“Political risk and banking sector performance in Nigeria”

Political risk is prevalent in Nigeria and tends to influence business outcomes and the stability of the banking system. As a result of this study, it was determined whether political risk matters to the performance of the banking sector in Nigeria. The effect of po- litical risk on different banks’ performance measures, such as return on assets, return on invested capital, credit risk and stock price, were examined in a panel of 12 selected commercial banks for the period 2006–2018. Data was analyzed using a two-stage system of generalized method of moments. The results provided evidence that the effect of political risk on bank performance depends on the performance proxies. Specifically, political risk was found to be negatively related to banks’ returns on invested capital and positively related to deteriorating credit risk. Hence, it can be concluded that political risk induces poor banking system performance in Nigeria. The study provides a critical insight into the management of a country’s political systems in terms of their potential to create unfavorable conditions for banking systems to thrive. in the number of banks from 89 in 2004 to 25 in 2005. Four years later, inefficient credit risk management rendered numerous banks insolvent as 9 out of the 24 strong banks were classified as failed ones (Nigeria Deposit Insurance Corporation, 2009). Nigerian Deposit Insurance Corporation published another list of six failed banks in 2011 due to excessive credit risk taking. Marshal (2017) observed that the problem of failures continues despite several amendments to banking regulations. While the issues of bank failures are often blamed on bank management, excessive risk taking and nonperforming loans (Sanusi, 2009) and poor regulatory framework (Marshal, 2017) among others, Mark and Nwaiwu (2015) explained that government actions and the Nigerian political environment could affect business outcomes.


INTRODUCTION
The banking sector is the focal point of money markets that are used for short-term funds and investments. As a result, the sector is a key component of the financial system and plays a fundamental role in the development and growth of nations (Fapetu & Obalade, 2015). The relevance of this sector qualifies it as the most regulated sector in Nigeria's economy, as in many other economies (Adaramola et al., 2018).
In the course of history, the Nigerian banking system has experienced series of banking failures resulted in the liquidation of 51 banks from 1936 to 2000 (Marshal, 2017). The unfavorable financial situation of the Nigerian banking system necessitated the 2004 consolidation policy, which led to the reduction in the number of banks from 89 in 2004 to 25 in 2005. Four years later, inefficient credit risk management rendered numerous banks insolvent as 9 out of the 24 strong banks were classified as failed ones (Nigeria Deposit Insurance Corporation, 2009). Nigerian Deposit Insurance Corporation published another list of six failed banks in 2011 due to excessive credit risk taking. Marshal (2017) observed that the problem of failures continues despite several amendments to banking regulations. While the issues of bank failures are often blamed on bank management, excessive risk taking and nonperforming loans (Sanusi, 2009) and poor regulatory framework (Marshal, 2017) among others, Mark and Nwaiwu (2015) explained that government actions and the Nigerian political environment could affect business outcomes.
The Nigerian political environment within which the banking system operates has been relatively unstable and affected by widespread corruption, absence of transparency and inferior quality of infrastructures. The implication was that political risk constitutes a major component of country risk in the Nigerian banking system. For example, banking turmoil, which led to the revocation of licenses of five banks and takeover of 18 banks by the CBN between 1994 and 1996, occurred following an unstable political environment engendered by the cancellation of the 1993 Presidential Election. In this context, Brownbridge (1998) attributed banking distress in African countries such as Nigeria, Kenya, Uganda and Zambia to moral hazard arising from political interference. . Political risk is a component of country risk and may be viewed as ''the exposure that a company or bank faces due to political events that might affect its profitability'' (Mokatsanyane, 2016, p. 3).
Although Tamadonnejad et al. (2013) revealed that country risk and political risks lead to deterioration of bank efficiency in Asian countries, research on banking industry profitability performance in the context of political risk exposures has rarely been conducted. Given the prevalence of political risk in Nigeria, this paper contributed to the risk-performance literature on the banking industry by examining the impact of political risk component of country risk on different deposit money banks' performance measures in Nigeria.

LITERATURE REVIEW
Extensive research has been conducted on various risk factors and bank performance. A review of the concepts and relevant theories and empirical studies is presented in this section, taking cognizance of various risk measures.

Conceptual review and theoretical framework
Country risks are usually measured by risk ratings provided by rating agencies (international credit risk guide (ICRG, 2017). It represents the general risk inherent in the domestic markets of individual countries. Country risk is proxied by three components, namely the political, financial and economic risk (ICRG, 2017), and affects the success of investment in a country as it is non-diversifiable (Perry, 2017). The factors of the political risk include government stability, socio-economic conditions, investment profile, internal and external conflicts, corruption, military in politics, religion in politics, law and order, ethnic tensions, democratic accountability, and bureaucratic quality . The financial risk index measures the level of foreign indebtedness, exchange rate stability, debt service, current account and international liquidity, while the economic risk index incorporates inflation, GDP per capita, GDP growth, budget balance and current account as a percentage of GDP (ICRG, 2017; Muzindutsi & Obalade, 2020). Kirikkaleli (2020) showed that political risk possesses the ability to explain economic and financial risks. For example, political instability reduces policymakers' and government's tenures, thereby resulting in unsuitable and unstable economic policies that unfavorably affect the soundness of the economy and the financial sector. Consequently, this study focused on political-risk effect as a proxy for country risk.
Theoretically, the capital asset pricing model (CAPM) illustrates the correlation between market risk and expected returns for securities, especially stocks (Sharpe, 1964;Lintner, 1965 2005) showed that bank size has no significant influence on performance. In this context, the effect of credit risk on bank performance has been thoroughly investigated. Generally, the results have been conflicting.

Bank performance and political risk
Few recent studies introduced country risk variables. For example, Mokatsanyane (2016) revealed through an ARDL panel that the effects of credit and political risks depend on the measure of performance of banks in South Africa. It was found that political risk has no significant effect on ROE, ROA, EPS and NIM in the short term. The longterm effect was similar, except for the positive effects of political risk in the long term. In China, the deteriorating effect of oil price shocks on the performance of the banking industry can be ameliorated by stable country risks through the political, economic and financial components using GMM for dynamic panels ( (2013) revealed that country risk and political risks lead to deteriorated bank efficiency in Asian countries using a Stochastic Frontier Analysis in ten East Asian countries. It can be seen from the review that the evaluation of country risk effect on bank performance is limited in the literature.

Credit risk and political risk
A number of studies showed that bank credit risk or nonperforming loans are usually triggered by bank-specific and macro-economic factors. In this context, country risk via its three components is usually ignored, while attention is devoted to factors such as inflation and GDP. For example, Thiagarajan (2013) and Bozga and Trenca (2018) showed that inflation contributes to rising credit risk, while GDP leads to its decline in Belize banks and 13 developed European countries, respectively. The reducing effect of GDP on credit risk also applied to banks in Pakistan (Kasana, 2016)

METHODOLOGY
The research approach adopted for this study was quantitative, involving the analyses of the effects of country risk and certain bank-specific and macroeconomic variables pertaining to the performance of Nigerian banking firms. There are three sections of the methodology, namely the sample and data, model specification and estimation technique. The main performance measures were ROA, return on invested capital (ROIC) and stock prices (STOCK). Literature (Aliabadi et al., 2013) supported the use of ROA as an established and valued accounting performance metric by industries for market and accounting performance, while ROIC were estimated for robustness purposes. In addition to internal performance measures, this study also examined the reaction of banking stock market prices to political risk as country risk is external to the banks and its effect on bank performance might pass through the external or market performance measures. The explanatory variables were credit risk (CRISK), bank size (SIZE), political risk (PRISK), gross domestic product (GDP) and inflation (INF), which were found to exert significant impacts on bank performance. Bank and macroeconomic variables were sourced from Bloomberg, while stock returns sourced from S&P capital IQ.

Model specification
Examining the impact of country risk on bank performance in Nigeria, four different models were estimated. These were estimated by using different internal profitability and market (stock price) measures of bank performance and by evaluating the effects of political risk on credit risk. To achieve the objective, this study implemented a generalized method of moments (GMM) that accounted for the influence of probable omitted regressor and time persistence of performance indicators and is also robust to cross-sectional dependence and missing data that could be inherent in the nature of this study. The dynamic model is given as follows: where ϕ represents the dependent variables ROA, ROIC, CRISK and STOCK, which are the ratio of 1 Pooled-OLS, fixed effect and random effect are not reported due to their deficiencies. For example, they fail most of the model diagnostic tests. However, they can be made available to the reviewers on demand.
net profit to total assets, ratio of net profit to invested capital, non-performing asset/total loan, and logarithm of stock price, respectively; i and t are notations representing individual banks (i = 1, ..., 12) and time (t = 2006, ..., 2018), respectively; X it denotes a vector of the regressors where PRISK, SIZE, GDP and INF symbolize political risk index, log of total assets, growth rate of gross domestic product and consumer price index in that order; β is a vector of the estimated coefficients, while λ i denotes unobserved bank-specific effects, ε it represents the error term.

Estimation techniques
The above equations were estimated using the pooled-OLS, fixed effect and random effect 1 , as well as GMM models. The study depended on the two-stage system GMM as a main estimation technique for notable reasons. First, pooled-OLS is prone to bias and inconsistency (Arellano & Bond, 1991;Baltagi, 2008). Secondly, fixed effect suffers from dynamic panel bias associated with perceived endogeneity of the lagged regressand and the fixed effects in the ε it , notably with limited t and large i panels (Judson & Owen, 1999). Thirdly, two-stage GMM is an asymptotically more efficient estimator when compared to the one-stage GMM (Arellano & Bond, 1991;Mpofu & Nikolasidou, 2018). In contrast with difference GMM, system GMM applied in this study generates high precision estimates with limited t and large i panels. Subsequently, the Pearson correlation matrix was used to ensure that there is no multicollinearity or correlation between the explanatory variables.

RESULTS
This study estimated four different models to evaluate the role of political risk components on the performance of the Nigerian banking sector using panel data approach. This was achieved by using different measures of bank performance (ROA, ROIC, STOCK) and credit risk (CRISK) as dependent variables. This study presents the analyses of the system GMM after the presentation and interpretation of preliminary analyses. To reveal the extent of correlation amongst regressors, the Pearson correlation matrix is given in Table  2. It can be observed that the correlations between certain variables were negative, while others were positive; generally, the observed values were below 40%. Moreover, the choice of data analysis techniques is robust to multi-collinearity issues in case.

Model estimation results
The results (coefficients and standard error) of the 2-step system GMM for the four estimated models are presented in Table 3. From the first model, the results indicated that ROA is strongly but negatively influenced by its own lag. This was shown by the negative and statistically significant coefficient of lagged ROA. Similarly, the macroeconomic determinants, namely the GDP and INF, contributed negatively to the ROA performance of Nigerian banking firms. In this context, a percent change in GDP and INF contributed 50% and 312% reduction in ROA, while SIZE had no significant impact. This suggested that the inverse effect of INF was pronounced compared to the GDP. However, the bank-specific variables are not statistically significant  in explaining the ROA. It can be seen that the PRISK 2 had a positive effect on ROA at the 5% level of significance. As such, a one percent increase in PRISK led to a 7.5% increase in ROA.
Subsequently, the second model presents the effects of the PRISK and other dependent variables on ROIC. The result in the second column of Table 3 shows that the previous period ROIC significantly influenced the current period ROIC. Amongst the bank-specific factors, CRISK is a positive and significant determinant of ROIC, while SIZE is insignificant. In addition, macroeconomic variables had no significant influence of the ROIC performance of the Nigerian banking sector. The effect of PRISK was negative on the ROIC albeit significant at a 10% level of significance.
The third model is estimated to determine whether country risk proxy by PRISK contributes to the CRISK. The result in the third column reveals that the CRISK is positively and significantly driven by its previous values at a 5% level of significance. Between the bank-specific factors, SIZE of the banking sector leads to a reduction in CRISK such that a one percent increase in SIZE leads to a 2.67% reduction in CRISK. Between the macroeconomic factors, INF exerted and exerts a negative influence on CRISK, given its negative coefficient, while GDP is not a statistically significant determinant of CRISK in the Nigerian banking sector. Subsequently, the effect of PRISK is positive on the CRISK albeit significant at a 10% level of significance. This suggested that an 0.163 percent rise in credit risk was attributable to a one percent increase in the political risk index.
While ROA and ROIC respectively measure bank performance in terms of management proficiency in asset and capital utilization, the market-wide effect of country risk could reflect on the stock market prices. Consequently, the result in the fourth column is meant to solve this assumption. The result showed that banking stock prices reduce with the increase in its own lag. This is demonstrated by negative and significant coefficients of L. STOCK. PRISK is significantly positive at a 5% level of significance. Specifically, a 1% increase in PRISK leads to a 49.4% increase in banking stock prices. 2 As per the international country risk guide, an increase in risk ratings reflects decreasing risk (Erb et al., 1996).
In addition, the macroeconomic factors portend positive signs for the STOCK. As a result, banking STOCK increases as the GDP and INF increase at a 5% level of significance.   Note: Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1. The unfavorable effect of the political risk can be linked to the effect of macroeconomic factors. It can be assumed that the growth rate of the economy and inflation have deteriorating effects on bank performance. These factors tend to increase the cost of business operations for banks and their customers. This informs the positive effect of inflation on credit risk and suggests that inflation increases nonperforming loans for the Nigerian banking industry. As most of these factors are beyond the control of the bank management, they could be the major reasons for the prolonged banking crisis, despite various government interventions and capital regulations in the Nigerian banking industry. As illustration, extending credit when confronted with rising country risks could result in loan defaults. In addition, banks may find it difficult to recover previously extended credits.

Discussion
Subsequently, the study found that CRISK is not a significant determinant of (ROA) performance in the Nigerian banking industry. This finding is not supported by most of the existing studies that estab- The observed differences in the findings relevant to literature emanate from a different estimation technique. However, the preference for GMM employed in this study was well articulated in literature. It was found that credit risk stimulated bank return on capital, which supports the majority of the existing studies, although prior studies employed return on equity. This study revealed that the effect of credit risk is dependent on performance measure. The differences in the impact of credit risk on return on assets relative to return on invested capital is acceptable as assets are larger than capital. No bank performance measure is affected by the size of Nigerian banking firms. This finding was proposed in literature (Almumani, 2013;Kasana, 2016).
Additionally, this study found that stock prices are determined by the macroeconomic condition in Nigeria. GDP growth and inflation rates and political risk index all exhibit positive significant effects on bank stock prices. This suggested that these factors pose favorable effects on bank performance in stock market.

CONCLUSION
This study has examined whether political risk matters to bank performance in the Nigerian banking sector. It was found that the effects of country risk and macroeconomic factors on bank performance were mostly unfavorable. It was concluded that political and macroeconomic risks mostly reduce performance or worsen credit risk situations of Nigerian banking firms. Through this article, the first empirical evidence on how political environment induces poor performance in Nigeria's banking system was provided. In terms of policy recommendation, there is a need to create tranquility in the political arena and develop the nation's political architectures. Subsequently, a general increase in business and economic outputs is necessary to improve the performance in the Nigerian banking system, and, as a result, policy decisions should be focused in this direction. The study of the effect of country risk on performance can be extended to nonbanking financial institutions. Finally, the effect of unexpected event such as COVID-19 on bank performance provides motivation for future empirical research.