Banque Nationale De Belgique Sa · EURONEXT:BNB
Banque Nationale De Belgique Sa — National Bank of Belgium — Technical Presentation (2010)
34 slides from Banque Nationale De Belgique Sa's technical presentation, Oct 26, 2010
Slide 1: The productivity and export spillovers of the internationalisation behaviour of Belgian firms
The productivity and export spillovers of the internationalisation behaviour of Belgian firms Michel Dumont (Federal Planning Bureau and UA) Bruno Merlevede (UGent and HUBrussel) Christophe Piette (National Bank of Belgium) Glenn Rayp (UGent, Sherppa)
Slide 2: Introduction
Introduction • Are there spillover effects on the decision to start exporting by Belgian domestic firms of the internationalisation behaviour of other firms in Belgium? • Firms will export if their productivity exceeds a threshold (Melitz, 2003) – Suggest two channels for spillovers • Productivity •
Slide 3: Firm A domestic, Firm B internationalised: TFP threshold
Firm A domestic firm Firm B internationalised firm TFP threshold A B
Slide 4: Firm A domestic, Firm B internationalised: TFP threshold
Firm A domestic firm Firm B internationalised firm TFP threshold A B
Slide 5: Pure productivity spillover
TFP Pure productivity spillover threshold A B A B
Slide 6: Pure threshold spillover
TFP Pure threshold spillover threshold A B A
Slide 7: Productivity and threshold spillover
TFP Productivity and threshold spillover threshold A B A
Slide 8: Productivity spillover effects
Productivity spillover effects • Do spillovers affect the productivity of domestic firms? • Ample research on impact of foreign MNE presence – Alvarez and Lopez (2008) focus on spillovers from domestic exporters • Within and between industries • Demonstration effects, competition effects, labour mob
Slide 9: Threshold spillover effects
Threshold spillover effects • Do spillovers affect the probability to start exporting controling for productivity? • Yes – Aitken et al. (1997): multinational exporters in Mexico – Greenaway and Kneller (2008): both exporters in same industry and in same region (independent) in the UK – Koenig et al
Slide 10: This paper
This paper • Spillover effects on – TFP – Decision to start exporting (controling for tfp) • Spillover at – industry level (economic distance) – regional level (geographic distance) • Spillover from – Exporters (domestic and foreign) – MNEs (domestic and foreign)
Slide 11: Data
Data • Sample of Belgian manufacturing firms with at least 5 employees • Detailed information at product-destination level for exports – data collapsed to 4 digits nomenclature • Because reporting thresholds were raised in 1998 and in 2006 with respect to foreign trade data we focus on the period 19
Slide 12: Data: Population of manufacturing firms in 2005
Data Table 1: Population of manufacturing firms in 2005 Firms with less Firms with at Firms with TFP estimates based Total population than5 1 least 5 : saute Radio employees employees method? Number of firms 24,027 15,250 8,777 6,114 Number of exporters 5,632 1180 4,452 3,324 Number of foreign firms
Slide 13: Data
Data • Crossroads bank: address(es) of firms and their establishments – Determine region • geographical unit = NUTS 3 level – Official classification used for dividing up economic territory of the EU – small regions for specific diagnoses with population between 150,000 and 800,000 – Belgium has 43
Slide 14: Plants of manufacturing firms by NUTS3 region
Plants of manufacturing firms* by NUTS3 region (data for 2005) (357,865] (254,357] (123,254] (58,123] [21,58] * Firms with at least 5 employees on average over the period 1998-2005
Slide 15: Plants of exporters by NUTS3 region
Plants of exporters* by NUTS3 region (data for 2005) (199,417] (128,199] (72,128] (25,72] [9,25] * Firms with at least 5 employees on average over the period 1998-2005
Slide 16: Plants of multinational firms by NUTS3 region
Plants of multinational firms* by NUTS3 region (data for 2005) (51,128] (28,51] (13,28] (4,13] [1,4] * Firms with at least 5 employees on average over the period 1998-2005
Slide 17: Productivity measures
Productivity measures • Index-approach as benchmark (Good et al., 1996): firm-level total factor productivity computed using index approach, representative firm (2-digit industry mean): t TFPit yit yt* ys* ys* 1 s 2 1 n n 1 t * Sijt S *jt xijt x*jt S j ,s S *j , s 1 x*j , s x*j , s 1 2j1 j 12s 2 • R
Slide 18: TFP distribution according to the type of involvement on foreign markets
TFP distribution according to the type of involvement on foreign markets (data for 2005) Total Factor Productivity (Index) Firms only involved on the Belgian market = ————- Exporters —-—-- Firms with foreign affiliates
Slide 19: Estimation
Estimation • Identifying TFP spillovers – OLS on first differenced model • Identifying threshold spillovers – Logit on subsample of firms facing the decision to start exporting (product p to destination d)
Slide 20: Spillover variables
Spillover variables • productivity and threshold spillover literature have measured spillover variables on a different basis – # exporters/MNEs likely affects prevalence of knowledge about foreign markets – volumes produced/sold likely affect the size and efficiency of linked industries • productivi
Slide 21: 1. Productivity spillovers
1. Productivity spillovers (1) (2) (3) (4) never non exporters exporters index ACF OLS index exporters' share in region va 0.130*** -0.028 0.099** 0.092** [0.038] [0.067] [0.041] [0.045] MNEs' share in region va -0.026 0.023 -0.017 -0.017 [0.017] [0.029] [0.018] [0.020] exporters' share in industry
Slide 22: Productivity spillovers
Productivity spillovers (1) (2) (3) (4) never non exporters exporters index ACF OLS index exporters' share in region va 0.130*** -0.028 0.099** 0.092** Firm productivity increases with [0.038] [0.067] [0.041] [0.045] • presence of exporting firms in MNEs' share in region va -0.026 0.023 -0.017 -0.01
Slide 23: Productivity spillovers
Productivity spillovers (1) (2) (3) (4) never non exporters exporters index ACF OLS index exporters' share in region va 0.130*** -0.028 0.099** 0.092** Firm productivity increases with [0.038] [0.067] [0.041] [0.045] • presence of exporting firms in MNEs' share in region va -0.026 0.023 -0.017 -0.01
Slide 24: 2. Decision to start exporting
2. Decision to start exporting • create dependent variable indicating the "new exporter" status on the basis of a firm’s export status • a firm is a new exporter in year t if it exports in t but was not exporting in t-1 and t-2 – continuous exporters are not part of the estimation sample – never-exp
Slide 25: Decision to start exporting firm-level
Decision to start exporting firm-level • Lagged productivity is the only robust determinant • No robust spillovers detected
Slide 26: Decision to start exporting - destination
Decision to start exporting - destination • In order to restrict the sample to computer-power feasible proportions we selected – combinations of NACE 5 digit manufacturing industries and destinations where we observe at least 1 export starter over the period 2000-2005 • 176 destinations in the estim
Slide 27: Decision to start exporting - destination
Decision to start exporting - destination (1) (2) (3) (4) (5) (6) NACE industry aggregation level 2-digit 3-digit 4-digit 2-digit 3-digit 4-digit lagged TFP (index, in logs) 0.873*** 0.885*** 0.891*** 0.820*** 0.820*** 0.828*** [0.057] [0.056] [0.056] [0.058] [0.058] [0.058] border-dummy -0.060 -0.0
Slide 28: Decision to start exporting - destination
Decision to start exporting - destination • lagged tfp and lagged # other destinations are important • positive and significant spillovers from exporters in – same region-same industry • especially other starters – same region-other industries • established exporters – other regions-same industry •
Slide 29: Decision to start exporting product-destination
Decision to start exporting product-destination • Restrict sample to computer-power feasible proportions: – combinations of destinations and 4 digits product categories that were exported by at least ten firms on average over the period 1998-2005 – then further reduced the sample by selecting combin
Slide 30: Decision to start exporting - product-destination
Decision to start exporting - product-destination (1) (2) (3) (4) (5) (6) NACE industry aggregation level 2-digit 3-digit 4-digit 2-digit 3-digit 4-digit lagged TFP-level (index, in logs) 0.814*** 0.814*** 0.813*** 0.814*** 0.810*** 0.812*** [0.034] [0.034] [0.034] [0.034] [0.035] [0.035] border-dum
Slide 31: Decision to start exporting product-destination
Decision to start exporting product-destination • lagged tfp and border dummy are significant • lagged # other destination-same product and same destination-other product are important – importance of within-firm learning from past export experience with other products and other destination markets
Slide 32: Conclusions
Conclusions • spillovers from the internationalisation behaviour of other firms in Belgian manufacturing in 1998-2005 on the decision to start exporting – Two possible channels • firm-level productivity • perceived threshold – Economic and geographic distance • in the same region • in the same, upst
Slide 33: Conclusions
Conclusions • productivity spillovers – increase with • presence of exporting firms in the same NUTS 3 region • supplying MNEs – indications of negative within-industry spillover effects • threshold spillovers on the decision to start exporting – firm-level: merely driven by productivity, not by spi
Slide 34: To do
To do • Put two and two together – How do the channels relate? • Marginal effects-type exercise
National Bank of Belgium — Technical Presentation (2010) on 6ix