





















Contents

Asian Finance & Banking Review
Vol. 1, No. 1; 2017

Published by Centre for Research on Islamic Banking & Finance and Business

Macroeconomic Aggregates and Retention Ratio of Quoted Firms in Nigeria
Henry Waleru Akani and Yellowe Sweneme


Macroeconomic Aggregates and Retention Ratio of Quoted Firms in Nigeria
Henry Waleru Akani1

Yellowe Sweneme1
1Department of Banking and Finance, Rivers State University, Nkpolu - Port Harcourt, Rivers State, Nigeria

Correspondence: Henry Waleru Akani, Department of Banking and Finance, Rivers State University, Nkpolu - Port Harcourt, Rivers State, Nigeria, Email: henryakani@yahoo.com
Received: October 14, 2017,        Accepted: October 19, 2017,     Online Published: October 26, 2017  

Abstract
This study examined the effects of macroeconomic aggregate on retention ratio of selected quoted manufacturing firms in Nigeria for the period 1981 to 2014. The study used secondary data. The technique adopted is the Ordinary Least Squares, Error correction mechanism and Autoregressive Distributed Lag (ARDL) Bounds approach to cointegration. The dynamic short-run estimate revealed that interest rate exerts a negative influence on retention ratio. The study also found that oil price exerts a positive and significant impact on retention ratio. Further, it revealed that capital market development exerts a positive influence on retention ratio, but financial sector's development showed a positive relationship with retention ratio, inflation rate appeared with an expected negative sign. Foreign exchange rate showed a positive relationship with retention ratio; money supply exhibited a positive influence on retention ratio of quoted firms in Nigeria. The error correction coefficients were significant with the expected sign. A long run relationship among the variables was established. Thus, the study concludes that macroeconomic variables have a significant influence on dividend policy. We recommend the need for firms to consider the operating macroeconomic framework in formulating dividend policy.

Keywords: Macroeconomic Variables, Retention Ratio, Macro.

1. Introduction
Dividend policy refers to the decision of management about the portion of income that is given to stakeholders in the form of dividend and this is an arguable issue for financial managers for decades (Toby, 2003). Dividend policy is also a critical finance management function that determines the proportion of corporate profit that is distributed to shareholders and the proportion that is retained. It is believed that dividends are highly sensitive to external events such as economic factors within the operating environment. A very important aspect of dividend policy is retention ratio. Retention ratio means the proportion of earnings kept back in the business as retained earnings. In other words, it is the percentage of net income that is retained to grow the business, rather than being paid out as dividends to shareholders. 

Despite various reforms, policies, structural changes the performance of corporate organizations remains abysmal and affects the retention ratio of firms. Unfortunately, a significant proportion of Nigerian firms has no consistent retention ratio policy over the past three decades, as they are largely influenced by unanticipated economic events with pervasive impact on dividend decisions of firms as it relates or affects return on the ratio. Also, there is huge scarce literature that identified the macroeconomic determinants of dividend decisions of firms. It is believed that macroeconomic policy directly or indirectly determines the value of firms retention ratio that shareholders desire to plow back.


In this case, behavioral finances play an important role in explaining the dividend policy in any organization. Miller (1986) presents a traditional argument against behavioral finance by contending that behavioral theories may be able to explain the micro factors, but rational theories suffice to explain the macroeconomic aggregates that determine dividend policy. This suggests that the factor that determines corporate retention ratio remains a matter of debate among scholars.


Knowledge on the nexus between dividend and macroeconomic variables is crucial to the investors in the equity market as well as to the policy makers. Therefore, it is important to examine the relationship between macroeconomic variables and retention ratio of quoted firms in Nigeria. 

2. Literature Review

Dividend Policy refers to a company's policy which determines a number of dividend payments and the amounts of retained earnings for reinvesting in new projects. Dividend policy has been of great interest to researchers and extensive empirical research has been carried out to identify the potential factors that influence the dividend decision of the firm. However, researchers are still unable to reach a consensus in this regards (Kim and Jang, 2010). It has remained a puzzle for financial economists (Black, 1976).


Duke, Nneji, and Nkemare (2015) examined the impact of dividend policy on share price valuation in Nigerian Banks, based on data from two banks operating in Nigeria. They found that dividend yield had a significantly negative effect on share price


Spyrou (2001) studied the relationship between dividend policy and inflation for the emerging economy of Greece. Spyrou (2001) in consistent with Kaul's results, found that inflation and dividend policy are negatively related to the year 1995, after which the relationship became insignificant. Spyrou accredited the change in the relationship to the increased role of monetary fluctuations in line with Marshalls (1992) argument, which states that the negative relationship between stock price returns will be less pronounced during the periods when inflation is generated by monetary fluctuations.


Ralph and Eriki (2001) conducted an empirical study on Nigerian stock market and found that a negative relationship exists between dividend policy and inflation. However, they also showed that the dividend policies are also strongly motivated by the level of economic activity measured by interest rate, money stock, GDP and financial deregulation.
Kalyanaraman and Al-Tuwajri (2014) examined the existence of a long run relationship among five macroeconomic variables of CPI, industrial output, money supply, exchange rate, oil price along with proxy of S&P 500 and the TASI (Saudi All stock index). They used monthly data from 1994 to 2013 and applied the time series analysis. They found an existence of a long run relationship among the five variables and all the five variables put an impact on stock price whereas S&P 500 index does not impact Saudi stock prices. They also found a two-way causality between stock prices and oil prices they also found that the industrial production shocks push up the stock prices while consumer price index shocks pull the stock price down. 

Mwangi (2013) tries to determine the effect of macroeconomic variables such as real exchange rate, GDP growth rate, the change in money supply (M3), average annual lending interest rates and inflation rate measured by annual percentage changes in the consumer price index (CPI) on financial performance proxied by Return on Assets (ROA) of aviation industry in Kenya. The results reveal that ROA has a weak positive insignificant correlation with gross domestic products growth rate and annual change in money supply while a weak negative insignificant correlation exists between ROA and exchange rate, annual average lending rate and annual average inflation. 

Singh, Tripathi, and Parashar (2013) examined the primary factors those are responsible for affecting the index (NIFTY) in National Stock Index in India. They took Exchange rate, Insurance Intermediation Premium (IIP), WPI as the independent variable and applied regression analysis and found that IIP, exchange rate, and WPI influences the stock prices.

Abedallat and Shabib (2012) examined the impact of macroeconomic indicators like a change in investment and gross domestic product (GDP) as the independent variables and the movement of Amman Stock Exchange index as the dependent variable for the data period of 1990- 2009. For the analysis of the above relationship, they used the multiple regressions. They found a relationship between the two macroeconomic indicators (the investment and GDP) and the Amman Stock Exchange index, and also between each of them separately and the stock index, which means that the movement of prices in the Amman Stock Exchange affected by the movement of these two variables, and there is the effect of both variables on the movement of Amman Stock Exchange index. Further, they found the impact of the change in investments was greater than the impact of the change in GDP on the Amman Stock Exchange index.

Basse and Reddemann (2011) examined inflation and dividend policy of US Firms and pointed that the neglecting of macroeconomic variables as the important reason why empirical tests often fail to support theories of dividend determination. He found a stable long-run relationship between dividend payments and real economic activity and price level.


3. Model and Estimation Techniques

The Ordinary Least Squares (OLS) has been successfully used in studies (Abedallat and Shabib, 2012; Singh, Tripathi, and Parashar, 2013; and Amadi, Oneyema and Odubo, 2000), hence, this technique is employed in estimating the specified equations, while the EViews 9 as a computing platform in the analysis. The study utilized annual time series secondary data (1981 – 2014). All the data set are obtained from the Nigerian Stock Exchange fact book, Central Bank of Nigeria (CBN), and annual reports and statement of accounts.
In line with Olugbenga (2011), Osa and Ikaibo (2002), Amadi, Oneyema and Odubo (2000), the regression model takes the form:
RR =β0 + β1INTR + β2OILP + β3FD + β4CD + β5M2 + β6INFR + β7RGDP + β8EXR + i…. (1)

Where:

RR   
=
Retention Rate

INTR
=
Interest Rate

OILP
=
Oil Price 

FD
=
Financial Sector Deepening 

CD
=
Capital Market 

MOS
=
Broad Money Supply

INFR
=
Inflation Rate

RGDP =
Real Gross Domestic Product

EXR
=
Exchange Rate 

i
=
Error Term

βi - β8 =
Coefficient of the Independent Variables 

β0
=
Regression Intercept

4. Empirical Results and Analysis
We present and analyze the estimated short-run model based on some goodness of-fit criterion, such as Akaike information criterion (AIC) and Schwartz information criterion (SIC) for selection of lag length in a model, hence; it is presented in the table below;

Table 1: Short-run Estimated Result showing the effects of Macroeconomic Variables on Dividend policy Indices in Nigeria
	Variable
	ΔRR

	C
	-8.712443   (-0.340313)

	ΔINTR
	-26.07145   (-5.129213)*

	ΔINTR(-1)
	-4.901002   (-1.338707)

	ΔINTR(-2)
	-7.099566   (-1.751070)

	ΔOIP
	1.378985    (3.242073)*

	ΔOIP(-1)
	1.424468    (3.075754)*

	ΔOIP(-2)
	2.485007    (4.534692)*

	ΔCD
	0.002173    (0.268275)

	ΔCD(-1)
	-0.026711   (-2.392656)

	ΔCD(-2)
	-0.011735   (-0.435949)

	ΔFD
	2.136948    (0.362346)

	ΔFD(-1)
	18.50266    (1.822995)

	ΔFD(-2)
	10.22963    (2.465365)*

	ΔINFR
	-0.005405   (-0.004556)

	ΔINFR(-1)
	-1.839189   (-1.030609)

	ΔINFR(-2)
	-0.697949   (-0.676941)

	ΔRGDP
	14.59239    (2.827198)*

	ΔRGDP(-1)
	18.72948    (2.626467)*

	ΔRGDP(-2)
	-37.52078   (-3.727010)*

	ΔEXR
	2.665646    (1.236714)

	ΔEXR(-1)
	-2.192780   (-1.236992)

	ΔEXR(-2)
	5.856717    (2.386177)*

	ΔMOS
	9.170141    (3.623753)*

	ΔMOS(-1)
	4.854979    (2.577289)*

	ΔMOS(-2)
	-0.041259   (-0.025766)

	ECM(-1)
	-1.540177   (-7.039320)*

	R-squared
	0.972314

	F-statistic
	7.023826     (0.019475)*

	Durbin-Watson stat
	1.294621

	Breusch-Godfrey Test
	0.398060     (0.7025)


Source: Eviews 9 Computation

Note: t-statistic in parenthesis ( ), *indicates significance at 5%
The above error correction model shows that the coefficient of determination (R2) was significantly high. That is, the selected macroeconomic variables explained 97% changes in retention ratio. Also the overall regression was significant at 5%. The error correction coefficients have the expected sign and relatively high, though RR is significant.
The a priori of the signs of explanatory variables in the RR equation were correctly sign, except one and two-period lag capital market development, current period exchange rate, lag two exchange rate and money supply respectively. RR responds negatively and significantly to current-period interest rate; it also responds positively and significantly to changes in oil price, two-period lag financial deepening also showed a positive and significant influence on RR, real GDP growth appeared positive and significant, exchange rate appeared positive at current period and lag two-period, though significant only at lag two, while money supply appeared with a positive and significant sign at current period and lag-one period respectively, this confirms the economic believe that the dependence of a variable on another is rarely instantaneous, but responds with a lapse of time (Gujarati, 2004).


This is an indication that certain economic indices do not reflect in dividend policy, and economic policies are not satisfactorily used in the direction of increasing shareholders returns. This finding may not be unconnected with the lack of prudence in the utilization of dividends. In sum, the result reveals that dividend policy indices in the period under review in Nigeria does adjust fairly to changes in macroeconomic variables.


Test for Serial Correlation

In essence, we employed the Durbin-Watson (DW) test for autocorrelation, it is based on the assumption underlying the Ordinary Least Squares (OLS), that the error term (μ) is assumed to be uncorrelated. The Breusch-Godfrey test shows evidence that the residuals are uncorrelated.

4.1 Test for Perfect Multicollinearity

Table 2: Test for Multicollinearity

	
	INTR
	OIP
	CD
	FD
	INFR
	RGDP
	EXR
	MOS

	INTR
	 1.000000
	
	
	
	
	
	
	

	OIP
	-0.076909
	 1.000000
	
	
	
	
	
	

	CD
	-0.018704
	-0.389686
	 1.000000
	
	
	
	
	

	FD
	 0.042121
	-0.356406
	 0.642279
	 1.000000
	
	
	
	

	INFR
	 0.026434
	 0.095067
	-0.315252
	-0.361966
	 1.000000
	
	
	

	RGDP
	 0.077516
	 0.115997
	 0.391266
	 0.349791
	-0.206831
	 1.000000
	
	

	EXR
	 0.306915
	-0.370523
	 0.753070
	 0.577952
	-0.228009
	 0.421499
	 1.000000
	

	MOS
	 0.362128
	 0.113533
	 0.112204
	-0.009332
	-0.120845
	-0.168355
	 0.018838
	 1.000000


Source: Author’s Computation Using EViews 9 Software
From the table 2, apart from the diagonal, the correlation between variables is not unity; this implies that the explanatory variables have no exact or perfect relationship.

4.2 Test for Model Stability



Figure 1: Model Stability Test for Retention Ratio Model
The CUSUM tests are used in this study to test for parameter stability, our graph shows that the plots of the residuals remain within the 5% critical bounds, therefore, we can accept that the parameters of the model are stable.


4.3 Test for Residual Normality

The Jarque-Bera (JB) test of normality is adopted in this study, purely to verify whether the residuals are normally distributed. It is conducted under the null hypothesis that the residuals are not normally distributed. From illustration below; the computed ρ value of the JB statistic (0.100786) under the normality assumption, we, therefore, reject the hypothesis that the error terms are not normally distributed. The diagram below shows that the residuals from the regression seem to be symmetrically distributed.


Figure 2:
Normality Test For Retention Ratio Model
4.4 Test for Heteroscedasticity
Table 3:
Test for Heteroscedasticity

	Heteroscedasticity Test for Retention Ratio

	F-statistic
	0.550246
	    Probability
	0.8792

	Obs*R-squared
	12.81713
	    Probability
	0.7483


Source: Author’s Computation
This test is conducted using white’s test, which involves either an auxiliary regression with no cross-terms or with cross terms. It also follows the F-distribution, from the table 3, since the ρrobability of F-value 0.8792 for retention ratio is not significant, we, therefore, conclude that there is homoscedasticity, that is to say, the variances are equal.

4.5 Unit Root Stationarity Test 
Table 4: Unit Root Stationarity Test


A Time Series Yt is Integrated of Order D, Denoted I(D) If ∆Dyt is Stationary. Then The Series Yt Has D Unit Roots.To Further Ascertain The The Stationarity of the Data Series, Hence;
Table 4: Augmented Dickey-Fuller (ADF) Unit Root Test of Stationarity Results


	Test
	Variables
	Levels
	
	Differences
	
	Order of Integration

	
	
	t- statistic
	Critical
	t- statistic
	Critical
	

	ADF
	RR
	
	
	-8.904988
	-3.653730
	I(1)

	
	INTR
	
	
	-6.384610
	-3.661661
	I(1)

	
	OIP
	-5.877960
	-3.646342
	
	
	I(0)

	
	CD
	
	
	-6.677589
	-3.653730
	I(1)

	
	FD
	
	
	-5.726112
	-3.661661
	I(1)

	
	INF
	
	
	-5.730629
	-3.670170
	I(1)

	
	RGDP
	-4.369344
	-3.646342
	
	
	I(0)

	
	EXR
	
	
	-6.006878
	-3.653730
	I(1)

	
	MOS
	
	
	-7.185088
	-3.653730
	I(1)

	


Note:
* Implies significance at 1%
Source: Author’s Computation based on data from Central Bank of Nigeria Publications
According to Maddala (1992), testing for unit roots is a formalization of the Box- Jenkins approach of differencing the time series after a visual inspection of correlogram. The analyzing and testing for unit root naturally lead to the theory of cointegration (Iyoha and Ekanem, 2002).

The summarized result presented in table 4 shows that at various levels of significance (1%, 5%, and 10%), all the variables were stationary, specifically, OIP and RGDP are integrated of order zero, I(0), whereas RR, INTR, CD, FD, INF, EXR, and MOS are integrated of order one, I(1). Hence, all the variable in this study is stationary.


4.6 ARDL Bounds Tests for Cointegration

Table 5a: Bounds Test for Cointegration Analysis 

	Critical value
	Lower Bound Value
	Upper Bound Value

	1%
	3.15
	4.43

	5%
	2.55
	3.68

	10%
	2.26
	3.34


Source: Pesaran et al. (2001)

	ARDL Bounds Test Test:
	

	
	Test Statistic
	Value  

	RR EQUATION
	F-statistic
	5.466722


Table 5b: ARDL Results for Cointegration Analysis

Source: Author’s Computation Using EViews Software
Pesaran and Shin (2001) showed that cointegrating systems can be estimated as ARDL models; it has the advantage to estimate cointegrating relationship on variables that are either I(0) or I(1).

According to Pesaran et al. (2001), the asymptotic distribution of the F-statistic is non-standard regardless of whether the regressors are I(0) or I(1), and provide two adjusted critical values that establish lower and upper bounds of significance.

4.7 Interpretation

Given a computed F statistics Value of 5.466722 for RR equations, the results of the bounds co-integration test, therefore, establish that the null hypothesis against its alternative is rejected at the various significance level. The computed F-statistic is greater than the lower and upper critical bound values at 1%, 5% and 10% respectively, thus indicating the existence of a steady-state long-run relationship among the variables. This suggests that the various selected macroeconomic variables have a long run relationship with dividend policy indices in Nigeria.

5. Concluding Remark

This study is an attempt to investigate macroeconomic variables and dividend policy of quoted firms in Nigeria. The overall results suggest that retention ratio responds to the dynamics of the macroeconomic environment; especially from interest rate, oil price, financial sector development, capital market development, foreign exchange rate, monetary policy instrument of money supply and the magnitude of growth in the economy, other check variables like inflation rate appeared with a slight stimulus on retention ratio.

Our result supports the outcry for firms to consider the existing macroeconomic environment, as essential to achieving quite good dividends policy. In this line, Chen, Roll, and Ross (1986) showed that in the United States, dividends have a strong relationship with macroeconomic variables. This study also supports the scholarly work of Olugbenga (2011) that the impact of macroeconomic indicators such as money supply, interest rate, exchange rate, inflation rate, oil price and gross domestic product have varying impact on stock prices of firms in Nigeria. This confirms that the dynamics in monetary policy and the oscillation in the external sector have an unpredictable influence on dividends policy of quoted firms in Nigeria. Our long-run estimate is similar to Basse and Reddemann (2011), based on the established long-run relationship found between dividend policy indices and the macroeconomic variables.


Remarkably, this study concludes that a long run relationship exists between retention ratio and the macroeconomic variables. These conclusions also offer valuable indications to policy makers when determining retention ratios, by considering the macroeconomic environment where the firm exists. Therefore, the need for firms to consider the operating macroeconomic framework in formulating dividend policy is recommended from this study.
References

Al-Abedallat, A.Z., & Al-Shabib, D.K. (2012). Effects of Macroeconomic Variables Non-Stock Market: Saudi Perspective, International Journal of Financial Research, 5(4),120-127


Black, F., (2001).The effects of dividend yields and dividend policy on common stock prices and return. Journal of financial economics 1(1),478-531. 

Brook, Y., Chalton, W., &  Hendershott, R., (1998). Do firms use dividends to signal large future cash flow increase? International journal of Finance management 9 (27), 46-57.

Basse, T., & Reddemann, S. (2011).Inflation and the Dividend Policy of US Firms. Economy watch 6 (4), 333-64

Black, F. (1976). The Dividend Puzzle, Journal of Portfolio Management, 2, 5-8

Brook, Y., Chalton, W. and Hendershott, R. (1998). “Do firms use dividends to signal large future cash flow increase?”, Financial Management, Autumn, 46-57.

Brooks. C. (2002). Introductory Econometrics for Finance, 2nd Edition, and Cambridge University Press.;

Chen, N, Roll, R .,   &  Ross, A., (1986). Economic Forces and the Stock Market: Testing the APT and Alternative pricing Theories. Journal of Business.l59 (17), 383-403.

Dickey, D., & Fuller, W., (1981). Likelihood ratio statistics for autoregressive time series with a unit root.Econometrica  17(49),  1057 – 1072.

Duke, S.B., Ikenna, N.D. and Nkamare, S.E.(2015). "Impact of Dividend Policy on Share Price Valuation in Nigerian Banks." Archive of Business Research, 3 (1) 156-170

Gordon, M. J. (1959). Dividends, Earnings, and Stock Prices. The Review of Economics and Statistics, 41(2), 99-105.

Gordon, M., (1962).The Savings, Investment, and Valuation of A Corporation. Review of Economics and Statistics, 37-51. 

Gordon, M. J., &  Shapiro,  E., (1956). Capital Equipment Analysis: The Required Rate of Profit. Management Science, 3 (1), 102-110.

Gujarati, N. (2004). Basic  Econometrics. Singapore: Mcgraw-Hill.

Kalyanaraman, L., & Al-Tuwajri, B. (2014). Macroeconomic Forces and Stock Prices: Some Empirical Evidence from Pakistan, Asain Journal 11(7) 89-102

Kaul, G., (1990). Monetary Regimes and the Relation between Stock Returns and Evidence from the Finnish Stock Market. The Finnish Journal of Business Economics, 49(2), 209-232. 

Maddala, G. S., & Wu, S. (1999). Cross-Country Growth Regressions: Problems of Heterogeneity, Stability, and Interpretation. Applied Economics, 32, 635–642.


Miller, M. H., & Modigliani, F. (1961).Dividend Policy Growth and the Valuation of Shares.The Journal of Business, 34, (4), 411-433.

Okpara, G. (2010).A Diagnosis of the Determinants of Dividend Payout Policy In Nigeria: A Factor Analytical Approaches. American Journal of Scientific Research, 8, 57-67.

Onoh, J.K.,(2002).Dynamics of Money Banking and Finance in Nigeria: An Emerging Market. Astra Meridian Publishers, Aba, Enugu, Lagos.
Copyrights 

Copyright for this article is retained by the author(s), with first publication rights granted to the journal.
55
63

_1542294018.bin

_1542294022.bin











-15
-10
-5
0
5
10
15
90929496980002040608101214
CUSUM5% Significance










0
1
2
3
4
5
6
7
8
9
-70-60-50-40-30-20-1001020304050
Series: Residuals
Sample 1984 2014
Observations 31
Mean      -3.40e-14
Median   1.987776
Maximum  47.66404
Minimum -61.85958
Std. Dev.   20.97054
Skewness  -0.551083
Kurtosis   4.529183
Jarque-Bera 4.589512
Probability 0.100786
