WP 10 Linkages with firm- level data 2nd EUKLEMS Consortium Meeting, 9-11 June 2005, Helsinki This project is funded by the European Commission, Research Directorate General as part of the 6th Framework Programme, Priority 8, "Policy Support and Anticipating Scientific and Technological Needs".
23
Embed
WP 10 Linkages with firm-level data 2nd EUKLEMS Consortium Meeting, 9-11 June 2005, Helsinki This project is funded by the European Commission, Research.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
WP 10 Linkages with firm-level data
2nd EUKLEMS Consortium Meeting, 9-11 June 2005, HelsinkiThis project is funded by the European Commission, Research Directorate General as part of the 6th Framework Programme, Priority 8, "Policy Support and Anticipating Scientific and Technological Needs".
Overview of presentation
• WP10: linkages with firm-level data• Use EUKLEMS data in micro-econometric analysis
• Industry price deflators and PPPs• Instruments from IO and/or trade matrices: Shea (94) , BCL(94)
• Add micro-aggregated indicators to EUKLEMS database• Higher moments, covariances, gross flows, etc.
• Integrate micro data sources into EUKLEMS statistical process• First: confrontation of different sources of productivity measures• Next: consistent, integrated, international data creation
• Paper: Bartelsman, Scarpetta, Haltiwanger (2005)• Creating higher moments as addition to EUKLEMS database
Measuring and Analyzing Cross-country
Differences in Firm Dynamics
Eric Bartelsman, Stefano Scarpetta, and John Haltiwanger
Free University Amsterdam and Tinbergen Institute; World Bank; University of Maryland and NBER
The firm-level project: a network of experts
• The firm-level project would have been impossible without extensive effort and support of many colleagues
• Mika Maliranta, Satu Nurmi, Jonathan Haskel, Richard Duhaitois, Pedro Portugal, Thorsten Schank, Fabiano Schivardi, Ralf Marten, Ylva Heden, Ellen Hogenboom, Mihail Hazans, Jaan Masso, John Earle, Milan Vodopivec, Kaplan, Maurice Kugler, Mark Roberts...
• The firm-level projects were funded by OECD, World Bank, various national government and NSOs
Distributed micro-data collection
• OECD sample• Demographics (entry/exit) for 10 countries
• Productivity decompositions for 7 countries
• Survival analysis 7 countries
• World Bank sample• Same variables, 14 Central and Eastern Europe, Latin
America and South East Asia
• EU Sample (10 countries), updates and a few new countries• Productivity decompositions
• Sample Stats and correlations by quartile
Data sources
• Business registers for firm demographics• Firm level, at least one employee, 2/3-digit industry
• Production Stats, enterprise surveys for productivity analysis
• Countries:• 10 OECD
• 5 Central and Eastern Europe
• 6 Latin America
• 3 East Asia
• Data are disaggregated by:• industry (2-3 digit);
• size classes 1-9; 10-19; 20-49; 50-99; 100-249; 250-499; 500+ (for OECD sample the groups between 1 and 20 and the groups between 100 and 500 are combined)
• Time (late 1980s – early 2000s)
Distributed micro data research
Provision of metadata.Approval of access.Disclosure analysis
Disclosure analysis of Publication
NSO
sR
esea
rche
r
Policy QuestionResearch Design Program Code
Publication
Net
wor
k Metadata
Networkmembers
Cross-countryTables
Measurement Error
• Three sources of error potentially affect comparability of indicators built from firm level data:
• Classical Error of firm-level measure
• Errors in observed firms (sample)
• Method of Aggregation of Indicator
• Aggregation is harmonized in our approach, but other errors may or may not cancel out in aggregation
*XX
*
fXAI f |
Case Variable Aggregator Disaggregation Potential Problems
1a Employment Mean/Sum Aggregate or Industry Industry misclassification, Sample selection
1b “” Mean/Sum Size Class Sample selection
1c
“”
Mean/Sum Firm Status (Continuer, Entrant, Exit)
Sample selection, Measurement error in longitudinal IDs
2a
“”
Variance Aggregate or Industry Sample selection, Classical measurement error
1a Productivity Mean Aggregate or Industry Industry misclassification, Sample selection,
1b Productivity Mean Productivity quartiles Sample selection, Classical measurement error
1c Prod change Mean Firm Status (Continuer, Entrant, Exit)
Sample selection, Measurement error in longitudinal IDs, Classical measurement error
Dynamic allocative efficiency: the importance of “technology factors”We decompose our data for manufacturing into a low technology group and a medium high tech group
Stronger contribution of entry to productivity growth in medium-to-high tech industries
-1.5
-1
-0.5
0
0.5
1
1.5
Argen
tina
Chile
Colom
bia
Eston
ia
Finlan
d
Franc
e
Korea
Latv
ia
Nethe
rland
s
Portu
gal
Sloven
ia
Taiwan UK
USA
Low tech industries Medium-high-tech industries
Contribution of entry to labor productivity growth, five year differencing, gross output
Labor Productivity Dispersion
ICT-producing ICT-using
Quartile US EU US EU
Top 123 118 74 58
3 88 87 51 48
2 61 72 40 46
Bottom 38 68 26 41
Units: Thousand US$ per worker
11
Dynamic efficiency : evolution of resource allocations over time, depending on initial productivity (by quartile)
Firm growth by Initial Productivity
-4
-2
0
2
4
6
8
0 20 40 60 80 100 120
LPV (US $, 1000/worker)
firm
gro
wth
(%
) FIN
FRA
GBR
NLD
SWE
USA
EUx
Firm growth is geometric mean of output and employment growth in 5-year period. LPV: labor productivity
In initial period of firms per productivity quartile, measured in value added (1000 $ ) per worker.
q4q3
q2
q1
Micro-aggregated indicators
• Distributed micro-data research is a practical way to exploit information in (confidential) firm-level datasets located at separate sites.
• While simple level comparisons may be problematic, difference-in-difference approach looks more promising
• There is significant cross-country variation in firm-level indicators that may be linked to differences in policy or market environment
Integrating micro-level statistical sources
• Using micro-level sources and integration framework is flexible way to generate customized statistics
• Micro-level sources may provide check on aggregate analytical indicators, such as output per worker• e.g. Nominal gross output per worker, aggregated from micro
data compared with same measure from National Accounts (STAN database).
• Different owing to: gross output a residual in N.A.; labor sources in N.A. with different industry distribution; sampling selectivity at micro level; unit of observation (firm/estab).
Further work in WP 10
• Survey paper with 2 components • Policy research using firm-level data
• Testing hypothesis
• Policy Evaluation
• Linkages with sectoral and micro data• Merging sectoral data into micro for econometric research
Much literature from US, and increasingly other OECD and global
• Using indicators built from micro data for sectoral research Theoretical in trade/IO/labor, some single country and BHS
• Statistics production from integrated micro-level sources