. Landais (LSE) Short Time Work 58/28
Appendix
Reallocation: Equilibrium Effects
Use spatial variation across more than 500 LLM
Specification, first difference firm / LLM fixed effects
∆Yij = ∆Tj + X ′j β + W ′
i γ + εij (5)
∆Tj = αZPREj + ηj (6)
Firm i , LLM j
Instrument: ZPREj fraction of eligible workers from size and INPS codes
in pre-recession period
Back
G. Giupponi C. Landais (LSE) Short Time Work 59/28
Appendix
Equilibrium Effects: Placebo
βRF=-.021 (.013)
.07
.09
.11
.13
Cha
nge
in lo
g fir
m s
ize
head
coun
t (M
arch
)
.1 .2 .3 .4 .5 .6Fraction of eligible workers 2000-2005
Back
G. Giupponi C. Landais (LSE) Short Time Work 60/28
Appendix
Equilibrium Effects: First-Stage
-.02
0.0
2.0
4.0
6Fr
actio
n of
em
ploy
ees
on C
IGS
2010
-201
3
0 .1 .2 .3 .4 .5Fraction of eligible workers 2005-2008
Back
G. Giupponi C. Landais (LSE) Short Time Work 61/28
Appendix
Table: Equilibrium Effects of STW on Non-Treated Firm Outcomes
OLS IV IV IV(1) (2) (3) (4)
Non-eligible FirmsLog Employment -0.327 -0.492 -0.918 -0.937
(0.080) (0.137) (0.216) (0.216)Inflows 0.136 -3.594 -4.406 -3.176
(1.060) (1.947) (2.380) (1.440)Labor Market
Log TFP 0.005 -1.332 -1.332 -0.368(0.159) (0.386) (0.286) (0.242)
ControlsLLM × ×Firm-level ×
Notes: Back
G. Giupponi C. Landais (LSE) Short Time Work 62/28
Appendix
Workers
Utility when employed in firm of productivity k ∈ {h, l}
W ek = u(ck , hk) + β(δW u + (1− δ)W e
k )
δ: exogenous separation rate
Utility when unemployed
W u = u(b, 0) + β(φ(θ)W e + (1− φ(θ))W u)
φ(θ): job finding probability
Workers only accept employment iff W u ≤W ek
Implicitly defines a lower bound constraint on hours offered by firms h = h(b, τ)back
G. Giupponi C. Landais (LSE) Short Time Work 63/28
Appendix
Workers’ Utility & Hours Constraint
5 10 15 20 25 30 35 40
Hours
0.85
0.9
0.95
1U
tility
Employed without STWUnemployed without STW
G. Giupponi C. Landais (LSE) Short Time Work 64/28
Appendix
Workers’ Utility & Hours Constraint: With STW
5 10 15 20 25 30 35 40
Hours
0.85
0.9
0.95
1U
tility
Employed with STWUnemployed with STWEmployed without STWUnemployed without STW
Back
G. Giupponi C. Landais (LSE) Short Time Work 65/28
Appendix
Firms
Production function F (h, n):
Diminishing returns to labor: Fh ≤ 0, Fn ≤ 0Hours and headcount employment not perfect substitute
Firms maximize profits, s.t. law of motion of employment:
Π(εt , nt−1) = εtF (ht , nt)− whtnt − c · v + ρEt [Π(εt+1, nt)] (7)
s.t.nt = (1− δ)nt−1 + v · q(θt)
Firms F.O.C w.r.t n:
{n} εtF ′n(ht , nt) = wht +c
q(θt)− ρEt(Πn(εt+1, nt)) (8)
Back
G. Giupponi C. Landais (LSE) Short Time Work 66/28
Appendix
Firms’ Labor Hoarding in Recession: High ε
1 2 3 4 5 6 7
Period
0.75
0.8
0.85
0.9
0.95
1La
bour
inpu
ts, r
elat
ive
to p
re-s
hock
Labour hoarding, high productivity firm
Employment responseHours response
= 0.041
Back
G. Giupponi C. Landais (LSE) Short Time Work 67/28
Appendix
Firms’ Labor Hoarding in Recession: Low ε, no STW
1 2 3 4 5 6 7
Period
0.75
0.8
0.85
0.9
0.95
1La
bour
inpu
ts, r
elat
ive
to p
re-s
hock
Labour hoarding, low productivity firm
Employment responseHours response
= -0.23
Back
G. Giupponi C. Landais (LSE) Short Time Work 68/28
Appendix
Firms’ Labor Hoarding in Recession: Low ε, with STW
1 2 3 4 5 6 7
Period
0.75
0.8
0.85
0.9
0.95
1La
bour
inpu
ts, r
elat
ive
to p
re-s
hock
Labour hoarding, low productivity firm with STW
Employment responseHours response
= 0.039
Back
G. Giupponi C. Landais (LSE) Short Time Work 69/28
Appendix
Firms’ Labor Demand: High ε
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Employment, n
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5La
bour
mar
ket t
ight
ness
, Labor Supply (steady state flows)Labor Demand High
G. Giupponi C. Landais (LSE) Short Time Work 70/28
Appendix
Firms’ Labor Demand: Low ε
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Employment, n
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5La
bour
mar
ket t
ight
ness
, Labor Supply (steady state flows)Labor Demand High Labor Demand Low
G. Giupponi C. Landais (LSE) Short Time Work 71/28
Appendix
Firms’ Labor Demand: Low ε with STW
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Employment, n
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5La
bour
mar
ket t
ight
ness
, Labor Supply (steady state flows)Labor Demand High Labor Demand Low Labor Demand Low with STW
G. Giupponi C. Landais (LSE) Short Time Work 72/28
Appendix
Equilibrium
Aggregate labor demand: nd = nl + nh
Steady-state equality of flows in/out of employment:
ns =φ(θ)
δ + φ(θ)
Gvt budget constraint:u · b + nl · τ(h − hl) ≤ tN
Equilibrium:nd(θ, b, τ) = ns(θ)
G. Giupponi C. Landais (LSE) Short Time Work 73/28
Appendix
Equilibrium
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Employment, n
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5La
bour
mar
ket t
ight
ness
, Labor Supply (steady state flows)Labor Demand High Labor Demand Low
G. Giupponi C. Landais (LSE) Short Time Work 74/28
Appendix
Equilibrium
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Employment, n
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5La
bour
mar
ket t
ight
ness
, Labor Supply (steady state flows)Aggregate Labour Demand
G. Giupponi C. Landais (LSE) Short Time Work 75/28
Appendix
Incidence of STW Policy
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Employment, n
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5La
bour
mar
ket t
ight
ness
, Labor Supply (steady state flows)Aggregate Labour DemandAggregate Labour Demand with STW
G. Giupponi C. Landais (LSE) Short Time Work 76/28
Appendix
Incidence of STW Policy
0.895 0.9 0.905 0.91 0.915 0.92 0.925 0.93
Employment, n
0.05
0.1
0.15
0.2La
bour
mar
ket t
ight
ness
, Labor Supply (steady state flows)Aggregate Labour DemandAggregate Labour Demand with STW
*STW
*
G. Giupponi C. Landais (LSE) Short Time Work 77/28
Appendix
Negative Employment Externalities for High ε Firms
0.8 0.9 1 1.1
Employment, n
0
0.05
0.1
0.15
0.2
0.25
0.3La
bour
mar
ket t
ight
ness
,
*
*STW
n*n*STW
G. Giupponi C. Landais (LSE) Short Time Work 78/28
Figure: Counterfactual STW Rate τ
A. Unemployment B. Hours
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Short time work subsidy,
0.106
0.108
0.11
0.112
0.114
0.116
0.118
0.12
0.122
0.124
0.126
Une
mpl
oym
ent
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Short time work subsidy,
15
20
25
30
35
40
Con
trac
ted
hour
s
Low productivity firm hoursHigh productivity firm hours
C. TFP D. Welfare
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Short time work subsidy,
0.98
0.982
0.984
0.986
0.988
0.99
0.992
0.994
0.996
0.998
1
Tot
al fa
ctor
pro
duct
ivity
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Short time work subsidy,
1
1.002
1.004
1.006
1.008
1.01
1.012
1.014
1.016
1.018
1.02
Wel
fare
Back
Appendix
Counterfactual UI Policy R
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Replacement rate, relative to full time worker
0.105
0.11
0.115
0.12
0.125
0.13
Une
mpl
oym
ent
Unemployment
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Replacement rate, relative to full time worker
15
20
25
30
35
40
Hou
rs
Hours schedule
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Replacement rate, relative to full time worker
0.998
1
1.002
1.004
1.006
1.008
1.01
1.012
1.014
1.016
1.018
Pro
duct
ivity
Total factor production
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Replacement rate, relative to full time worker
0.975
0.98
0.985
0.99
0.995
1
1.005
Wel
fare
Total welfare
BackG. Giupponi C. Landais (LSE) Short Time Work 80/28
Appendix
Counterfactual STW Rate τ : Employment
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.91
1.005
1.01
1.015
1.02
1.025
1.03E
mpl
oym
ent
Total employment
Back
G. Giupponi C. Landais (LSE) Short Time Work 81/28
Appendix
Counterfactual STW Rate τ : Productivity
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90.98
0.985
0.99
0.995
1
1.005P
rodu
ctiv
ityTotal factor productivity
Back
G. Giupponi C. Landais (LSE) Short Time Work 82/28
Appendix
Counterfactual STW Rate τ : Welfare
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.91
1.005
1.01
1.015
1.02
1.025W
elfa
reTotal welfare
Back
G. Giupponi C. Landais (LSE) Short Time Work 83/28
Appendix
Reallocation effects: Selection and Heterogeneity
1. : Firms taking up STW more likely to layoff workers absent STW Graphs
2. : Liquidity constrained firms more likely to take-up STW Graphs
3. : Low productivity firms select more into STW Graphs
G. Giupponi C. Landais (LSE) Short Time Work 84/28
Appendix
Reallocation effects: Selection and Heterogeneity
1. : Firms taking up STW more likely to layoff workers absent STW Graphs
STW well targeted = firms that take it up most would have laid offworkers absent STW
2. : Liquidity constrained firms more likely to take-up STW Graphs
3. : Low productivity firms select more into STW Graphs
G. Giupponi C. Landais (LSE) Short Time Work 84/28
Appendix
Reallocation effects: Selection and Heterogeneity
1. : Firms taking up STW more likely to layoff workers absent STW Graphs
2. : Liquidity constrained firms more likely to take-up STW Graphs
If financially constrained firms select more, STW injects liquidity andmay prevent excessive layoffs (Schoefer [2016])
3. : Low productivity firms select more into STW Graphs
G. Giupponi C. Landais (LSE) Short Time Work 84/28
Appendix
Reallocation effects: Selection and Heterogeneity
1. : Firms taking up STW more likely to layoff workers absent STW Graphs
2. : Liquidity constrained firms more likely to take-up STW Graphs
3. : Low productivity firms select more into STW Graphs
If low-productivity firms select more, then STW subsidizeslow-productivity matches and prevents efficient reallocation of labor
G. Giupponi C. Landais (LSE) Short Time Work 84/28
Appendix
Related Literature
Following earlier cross-country empirical analyses (Abraham and Houseman, 1993),
renewed interest in STW at the onset of the crisis:
Cross-country studies: positive effect of STW on employment and a negativeon hours (Hijzen and Venn, 2010; Boeri and Bruecker, 2011; Cahuc andCarcillo, 2011; Hijzen and Martin, 2013)
Analysis at firm-level remains scarce and inconclusive due to limited data
availability and credible exogenous variation (Boeri and Bruecker, 2011;
Brenke et al., 2013; Calavrezo et al., 2009)
Early theoretical literature: STW reduces layoffs, but generates distortions at theintensive margin (Burdett and Wright, 1989)
Recent theoretical work: STW decreases allocative efficiency (Cooper et al., 2017)
G. Giupponi C. Landais (LSE) Short Time Work 85/28