(Luận văn) ownership structure, audit quality and information disclosure, an approach at firm level in vietnam

t to UNIVERSITY OF ECONOMICS HO CHI MINH CITY VIETNAM ng INSTITUTE OF SOCIAL STUDIES THE HAGUE THE NETHERLANDS hi ep w n lo ad VIETNAM - NETHERLANDS PROGRAM FOR M.A IN DEVELOPMENT ECONOMICS ju y th yi pl ua al n OWNERSHIP STRUCTURE, AUDIT QUALITY AND INFORMATION DISCLOSURE: AN APPROACH AT FIRM LEVEL IN VIETNAM n va ll fu oi m at nh z z jm ht vb By k QUACH MANH HUNG om l.c gm n a Lu MASTER OF ARTS IN DEVELOPMENT ECONOMICS n va y te re th Ho Chi Minh City, March 2014 t to ng ABSTRACT hi ep This paper investigates the relationship between the ownership concentration, audit quality and the amount of firm-specific information captured in the stock price w n The study covers 195 Vietnamese listed firms on the Hochiminh stock market from lo ad 2006 to 2011 The findings confirm the positive relationship between the ownership of ju y th the largest shareholder and the stock price informativeness of that firm Moreover, this relationship is not significant when the largest owner is related to the state Further, the yi pl largest owner would like to disclose more information to the market if his asset, tied al ua with the stock price, is large The reason is that the asset value could be reduced when n the investors punish the owner bad behavior by discounting the stock price In addition, va n the results also show a positive impact of audit quality on the disclosure of firm fu ll specific information Moreover, a trend of better information environment for m oi information disclosure is also discovered Finally, this research also finds out that the at nh ownership concentration of top largest shareholders also has a similar positive impact z on the corporate disclosure of firm-specific information z vb k jm ht Key words: ownership concentration, state joint stock firms, audit, disclosure om l.c gm n a Lu n va y te re th i t to ACKNOWLEDGMENT ng hi Foremost, I would like to sincerely and gratefully thank Dr Pham Thi Bich ep Ngoc, my enthusiastic supervisor, for her great support, crucial advice and precious w n guidance during my thesis completion lo ad Besides, I would like to thank Dr Truong Dang Thuy who also gave me useful y th ju assistance for my thesis More special thanks go to the Vietnam – Netherlands yi pl program, especially professors, staffs and classmates, for their helps, encouragement al n ua and wonderful knowledge n va Last but not least, I would like to thank my family for their sacrifices for fu ll supporting me in not only the thesis writing but also my whole life oi m at nh z z k jm ht vb om l.c gm n a Lu n va y te re th ii t to ng hi TABLE OF CONTENTS ep ABSTRACT .i w ACKNOWLEDGMENT ii n lo TABLE OF CONTENTS iii ad LIST OF TABLES v y th LIST OF FIGURES vi ju INTRODUCTION yi Problem statement .1 1.2 Research objectives 1.3 Research questions 1.4 Data and methodology 1.5 Thesis structure pl 1.1 n ua al n va ll fu LITERATURE REVIEW Some concepts nh 2.1 oi m Corporate disclosure 2.1.2 Ownership concentration 2.1.3 Audit quality 2.1.4 State joint stock firm at 2.1.1 z z jm ht vb Theoretical literature k 2.2 gm Theory of informed trading, asymmetric information and corporate disclosure 2.2.2 The relationship between ownership concentration and the corporate disclosure 13 2.2.3 The relationship between government ownership and the corporate disclosure 15 2.2.4 The impact of audit quality on the corporate disclosure 16 om a Lu Empirical literature 18 n 2.3 l.c 2.2.1 va Ownership concentration and the information disclosure 18 2.3.2 Government ownership and the information disclosure 22 2.3.3 Audit quality on the information disclosure 22 th iii y Hypothesis construction 23 te re 2.4 n 2.3.1 t to ng hi 2.4.1 Ownership concentration and information disclosure 23 2.4.2 Auditor quality and information disclosure 23 2.4.3 Public firm and information disclosure 24 ep DATA AND METHODOLOGY 25 Data source 25 w 3.1 n lo 3.2 Model specification 25 ad 3.2.1 Constructing the dependent variable 28 y th 3.2.2 ju 3.3 Models 31 Estimation strategy and correction method 35 yi Ordinaly least squares 35 3.3.2 Fixed effects model (FE) 36 3.3.3 Random effects model (RE) 36 3.3.4 Pooled OLS, Fixed effects or Random effects 37 pl 3.3.1 n ua al n va OVERVIEW, STATISTICS SUMMARY, EMPIRICAL RESULTS AND ANALYSIS 40 ll fu Overview: 40 4.2 Statistics summary 41 4.3 Empirical results and analysis: 50 oi m 4.1 at nh Ownership concentration, audit quality and corporate disclosure: 50 4.3.2 State joint stock firms and information disclosure 56 4.3.3 Robustness check 61 z 4.3.1 z jm ht vb CONCLUSION 71 k gm Main findings 71 5.2 Policy implications 72 5.3 Limitations and further researches 73 om l.c 5.1 a Lu REFERENCE 76 n APPENDIX 79 Firms and industry category summarize 79 B Regression results 80 n va A y te re th iv t to LIST OF TABLES ng Table 2.1: Empirical researches about ownership concentration and corporate disclosure 21 hi ep Table 2.2: Empirical researches about government ownership and corporate disclosure 22 Table 2.3: Empirical researches about the audit quality and corporate disclosure 23 w Table 3.1: Coefficients and expected sign 35 n lo Table 4.1: Descriptive statistic 41 ad Table 4.2: Correlation table of variables 49 y th Table 4.3: VIF results 49 ju yi Table 4.4: Ownership concentration, audit quality and corporate disclosure 52 pl Table 4.5: State joint stock firms and corporate disclosure 58 al n ua Table 4.6: Weighted regression by FE and RE model 63 Table 4.7: Regression results of Independent variable HSUM with FE and RE model 67 va n Table 4.8: Regression results of Independent variable H with FE and RE model 69 ll fu oi m at nh z z k jm ht vb om l.c gm n a Lu n va y te re th v t to LIST OF FIGURES ng Figure 2.1: Capital and information flows in the financial market economy hi ep Figure 3.1: Empirical study stages 27 Figure 3.2: Diagram of the comparison progress between FE & RE models 38 w n Figure 4.1: Informativeness level by year 43 lo ad Figure 4.2: Ownership concentration by year 44 y th Figure 4.3: Fitted value of Corporate disclosure against Ownership concentration 45 ju Figure 4.4: Fitted of INFO against TOPHOLD of State joint stock firms 46 yi pl Figure 4.5: Fitted value of INFO against TOPHOLD of Private joint stock firms 48 n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm n a Lu n va y te re th vi t to ng hi INTRODUCTION ep 1.1 Problem statement w n From the first trading day at 28/07/2000 with only two listed stocks, the lo ad Hochiminh stock market just have its 13th birthday with 302 listed stocks, funds and y th 28 bonds With the capitalizing value around 40 billion of USD (hsx.vn), the largest ju yi stock market in Vietnam only accounts 23,5% of the GDP (170 billion of USD) pl Hence, there still a large room for developing However, although much effort of the al n ua Vietnamese state and firms in improving the domestic investment environment to make va it more attractive to investors, both foreign and domestic, it is not an easy task One n major obstacle for the development of the stock market is the Vietnamese fu ll informativeness environment especially the corporate disclosure practices m oi In fact, the stock market is always considered the best way to forecast the nh economic performance The main reason is that the stock market reflects much at z economic information through its stock pricing system Not only the macro events but z ht vb also the firm specific information is captured in the market movements However, the jm reflection of those firm specific information is not always good because the corporate k disclosure problem Bad disclosure blocks the information to be captured in the stock gm price and creates the asymmetric information between the insiders and uninformed om l.c outside investors Due to this market failure, the cost for acquiring information becomes higher for the outside investor (Jiang, et al, 2011) The more severe the a Lu situation is, the less attractive the investment environment is Thus, the fund attraction n va progress for the capital market in order to lower the domestic cost of capital could be in n vain if the corporate disclosure problem is not resolved (Jiang, et al, 2011; Lawrence, y te re 2013) Hence, analyzing this problem is important for Vietnam situation when we are th in the urge of attracting capital flow for developing the domestic financial market t to According to Morck, Yeung, and Yu (2000), low level of informativeness is a ng hi common phenomenon in emerging financial markets The problem is the consequence ep of many serious corporate and institution structure problems Firstly, according to Ball w (2001); Chan and Hameed (2006) the bad enforcement of disclosure regulations in n lo emerging markets makes the insiders and managers have less incentive to release more ad information for the outside investors Secondly, according to Gul, Kim and Qiu (2010), y th ju the corporate structure is considered as a main cause for the problem In emerging yi economies, the ownership concentrates in hands of the entrepreneurs and their pl ua al relatives Meanwhile, in the case of transitions economies, the largest shareholders are related to the state The high ownership concentration in a small number and powerful n n va owners certainly has a large influence on the disclosure policy of the firms ll fu In this study, the impact of the ownership structure on the disclosure activities is oi m the main research objective The first argument behind this suggestion is based on the agency problem mentioned by Jensen and Meckling (1976), which states that the nh at controlling shareholder would hide corporate information from outside investors for z z their private benefit, such as covering their corruption or inside trading Therefore, the ht vb ownership concentration, especially state ownership, in Vietnam doesn’t favorite the k jm information disclosure system Hence, the ownership level would have a negative gm relationship with the corporate disclosure level However, there could be another argument, by Healy and Palepu (2001), that the high holding rate of a firm stock is a l.c om strong commitment for the owner faith The investors could easily punish his bad a Lu behavior by discounting his asset easily and directly on the stock market Thus, the effect of ownership concentration on the corporate disclosure is important but with n n va mixed results Moreover, Gul, Kim and Qiu (2010) also argue that auditors with much th positive relationship with the information disclosure level y role for the transparency of the market Therefore, good audit quality could have a te re experiences and skills in analyzing financial information will contribute an important t to Many previous studies such as Gul, Kim and Qiu (2010), Fernandes and ng hi Ferreira (2008),… intends to find the relationship between ownership concentration, ep audit quality and the information disclosure However, such empirical researches had w never been performed in Vietnam Therefore, by analyzing the firm-specific return n lo variation and its relationship with ownership concentration and audit quality, this study ad aims to shed a light to the situation of transparency of Vietnam stock market By y th ju drawing the results, this study aims to contribute a better understanding about the yi importance information disclosure and ownership concentration in the case of pl ua al emerging country like Vietnam Then some findings about the Vietnamese corporate governance structure and governance could be drawn out A contribution to the debate n n va between managerial entrenchment and incentive alignment of large shareholder is also ll fu expected A better glance about the audit performance could also be acquired at the end oi m of this study Research objectives at nh 1.2 z Based on all the important points mentioned earlier, this study’s main objectives z vb are to analyze: jm ht - The relationship between ownership concentration and the information disclosure of Vietnamese firms k gm - The difference disclosure quality between state ownership and private - The impact of audit quality on the stock price informativeness om l.c ownership n a Lu Research questions n va 1.3 y answered: te re For reaching those objectives, the followings research questions must be th t to Table 56: Hausman test result ng hi Coefficients (b) (B) fe re ep w HSUM SIZE INDSIZE VOL 2007bn.year 2008.year 2009.year 2010.year 2011.year n lo ad ju y th 815445 -9.68e-09 -1.22e-11 -2.17e-06 2916031 5294015 8533081 1.703316 1.583907 yi 898652 -8.70e-09 -3.44e-10 -1.44e-06 456165 6855333 1.161037 2.114964 1.997783 (b-B) Difference sqrt(diag(V_b-V_B)) S.E pl 0832069 9.73e-10 -3.32e-10 7.28e-07 1645619 1561318 3077286 4116488 4138756 3862692 6.48e-09 2.55e-10 3.00e-07 137795 1814633 2697972 3542047 390037 al n ua b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg difference in coefficients not systematic n Ho: va Test: fu ll chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 5.89 Prob>chi2 = 0.4352 oi m at nh z z k jm ht vb om l.c gm n a Lu n va y te re th 118 t to Table 57: Fixed effects regression result ng hi ep Fixed-effects (within) regression Group variable: id R-sq: Number of obs Number of groups w within = 0.5277 between = 0.3879 overall = 0.4457 n = = 306 100 Obs per group: = avg = max = 3.1 lo ad F(8,99) Prob > F = 0.0613 ju y th corr(u_i, Xb) = = (Std Err adjusted for 100 clusters in id) yi pl Coef ua al INFO Robust Std Err P>|t| 898652 -8.70e-09 -3.44e-10 -1.44e-06 year 2007 2008 2009 2010 2011 456165 6855333 1.161037 2.114964 1.997783 2471865 3096658 4538388 5786566 6464298 _cons -.0702203 3450618 sigma_u sigma_e rho 70030323 55074052 61786577 (fraction of variance due to u_i) n va 1.99 -1.08 -0.90 -3.64 [95% Conf Interval] HSUM SIZE INDSIZE VOL n 4509632 8.08e-09 3.81e-10 3.96e-07 t 0038431 -2.47e-08 -1.10e-09 -2.23e-06 1.793461 7.33e-09 4.11e-10 -6.57e-07 0.068 0.029 0.012 0.000 0.003 -.0343066 0710891 260522 966784 7151257 9466366 1.299977 2.061551 3.263145 3.28044 -.7548978 6144573 ll fu 0.049 0.284 0.368 0.000 oi m at nh z vb 0.839 k jm ht -0.20 z 1.85 2.21 2.56 3.65 3.09 om l.c gm n a Lu n va y te re th 119 t to Table 58: Random effects regression result ng hi ep Random-effects GLS regression Group variable: id R-sq: Number of obs Number of groups w within = 0.5246 between = 0.6003 overall = 0.5444 n = = 306 100 Obs per group: = avg = max = 3.1 lo ad Wald chi2(15) Prob > chi2 = (assumed) ju y th corr(u_i, X) = = (Std Err adjusted for 100 clusters in id) yi pl Coef ua al INFO Robust Std Err P>|z| 8662997 -1.30e-08 -2.66e-10 -1.85e-06 year 2007 2008 2009 2010 2011 3900496 6706905 1.11998 2.079784 1.9909 1971861 2450927 3540403 442844 4950907 indus 10 11 3831397 354249 -.7882245 -.0207191 4581948 4354034 -.1368309 3999631 4312873 2551421 (omitted) 244547 2716699 2524678 2610414 0.96 0.82 -3.09 0.338 0.411 0.002 -0.08 1.69 1.72 -0.52 _cons -.1440456 3841713 -0.37 sigma_u sigma_e rho 47025049 55074052 42165173 (fraction of variance due to u_i) n va 2.89 -3.53 -0.85 -5.85 [95% Conf Interval] HSUM SIZE INDSIZE VOL n 2997002 3.67e-09 3.14e-10 3.17e-07 z 2788982 -2.01e-08 -8.80e-10 -2.48e-06 1.453701 -5.77e-09 3.49e-10 -1.23e-06 0.048 0.006 0.002 0.000 0.000 0035719 1903176 4260736 1.211826 1.02054 7765273 1.151063 1.813886 2.947742 2.96126 -.4007737 -.4910586 -1.288294 1.167053 1.199557 -.2881551 0.932 0.092 0.085 0.600 -.5000225 -.0742684 -.0594244 -.6484627 4585842 9906579 9302312 3748009 0.708 -.8970076 ll fu 0.004 0.000 0.397 0.000 oi m at nh z z 1.98 2.74 3.16 4.70 4.02 k jm ht vb om l.c gm a Lu 6089163 n n va y te re th 120 t to 12 Robustness check of the independent variable H with the total sample ng hi ep w n Table 59: F-test result lo F(194, 519) = ad F test that all u_i=0: 3.57 Prob > F = 0.0000 ju y th yi pl Table 60: LM test result ua al n Breusch and Pagan Lagrangian multiplier test for random effects va n INFO[id,t] = Xb + u[id] + e[id,t] ll fu oi m Estimated results: 1.000952 3192878 2419721 1.000476 5650555 4919066 at z z vb Var(u) = 98.79 0.0000 k chibar2(01) = Prob > chibar2 = jm ht Test: sd = sqrt(Var) nh INFO e u Var om l.c gm n a Lu n va y te re th 121 t to Table 61: Hausman test result ng hi Coefficients (b) (B) fe re ep (b-B) Difference sqrt(diag(V_b-V_B)) S.E w n H SIZE INDSIZE VOL 2007bn.year 2008.year 2009.year 2010.year 2011.year lo ad ju y th 5097383 -8.82e-09 1.42e-10 -2.20e-06 0202575 1980403 5339979 1.176213 1.216343 yi 1.114545 -1.40e-08 -3.74e-11 -1.46e-06 0643226 2095882 6268287 1.361809 1.407247 pl 6048065 -5.15e-09 -1.80e-10 7.39e-07 0440651 0115479 0928308 1855963 190904 4926333 3.58e-09 1.21e-10 2.04e-07 0354349 059104 0995519 1352844 1504128 ua al n b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg Ho: n va Test: difference in coefficients not systematic ll fu oi m chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 12.00 Prob>chi2 = 0.0621 at nh z z k jm ht vb om l.c gm n a Lu n va y te re th 122 t to Table 62: Fixed effects regression result ng hi ep Number of obs Number of groups = = 723 195 R-sq: Obs per group: = avg = max = 3.7 w Fixed-effects (within) regression Group variable: id n within = 0.4947 between = 0.3784 overall = 0.4105 lo ad y th corr(u_i, Xb) F(8,194) Prob > F = 0.0341 = = ju (Std Err adjusted for 195 clusters in id) yi pl INFO Coef Robust Std Err t P>|t| [95% Conf Interval] ua al 1.114545 -1.40e-08 -3.74e-11 -1.46e-06 year 2007 2008 2009 2010 2011 0643226 2095882 6268287 1.361809 1.407247 1093556 1378714 1939668 2428164 2672518 _cons 4270946 1373722 sigma_u sigma_e rho 64676143 56505554 56711916 n H SIZE INDSIZE VOL n va 4335053 6.18e-09 1.83e-10 3.31e-07 0.011 0.025 0.838 0.000 2595564 -2.62e-08 -3.98e-10 -2.11e-06 1.969533 -1.78e-09 3.23e-10 -8.05e-07 0.59 1.52 3.23 5.61 5.27 0.557 0.130 0.001 0.000 0.000 -.1513559 -.0623311 2442743 8829102 8801545 2800011 4815076 1.009383 1.840708 1.934339 0.002 1561598 6980294 ll fu 2.57 -2.26 -0.20 -4.40 oi m at nh z z 3.11 jm ht vb k (fraction of variance due to u_i) om l.c gm n a Lu n va y te re th 123 t to Table 63: Random effects regression result ng hi ep Random-effects GLS regression Group variable: id R-sq: Number of obs Number of groups w within = 0.4912 between = 0.5096 overall = 0.4785 n = = 723 195 Obs per group: = avg = max = 3.7 lo ad = (assumed) ju y th corr(u_i, X) Wald chi2(16) Prob > chi2 = = (Std Err adjusted for 195 clusters in id) yi pl Coef ua al INFO Robust Std Err P>|z| H SIZE INDSIZE VOL 5590321 -1.12e-08 2.33e-11 -2.02e-06 year 2007 2008 2009 2010 2011 0452182 234053 6235705 1.319256 1.375436 1032975 1295821 1801786 2189657 2417996 indus 10 11 1895199 1188107 -.4079051 0598566 -.063971 6907432 348476 -.1796475 2299976 222947 2189176 2473642 1696671 2084483 198946 1825421 0.82 0.53 -1.86 0.24 -0.38 3.31 1.75 -0.98 0.410 0.594 0.062 0.809 0.706 0.001 0.080 0.325 _cons 4498418 1945851 2.31 0.021 sigma_u sigma_e rho 48214044 56505554 42131509 n 2.15 -3.89 0.14 -6.81 n va 2602148 2.88e-09 1.70e-10 2.97e-07 z [95% Conf Interval] 0490205 -1.69e-08 -3.10e-10 -2.60e-06 1.069044 -5.58e-09 3.57e-10 -1.44e-06 0.662 0.071 0.001 0.000 0.000 -.1572413 -.0199233 2704269 8900911 9015178 2476777 4880292 9767141 1.748421 1.849355 -.261267 -.3181574 -.8369757 -.4249682 -.3965123 282192 -.0414509 -.5374235 6403068 5557789 0211655 5446815 2685704 1.099294 7384029 1781285 ll fu 0.032 0.000 0.891 0.000 oi m at nh z z 0.44 1.81 3.46 6.02 5.69 k jm ht vb om l.c gm a Lu 068462 8312216 n n va y te re (fraction of variance due to u_i) th 124 t to 13 Robustness check of the independent variable H with the state joint stock ng hi firms ep w n Table 64: F-test result lo F(113, 294) = ad F test that all u_i=0: 3.31 Prob > F = 0.0000 ju y th yi pl ua al Table 65: LM test result n Breusch and Pagan Lagrangian multiplier test for random effects va n INFO[id,t] = Xb + u[id] + e[id,t] ll fu Var sd = sqrt(Var) oi m Estimated results: 9560807 3279339 2502616 at z z 9777938 5726551 5002615 Var(u) = 40.67 0.0000 k om l.c gm chibar2(01) = Prob > chibar2 = jm ht vb Test: nh INFO e u n a Lu n va y te re th 125 t to Table 66: Hausman test result ng hi ep Coefficients (b) (B) fe re (b-B) Difference sqrt(diag(V_b-V_B)) S.E w n lo H SIZE INDSIZE VOL 2007bn.year 2008.year 2009.year 2010.year 2011.year ad ju y th 3478207 -8.12e-09 2.22e-10 -2.32e-06 -.0611693 0896008 4510582 9453126 1.124813 yi 1.092615 -1.36e-08 5.72e-11 -1.43e-06 -.0768717 0435191 501152 1.086756 1.268362 pl 1.884636 4.57e-09 1.52e-10 2.78e-07 0558458 0799957 1255628 1632589 1782723 ua al 744794 -5.50e-09 -1.64e-10 8.92e-07 -.0157023 -.0460818 0500937 1414429 1435489 n b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg difference in coefficients not systematic ll fu Ho: n va Test: oi m chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 11.59 Prob>chi2 = 0.0718 at nh z z k jm ht vb om l.c gm n a Lu n va y te re th 126 t to Table 67: Fixed effects regression results ng hi ep Fixed-effects (within) regression Group variable: id R-sq: Number of obs Number of groups w within = 0.4846 between = 0.3596 overall = 0.3876 n = = 417 114 Obs per group: = avg = max = 3.7 lo ad F(8,113) Prob > F = -0.0134 ju y th corr(u_i, Xb) = = (Std Err adjusted for 114 clusters in id) yi pl Coef ua al INFO Robust Std Err 1.092615 -1.36e-08 5.72e-11 -1.43e-06 1.890776 8.88e-09 2.38e-10 4.58e-07 year 2007 2008 2009 2010 2011 -.0768717 0435191 501152 1.086756 1.268362 1304244 1696987 2342988 2882399 3129507 _cons 4356781 4065155 sigma_u sigma_e rho 65282971 57265512 56514393 n H SIZE INDSIZE VOL t P>|t| [95% Conf Interval] n va 0.565 0.128 0.811 0.002 -2.653354 -3.12e-08 -4.15e-10 -2.34e-06 4.838584 3.98e-09 5.29e-10 -5.20e-07 -0.59 0.26 2.14 3.77 4.05 0.557 0.798 0.035 0.000 0.000 -.3352659 -.2926846 0369637 5157003 6483499 1815226 3797227 9653402 1.657811 1.888373 -.3697025 1.241059 ll fu 0.58 -1.53 0.24 -3.12 oi m at nh z z 0.286 k jm ht vb 1.07 (fraction of variance due to u_i) om l.c gm n a Lu n va y te re th 127 t to Table 68: Random effects regression result ng hi ep Random-effects GLS regression Group variable: id R-sq: Number of obs Number of groups w within = 0.4791 between = 0.4742 overall = 0.4535 n = = 417 114 Obs per group: = avg = max = 3.7 lo ad Wald chi2(15) Prob > chi2 = (assumed) ju y th corr(u_i, X) = = (Std Err adjusted for 114 clusters in id) yi Robust Std Err pl Coef ua al INFO z [95% Conf Interval] H SIZE INDSIZE VOL 2424973 -9.21e-09 1.41e-10 -2.21e-06 year 2007 2008 2009 2010 2011 -.0416693 1150614 5092913 1.040978 1.228342 1291698 1607026 2218143 2679945 2939735 indus 10 11 3451194 3621347 0743096 3743605 1885247 5279751 0551435 2867646 2638486 3583612 277295 2246753 (omitted) 2494274 2387183 1.20 1.37 0.21 1.35 0.84 0.229 0.170 0.836 0.177 0.401 2.12 0.23 0.034 0.817 0391063 -.4127359 _cons 2939015 239437 1.23 0.220 -.1753864 sigma_u sigma_e rho 49703672 57265512 42965976 n n va 3907795 4.10e-09 2.21e-10 5.33e-07 P>|z| 0.535 0.025 0.524 0.000 -.5234165 -1.72e-08 -2.92e-10 -3.26e-06 1.008411 -1.18e-09 5.74e-10 -1.17e-06 -0.32 0.72 2.30 3.88 4.18 0.747 0.474 0.022 0.000 0.000 -.2948375 -.1999099 0745433 5157186 6521649 2114988 4300327 9440393 1.566238 1.80452 -.2169289 -.1549991 -.6280653 -.1691277 -.2518308 9071678 8792685 7766846 9178487 6288803 ll fu 0.62 -2.25 0.64 -4.15 oi m at nh z z k jm ht vb l.c gm om 1.016844 5230228 n a Lu 7631893 n va (fraction of variance due to u_i) y te re th 128 t to 14 Robustness check of the independent variable H with the private joint stock ng hi firms ep Table 69: F-test result w n F test that all u_i=0: F(99, 197) = 3.15 Prob > F = 0.0000 lo ad ju y th yi Table 70: LM test result pl Breusch and Pagan Lagrangian multiplier test for random effects ua al INFO[id,t] = Xb + u[id] + e[id,t] n va Estimated results: n Var sd = sqrt(Var) 1.041377 302181 230275 1.020479 54971 4798697 oi m at nh Test: ll fu INFO e u Var(u) = z z chibar2(01) = Prob > chibar2 = k jm ht vb 41.37 0.0000 om l.c gm n a Lu n va y te re th 129 t to Table 71: Hausman test result ng hi Coefficients (b) (B) fe re ep (b-B) Difference sqrt(diag(V_b-V_B)) S.E w n H SIZE INDSIZE VOL 2007bn.year 2008.year 2009.year 2010.year 2011.year lo ad ju y th yi 1.264477 -8.10e-09 -3.35e-10 -1.45e-06 4440623 6677449 1.132576 2.081061 1.968005 pl 9300923 -9.35e-09 -5.35e-12 -2.25e-06 2528539 5015591 8204382 1.656618 1.549754 3343851 1.25e-09 -3.30e-10 7.95e-07 1912084 1661858 3121381 4244429 4182515 5041744 6.46e-09 2.54e-10 3.03e-07 135479 1793219 2679428 35199 3878114 ua al n b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg difference in coefficients not systematic n Ho: va Test: fu ll chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 6.95 Prob>chi2 = 0.3252 oi m at nh z z k jm ht vb om l.c gm n a Lu n va y te re th 130 t to Table 72: Fixed effects regression result ng hi ep Fixed-effects (within) regression Group variable: id R-sq: Number of obs Number of groups w n lo within = 0.5295 between = 0.3704 overall = 0.4336 ad Obs per group: = avg = max = 3.1 = = (Std Err adjusted for 100 clusters in id) yi pl Robust Std Err al INFO 306 100 F(8,99) Prob > F = 0.0502 ju y th corr(u_i, Xb) = = Coef t P>|t| [95% Conf Interval] n ua 1.264477 -8.10e-09 -3.35e-10 -1.45e-06 4493588 7.79e-09 3.68e-10 3.92e-07 year 2007 2008 2009 2010 2011 4440623 6677449 1.132576 2.081061 1.968005 218975 2808576 4197617 5397837 6076391 _cons 1216201 2220908 sigma_u sigma_e rho 70928414 54970996 62474337 2.81 -1.04 -0.91 -3.71 n va H SIZE INDSIZE VOL 372852 -2.36e-08 -1.06e-09 -2.23e-06 2.156103 7.36e-09 3.94e-10 -6.76e-07 0.045 0.019 0.008 0.000 0.002 0095684 1104625 299678 1.010013 7623173 8785563 1.225027 1.965475 3.152109 3.173693 -.3190562 5622963 ll fu 0.006 0.301 0.364 0.000 oi m at nh 2.03 2.38 2.70 3.86 3.24 z z ht vb 0.55 0.585 k jm gm (fraction of variance due to u_i) om l.c n a Lu n va y te re th 131 t to Table 73: Random effects regression result ng hi ep Random-effects GLS regression Group variable: id R-sq: Number of obs Number of groups w within = 0.5252 between = 0.5990 overall = 0.5419 n = = 306 100 Obs per group: = avg = max = 3.1 lo ad Wald chi2(15) Prob > chi2 = (assumed) ju y th corr(u_i, X) = = (Std Err adjusted for 100 clusters in id) yi Robust Std Err pl Coef ua al INFO z [95% Conf Interval] H SIZE INDSIZE VOL 1.058885 -1.29e-08 -2.53e-10 -1.92e-06 year 2007 2008 2009 2010 2011 3523899 6404993 1.081128 2.029794 1.952925 1758489 2246257 3323303 4190481 4727486 indus 10 11 3776159 3593328 -.8409514 -.0343544 5169926 4350747 -.0943465 3935642 4042144 2543532 (omitted) 2334348 2625907 2556728 260248 0.96 0.89 -3.31 0.337 0.374 0.001 -0.15 1.97 1.70 -0.36 0.883 0.049 0.089 0.717 -.4918783 0023242 -.0660348 -.6044232 _cons 0974481 3216922 0.30 0.762 -.533057 sigma_u sigma_e rho 4693801 54970996 42166167 n n va 3433227 3.64e-09 3.07e-10 3.04e-07 P>|z| 0.002 0.000 0.410 0.000 385985 -2.01e-08 -8.54e-10 -2.51e-06 1.731785 -5.80e-09 3.48e-10 -1.32e-06 2.00 2.85 3.25 4.84 4.13 0.045 0.004 0.001 0.000 0.000 0077323 200241 4297724 1.208475 1.026355 6970474 1.080758 1.732483 2.851113 2.879495 -.3937556 -.4329129 -1.339475 1.148988 1.151579 -.3424283 ll fu 3.08 -3.55 -0.82 -6.29 oi m at nh z z k jm ht vb om l.c gm 4231694 1.031661 9361842 4157302 a Lu 7279532 n n va y te re (fraction of variance due to u_i) th 132