Determinants of R&D Investment in Emerging Markets: The Role of Ownership and Financial Development

Transcription

1 Philipps-Universität Marburg School of Business and Economics Accounting and Finance Area Management Accounting Research Group Determinants of R&D Investment in Emerging Markets: The Role of Ownership and Financial Development Marc Steffen Rapp a,b,c and Iuliia A. Udoieva a a Accounting and Finance Area, School of Business and Economics, Philipps-Universität Marburg b Marburg Centre for Institutional Economics (MACIE) c Center for Corporate Governance, HHL Leipzig Graduate School of Management Leipzig, September 26 th, 2015 We appreciate financial support from Fritz Thyssen Foundation and the Center for Corporate HHL Das elektronische und mechanische Vervielfältigen, Aufzeichnen und Speichern der vorliegenden Präsentation als Ganzes oder in Teilen sowie die Weitergabe an Dritte bedarf der vorheriger expliziten Zustimmung des Autors. No parts of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form or any means electronic, mechanical, photocopying, recording, or otherwise without the explicit permission of the author. Copyright 2015 by Marc Steffen Rapp. All rights reserved.

2 Tuesday, September 29, 2015 Corporate Governance in Emerging Markets: How to gain competitive advantage? A Corporate Governance Competitive Advantage B Corporate Governance Business model Competitive Advantage - 1 -

3 Importance of R&D investment is widely acknowledged in the business sector all over the world Tuesday, September 29, 2015 Percentage of CEOs considering R&D and innovation among top three investment priorities over the next 12 months 27% 31% 38% 30% 17% 44% 42% 47% 43% 31% 48% 34% 7% 22% 44% 17% 48% 40% 25% 36% 21% 29% 18% 16% 15% All CEOs Asia Pacific Western Europe North America Latin America Africa CEE Source: Own representation based on PwC 16 th Annual Global CEO Survey (2013). The sample used for the survey comprises 1,330 CEO respondents from 68 countries worldwide - 2 -

4 2012* 1998* Over last decades, emerging market economies have excelled some advanced economies in the level of business enterprise R&D Tuesday, September 29, 2015 Business enterprise R&D (BERD) as % of GDP Top-5 R&D intensive economies BERD (% GDP) Sweden 2,5 Japan 2,1 Israel 2,0 [ ] ( ] ( ] ( ] (>2.0] No data Finland United States 1,9 1,8 BERD (% GDP) Israel 3,5 South Korea 3,1 Japan 2,6 [ ] ( ] ( ] ( ] (>2.0] No data * Or nearest year available Source: Own representation based on OECD Research and Development Statistics and IMD World Competitiveness Yearbook (based on data from UNESCO and national sources) Finland Taiwan 2,4 2,3-3 -

5 The growth potential of emerging markets manifested itself in escalating levels of R&D investment, both in absolute and relative terms Tuesday, September 29, 2015 Total R&D investment by public listed firms (in inflation-adjusted 2010 USD, billions) Share of emerging market firms among Top 1300 R&D investors (in absolute and relative terms) OECD +24% +35% US +38% +20% JP +19% +99% Share in the Top 1300 list Number of Top R&D-investing firms EU15 EM % +1% % +319% Source: Own calculation based on Thomson Reuters Datastream Source: Own calculation based on the EU Industrial R&D Investment Scoreboard - 4 -

6 Tuesday, September 29, 2015 Recently, many new brands emerged from emerging markets The distribution of Top R&D-spending companies among emerging market economies (based on the ranking of Top 1300 world R&D-spenders in 2012) R&D investment, M EUR Number of Top R&D-spending firms China Taiwan South Korea India Israel Brazil Hong Kong Russia Singapore Turkey Hungary Mexico South Africa Thailand Venezuela Source: Own representation based on the EU Industrial R&D Investment Scoreboard - 5 -

7 Tuesday, September 29, 2015 Agenda Motivation Sample description Results Robustness

8 Tuesday, September 29, 2015 Agenda Motivation Sample description Results Robustness

9 Taking a closer look at the data, we find an interesting association between R&D intensity and ownership concentration Tuesday, September 29, 2015 Industry: Business Software Industry: Electronics 38-16% % % % % % % 31-37% Block RD/Sales Block RD/Sales Block RD/Sales Block RD/Sales

10 Summary: We examine corporate R&D intensity in emerging markets by analyzing ownership concentration and financial development Tuesday, September 29, ) Over , ownership concentration has been significantly negatively associated with the level of R&D investment in 24 emerging market economies What we find 2) Simultaneously, R&D intensity is increasing in the level of financial development 3) Finally, an economy s financial structure moderates the ownership effect: Widely-held firms engage more in RnD, whenever then economy has relatively developed financial institutions. Contribution to literature We use an extensive panel data set from emerging markets to analyze R&D investment behavior Combine country- and firm-level perspective Apply various econometric methods to deal with selection and endogeneity issues Contribution to public debate Provide evidence on relationship between R&D and ownership concentration conditional on financial development Highlights importance of risk allocation opportunities for decisions which involve uncertainty, in particular, financing of R&D activities - 9 -

11 Tuesday, September 29, 2015 Agenda Motivation Sample description Results Robustness

12 Tuesday, September 29, 2015 To start with, we define and identify emerging markets (EM) Definition Instead of being peripheral, as they have been since the first Industrial Revolution, key economies of the former Third World will soon reemerge as the dominant economies of the future. (A.W.van Agtmael, 2007) A.W.van Agtmael Dr. V. Kvint [a]n emerging market country can be defined as a society transitioning from a dictatorship to a free marketoriented economy, with increasing economic freedom, gradual integration within the global marketplace, an expanding middle class, improving standards of living and social stability and tolerance, as well as an increase in cooperation with multilateral institutions. (Dr. V. Kvint, 2008) Methodology Advantages: Classification is based on a set of comprehensive criteria with respect to economic development, size, liquidity and accessibility of markets Persistency over time Correspondence to other organizations and recent studies (Braga-Alves & Morey et al. (2012), Kearney (2012), Ibragimov et al. (2013), Jaslowitzer & Rapp (2014) to name but a few) EM ASIA EM LATAM EM EMEA China Hong Kong India Indonesia Malaysia Pakistan Philippines Singapore South Korea Sri Lanka Taiwan Thailand Argentina Brazil Chile Colombia Mexico Peru Venezuela Czech Republic Egypt Hungary Israel Jordan Morocco Poland Russia South Africa Turkey MSCI 1997 MSCI 2012 The Economist Sample Dow Jones BBVA IMF BRICS + Next Eleven FTSE S&P Columbia University EMGP

13 As a result of data cleaning process, the final sample includes 24 emerging market economies over Tuesday, September 29, 2015 Sample construction Data cleaning 1 Sample Countries: 24 emerging market economies Period: Firms % Firm-year observations Initial sample of emerging markets 14, ,354 - missing SIC code 1 0.0% 2 0.0% % Firms: non-financial non-utility firms - negative total assets / net sales % % - negative R&D 0 0.0% 4 0.0% 2 3 Data sources Selection criteria Thomson Reuters Datastream: accounting data Thomson One Banker: ownership data IMF, World Bank: financial indicators Identified industry affiliation Non-negative total assets Non-negative net sales Non-negative R&D No bankruptcy Available accounting information At least 2 consecutive years of information - bankruptcy 2 0.0% % - missing accounting data (except for R&D) 1, % 21, % - less than 2 consecutive years of information 1, % 3, % Final sample 11, % 75, % - missing R&D 8, % 45, % Final sample with reported R&D 7, % 29, % EM 35-33% OECD 52 JP 59 US 71 R&D reporting, % Selection => Heckman!

14 The 24 emerging market economies represent a quite heterogeneous set, in terms of R&D intensity, ownership concentration and financial markets development Tuesday, September 29, 2015 Definitions of key variables Sample composition by country R&D investment: R&D expense deflated by net sales Block: total control stake held by large investors (the block threshold is set at 5%) Financial development index (FD): an aggregate index based on the measures of depth, activity and efficiency of financial institutions (FI) and financial markets (FM). Financial institutions index (FI): an aggregate index of financial institutions development based upon: amount of private-sector credit, mutual/ pension fund assets, and insurance premiums (depth); number of commercial banks branches and ATMs per 100,000 adults (access); net operating margin, lending-deposits spreads, noninterest income, overhead, ROA, and ROE (efficiency) Financial markets index (FM): an aggregate index of financial markets development composed of: stock market capitalization, value traded, and debt securities statistics (depth); market capitalization outside of top 10 largest firms, and total number of issuers of debt (access); stock market turnover ratio (efficiency) Economy Firms Firm-years % R&D/Sales Block FD FI FM Brazil 217 1, % Chile 131 1, % China 1,319 7, % Colombia % Czech Republic % Egypt % Hong Kong 988 8, % Hungary % India 1,915 10, % Indonesia 287 1, % Israel 272 1, % Malaysia 891 6, % Mexico % Morocco % Peru % Philippines % Poland 290 1, % Russian Federation % Singapore 577 4, % South Africa 308 2, % South Korea 1,455 9, % Taiwan 1,335 11, % Thailand 425 2, % Turkey 135 1, % Total 11,095 75, % Min % Max 1,915 11, % Median 245 1, %

15 The distribution of R&D reveals a censoring problem, while correlation matrix indicates no concern Tuesday, September 29, 2015 N Mean P50 Sd Min P5 P25 P75 P95 Max Dependent variable R&D/Sales 75, Explanatory variables Control variables Exclusion criteria H1 H2 H3 H4 Block 75, FD 63, FI 63, FM 63, Cash flow 75, Sales growth 75, Size 75, Leverage 75, Tangibility 75, Business R&D personnel 74, ,017 80, , ,294 11, ,777 1,647,454 2,486,400 Industry R&D/ Value Added 67, (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Pearson correlation matrix R&D/Sales (1) Block (2) FD (3) FI (4) FM (5) Cash flow (6) Sales growth (7) Size (8) Leverage (9) Tangibility (10) Business R&D personnel (11) Industry R&D/ Value Added (12) Note (1): Data on financial development indices is not available for Taiwan. Note (2): In analysis that follows, Business R&D personnel is logarithmized

16 Tuesday, September 29, 2015 Agenda Motivation Sample description Results Robustness

17 Tuesday, September 29, 2015 The empirical analysis applies various econometric techniques Estimators Censoring Tobit Selection Heckman two-step estimator Persistent R&D time series OLS with the lag of dependent variable Unobserved heterogeneity Endogeneity System GMM Reverse causality

18 To start with, we document that ownership concentration is significantly negatively associated with R&D investment Tuesday, September 29, 2015 Dataset Type Panel data Correlation analysis Beyond correlation analysis (1) (2) (3) (4) (5) (6) Model Specification Dep. Variable: R&D expenditures deflated by net sales Expl. Variable: ownership concentration Method: Tobit, Heckman 2-step, OLS with the lag of R&D, S-GMM SE: (i) robust, clustered by firm, (ii) robust, with Windmeijer correction Findings Ceteris paribus, ownership concentration in emerging markets is negatively associated with R&D investment (support for H1b) The difference in R&D intensities of a widely-held firm and a firm with 50% block ownership amounts to 0.5% Tobit Tobit Heckman OLS S-GMM missing R&D set to 0 missing R&D dropped 2nd step 1st step 1st lag of R&D 1st lag of R&D R&D/Sales R&D/Sales R&D/Sales R&D positive R&D/Sales R&D/Sales Block, t *** *** *** *** *** ** (-6.47) (-6.40) (-8.29) (-5.84) (-2.80) (-2.17) Cash flow, t *** *** 0.506*** 0.011*** (0.35) (-2.76) (-6.48) (7.16) (6.12) (1.55) Sales growth, t *** *** *** *** (-2.63) (-3.86) (-4.92) (-0.65) (3.25) (-0.76) Size, t *** *** *** 0.137*** *** ** (4.29) (-3.83) (-14.04) (29.63) (-2.82) (-1.98) Leverage, t *** *** *** *** (-5.83) (-3.97) (-4.45) (-9.63) (-0.08) (0.72) Tangibility, t *** *** *** *** *** (-6.52) (-5.03) (-13.43) (-6.73) (-7.12) (-0.57) BRD personnel, t *** (14.68) Industry RD, t *** (27.71) IMR 0.004** (2.09) Dependent, t *** 0.740*** (113.12) (19.49) Year FE YES YES YES YES YES YES Country FE YES YES YES YES YES YES Industry FE YES YES YES YES YES YES Observations 63,000 21,708 20,495 62,415 21,708 17,187 Firms 11,095 4,416 4,450 10,927 4,416 2,362 Hansen p-value M2 p-value R-squared

19 In contrast, higher level of financial development is associated with higher R&D investment in an average emerging market economy Tuesday, September 29, 2015 Dataset Type Panel data Correlation analysis Beyond correlation analysis (1) (2) (3) (4) (5) (6) Model Specification Dep. Variable: R&D expenditures deflated by net sales Expl. Variable: financial development index (FD), centered around a sample mean Method: Tobit, Heckman 2-step, OLS with the lag of R&D, S-GMM SE: (i) robust, clustered by firm, (ii) robust, with Windmeijer correction Findings Ceteris paribus, financial development in emerging markets is positively associated with the level of corporate R&D investment (support for H2) If financial development of Colombia would be the same as in South Korea, the average R&D investment will rise by 1 % Tobit Tobit Heckman OLS S-GMM missing R&D set to 0 missing R&D dropped 2nd step 1st step 1st lag of R&D 1st lag of R&D R&D/Sales R&D/Sales R&D/Sales R&D positive R&D/Sales R&D/Sales Block, t *** ** ** *** ** *** (-4.97) (-2.34) (-2.15) (-7.69) (-2.10) (-3.23) FD, t *** 0.051*** 0.020*** 2.360*** 0.005*** 0.015* (15.45) (9.12) (3.58) (32.83) (5.45) (1.91) Cash flow, t *** *** 0.417*** 0.007*** 0.007* (-1.30) (-5.67) (-13.20) (5.81) (3.00) (1.71) Sales growth, t *** (-0.51) (0.29) (0.55) (-1.58) (3.69) (-0.14) Size, t *** 0.001* *** 0.134*** (8.82) (1.81) (-5.44) (30.16) (-0.50) (-0.03) Leverage, t ** *** *** ** (-2.48) (-4.96) (-3.28) (-2.37) (-0.71) (-0.28) Tangibility, t *** *** *** *** *** (-4.63) (-4.75) (-13.44) (-3.93) (-5.05) (0.58) BRD personnel, t *** (13.93) Industry RD, t *** (22.22) IMR 0.007*** (3.08) Dependent, t *** 0.734*** (72.38) (13.00) Year FE YES YES YES YES YES YES Country FE NO NO NO NO NO NO Industry FE YES YES YES YES YES YES Observations 52,686 14,336 12,997 52,101 14,336 10,343 Firms 9,760 3,357 3,383 9,592 3,357 1,521 Hansen p-value M2 p-value R-squared

20 Meanwhile, financial development is a significant moderator of the relationship between ownership concentration and R&D investment Tuesday, September 29, 2015 Dataset Type Panel data Correlation analysis Beyond correlation analysis (1) (2) (3) (4) (5) (6) Model Specification Dep. Variable: R&D expenditures deflated by net sales Expl. Variable (1): FD index Expl. Variable (2): interaction term between Block and FD index Method: Tobit, Heckman 2-step, OLS with the lag of R&D, S-GMM SE: (i) robust, clustered by firm, (ii) robust, with Windmeijer correction Findings Ceteris paribus, financial market development moderates the negative association between ownership concentration and R&D investment The negative effect of block ownership is 3-5 times larger in environment with high financial development (support for H3b) Tobit Tobit Heckman OLS S-GMM missing R&D set to 0 missing R&D dropped 2nd step 1st step 1st lag of R&D 1st lag of R&D R&D/Sales R&D/Sales R&D/Sales R&D positive R&D/Sales R&D/Sales Block, t *** *** *** *** *** *** (-5.94) (-6.18) (-7.48) (-5.18) (-3.34) (-2.62) Block, t-1 * FD, t *** * * (0.06) (-1.34) (-3.59) (0.37) (-1.68) (-1.89) Cash flow, t * *** *** 0.233*** 0.007*** (-1.94) (-4.84) (-11.40) (2.99) (3.25) (1.53) Sales growth, t *** (0.38) (0.06) (0.17) (-0.09) (3.54) (-0.22) Size, t *** *** 0.170*** (8.60) (0.67) (-4.99) (33.62) (-1.41) (-0.08) Leverage, t *** *** *** *** (-6.08) (-4.11) (-4.90) (-10.24) (-0.46) (0.08) Tangibility, t *** *** *** *** *** (-5.25) (-3.31) (-9.37) (-6.09) (-4.29) (0.74) BRD personnel, t *** (15.41) Industry RD, t *** (22.27) IMR 0.006*** (3.05) Dependent, t *** 0.700*** (66.22) (12.19) Year FE YES YES YES YES YES YES Country FE YES YES YES YES YES YES Industry FE YES YES YES YES YES YES Observations 52,686 14,336 12,977 52,101 14,336 10,343 Firms 9,760 3,357 3,383 9,592 3,357 1,521 Hansen p-value M2 p-value R-squared

21 The moderating role of financial development for R&D investment in emerging markets is driven by development of financial institutions Tuesday, September 29, 2015 Dataset Type Panel data Correlation analysis Beyond correlation analysis (1) (2) (3) (4) (5) (6) Model Specification Dep. Variable: R&D expenditures deflated by net sales Expl. Variable (1): FI index Expl. Variable (2): interaction term between Block and FI index Method: Tobit, Heckman 2-step, OLS with the lag of R&D, S-GMM SE: (i) robust, clustered by firm, (ii) robust, with Windmeijer correction Findings In line with the previous findings, we document a significant negative effect of concentrated ownership on R&D investment that increases in countries with relatively more developed financial institutions (support for H4b) Tobit Tobit Heckman OLS S-GMM missing R&D set to 0 missing R&D dropped 2nd step 1st step 1st lag of R&D 1st lag of R&D R&D/Sales R&D/Sales R&D/Sales R&D positive R&D/Sales R&D/Sales Block, t *** *** *** *** *** *** (-6.39) (-6.16) (-8.42) (-5.70) (-3.44) (-2.87) Block, t-1 * FI, t * ** *** *** *** (-1.72) (-2.42) (-6.72) (-0.65) (-2.69) (-3.51) Cash flow, t * *** *** 0.249*** 0.008*** (-1.74) (-4.82) (-11.19) (3.18) (3.30) (1.43) Sales growth, t *** (0.27) (0.16) (0.17) (-0.13) (3.58) (-0.40) Size, t *** *** 0.171*** (8.77) (0.63) (-5.39) (33.87) (-1.44) (-0.27) Leverage, t *** *** *** *** (-6.07) (-4.05) (-4.57) (-10.16) (-0.38) (-0.22) Tangibility, t *** *** *** *** *** (-5.30) (-3.32) (-9.35) (-6.19) (-4.35) (0.72) BRD personnel, t *** (14.92) Industry RD, t *** (22.31) IMR 0.005*** (2.58) Dependent, t *** 0.699*** (66.11) (12.13) Year FE YES YES YES YES YES YES Country FE YES YES YES YES YES YES Industry FE YES YES YES YES YES YES Observations 52,686 14,336 12,977 52,101 14,336 10,343 Firms 9,760 3,357 3,383 9,592 3,357 1,521 Hansen p-value M2 p-value R-squared

22 Tuesday, September 29, 2015 While analyzing economic significance, we observe further compelling findings Economic significance Illustration Ownership mean 37% SD 28% mean - SD mean mean + SD Ownership 10% 37% 65% Panel A: Financial development Coefficients Financial Model Block Structure Interaction min mean - SD mean mean + SD max Dependent variable (Financial development): GMM Widely held firm (MEAN OWC - SD OWC): -0.38% -0.31% -0.16% -0.01% 0.10% Average firm (MEAN OWC): -0.34% -0.30% -0.20% -0.10% -0.03% Closely held firm (MEAN OWC + SD OWC): -0.31% -0.29% -0.24% -0.19% -0.16% Panel B: Financial institutions Coefficients Financial Model Block Structure Interaction min mean - SD mean mean + SD max Dependent variable (Financial institutions): GMM Widely held firm (MEAN OWC - SD OWC): -0.38% -0.32% -0.15% 0.02% 0.16% Average firm (MEAN OWC): -0.31% -0.28% -0.20% -0.12% -0.06% Closely held firm (MEAN OWC + SD OWC): -0.24% -0.24% -0.25% -0.27% -0.27% 1 2 Financial Development 0% 0% 0% % % 0% 0% -1% Widely held firm (MEAN OWC - SD OWC): Average firm (MEAN OWC): Closely held firm (MEAN OWC + SD OWC): Financial Institutions 0% 0% 0% % % 0% 0% -1% Widely held firm (MEAN OWC - SD OWC): Average firm (MEAN OWC): Closely held firm (MEAN OWC + SD OWC):

23 Tuesday, September 29, 2015 Agenda Motivation Sample description Results Robustness

24 Tuesday, September 29, 2015 To examine robustness we perform a battery of alternative tests Extensions Robustness Tests Results 1 Considering owner s type Strategic owners Institutional owners (0) 2 Alternative proxies for ownership concentration An aggregate share of three largest shareholders An aggregate stock of all shareholders with 10% and more 3 Alternative proxies for financial development The level of FI development relative to the overall FD The level of FI development relative to FM development 4 Sampling Random sampling for firms from South Korea and India Exclusion of transition economies

25 Tuesday, September 29, 2015 Summary 1) Over , ownership concentration has been significantly negatively associated with the level of R&D investment in 24 emerging market economies What we find 2) Simultaneously, R&D intensity is increasing in the level of financial development 3) Finally, an economy s financial structure moderates the ownership effect: Widely-held firms engage more in RnD, whenever then economy has relatively developed financial institutions. Contribution to literature We use an extensive panel data set from emerging markets to analyze R&D investment behavior Combine country- and firm-level perspective Apply various econometric methods to deal with selection and endogeneity issues Contribution to public debate Provide evidence on relationship between R&D and ownership concentration conditional on financial development Highlights importance of risk allocation opportunities for decisions which involve uncertainty, in particular, financing of R&D activities

26 Tuesday, September 29, 2015 Thank you for your attention!

27 Tuesday, September 29, 2015 Backup: R&D in Emerging Markets

28 Among determinants of R&D, ownership plays a special role, although the nature of relationship still remains a puzzle Tuesday, September 29, 2015 Ownership Innovation Ownership Innovation Cost of capital (financing constraints, insufficient diversification) Agency II conflict Managerial Entrenchment Negative impact on internal and external corporate governance mechanisms (managerial labor market) Incentive Alignement Agent-principal interest alignment Substitution for weak legal and institutional environment Signaling of commitment to bailout and limit expropriation Concentration Strategic Institutional Concentration Strategic Institutional Battaggion & Tajoli (2001) Ortega-Argilés et al. (2003) Czarnitzki & Kraft (2004) Zeng & Lin (2011) Seifert&Gonenc (2012) Chen & Hsu (2009) Munari et al. (2010) Mahlich & Yurtoglu (2011) Munoz-Bullon & Sanchez-Bueno (2011) Anderson (2012) Tribo et al. (2007) Ayyagari et al. (2011) Driver & Guedes (2012) Brossard et al. (2013) Lacetera (2001) Hosono et al. (2004) Tribo et al. (2007) Ayyagari et al. (2011) Choi & Lee (2011) Zeng & Lin (2011) Lee (2012) Lodh et a. (2014) Wahal & McConnell (2000) Eng & Shackell (2001) Choi & Lee (2011) Aghion et al. (2013) Brossard et al. (2013) López -Iturriaga & López -Millán (2015) Source: Own elaboration based on Wang & Shailer (2015)

29 Herewith, we develop four hypotheses, which we aim to tackle with various econometric techniques Tuesday, September 29, 2015 Hypotheses Estimators Ownership concentration H1a: + H1b: - R&D investment Censoring Tobit H3a: + H3b: - H2: + Financial development Selection Heckman two-step estimator H4a: + H4b: - Financial institutions Persistent R&D time series OLS with the lag of dependent variable Unobserved heterogeneity Following the control rights theory and pecking order, large block holders choose debt as the primary source of external funding, in particular, if R&D investment is incremental (Aghion et al., 2004) While financial institutions grow in size and depth and private credit becomes more accessible, the threat to stability raises, what, in turn, makes undiversified shareholders more risk-averse (IMF, 2015) Endogeneity Reverse causality System GMM