Embed Size (px)
Citation preview
THE DYNAMICS OF
EMPLOYMENT GROWTH:
EVIDENCE FROM THE OECD
DYNEMP PROJECT
Carlo Menon Structural Policy Division Directorate for Science, Technology and Innovation Based on joint work with Flavio Calvino, Chiara Criscuolo, Peter Gal Séminaire de Recherche en Développement, 17th October 2016 l’Institut des Hautes Etudes de l’Amérique Latine
DYNEMP: A NEW TOOL TO EXPLOIT
EXISTING DATA
Criscuolo, C., Gal, P. N., & Menon, C. Dynemp: A routine for distributed microdata analysis of business dynamics. The Stata Journal, Volume 15 Number 1: pp. 247-274.
• Ongoing Project :
- Micro-aggregated harmonized database via distributed microdata approach
– 20+ countries
• Aim:
– To provide cross-country evidence on employment dynamics from confidential national business registers
What is DynEmp
• Based on firm level data; cross-country; longitudinal; representative; sector, age and size dimensions
• Challenges:
• Commercial data with well known shortcomings
• National surveys and Business Registers: access / confidentiality
• Solution: metadata collection and common Stata routine run locally by a network of experts on national firm-level data
• Outputs: Dynemp express database and Dynemp v.2 database
Data: needs, challenges, solution,
outputs
1. Where are most jobs?
2. Where are new jobs coming from?
3. Can we deconstruct job creation?
4. How is the dynamics of (micro) startups?
5. How much can micro start-up up-scale across countries?
6. What has been the impact of the crisis on job creation and growth for startups and others?
7. What is the role of national policies?
Policy Questions
WHERE ARE EXISTING JOBS?
Criscuolo, C., P. N. Gal and C. Menon (2014). The Dynamics of Employment Growth: New Evidence from 18 Countries. OECD Science, Technology and Industry Policy Papers, No. 14, OECD Publishing.
Most firms are small
Share of firms by size class
But more than half of the employment
is in medium and large firms Share of employment by firm size class
Role of small firms across
countries: two polar cases Magnified
0
10
20
30
40
50
60
70
80
90
100
ITA USA ITA USA ITA USA ITA USA
Micro (1-9) Small (10-49) Medium (50-249) Large (250+)
% Firms Employment
0.0
0.2
0.4
0.6
0.8
1.0
ITA USA ITA USA
Medium (50-249) Large (250+)
% 3.8
Source: OECD, Dynemp project. Preliminary results.
The age profile of small firms
vary across countries Average over time, firms below 50 employees
WHERE ARE NEW JOBS COMING FROM?
Criscuolo, C., P. N. Gal and C. Menon (2014). The Dynamics of Employment Growth: New Evidence from 18 Countries. OECD Science, Technology and Industry Policy Papers, No. 14, OECD Publishing.
• Evidence - mainly from the US - suggests:
– Young (which are small) rather than (all) small firms create jobs (Kane, 2010; Haltiwanger, 2011; Haltiwanger, Jarmin and Miranda, 2013)
• To what extent can this be generalized across countries?
Existing evidence
No matter their size, young firms are
the ones which create jobs
Source: OECD, Dynemp Express project
0%
10%
20%
30%
40%
50%
60%
Young (1-5) Old (>5) Young (1-5) Old (>5)
Small (1-249) Large (250+)
Perc
en
tag
e s
hare
in
to
tal em
plo
ym
en
t, t
ota
l jo
b
destr
ucti
on
an
d t
ota
l jo
b c
reati
on
Contribution to employment Contribution to job destruction Contribution to job creation
Young firms are job creators in all
countries Contribution to employment, job destruction and job creation by small young firms
The share of start-up is
declining in most countries Share of start-ups among all firms
Note: The graph provides data on entry rates (calculated as the number of entrants with positive employment over the total number of units with positive employment). The figures report averages for the periods 1998-2000; 2001-04; 2005-08 and 2009-13 conditional on the availability of data. Sectors covered are: manufacturing, construction, and non-financial business services. The period between 2005 and 2008 has been excluded for the Netherlands due to a redesign of the business register in 2006. Figures for Chile are preliminary. Owing to methodological differences, figures may deviate from officially published national statistics. Source: OECD DynEmp v.2 database, see http://oe.cd/dynemp.
HETEROGENEOUS IMPACT OF
GREAT RECESSION
Criscuolo, C., P. N. Gal and C. Menon (2014). The Dynamics of Employment Growth: New Evidence from 18 Countries. OECD Science, Technology and Industry Policy Papers, No. 14, OECD Publishing.
Young firms suffered more from the crisis, but
recovered more quickly
Average net growth rate over participating countries, total private sector
During the financial crisis young firms are
still net job creators
-8%
-6%
-4%
-2%
0%
2%
4%
6%
By young (0-5) By old (>5) Net growth rate
Contributions to aggregate net job creation
Adjustment via lower entry and growth of
young firms and job shedding
-8%
-6%
-4%
-2%
0%
2%
4%
6%
Young (entry-exit) Young (incumbents) Old (exits)
Old (incumbents) Total
Contributions to aggregate net job creation
CAN WE DECONSTRUCT JOB CREATION?
Calvino, F., C. Criscuolo and C. Menon (2015). Cross-country evidence on start-up dynamics, OECD Science, Technology and Industry Working Papers, 2015/06, OECD Publishing, Paris.
• The net job contribution by a particular group of firms can be decomposed into four components
• Cross-country differences across countries
21
Introduction: Deconstructing job
creation
Significant cross-country heterogeneity in each element as well as within-country differences between incumbents and entrants
Net job contribution
Number of units
(over total employment)
Average empl.
growth
Survival share
Average size
• We decompose normalized net job creation by surviving entrants as follows
𝐸𝑀𝑃𝑎𝑐𝑡𝑠𝑢𝑟𝑣 𝑡 + 𝑗
𝐸𝑀𝑃𝑐𝑡(𝑡)=
𝐸𝑀𝑃𝑎𝑐𝑡𝑠𝑢𝑟𝑣 𝑡 + 𝑗
𝐸𝑀𝑃𝑎𝑐𝑡𝑠𝑢𝑟𝑣 𝑡
∗𝐸𝑀𝑃𝑎𝑐𝑡
𝑠𝑢𝑟𝑣 𝑡
𝑁𝑟𝑈𝑛𝑖𝑡𝑠𝑎𝑐𝑡𝑠𝑢𝑟𝑣(𝑡)
∗𝑁𝑟𝑈𝑛𝑖𝑡𝑠𝑎𝑐𝑡
𝑠𝑢𝑟𝑣(𝑡)
𝑁𝑟𝑈𝑛𝑖𝑡𝑎𝑐𝑡(𝑡)∗
𝑁𝑟𝑈𝑛𝑖𝑡𝑎𝑐𝑡(𝑡)
𝐸𝑀𝑃𝑐𝑡 𝑡
22
The growth decomposition (entrants)
Average post-entry growth rate
Average size at entry
Survival share
Start-up ratio
Note: NrUnitsactsurv(t) identifies the number of entrants in country c surviving between time t and t + j
and NrUnitact(t) identifies the total number of entrants in country c at time t. Source: Calvino, Criscuolo and Menon (2015)
23
Net job creation by surviving entrants
relative to total employment
Source: OECD (2016)
24
Net job creation by surviving entrants is a
combination of four elements
Start-up ratio
Survival rate (after 3 years)
Average size at entry
Average post-entry growth
Source: OECD (2016)
Em
plo
yee
s
Un
its
per
10
00
Em
plo
yee
s
Fin
al/
init
ial
emp
l. (
%)
Sh
are
of
surv
ivin
g u
nit
s (%
)
WHAT IS THE GROWTH DYNAMICS OF THESE
YOUNG FIRMS?
Calvino, F., C. Criscuolo and C. Menon (2015). Cross-country evidence on start-up dynamics, OECD Science, Technology and Industry Working Papers, 2015/06, OECD Publishing, Paris.
The majority of micro start-ups do not
grow Share of units in all micro (0-9 employees) entrants by final size class 5 years later
But those who do create a lot of jobs
Share of units in all micro (0-9 employees) entrants by final size class 5 years later and their contribution to total net job creation
Exit probability peaks around the age of 3
– more so during the recession
Probability of exit at different ages in recession and no-recession periods in Belgium (coefficients from distributed regressions)
NO COUNTRY FOR YOUNG FIRMS?
Calvino, F., C. Criscuolo and C. Menon (2016). No Country for Young Firms?: Start-up dynamics and National Policies, OECD Science, Technology and Industry Policy Papers, No. 29, OECD Publishing, Paris.
• National policies are likely to play a significant role in explaining cross-country differences – Importance of access to finance, bankruptcy
regulation and contract enforcement
• Limited work analyzed different effects of policies on entrants vs. incumbents – Incumbents may have higher opportunities to
influence policy-makers
30
Motivation: little is known about the
(heterogeneous) effect of policies
• How bankruptcy regulation, contract enforcement and access to finance affect young (vs. incumbent) firms?
– …especially in risky and finance-dependent sectors (high growth potential)
– …along the different elements of the decomposition through which firms affect net job creation (number of units, average size, survival share, employment growth)
31
Research questions
• Some sectors or groups of firms are more exposed to certain national policies – Do “horizontal” policies really exist?
• This implies that there are three dimensions along which the role of policies can vary 1. across sectors within the same country (because of
variation in the exposure)
2. across countries within the same sector (because of differences in the policy setting)
3. between entrants and incumbents
32
Framework of analysis
References: Rajan and Zingales (1998)
• Volatility (DynEmp v.2) – average within-firm variation of employment
growth rates over time
• Growth dispersion (DynEmp v.2) – measure of between-firm (cross-sectional)
variation of employment growth rates at a given time
• Financial input intensity (I/O tables) – proxies the extent to which an industry relies on
external finance
33
Industry variables (sectoral exposure)
34
Volatility of entrants and the growth
decomposition Start-up ratio Survival rate (after 3 years)
Average size at entry Average post-entry growth
Source: Calvino, Criscuolo and Menon (2016)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.00 0.20 0.40 0.60 0.80
Un
its
pe
r 1
,00
0 e
mp
loye
es
Volatility
20
30
40
50
60
70
80
90
100
0.00 0.20 0.40 0.60 0.80
Sh
are
of
su
rviv
ing
un
its
(%
)
Volatility
0
5
10
15
20
25
30
35
40
45
50
0.00 0.20 0.40 0.60 0.80
Em
plo
ye
es
Volatility
0
50
100
150
200
250
300
350
400
450
0.00 0.20 0.40 0.60 0.80
Fin
al
/ in
itia
l e
mp
loym
en
t (%
)
Volatility
Em
plo
yee
s
Un
its
per
10
00
Em
plo
yee
s
Fin
al/
init
ial
emp
l. (
%)
Sh
are
of
surv
ivin
g u
nit
s (%
)
• Access to finance
– Seed or early stage policies (equity-based vs. tax-based, OECD)
– Venture Capital availability (WEF)
– Government controlled banks (WB)
– Easiness to access bank loans (WEF)
– Percentage of foreign banks (WB)
– Independence of banking supervision (WB)
• Bankruptcy regulation
– Resolving insolvency (time, WB)
• Contract enforcement
– Enforcing contracts (time, WB)
– Indicator of courts’ specialization (OECD)
• Challenges with policy variables
35
Policy variables (country-specific)
Note: WB stands for World Bank and WEF for World Economic Forum
• 𝑌𝑐𝑠𝑡 =∝ +𝛽 ∗ 𝑃𝑜𝑙𝑐𝑡 ∗ 𝐸𝑥𝑝𝑠𝑡 + 𝜃𝑡 + 𝜅𝑐 + 𝛾𝑠 + 𝜖𝑐𝑠𝑡
• The role of policies is assessed along each component of the growth decomposition – Y is : i) net job contribution, ii) average size, iii) number of
units, iv) average employment growth, v) survival share
– Separate estimates for entrants and incumbents
– All specifications include country, sector and year dummies
– SUR, winsorization, additional controls for average employment growth and survival share
36
Empirical strategy: “difference-in-
difference”
37
Venture capital availability in most
volatile sectors
Policy / Dep. var.
Net job contrib.
Average size
No. of units
Empl. growth
Survival share
VC availability
Incumbents 0.0829*** -0.0545*** 0.134*** 0.0147*** -0.00362
Entrants 0.193*** 0.0292 0.137*** 0.0355*** -0.00041
Diff. 0.110*** 0.084*** 0.003 0.021* 0.003
• In most volatile sectors, VC availability is positively related to net job contribution by surviving entrants
• Especially via number of units and employment growth
• Significant difference between entrants and incumbents
38
Enforcing contracts in most volatile
sectors
Policy / Dep. var.
Net job contrib.
Average size
No. of units
Empl. growth
Survival share
Enforcing contracts time
Incumbents -0.0136 -0.00107 -0.0126 -0.00561** 0.0013
Entrants -0.0820*** -0.0883*** 0.015 -0.0177* 0.00565
Diff. -0.068** -0.087*** 0.028 -0.012 0.004
• In most volatile sectors, weaker contact enforcement is negatively related to net job contribution by surviving entrants
• Especially via lower average size and post-entry employment growth
• Significant difference between entrants and incumbents
39
Resolving insolvency in financially
dependent sectors
Policy / Dep. var.
Net job contrib.
Average size
No. of units
Empl. growth
Survival share
Resolving insolvency
time
Incumbents 0.0464*** 0.0391*** 0.00833 -0.00565** 0.00516**
Entrants -0.0384 -0.0113 -0.0209 -0.0172* 0.00523
Diff. -0.085*** -0.050* -0.029 -0.012 0.000
• In most financially dependent sectors, longer bankruptcy procedures are positively related to the survival share of incumbents
• Negative employment growth and higher average size yield an overall positive effect on incumbents
• Tendency to slow down reallocation
• Start-ups in volatile sectors and in sectors that have higher dispersion are more exposed to national policies and framework conditions
– Promoting policies that lower risk and tackle policy failures that impose extra-cost to risk
• Start-ups appear systematically more exposed to the policy environment than incumbents especially in more volatile, dispersed and financially dependent sectors
40
Summary of results (i)
• Employment growth, average size and number of units generally respond more for entrants than incumbents
• Survival share of entrants does not seem particularly related to policies. Survival of incumbents does, especially in financially dependent industries
– Some policies (especially inefficient bankruptcy regulations) seem to slow down the reallocation process
• Timely bankruptcy procedures, strong contract enforcement and access to finance seem to be key for an environment conducive to employment growth
41
Summary of results (ii)
42
Reform simulation: bankruptcy
regulation and civil justice Italy: estimated effect in the information technology and other information services sector
Notes: The bars show the effect ceteris paribus of policy changes on the response variable if the econometric estimates reported in the paper are interpreted causally. Dashed lines show 95% confidence intervals. The response variable is final over initial employment ratio.
THANK YOU!
For any additional information on DynEmp
please email: [email protected]
and visit oe.cd/dynemp
