Dividend Policy and Political Uncertainty: Does Firm Maturity Matter?

Article history: Received date: 25 May 2020 Received in revised form: 27 June 2020 Accepted: 3 August 2020 Available online: 18 September 2020 Previous evidence has shown that numerous factors influence dividend policy, but how political uncertainty affects a firm’s cash dividend policy remains blurry. This study examines the relationship between cash dividends and political uncertainty in Nigeria. More so, the study analyses whether this relationship prevails on matured and non-matured firms. The study employed ordinary least squares dummy variable (LSDV) approach with robust standard error on a data set of non-financial listed Nigerian firms. The results revealed that political uncertainty strongly influences firm’s cash dividend, and a matured firm tends to pay greater dividends than nonmatured firms (firms with more growth options). Thus, this finding suggests that matured firms pay more dividends during period of political uncertainty. Consequently, the study supported the agency theory and the life cycle theory. Kebijakan Dividen dan Ketidakpastian Politik: Apakah Maturitas


Introduction
Researchers around the world have attempted to explore how politics and corporate activities are related. Growing evidence has indicated that political uncertainty affects corporate financial policy and one of these policies is dividend payout (Farooq & Ahmed, 2019;Huang, Wu, & Yu, 2015;Tran, 2020). Dividend payout is critically important to the stakeholders of firms. This is not surprising as dividend policy appears to be at the top of the most debated field in finance (Baker & Weigand, 2015). This study analyses the political influence on a firm's dividend policy in Nigeria as well as how this relationship prevails in matured firms.
A presidential or general election in a country could also provide an insight into the dividend puzzle. This is because every political party has its unique way of dealing with economic issues through the issuance of a new policy. For instance, in the developing world, the assumption is that a conservative party politician serving as a president may be stricter on foreign importation policies compared with a member of a liberal party. So, if a firm depends on sourcing its raw materials from another nation, then this may affect its operations as it exposes the firm to uncertainty and, hence, impacts its profitability, which will, in turn, have an influence on its dividend policy. Brav, Graham, Harvey and Michaely (2005) asserted that a prevailing economic setting tends to influence a firm's dividend policy. Uncertainty may prevail, particularly when a new president is expected to be elected. Buchanan, Cao, Liljeblom and Weihrich (2017) indicated that policy shocks such as economic and regulatory reforms might affect a firm on the one hand and the benefits of shareholders on the other hand. Hence, this political uncertainty leads a firm to reconsider its financial policies, such as a dividend policy as a precautionary measure.
Prior works have indicated the significant effect of uncertainty on cash dividends, from the United States (Buchanan et al., 2017;Farooq & Ahmed, 2019) and other countries (Huang et al., 2015). However, evidence on how political uncertainty impacts dividend policy from a developing economy, for instance, Nigeria remains scanty. For example, Farooq and Ahmed (2019) reported that firms pay more dividends during the years of a US presidential election. First, unlike the United States, Nigeria has experienced political instability. Therefore, the result obtained in the US market may not apply to the Nigerian context because of differences in their environmental settings and regulatory frameworks. Second, since the inception of the fourth republic in 1999, only one party was in power from 1999 to the first quarter of 2015, and this party was considered liberal. Third, it was a coalition or alliance of a group of parties that ousted the long-term ruling party. Therefore, a period of high uncertainty as to economic and regulatory reforms came into being, which could affect the listed firms.
Fourth, other studies, for example, Huang et al. (2015), have considered international crises as a source of political uncertainty, this may not provide a clear understating of the phenomenon in view since political crises varies from country to country. For instance, the political crises in UK may not be the same for US let alone the political crises of an advanced country with that of developing country. Lastly, to the best of our knowledge there is no single study that have tested the relationship between political uncertainty and dividend policy on matured firms.
Hence, this paper contributes to the dividend policy literature in a newly democratised system of government compared to the United States and other countries of the developed world that political uncertainty has greater influence on dividend policy. Second, the current study also found, consistent with the life cycle theory, that matured firms pay more dividends as compared with growing firms. The result indicates that a mature firm may not be influenced by the shock that arises as a result of a national election. Therefore, shareholders of a matured firm are assured of an inflow of returns in the form of dividends.
The rest of the paper is organised as follows. Section two discusses the relevant literature and hypothesis development. The methodology is situated and discussed in section three, while section four discusses the findings. The last section concludes the study.

Literature review and hypotheses development
Prior studies have attempted to link the dividend policy puzzle to various theories. One of them is agency theory. This theory predicts that a conflict of interest exists between managers and shareholders (Jensen & Meckling, 1976). Often, shareholders who happen to be outside investors tend to receive fewer benefits from the capital that they have invested in a firm because the managers prefer to maintain the cash or use it for perquisite consumption (Cao, Du, & Ørding, 2017). In this instance, a dividend payment remedies the agency conflict associated with squandering the available free cash flow in a firm (Easterbrook, 1984;Jensen, 1986). It is also argued by Choy, Gul, and Yao (2011) that agency problems were severe in poorer shareholder protection and proportionalelectoral countries because minority shareholders are unable to exercise their rights, let alone to address agency-related problems.

Political uncertainty and dividend policy
Political forces are part of the forces that either strengthen or weaken the economic activities of countries. For instance, Farooq and Ahmed (2019) affirmed that politics significantly affected corporate managerial decisions in the United States. Moreover, the operating environment of a firm is often altered as a result of national elections, thereby leading to uncertainty, which, in turn, affects future policy (Baloria & Mamo, 2017). Similarly, Chay andSuh, 2009 andHoberg andPrabhala (2009) have provided evidence that uncertainty at the firm level impacts its payout policy. This finding implies that examining how uncertainty relates to cash dividend decision will be an interesting issue. Using a data sample from the United States between 1996 and 2016, Farooq and Ahmed (2019) documented a positive and statistically significant relation between dividend payout and political uncertainty proxied by a presidential election. The results indicated that firms pay a higher dividend during a presidential election year as compared to a non-election period. Therefore, attesting that the firms can withstand the shocks that may prevail as a consequent of economic policy changes. Huang et al. (2015), while studying 35 countries, revealed that firms retained more cash during political uncertainty to provide a cushion or preventive measures against future political shocks thus, corroborating the evidence of Buchanan et al. (2017) that firms in the United States were less likely to initiate dividends during a period of policy uncertainty.
On the other hand, firms are likely to initiate a dividend or increase their existing dividends in expectation of a policy change such as tax increase. The increase or decrease in tax expectations is a mechanism through which firms respond in advance of the real changes in taxes (Farooq & Ahmed, 2019). Buchanan et al. (2017) found that firms in the United States reacted differently regarding their regular and special dividend payout policy with respect to 2010 and 2012 tax policies. Awotundun (2018) also confirmed these findings by documenting a negative and statistically significant relationship between the political factor and dividend payout while investigating the listed commercial banks in Nigeria between 2004 and 2014. More recently, Tran (2020) used data from US and showed that banks decreased their dividend payout as a result of high uncertainty. These findings negated the evidence documented from US market (Farooq & Ahmed, 2019). Summarily, National election could serve as a pipeline through which economic outcomes are influenced. Consistent with the agency theory that shareholders may demand a dividend because of uncertainty in the managerial behaviour that may lead to perquisite consumption. Hence, we posited that: H1: Political uncertainty is positively associated with cash dividend payout.
Prior literature has suggested the importance of retained earnings in dividend payout policy. This variable (retained earnings) is widely used as a proxy for firm maturity. For instance, DeAngelo, DeAngelo and Stulz (2006) augured that dividends are paid by matured (retained earnings as a proxy of firm maturity) and established firms. Conceivably reflecting the financial life cycle of firms, matured firms tend to pay dividends as they have higher profitability and are less attractive for new investment opportunities. Implicitly, paying out dividend could be a means of addressing an agency problem as these firms may have cash in abundance, and if not distributed to the shareholders, these will result in an adverse agency conflict (Jensen, 1986).
Therefore, retained earnings may provide a clue as to whether during uncertainty period firms may disgorge more cash as dividends to the owners or otherwise. Farooq and Ahmed (2019) reported that larger firms pay a higher dividend during periods of uncertainty. This is so because the size of a firm is one of the characteristics of dividend-paying firms see, for example, (Adamu, Ishak, & Hassan, 2019;DeAngelo et al., 2006;Fama & French, 2001;Hoberg & Prabhala, 2009;Jiraporn, Kim, & Kim, 2011). This study argues that firms with more retained earnings at their disposal may pay more dividends than growing firms. DeAngelo et al. (2006) showed that a significant number of firms with more retained earnings pay a dividend in the US market.
Both Coulton and Ruddock (2011), and Yarram and Dollery (2015) using data from the Australian market also supported the lifecycle theory that dividend payers are firms with fewer growth options at their disposal. More so, Denis and Osobov (2008) revealed that among the top features of dividend-paying firms was retained earnings in countries like Canada, France, Germany, Japan, the United Kingdom and the United States. The study found that dividend payers in these countries were large firms, with more amounts of retained earnings. Recently, Adamu et al. (2019) from the Nigerian market concurred with the prior evidence that the decision to pay a dividend is greater in mature firms. Thus, suggesting that matured firms have higher accumulated cash and used it in paying cash dividend.
Therefore, since maturity is among the features of dividend-paying firms, it is expected that matured firms will continue to disgorge cash in the form of dividend even during uncertainty periods such as during presidential election. It depends on the availability of cash at their disposal and to maintain their reputation by paying a dividend. Based on the life cycle theory and prior literature, the following hypothesis can be stated as: H2: The influence of political uncertainty on dividend policy in matured firms is greater than in non-matured firms

Research method
This study uses secondary data to analyze the listed firms of the Nigerian stock exchange market (NSE) between 2011 and 2015. This crucial period marks the final year of the ruling party since the inception of the fourth republic. It is also the start of the new political party that emerged from an alliance of parties. The sample covers nonfinancial firms listed on the main floor of the NSE. Financial firms were not included as their operations are governed by different regulation than other sectors, and they have different requirements in terms of dividend payout policy. Additionally, prior studies (Al-Najjar & Kilincarslan, 2016;Farooq & Ahmed, 2019;Huang et al., 2015) have excluded financial firms in their final sample.
The availability of information related to corporate governance and ownership needed for the analysis led to the selection of 89 firms and, hence, 445 firm-year observations from 2011-2015. Of the 445 firm-year observations, 250 observations paid a dividend during the period of the study, while 195 did not pay a dividend. The period of the study (2011)(2012)(2013)(2014)(2015) was selected because of the downward trend of dividend payment in the NSE market. Abdulkadir (2015) who posited that among the major challenges confronting the NSE market relates to the issue of non-payment phenomenon. An investigation into the history of firms paying dividend indicated that only 18 listed firms consistently paid dividends to their shareholders between September 2011 and September 2016 (Awoyemi & Bagga, 2016). From a survey, Nwidobie (2011) reported that dividend payment by listed firms in Nigeria is falling below the expectation of investor in the NSE market.
Therefore, these samples were used simultaneously throughout the estimations of the paper. The firm-specific characteristics used for this study were extracted from the Datastream database, while corporate governance and ownership variables were obtained from the annual reports of the firms under review. Last, information about the presidential election years was retrieved from Independent National electoral commission official website (INEC). Following AL-Dhamari, Ku Ismail, and Al-Gamrh (2016), Farooq and Ahmed (2019), and Huang et al. (2015), the model in the current study was estimated using ordinary least squares dummy variable fixed effects with robust standard error to account to control heteroskedasticity issues.

Variable measurement Dependent variable
The dependent variable for this study was dividend yield (DIY), which was measured as the ratio of dividend per share to the price per share for a firm (AL-Dhamari et al., 2016;Al-Najjar & Kilincarslan, 2016;Farooq & Ahmed, 2019).

Independent variable
The independent variable of interest in this study was political uncertainty proxied by presidential elections. Following Farooq and Ahmed (2019), the variable took the value of "1" for years in which presidential elections were held and "0" if otherwise. Baloria and Mamo (2017) argued that elections should be treated as an event that can change the existing environment in which firms operate and after that, provide an avenue for the rise of uncertainty on the outcome of future policy (Farooq & Ahmed, 2019).
Presidential elections in Nigeria are conducted after every four years, and this study covers only presidential elections 2011 and 2015. Retained earnings (RET) was used as a proxy for firm maturity. Following Francis, Hasan, John and Song (2011) retained earnings scaled by total capital. Prior evidence has shown that retained earnings are among the main features of a dividend-paying firm (for example, DeAngelo et al., 2006;Francis et al., 2011;Jiraporn et al., 2011).

Control variables
Consistent with the literature on dividend policy, this study used control variables that included return on assets (ROA) as a proxy for profitability, firm leverage (LEV) for indebtedness, sales growth (SGW) for firm growth, block-holders ownership (BLK) to control for ownership structures and as an indication of whether a firm was closely held or otherwise and total assets, (FZE) to capture the effect size of the firm and whether was is a large or small firm, and board size (BSZ) to control for corporate governance (Adamu et al., 2019;Al-Najjar & Kilincarslan, 2016;Farooq & Ahmed, 2019;Francis et al., 2011). ROA was measured as net income to total assets; LEV represented the total debt divided by total assets (Farooq & Ahmed, 2019;Francis et al., 2011). SGW was measured by current sales less previous sales divided by previous sales (DeAngelo et al., 2006).
BLK was the fraction of shares owned by owners of at least 5% shares of the firm scaled by total shares in issue (Huang et al., 2015). FZE was the natural logarithm of total assets to proxy for firm size (Farooq & Ahmed, 2019

Results and discussion
Descriptive statistics Table 1 reports the descriptive statistics of the variables used in the study. On the average, the dividend yield was 2.6%, which was higher than 1.9% as previously reported (Al-Najjar & Kilincarslan (2016) from Turkish firms and is less than 3.68% as documented by Awotundun (2018) while analysing the banking sector in Nigeria. The values imply that the shareholders of the sampled firms earned N2.60 relative to the market price of the shares of the firms.  Table 2 indicates the descriptive statistics for the three dummy variables (POL, RET and BLK) used in the study. Frequencies and percentages are used for interpreting the dummy variables. A total of 104 firm year-observations, (41.60% of the sample firms) represent the presidential elections of Nigeria, while 146 (58.40% of the sample firms) were period of non-presidential elections. A total of 119 firm year-observations (representing 47.60% of the sample firms) were having retained earnings higher than the within-sample median whereas 131 firm year-observations of the study (52.40% of the sample firms) were having lower than the withinsample median value. Thus, indicating that the number of matured firms fall below the number of growing firms in this study. Lastly, a total of 127 firm year-observations (50.80% of the sample firms) have block holders greater than the withinsample median whilst 123 observations of the sample firms (49.20% of the sampled firms) have less than the within-sample median. The Variance-Inflation-Factor (VIF) was also estimated, and the result is presented in Table 3. The results show that none of the variables was above the threshold of 10. The highest VIF in this study was retained earnings (RET of 1.54), which is below the upper boundary of 10. Hence, no evidence exists of a multicollinearity problem (Gujarati, 2004). Furthermore, the correlation matrix result is documented in Table 3. The correlation matrix between the pairs was relatively low and less than 0.6. None of the correlation coefficients was greater than 0.6, which may call for multicollinearity concerns. Therefore, the model does not suffer from multicollinearity issues. The variables of interest in this correlation matrix were political uncertainty (POL) and Retained earnings (RET). As can be seen, DIY was positively related to POL (r=0.158) and RET (r=0.383). The sign of the correlation coefficient of the matrix may suggest the probable direction of the relationship between the dependent and independent variable in the main regression equation.  Table 4 shows the total dividend per share paid during the presidential election and nonpresidential election years in Naira. As can be seen, the amount paid (N64.94) in the election year, for instance, 2011 which was a year of a national election in Nigeria, was higher than the amount paid in non-election (2012: N58.045 and 2014: N57.685) year except for 2013. The second presidential election captured in this study was 2015. When the amount of dividend paid in this year is compared with other non-election years, the amount paid (N81.03) was much higher than in other non-election years. Thus, this may also provide a clue to the fact that uncertainty surrounding the election period may impact the dividend paid by a firm particularly because the ruling party was forecast be the loser in the 2015 presidential election.  Table 5 shows the result regarding the prediction on the relationship between dividend policy, political uncertainty and how this relationship prevails in matured firms. The estimations were based on ordinary least squares with robust standard errors to control the effect of heteroskedasticity related issues. From Table 5 Panel A, column 1 is the result of only dividend payers during the study period while column 2 in Panel A is the result of both payers and non-dividend payers. The variable of interest in this current study was political uncertainty. Consistent with the hypothesis, the result shows that political uncertainty was positive and statistically significant in Panel A column 1 and 2. The positive and significant coefficient suggests that relative to non-election year (national election), firms pay more dividends to their shareholders This finding has several possible explanations. First, the result could imply that shareholders demand more cash dividends from firms during high uncertainty associated with the presidential election, hence, agreeing with the agency theory. Alternatively, the result could be that firms use a cash dividend payout to enhance the confidence of the investors' perceptions that uncertainty emanating from policy shocks may not hinder a firm from paying a cash dividend

Regression analysis
The result in this current study agrees with the findings of recent study (Farooq & Ahmed, 2019)that firms pay a higher ratio of their profits as dividends during the year of the national election. This study also concurs with the prior evidence (Buchanan et al., 2017;Choy et al., 2011) who found evidence on the increase of dividend during uncertainty in a tax environment. The literature also revealed that firms in a majoritarian system tend to pay more dividends than firms in a proportional system of election respectively. However, the results failed to agree with Awotundun (2018), who reported an inverse association between dividend policy and political factors among listed Nigerian banks.
For the second hypothesis, firms were divided into two categories based on retained earnings. These categories were high and corded as "1" if the retained earnings of a firm were higher than the sample median and low coded as "0" if otherwise. The regression result is also presented in Table 5, Panel B and C column 3-6. The result indicates that political uncertainty in high retained earnings firms was positive and statistically significant in Panel B column 4 and Panel C column 5. Hence, this was consistent with hypothesis 2 that matured firms pay more dividends than growing firms; therefore, the results confirm the life cycle theory of dividend policy.
The result, therefore, concurs with the prior evidence that matured firms have more tendency in paying a dividend as compared with firms with more growth option (Al-Najjar & Kilincarslan, 2016;Coulton & Ruddock, 2011;DeAngelo et al., 2006;Hoberg & Prabhala, 2009;Huang et al., 2015;Yarram & Dollery, 2015). However, it does not support earlier evidence (Abdulkadir, Abdullah, & Woei-Chyuan, 2015) that showed firms with more retained earnings are less likely to pay a dividend. The results of the current study also suggest that political uncertainty may not hinder matured firms from paying a cash dividend to their shareholders. They do so possibly to assure the existing investors or to attract potential investors because politics may not affect their dividend policy irrespective of the political party that is in power.

Robustness tests
Logistic regression was used to check for the robustness of the results. The firms were categorised based on their retained earnings. If the retained earnings were more than the median of the sample, then a firm was assigned a "1," representing a high dividend yield firm. If the retained earnings fell below the median score, then a firm was assigned "0," representing a low dividend yield firm. The results are reported in Table 6 Panel A to C.
The results show that the coefficient of the variable of interest POL was positively and statistically significant in columns 1 to 6, as reported in Table 5. However, POL in Panel C column 5 to 6 was positive but not significant. Consequently, the results suggest that the findings were not sensitive to the alteration of the dependent variable, as well as the estimation method.

Conclusions
The paper tested the association between dividend policy and political uncertainty, it also explored how this relationship influence matured and non-matured firms in Nigeria. Empirically the analysis indicated that political uncertainty significantly affected dividend policy, the effect is more pronounced in matured than non-matured firms. Accordingly, the findings provide support to prior evidence on political uncertainty and dividend policy and are consistent with agency theory and life cycle theory.
The findings of the current study may be useful to both existing and potential investors, particularly for those that favour a cash dividend against a capital gain in the NSE market. Also, mature firms are worth investing in because the national election in Nigeria did not affect these categories of firms.
The study may also be of benefit to regulatory bodies such as the NSE and Securities and Exchange Commission of Nigeria because the results show, that besides the traditional determinant of dividends such as corporate governance variables and firm-specific features, a national election is also an essential driver of a dividend that should be given more attention when designing policy concerning dividend pay-outs for listed firms in the country Lastly, future research may extend this study by capturing more years and dwelling more on the features of growth option firms. Also, other researchers may incorporate more countries from sub-Saharan Africa.