What explains the productivity gap of less developed regions? Carlo Menon

Presentation by Carlo Menon, Economist, Co-ordinator of the Spatial Productivity Lab, CFE, OECD at the 25th Spatial Productivity Lab meeting, held online on 26 February 2025.

More info: https://oe.cd/SPL

1. @OECD_local www.linkedin.com/company/oecd-local www.oecd.org/cfe
WHAT DRIVES THE PRODUCTIVITY GAP IN LESS
DEVELOPED REGIONS?
25
th
Meeting of the Spatial Productivity Lab, February the 26
th
, 2025
Carlo Menon, Wessel Vermeulen
Spatial Productivity Lab, OECD Trento Centre for Local Development
OECD Centre for Entrepreneurship, SMEs, Regions and Cities
Insights from firm-level data

2. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre 2
Within-country variation in GDP per worker can be as sizable as
between countries
Note: Regions with GDP per worker above USD200000 are suppressed. GDP deflators and purchasing-power indicators are not typically available at subnational level, which dictates caution in interpreting relatively small differences in regional
productivity.
Source: https://www.oecd.org/regional/regional-statistics/.
GDP per worker (USD2015 PPP), by TL2 region, 2022

3. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
Italian and Spanish less developed region (LDRs) have not narrowed
their gap
3
GDP per worker (USD2015, ppp)Molise
Campania
Apulia
Basilicata
Calabria
Sicily
Sardinia
Castile-
La Mancha
Extremadura
Andalusia
Ceuta
Melilla
Italy Spain
2000 2010 2020 2000 2010 2020
80 000
90 000
100 000
110 000
120 000
130 000
Less developedMore developed Transition
Source: https://www.oecd.org/regional/regional-statistics/.

4. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
Overview
4
•Productivity gaps of less developed regions in (Southern) Europe are large and
persistent
•Is the gap equally distributed across sectors and groups of firms?
•Is the gap driven by too many low-productivity firms or too few top
performers?
The issue
•We use firm-level data to disentangle the productivity gaps of “less developed”
regions (LDRs) with respect to “developed” regions (DRs) in Italy and Spain
What this paper is doing

5. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
Tradable sectors account for a lower share of employment in LDRs35%
24%
17%
25%
3%
13%
5%
9%
25% 24%
16% 18%
4%
10%
8%
11%
Italy Spain
Less developed More developed Less developed More developed
0.0
0.2
0.4
0.6
Employment shares
Advanced servicesMedium-high tech mfgMedium-low tech mfg & extractiveOther services tradable
60%
40%
20%
0%
Source: Authors’ elaboration on Moody’s-BVD ORBIS

6. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
Firms in LDRs are 20-30% less productive and smaller across all
sectors
Average productivity (MFP) and employment gaps of LDRs compared to DRs by sector group,
conditional on 2-digit sector, size and age classesItaly Spain
-0.5-0.4-0.3-0.2-0.10.00.1-0.5-0.4-0.3-0.2-0.10.00.1
Medium-high tech mfg
Medium-low tech mfg & extractive
Advanced services
Other services tradable
Estimate
Dependent variableMFP Employment
Note: the charts show the point estimates (dots) and 5% confidence interval (whiskers) from firm-level regressions of MFP or employment on a dummy variable for LDRs, controlling for age and size classes.
Source: Authors’ elaboration on Moody’s-BVD ORBIS.

7. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
The gap is comparable across firms of different sizesItaly Spain
-0.3 -0.2 -0.1 0.0 -0.3 -0.2 -0.1 0.0
0-9
10-19
20-49
50-99
100+
0-9
10-19
20-49
50-99
100+
Estimate
Average productivity (MFP) gaps of LDRs regions compared to DRs by size class,
conditional on 2-digit sector, size and age classes
-30% -20% -10% 0 -30% -20% -10% 0
Note: the charts show the point estimates (dots) and 5% confidence interval (whiskers) from firm-level regressions of MFP on a dummy variable for LDRs, controlling for age classes and 2-digit sector dummies.
Source: Authors’ elaboration on Moody’s-BVD ORBIS.

8. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
Productivity gaps are reflected in profits and wages
Dep. Var: EBIDTA Italy Spain
Size class Coeff.
median
EBIDTA Coeff.
Median
EBIDTA
100+ -372,409*** 1,503,000 -65,143*** 1,446,000
50-99 -265,183*** 711,956 -165,242*** 554,362
20-49 -121,005*** 264,900 -39,442*** 203,212
10-19 -30,015*** 98,512 -22,130*** 65,832
0-9 7,482*** 24,110 -5,652** 12,339
Observations 211,445 106,540
R-squared 0.452 0.491
Note: all regressions include polynomials of gross output, employment and fixed assets, as well
as fixed effects for size class, age class, and sectors. EBIDTA is winsorised at the 1% level.
Source: Authors’ elaboration on Moody’s-BVD ORBIS.
Average worker compensation gap in LDRs by firm size class, conditional on
2-digit sector, size and age classes
Average profit (EBIDTA) gap in LDRs by firm size class, conditional on 2-digit
sector, size and age classes0-9
10-19
20-49
50-99
100+
-0.4 -0.3 -0.2 -0.1 0.0
Estimates on ln(WL/L)
Italy Spain
-40% -30% -20% -10% 0
Note: the charts show the point estimates (dots) and 5% confidence interval (whiskers) from
firm-level regressions of total payroll per employee on a dummy variable for LDRs, controlling
for age classes and 2-digit sector dummies.

9. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
Are there some places in LDRs that “escape” the gap?
Distributions of average productivity across small regions (TL3)
Italy Spain
Note: the plotted values are TL3 dummies from a firm-level regression of TFP on sector 2-digit dummies.
Source: Authors’ elaboration on Moody’s-BVD ORBIS

%
%

10. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
Shifting, selection, or dilation?
•Shifting: All firms in LDRs have the same productivity gap, irrespective of whether they are low-productivity or
high-productivity firms (relative to others in the same region)
•Selection: There are more low-productivity firms in LDRs, as there is less competition and thus less selection
(“survival of the fittest”)
•Dilation: There are fewer high-productivity firms in LDRs
Note: Simulated examples

11. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
A lot of shifting, some positive dilation, no selection
Country:
Italy Spain
Sectors:
All tradable
sectors
Manufacturing
Tradable
services
All tradable
sectors
Manufacturing
Tradable
services
Shifting (all
firms)
-30.6% -40.2% -26.1% -29.1% -36% -26.1%
Dilation (top
performers)
+4.1% +10.6% = = +9.4% -4.8%
Selection
(bottom
performers)
= = = = = =
Source: Authors’ elaboration on Moody’s-BVD ORBIS

12. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
Summing up…
•Sectoral composition is markedly different, with manufacturing and tradable services
weighting less in LDRs
•20-30% multi-factor productivity (MFP) gap in LDRs, remarkably constant across
(tradable) sectors and size classes
•The gap is fully reflected in payroll per worker and profits
•The gap is not due to weaker market selection
•In manufacturing, the gap is smaller for the top performers
•In Spain the size (and capital deepening) gap is smaller than in Italy

13. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre
Policy discussion
The gap is driven by a business environment that limits the incentives to invest
and grow for all types of businesses → the priority is removing barriers
Differences in institutional setting, e.g. decentralised wage bargaining, may
explain some of the differences between Italy and Spain (work in progress!)
Local development policies should prioritise tradable sectors
Local links are an asset: firms are not moving despite lower salaries and profits
There is more dispersion in LDRs: which are the high-productivity firms in LDRs?

14. © OECD | Centre for Entrepreneurship, SMEs, Regions and Cities | @OECD_Local| OECD Trento Centre

www. oecd. org /cfe www. trento. oecd. org
Twitter: @OECD_local
LinkedIn: www.linkedin.com/company/oecd-local
Website: www.oecd.org/cfe & www.trento.oecd.org
Blog: oecdcogito.blog
Thank you for your attention
[email protected]
[email protected]
Do you want to know more?
•Buzzacchi L.,A. De Marco, M. Pagnini (2024), Agglomeration and the Italian North–South divide, Journal of Economic
Geography, 24:5, p.p. 707–728, https://doi.org/10.1093/jeg/lbae025
•Combes, P.-P., Duranton, G., Gobillon, L., Puga, D. and Roux, S. (2012), The Productivity Advantages of Large Cities:
Distinguishing Agglomeration From Firm Selection. Econometrica, 80: 2543-2594.
https://doi.org/10.3982/ECTA8442
•OECD (2023), OECD Regional Outlook 2023: The Longstanding Geography of Inequalities, OECD Publishing, Paris,
https://doi.org/10.1787/92cd40a0-en
•OECD (2018), Productivity and Jobs in a Globalised World: (How) Can All Regions Benefit?, OECD Regional
Development Studies, OECD Publishing, Paris, https://doi.org/10.1787/9789264293137-en
•Rice, P and A Venables (2023), Tradability, Productivity, and Regional Disparities: theory and UK evidence, CEPR
Discussion Paper No. 17806, https://cepr.org/publications/dp17806