Structural Reforms: Need and Impact The case of the product market - - PowerPoint PPT Presentation

Structural Reforms: Need and Impact

The case of the product market reforms

GILBERT CETTE

BANQUE DE FRANCE AND UNIVERSITÉ OF AIX-MARSEILLE

Luxembourg, December 2012 1

General introduction

• After recovery plans, in 2008-2009, European countries have

now engaged public finance consolidation plans

• Since 2008, numerous papers on the size of the fiscal

multipliers Among others, Romer and Romer (2010), Perotti (2012), Ramey (2011), Alesina and Ardagna (2012), Parker (2011), Corsetti, Meier and Gernot (2012), Auerbach and Gorodnichenko (2012), Alesina, Favero and Giavazzi (2012)...

• Incertainties, in particular in times of recessions

General introduction

Luxembourg, December 2012

Gilbert Cette

2

General introduction

The three main questions are now:

• What is the GDP impact of a coordination of country

consolidation plans? Particularly in Europe?

• What is the public finance impact of a coordination of

country consolidation plans? Particularly in Europe?

• What is the best consolidation plan schedule (less GDP

costly)? My presentation deals with the two first questions.

General introduction

Luxembourg, December 2012

Gilbert Cette

3

General introduction

Four questions are raised in particular in this presentation:

• The two previous ones

For this, I will use a multinational modelisation tool, named

• MacSim. This tool is a software to simulate multinational

macroeconomic policies, built by Brillet, Cette, Gambini and Lagoarde-Segot.

• What could be other sources of growth?

I will show that product market structural reforms could be

• ne of these sources. Labour market also, but I will stay on the

product market

• How would it be possible to articulate product market

structural reforms and fiscal consolidation?

General introduction

Luxembourg, December 2012

Gilbert Cette

4

General introduction

In all presented simulations:

One MU including: Germany, Belgium, France, Italy, Luxembourg, The Netherlands, OEA (for Other countries of the Euro Area) Interest rates: Taylor rule (Option 3) Risk premium coefficient = 0.2 Exchange rate: Purchasing Power Parity (Option B) External trade price elasticities: Supposed equal to one Result presentation: Differences with a reference situation without the policy GDP and consumer price: difference in % Public finance balance and current account: differences in GDP points

General introduction

Luxembourg, December 2012

Gilbert Cette

5

General introduction

Fiscal consolidation Public demand shock: Dicrease of the public demand by one GDP point 2 scenarios:

• Scenario 1: Consolidation in one country
• Scenario 2: Simultaneous consolidation in all industrialised

countries

MacSim simulation results General introduction

Luxembourg, December 2012

Time France Luxembourg GDP Cons. Price Cur. Acc. Public acc. GDP Cons. Price Cur. Acc. Public acc. Scenario 1 Year 1

• 0.70

0.12 0.57 0.71

• 0.16

0.23 0.54 1.10 Year 8

• 0.57
• 0.57

0.61 0.63

• 0.06

0.18 0.80 0.92 Scenario 2 Year 1

• 1.39
• 0.05

0.27 0.50

• 1.22
• 0.10

0.54 0.62 Year 8

• 1.56
• 2.83

0.17 0.10

• 1.69
• 2.18

0.40

• 0.59

Gilbert Cette

6

General introduction

• Standard low multipliers for scenario 1

8 years: France 0.6 and Luxembourg 0.1

• In the scenario 2, high multipliers

8 years: France 1.6 and Luxembourg 1.7

• Public finance account consolidation impact:

Standard in scenario 1 8 years: France 0.6 and Luxembourg 0.9 Nill or negative in scenario 2 8 years: France 0.1 and Luxembourg -0.6

• Fiscal consolidation is inefficient in case of coordination of

the fiscal consolidation

• Other sources of growth have to be associated to the fiscal

consolidation

• Structural reforms

Here, product market ones

Gilbert Cette

7

General introduction

Luxembourg, December 2012

Upstream Product Market Regulations, ICT, R&D and Productivity

by

Gilbert Cette Banque de France, DEFI Jimmy Lopez Banque de France, LEG Jacques Mairesse CREST-INSEE, UNU-MERIT, NBER, BdF

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

8

Motivation

• Competition is an important determinant of productivity growth:

According to endogenous growth theory, the link between competition and productivity growth relies on incentives to escape peer pressure and reap temporary rents by improving efficiency through innovation/adoption/imitation

• Most
• f

previous empirical research focus

• n

competitive conditions within each sector (or market):

Nicoletti and Scarpetta 2003; Inklaar et al. 2008; Buccirossi et al. 2009; Nickell, 1996; Nickell et al. 1997; Blundell et al. 1999; Griffith et al. 2002; Aghion et al. 2004 ; Haskel et al. 2007; Conway et al. 2005; Aghion et al. 2009.....

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

9

Motivation

• Only few papers are accounting for indirect effect of lack of upstream

competition, using data on:

cross-section: Allegra et al. (2004), Faini et al. (2006), Barone and Cingano (2008) single country: Arnold et al. (2006) and Forlani (2010) cross country/industry panel: Conway et al. (2005), Bourlès et al. (2010, RES forthcoming)

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

10

Bourlès, Cette, Lopez, Mairesse and Nicoletti (2010)

• Underlines

two channels from upstream competition to downstream productivity:

upstream market power is used to grab part of the downstream rents barriers to competition upstream generate less entry and weaker competition downstream

• Examine empirically this link:

using industry-level/cross-country panel data proxying upstream competition with the OECD indicators of sectoral product market regulation, focusing

• n

non- manufacturing

• Find that anticompetitive regulations upstream curb significantly

MFP growth

Motivation

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

11

Purpose and findings of paper

• Go in detail on the channels of the upstream competition impact
• Estimate the impact of upstream regulations on ICT and R&D

investments downstream

• Evaluate the importance of these channels relatively to the whole

impact of upstream regulations on productivity

• Using,

as Bourlès et al. (2010, RES forthcoming), industry- level/cross-country panel data and OECD upstream regulations indicators

• Main findings
• An important significant negative impact on R&D accumulation
• A smaller but significant impact on ICT accumulation
• The R&D and ICT channels explain between 30% and 60% of the

whole impact on productivity

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

12

Outline

1. Empirical specifications 2. Data 3. Main empirical results 4. Economic significance of estimates: policy simulations 5. Further analysis 6. Sensibility analysis 7. Conclusion

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

13

Productivity specification

• The cointegrated relationship between domestic and leader MFP is

augmented by the ‘Regulatory Burden’ indicator:

With: small letters for logarithm; sector (s), country (c), and year (y) indices omitted

• The MultiFactor Productivity defined by:

With: Y the value added; L the level of employment; C, Cn and K the ICT, non-ICT and Knowledge capital stocks, respectively

• In order to estimate the elasticities, we introduce the accounting

MFP into the cointegrated relationship:

• With π = β + α + γ the return to scale

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

14

Productivity specification

• The benchmark specification assumes:
• Constant return to scale (π = 1)
• The non-ICT capital elasticity is equal to its share in total costs

in the US, specific by sector (noted , sector specific)

• The MFP gap with the US is bounded on the long-run (ρ = 1)

We check the sensibility of the estimates to these assumptions

• Finally the estimated long-term productivity equation is:

With uscy = θs + θcy + εscy in our favourite specification 5 sectors among our estimation sample are almost not investing in R&D. For these sectors we assume τ = 0

α ~

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

15

Capital demand specifications

• We use a common factor demand specification:
• With X a production factor (X = L, Cn, C, K)

This common specification come from an intertemporal maximization of profit for a firm with a CES production function

• The specifications of ICT and R&D intensity relatively to labor are

then deduced and augmented by the ‘Regulatory burden’ indicator:

With ui,scy = θi,s + θi,cy + εi,scy , i = { k, c }, in our favourite specifications With a lagged employment level because of simultaneity issues Note that we relax the assumption of equal factors elasticity of substitution between K and C

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

16

∑

=

=

K k T s k T s k t c t cs

• utput

input NMR

1 , , , ,

* Reg

• To proxy for upstream competition:

we use the anticompetitive Non-Manufacturing Regulations (NMR) indicators build by the OECD 6 sectors are covered: Energy, transport, communication, retail services, banking sector and professional services Advantages: minimize endogeneity issues and provide direct link with policies

• The ‘Regulatory burden’ indicator (Reg):

With: K = 6 non-manuf. Sectors; T0 = 2000: inputs from the US I-O table Note that when introducing country*year fixed effect, as in our favourite specifications, we test the conjecture that the impact of the upstream regulations is growing with the intensity of use of the regulated inputs. In this respect, our estimation method is a ‘diff-in- diff’ approach

“Regulatory burden” indicator

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

17

NMR indicators Sector component contribution in 2007

Scale 0-6 for each sector, 0 for the most pro-competitive regulations

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

18

“Regulatory burden” indicator Industry sample average

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

19

“Regulatory burden” indicator

Country sample average

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

20

Data: labor productivity and capital stocks

• Data sources:
• n value added and total employment from STAN (OECD)

R&D expenses from ANBERD (OECD) Physical investments – ICT and non-ICT – from EUKLEMS

• Capital stocks calculated with the P.I.M.:

X = ( 1 – δ.X ) * X-1 + Ix,-1 with: X: R&D, hardware, software, com. eq., non-ICT eq., non-residential structures δ the depreciation rate of capital (25% for R&D, 30% for hardware and software, 15% for com. eq., 10% for non-ICT eq. and 5% for non-

• resid. struct.)

The R&D price is approximated by the GDP price

• The US gaps are calculated using the OECD 2000 PPP’s

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

21

ICT and R&D investment coefficient industry sample averages (2001-2005 period)

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

22

List of sectors

SECTORS ISIC rev. 3 code

FOOD PRODUCTS, BEVERAGES AND TOBACCO 15-16 TEXTILES, TEXTILE PRODUCTS, LEATHER AND FOOTWEAR 17-19 WOOD AND PRODUCTS OF WOOD AND CORK 20 PULP, PAPER, PAPER PRODUCTS, PRINTING AND PUBLISHING 21-22 CHEMICAL, RUBBER, PLASTICS AND FUEL PRODUCTS 23-25 OTHER NON-METALLIC MINERAL PRODUCTS 26 BASIC METALS AND FABRICATED METAL PRODUCTS 27-28 MACHINERY AND EQUIPMENT, N.E.C. 29 ELECTRICAL AND OPTICAL EQUIPMENT 30-33 TRANSPORT EQUIPMENT 34-35 MANUFACTURING NEC; RECYCLING 36-37 ELECTRICITY GAS AND WATER SUPPLY (Energy) 40-41 CONSTRUCTION 45 WHOLESALE AND RETAIL TRADE; REPAIRS (Retail services) 50-52 HOTELS AND RESTAURANTS 55 TRANSPORT, STORAGE (Transport), POST AND TELECOMMUNICATIONS (Communication) 60-64 FINANCIAL INTERMEDIATION (Financial Services) 65-67 RENTING M&EQ AND OTHER BUSINESS ACTIVITIES (Business Services) 72-74

• Sectors with almost no ICT and R&D investments (then excluded)
• Sectors with almost no R&D investments (but some in ICT)
• When

a downstream sector is also an upstream sector its abbreviation is indicated between parenthesis

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

23

Sample

• The sectors almost not investing in ICT and R&D (5 sectors) are

excluded from our estimation sample

• The resulting estimation sample include:

2612 observations 18 years: 1989-2006 14 countries (+ US for the gap calculations) 13 manufacturing and market service sectors

• The R&D demand is estimated on the 8 sectors investing sensibly

in R&D only (1490 observations)

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

24

Labor productivity ratio with the US country sample 2001-2005 average

The low R&D sectors are excluded from the sample - The panel isn’t balanced Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

25

Labor productivity Histogram of log growth rate

Average growth rate: 3.61% Median growth rate: 2.65%

2 4 6 8 10 Density

• .4
• .2

.2 .4 .6 Labor Productivity Growth

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

26

Descriptive statistics

Level (2) Annual growth rate Q1 Median Q3 Mean Q1 Median Q3 Mean Labor productivity 4.41 5.51 7.24 6.54

• 3.10%

2.65% 6.24% 3.61% Non-ICT capital intensity 1.8 2.35 3.89 2.98

• 0.29%

2.08% 4.65% 2.46% ICT capital intensity 5.30 5.96 6.74 6.01 5.93% 10.39% 15.55% 11.34% R&D capital intensity (1) 5.63 6.52 7.65 6.54 1.06% 5.12% 10.22% 5.85% Labor productivity gap

• 0.54
• 0.39
• 0.26
• 0.42
• 4.15%
• 0.29%

3.61%

• 0.19%

ICT capital intensity gap

• 1.1
• 0.75
• 0.27
• 0.73
• 52.2%
• 0.13%

5.3% 0.28% R&D capital intensity gap (1)

• 1.28
• 0.54
• 0.04
• 0.62
• 4.94%

1.01% 7.02% 1.55% Non-ICT capital intensity gap

• 0.39

0.13 1.44 0.69

• 2.95%

0.06% 3,00%

• 0.01%

Regulatory burden indicator 0.4 0.65 0.88 0.65

• 4.75%
• 2.62%
• 1.17%
• 3.33%

ICT capital relative cost 0.87 1,00 1.09 1.01

• 2.7%
• 1.39%

5.04% 0.9% R&D capital relative cost (1) 0.91 1,00 1.07 1.02

• 2.15%

1.22% 3.54% 0.17% Labor relative cost 0.9 0.98 1.03 0.96

• 0.25%

1.98% 4.9% 3.54%

(1): the low R&D sectors are excluded from the sample (2): the variables are in logarithm, except the ‘regulatory burden’ indicator Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

27

Main estimates

Table 1: Production function

*** significant at 1%; ** significant at 5%; *significant at 10%

Sector and country*year fixed effects are included in all specifications Dependent variable: MFP gap (labor productivity minus the calibrated impact of non-ICT capital stock) DOLS estimations with one lag and one lead (their coefficients are not presented)

(1) (2) (3) (4) (5) (6) (7) (8) Gap in ICT capital intensity 0.052*** 0.085*** 0.074*** 0.101*** [0.009] [0.009] [0.009] [0.009] Gap in R&D capital intensity 0.078*** 0.092*** 0.069*** 0.088*** [0.007] [0.007] [0.007] [0.007] Regulatory burden indicator-1

• 0.234***
• 0.219***
• 0.257***
• 0.253***
• 0.064
• 0.109
• 0.083
• 0.155**

[0.055] [0.056] [0.056] [0.057] [0.067] [0.068] [0.069] [0.071] Observations 2612 2612 2612 2612 2612 2612 2612 2612 R-squared 0.565 0.556 0.540 0.518 0.646 0.631 0.624 0.596 RMSE 0.1821 0.1838 0.1870 0.1911 0.1720 0.1756 0.1771 0.1835 Sector*year fixed effects N N N N Y Y Y Y

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

2 8

Main estimates

Table 2: R&D demand

*** significant at 1%; ** significant at 5%; *significant at 10%

Sector and country*year fixed effects are included in all specifications Dependent variable: R&D capital intensity (ln (K/L)) DOLS estimations with one lag and one lead (their coefficients are not presented)

(1) (2) (3) (4) R&D capital over labor costs

• 0.628***
• 0.619***

[0.128] [0.135]

Regulatory burden indicator-1

• 1.395***
• 1.261***
• 0.868**
• 0.758*

[0.385] [0.382] [0.425] [0.417]

Observations

1478 1478 1478 1478

R-squared

0.801 0.795 0.810 0.805

RMSE

0.6599 0.6672 0.6776 0.6843

Sector*year fixed effects N N Y Y

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

29

Main estimates

Table 3: ICT demand

*** significant at 1%; ** significant at 5%; *significant at 10%

Sector and country*year fixed effects are included in all specifications Dependent variable: ICT capital intensity (ln (C/L)) DOLS estimations with one lag and one lead (their coefficients are not presented)

(1) (2) (3) (4) ICT capital over labor costs

• 0.758***
• 0.728***

[0.041] [0.045]

Regulatory burden indicator-1

• 0.263**
• 0.582***
• 0.342**
• 0.614***

[0.125] [0.133] [0.164] [0.174]

Observations

2612 2612 2612 2612

R-squared

0.863 0.842 0.871 0.853

RMSE

0.4139 0.4443 0.4220 0.4487

Sector*year fixed effects N N Y Y

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

30

Economic significance Implied MFP gains from reforms

• Proxying upstream competition with the OECD indicators makes it

possible to link policies to MFP

• Different experiments are possible: simpler one is instantaneous

alignment in 2007 on sectoral “lightest practice” observed

• ‘Lightest practice” is defined for each sector as the ‘Regulatory

burden’ indicator resulting from the average of the three smallest values of the NMR in each upstream sector

• Calculation using domestic I-O tables and value added shares
• Compute the long term R&D, ICT and productivity changes according

to: A: the preferred estimates (col. (1) of tables 1, 2 and 3) B: the alternative introducing sector*year fixed effects (col. (5) of table 1, col. (3) of tables 2 and 3)

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

31

Economic significance

R&D capital intensity gains from reforms

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

32

Economic significance

ICT capital intensity gains from reforms

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

33

Economic significance

Productivity gains from reforms

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

34

Conclusions

• Negative significant impacts of upstream regulations on:

R&D and ICT capital stocks Productivity (in addition to the previous channels)

• According to the simulation policies:

The average impact of the simulated policy reforms on R&D capital is more than three time higher than the impact on ICT capital The whole impact of the reforms on productivity would be of 6.7% on average The R&D and ICT channels explain 30% of the whole impact

• However, when sector*year fixed effects are introduced:

The additionnal direct impact on productivity is smaller and

• ften not significant…

…whereas the impacts on the channels are not very sensitive Therefore, the whole simulated impact of the policy reforms is strongly reduced (to 2.6% on average), and is explained at 60% by the R&D and ICT channels

Upstream Product Market Regulations,…

Luxembourg, December 2012

Gilbert Cette

35

General conclusion

What could be the impact of coordinated product market structural reforms? 1 scenario: TFP acceleration by 0.5 points per year in the business sector for all industrialised countries

MacSim simulation results GDP impact General conclusion

Luxembourg, December 2012

France Luxembourg Year 1 0.08 0.20 Year 8 0.90 0.40

Gilbert Cette

36

General conclusion

The implementation

• f

product market structural reforms could avoid the negative GDP impact of the fiscal consolidation Urgency of this implementation

General conclusion

Luxembourg, December 2012

Gilbert Cette

37