Part 1: Previous research evidence
1.8. Conclusions
independent audit committee, block holding, leverage and firm size. The summary statistics show that on the average the board size, outside board directors and directors shareholdings are 9.9217, 65.1718 and 7.5264 respectively while that of independent audit committee, block holding, leverage and firm size are 49.6502, 43.7460, 0.6713 and 15.3978 correspondingly. The standard deviation of gauges of corporate governance mechanisms are 3.9292, 14.9493, 15.3702 and 5.4471 for board size, outside board directors, directors shareholding and independent audit committee while block holding, leverage and firm size posted 26.0204, 0.3392 and 2.3067 correspondingly. It is also noted that the identified corporate governance mechanisms are diverse over time. For example the minimum for board size, outside board directors, directors‟ shareholding, independent audit committee and block holding is zero (0) while leverage and firm size have -0.42 and 10.33 respectively as their minimum. On the other hand, the maximum for board size, outside board directors, directors‟ shareholding and independent audit committee are: 25, 88.24%, 86.94% and 75% correspondingly while block holding, leverage and firm size recorded 91.36, 3.88 and 22.19 respectively
Hausman test is significant in almost all the regression which indicates that Fixed Effects results are preferred. Hence, Random Effect results are reported at the Appendix.
In addition, theoretically the superiority of the fixed-effect models to the random effect models in this kind of scenario is expected as N (the number of cross-section units, in this case 43 firms) is relatively larger than T (the number of time period, 26 years).
Note that statistical inference under the fixed-effect is conditional on the observed sectional units in the sample as against the random effect that assumes that the cross-sectional units are randomly drawn (Gujarati, 2003). Further, the fact that our samples are only for selected sectors listed on the Nigerian Stock Exchange on which data are available, makes the assumption of random drawing ineffective, thereby further bracing the superiority of the fixed effect. Therefore, Fixed Effect results of the five corporate performance measures are reported in Tables 4.3 to 4.7, while the results of Random Effects are reported in Tables 4.8 to 4.12 respectively.
Regressions are estimated on the total sample (All Firm). Specifically, Table 4.3 reports Fixed Effects results of the regression on Return on Assets, while Tables 4.4, 4. 5, 4.6 and 4.7 report results on Return on Equity, Price Earnings Ratio, Tobin‟s q and Labour Productivity respectively.
Table 4.3: Fixed Effect Regression Results of Effect of Corporate Governance Mechanisms on Return on Assets
VARIABLES Selected Firms
BS 0.0325*
(0.0171)
OBD -0.000666
(0.00147)
DSH 0.000259
(0.00184)
IAC 0.00122
(0.00375)
BH -0.00311*
(0.00176)
L -0.327***
(0.0693)
FS -0.0398***
(0.0135)
BS2 -0.00191**
(0.000915)
Constant 0.939***
(0.312)
Observations 1,118
R-squared F-statistic
0.084 5.21***
Number of company Hausman test
43
42.2***
Source: as in table 4.1
From table 4.3 result for the entire sample (All Firms), five variables are significant which include board size (BS), block holding (BH), leverage (L), firm size (FS) and squared of board size (BS2). While leverage (L), firm size (FS) and squared of board size (BS2) are significant at 1%, board size (BS) and block holding (BH) are significant only at 10%. These five variables were also found to be significant in the studies of Rostami et al, 2016; Almatari 2014; Olowookere 2008; Sanda et al 2005).
Other variables like board independence measured by number of outside board directors,
director shareholding and independence of audit committee have no significant impact on Return on Assets (ROA).
Out of the five variables, only BS has positive effect on return on asset, (Akpan &
Riman, 2012; Nidhi et al, 2016) while the remaining four have negative effect. The negative relationship between leverage and firm performance conforms to the studies of (Mule & Mukras, 2015; Hussain, Ashfaq & Muhammad, 2016), whereas (Olowookere, 2008; Cheema & Muhammad 2013) observed positive relationship between Leverage and return on assets. Similarly, the negative effect of firm size on firm performance observed in this study contradicts the findings of Hussain, Ashfaq, & Muhammad (2016) which observed a positive link between firm size and firm performance.
The negative coefficient of squared of BS confirm the fact that board size does not have a linear relationship with firm performance as measured by return on asset, rather a quadratic relationship. At the initial stage increase in BS will lead to increase in performance (Olowookere, 2008) but after certain threshold level of board size increase in number of board members will reduce performance (Eisenberg 1998; Guest, 2009)
The prior expectation as concerns the influence of board size on performance is met as board size positively affects return on asset to a level after which the increase in the latter achieve a corresponding decline in the former. This shows that at lower board size, a marginal increase in the number of members will initially raise ROA as the board‟s capacity for monitoring is enhanced (Donaldson et al, 2013). However, at some higher board size, increasing the board further may distort its operations.
Table 4.4: Fixed Effect Regression Results of Effect of Corporate Governance Mechanisms on Return on Equity
All Firms VARIABLES
BS -1.257**
(0.528)
OBD -0.0542
(0.0454)
DSH -0.0775
(0.0567)
IAC 0.175
(0.116)
BH -0.101*
(0.0543)
L 3.932*
(2.138)
FS -0.408
(0.415)
BS2 0.0645**
(0.0282)
Constant 9.322
(9.630)
Observations 1,118
R-squared F-statistic
0.063 3.55***
Number of company Hausman test
43 15.19*
Source: as in table 4.1
Only Board size (BS) and square of Board size (BS2) are statistically significant at 5%
level in the result for All Firms. However, both came out with wrong signs, while BS has negative sing, BS2 has positive sign. This is contrary to expectation as BS which is expected to be positive turned to be negative while BS2 which is expected to be negative but recorded positive. The result is consistent with the conclusions drawn by (Uadiale, 2010; Bebeji, et al 2015; Anca- Elena, 2015) who reported a significant negative
relationship between board size and the performance of a firm measured by ROE. L is positive and statistically significant at 10%, contrary to the findings of (Hussain et al 2016; Mule, et al 2015) who detected an inverse relation between leverage and return on equity. In the same vein BH is significant but negative at 10% level of significant. This also negates the findings of (Manawadugbe, et al 2013) who observed a positive relation.
Table 4.5: Fixed Effect Regression Results of Effect of Corporate Governance Mechanisms on Price Earnings Ratio
VARIABLES All Firms
BS -0.00142
(0.0443)
OBD -0.00788**
(0.00380)
DSH 0.00569
(0.00475)
IAC 0.0124
(0.00972)
BH -0.00735
(0.00456)
L 0.0277
(0.179)
FS 0.122***
(0.0348)
BS2 0.000509
(0.00237)
Constant -1.580*
(0.808)
Observations 1,118
R-squared F-statistic
0.054 2.77***
Number of company Hausman test
43 15.3*
Source: as in table 4.1
The sample of all firms from table 4.5 shows that only OBD and FS have significant impact on firm performance measured by Price earnings ratio. Outside board
size is negatively statistically significant at 5% level which contradicts the findings of (Sanda, et. al (2005) and Olowookere, 2008) who found no significant relationship between OBD and firm performance measured by PE ratio. Firm size is statistically significant at 1% level with positive signs. This is inconsistent with the study of Olowookere, (2008) who observed a negative linkage between firm size and price earnings ratio. Other independent variables such as BS, BH, DSH, IAC and L do not have significant relationship with firm‟s performance gauged by PE ratio. This is contrary Olowookere, (2008) who found positive correlation between BS; IAC; BH and price earnings ratio. He also observed relations between FS; square of board size and price earnings ratio inverse relationship between BH and firm performance measured by PE ratio
Table 4.6: Fixed Effect Regression Results of Effect of Corporate Governance Mechanisms on Tobin’s Q
All Firms VARIABLES
BS 0.200**
(0.0889)
OBD -0.0166**
(0.00764)
DSH -0.00393
(0.00954)
IAC 0.0175
(0.0195)
BH -0.00716
(0.00915)
L -0.129
(0.360)
FS -0.0152
(0.0699)
BS2 -0.0111**
(0.00475)
Constant 1.790
(1.621)
Observations 1,118
R-squared F-statistic
0.090 1.36 Number of company
Hausman test
43 21.48***
Source: as in table 4.1
From the regression result on table 4.6 board size (BS), square of BS and OBD as a measure of corporate governance are statistically significant at 5% level respectively;
all other independent variables are insignificant. Board size is positively and square of board size is inversely related with Tobin‟s q as a measure of firm performance. This indicates that the performance measured by Tobin‟s q of a firm continues to appreciate with additional board member up to a remarkable extent before it starts impacting negatively. This confirms the quadratic relationship between Board size (BS) and firm performance which is measured as Tobin Q. This is in conformity with the findings of (Olowookere, 2008; Rashid, De Zoysa, Lodh, & Rudkins, 2010; Al-Matar et al 2014). On the other hand, OBD is inversely correlated with firm performance measured by Tobin‟s Q. This implies that as more member of OBD is added, Tobin‟s Q as a measure of firm performance declines. This is supported by the findings of (Rashid, De Zoysa, Lodh, &
Rudkins, 2010; Fauzi et al 2012). However, Olowookere (2008) and Bruno, (2013) observed a positive relationship between OBD and Tobin‟s Q
Table 4.7: Fixed Effect Regression Results of Effect of Corporate Governance Mechanisms on Labour Productivity
All Firms VARIABLES
BS 0.152***
(0.0402)
OBD 0.00485
(0.00345)
DSH 0.00429
(0.00431)
IAC 0.0102
(0.00882)
BH 0.0151***
(0.00414)
L -0.317*
(0.163)
FS 0.846***
(0.0316)
BS2 -0.00591***
(0.00215)
Constant -8.261***
(0.733)
Observations 1,118
R-squared F-statistic
0.619 120.65***
Number of company Hausman test
43 47.1***
Source: as in table 4.1
From the sampled firms, the board size (BS) is positively statistically significant at 1% while square of board size (BS2) is negatively statistically significant at the level of 1% with labour productivity. This shows that an increase in board membership will continue to enhance performance measure by labour productivity up to a certain extent thereafter, it will start declining. This equally confirms that BS does not have linear
relationship with firm performance measured by labour productivity rather a quadratic relationship. The result of this study is supported by (Olowookere, 2008). Also, Block Holding (BH) is positively statistically significant at 1% level of significance with labour productivity. It reveals that an increase in Block Holding will lead to an improvement in performance measured by labour productivity. The result this study is affirmed by (Olowookere, (2008) while Xu et al observed an inverse relation between BH and labour productivity. However, leverage (L) is negatively statistically significant at 10% level of significance. It shows that an increase in the level of leverage will result into decrease in the firm performance gauged by labour productivity. The finding of this study negates the study of Olowookere, (2008) who found a positive significant relationship between leverage and labour productivity. Lastly Firm size (FS) as a measure of corporate governance indicator is positively significant at 1% with labour productivity. This reflects that as the firm size increases the performance level measured by labour productivity will continue to rise. This is contrary to Olowookere, (2008) who found an inverse relationship between firm size and labour productivity.
4.3 Hypotheses Testing