The Impact of Macroeconomic Factors on Profitability of Commercial Bank in the UK

Due to the vast contribution of commercial banks to the economic development in the UK, this study examined the effect of macroeconomic variables on profitability of Commercial Banks listed at the London Stock Exchange (LSE) for years 2015 to 2019. The objective was to determine the effect of macroeconomic factors on profitability. Panel data for the 23 listed banks in the LSE was utilized. Panel data regression analysis with random effects was utilized on the data to examine the effects of these five macroeconomic variables which included: Gross Domestic Product (GDP), Real interest rate, Inflation rate, unemployment rate, and Exchange rate on Return on Asset (ROA) and Return on Equity (ROE) which proxies’ profitability. The study findings indicated that real GDP growth rate had a negative significant effect on profitability of commercial banks as measured through Return on Assets (ROA) and Return on Equity (ROE). While the rest of the macro-indicators do not have a significant impact on profitability measured by ROA and ROE. It is therefore indicating that internal factors are impacting the profitability of banks in the UK rather than external factors.


Introduction
The banking industry in the UK has made a huge contribution to the economy by providing 1.6% of employment opportunities for the banking industry and more than 40% of the employment opportunities for the entire financial industry (Burgess, 2011).In addition, the banking industry's contribution to GDP is close to 3.7%, which exceeds 50% of the entire financial market (Burgess, 2011).Therefore, it is inherent to investigate factors that directly and/or indirectly affect the profitability of British banks.In this regard, this research aims to study the possible determinants of the profitability of British banks, with a particular focus on the period before, during and after the referendum of leaving the EU.
Due to Brexit, It is reasonable to assume that it poses a huge challenge to British banks because their business environment changes rapidly, which affects their performance.Therefore, The purpose of this study is to investigate the impact of macroeconomic indicators on profitability of commercial banks in the UK during 2015-2019 as commercial banks in the UK witnessed significant growth and changes after the referendum for leaving the EU of the UK banking industry.In the following literature, the study has found many researches on the determinants of the profitability of banks in other countries, while previous studies on British banks have focused on other aspects of bank performance.For example, Drake (2001) and Webb (2003) studied the efficiency of the UK banking industry.Holden and El-Bannany (2004) studied the importance of information technology development to the profits of major UK banks.Kosmidou et al. (2006) Analyze performance factors to identify salient features of British foreign and domestic bank profits.Other research on bank profitability also considers British banks as part of a larger sample aggregated in multiple countries (e.g.Molyneux and Thorton, 1992;Pasiouras and Kosmidou, 2006).
However, this research differs from the above research in three main aspects.First, this study only focuses on the profitability of UK banks.Secondly, we considered representative samples of large and small banks in the near term, thus providing more appropriate and up-to-date empirical evidence.Third, this study's empirical analysis separates the specific influence of internal factors from external factors (macro indicators), so that this study can investigate the impact of the macro indicators on the profitability of commercial banks.

Literature Review
The financial performance (Profitability) can be affected by internal and external factors.However, this study will only focus on external factors which are macroeconomic factors such as GDP, inflation, interest rate, unemployment rate, and exchange rate that affect the bank's profitability.In theory, real GDP growth has a positive impact on banking performance through three main channels: net interest income, loan loss improvement and operating costs (Jiménez et al., 2009;Bolt et al., 2012;Calza et al., 2006).During the period of economic expansion, the company's profitability will increase, and during the period of economic recession it will decline.Therefore, higher GDP growth will lead to an increase in corporate loans and deposits and improve banks' net interest income and loan losses.In addition, higher GDP growth means higher disposable income, lower unemployment rates and fewer consumer loan defaults.Therefore, net interest income and loan losses are cyclical with GDP growth.However, the relationship between bank operating costs and GDP growth is unclear.Bolt et al. (2012) indicates that unfavorable economic conditions, such as lower GDP growth rates, may reduce deposits and loans and their management costs.These conditions may also increase the cost of collecting loan repayments.Revell (1979) introduced the relationship between inflation and bank performance in the theory.He showed that inflation will affect the profitability of banks by affecting indirect expenses (especially wages and operating costs).If the inflation rate rises, it may increase wages and operating costs, thereby reducing the profitability of banks.However, if the bank's management fully anticipates the inflation rate, then the bank can adjust interest rates appropriately so that revenue growth is faster than cost growth, which will have a positive impact on profitability (Trujillo-Ponce, 2012).Samuelson (1945) studied the impact of interest rates on bank profits.The results show that, in general, bank profits increase with the rise of internal interest rates."Rising interest rates is an invaluable help to the banking system as a whole, rather than hindering it.Commercial banks will make more profits than savings banks" (Samuelson, 1945). Further, Samuelson (1945) pointed out that raising interest rates actually has an impact on borrowers, but it has no impact on bank performance.Borrowers will tolerate the impact of high interest rates, and the performance of banks will not be affected by high interest rates.Gelos ( 2006) used bank and country-level data to study the determinants of profitability of Latin American banks.They found that relatively high interest rates (which can replace macroeconomic risks) have a great impact on bank efficiency.As pointed out by Adler and Dumas (1980), the activities of banks are affected by exchange rates because asset values fluctuate with exchange rates.Chamberlain et al. (1997) showed that the most direct impact of exchange rate is those banks that conduct foreign exchange transactions and conduct foreign business.Even if there is no such activity, the exchange rate will be affected.Banks can be indirectly affected by affecting foreign competition, loan demand, and other aspects of bank conditions.The lower exchange rate improves the competitiveness of the company, because domestic ex-factory prices of goods fall and foreign demand rises (Luehrman, 1991).As a result, loans and deposits and bank profits have increased.However, a lower exchange rate may also reduce the purchasing power of domestic consumers as imported goods become more expensive.This situation may increase loan losses and may have a negative impact on the bank's profitability.Economic factors such as financial crises could affect the unemployment rate.Tanveer et al., (2012) proved that the financial crises in 2007-2008 has impacted on the unemployment rate in different groups of countries.Increasing unemployment has an impact on the decrease of income per capita which results in bank profitability.
On the other hand, macroeconomic factors are determinants that are related to economic, industrial and legal environments that are out of a bank's control (Ongore & Kusa, 2013).Macroeconomic determinants comprise variables such as GDP, inflation rate, interest rate and exchange rate (Acaravci & Calim, 2013;Chowdhury & Rasid, 2017;Jara-Bertin et al., 2014;Marijana et al., 2012;Masood & Ashraf, 2012;Menicucci & Paolucci, 2016;Saona, 2016;Pasiouras & Kosmidou, 2007).Although extensive research has been conducted on banks' profitability determinants in different countries, comprehensive empirical evidence from emerging and developing countries have either yielded ambiguous evidence or mixed results (Almaqtari et al., 2018).With regard to banks' profitability determinants studies in the UK context, there are not too many studies that investigate this issue.Kosmidou, Tanna, and Pasiouras (2005) have examined bank-specific and macroeconomic factors that determined the profitability of UK banks.The results revealed that macroeconomic factors such as GDP, and inflation rate have positive relation with ROA.Accordingly, the current research aims to investigate the impact of Macroeconomic factors on UK commercial banks' profitability.More specifically, it empirically evaluates only macroeconomic determinants that may have an impact on UK commercial banks' profitability as measured by ROA and ROE.The present study bridges a serious gap in UK banks' profitability literature.Furthermore, the current study extends, contributes and builds on the work of Kosmidou et al. (2005) who ignored a major proxy of banks' profitability namely; Return on equity (ROE).The current study also uses different econometric techniques for analysis of data, which give more sound results.

Research Methodology
The main research question will be answered by using deductive research approach, descriptive research design, and secondary data collection which will be supported by panel data regression analysis.The full dataset will contain strong balanced panel data on 23 banks for the period 2015-2019.The main research question will be estimated by a mostly used panel data regression model such as Pooled OLS, Fixed effects, and Random effects model for appropriate statistical decision.In terms of getting appropriate results, the whole data will be analyzed by different kinds of diagnostic tests such as normality test, multicollinearity test, and Heteroscedasticity Test.Furthermore, the Hausman test will be run to choose between fixed effect and random effects if pooled OLS is not used.
In regression analysis ROA and ROE profitability measures will be used as dependent variables and GDP, inflation, interest rate, exchange rate, unemployment rate as independent variables to evaluate their impact on each measure of profitability.Most importantly, this research ignored the net interest margin due to data absence of sampled banks.
The sample period is from 2015 to 2019.This study implemented an analysis of the year 2015-2019 as there are a number of studies that have been done up to the year 2015.Since 2015, there have been lots of changes such as Brexit, technological innovation, and economic downfall which directly and indirectly impacted the UK banking sector.

Analysis and Discussion
This chapter introduces the data analysis, interpretation and discussion of the research results.After analysis and interpretation of the data collected from secondary sources, it is in line with the purpose of the research, that is, to determine, examine, evaluate and determine the impact of macroeconomic variables on the 23 listed commercial banks on the LSE.This chapter will cover descriptive statistics, Correlation analysis, normality tests, diagnostic tests and panel regression analysis.
Summary Statistics:The descriptive statistics for the dependent and independent variable for the chosen studies.Descriptive statistics can provide initial indications of variables, can be used for regression analysis, and provide multiple summary statistics about variables, such as mean, standard deviation, and lowest and highest observed values.

Table 1:Descriptive Statistics
The descriptive statistics in Table 1 show that the total number of data analyzed (n) is 115, representing 5 years of data from 2015 to 2019 for 23 listed banks on the LSE in the UK.According to Gujarat (2003), the standard deviation represents a measure of the dispersion from the mean, which indicates that the smaller the standard deviation, the less variability, the more accurate the future prediction may be.In Table 1, the results show that the bank's profitability, return on assets (ROA), and Return on Equity (ROE) fluctuations do not differ much from the mean; however, both variables have much more standard deviation than mean.Therefore, the future prediction will not be accurate.
GDP growth rate, interest rate, inflation rate, exchange rate, and Unemployment rate fluctuations are not much different from the mean, and the standard deviation of the variables is small, so the future forecast is more accurate.The average ROA is 1.277 and the standard deviation is 1.796, and the average ROE is 10.617 and the standard deviation is 15.412 which means that the data are not clustered around the means.The average GDP is 1.801, the standard deviation is 0.381, the average interest rate is 0.559, the standard deviation is 0.218, the average inflation rate is 1.593, and the standard deviation is 0.814, the average exchange rate is 1.332, and the standard deviation is 0.081, and the average unemployment rate is 4.459, and the standard deviation is 0.538.Moreover, the table shows that the mean of each variable is between the minimum and maximum values, which enhances the accuracy of the data considered.
Correlation Analysis: According to the statistical data in Table 2, most variables are weakly correlated or negatively correlated with correlation coefficients less than 0.55, but among the dependent variables, ROE and ROA are highly correlated with the value 0.796.And the correlation between macroeconomic factors; GDP growth rate and Unemployment rate, and Inflation rate and Unemployment rate are highly positive and negative correlated respectively, with a value higher than 0.80 which could possibly create a problem of multicollinearity in the regression analysis, as Kennedy ( 2008) pointed out that when the correlation value is higher than 0.80 or 0.90, the correlation is high.Bryman and Cramer (2001) believe that multicollinearity is obviously possible when the correlation exceeds 0.80, while Anderson, Sweeney and Williams (1990) use 0.70 as the standard point.By considering all the views, there is multicollinearity between two of the dependent and independent variables, which may change the results.In order to obtain accurate output, one of the variables could be removed from the model if needed.

Table 2:Correlation Analysis
Normality Test: According to Gujarati (2003), if skewness lies between -3 and +3 and kurtosis for all the variables are positive then it is implied that the variables have normal curves with lighter tails.The below skewness/kurtosis analysis shows that skewness of all variables is between -3 and +3, and all the variable values of kurtosis are positive which indicates that data are normally distributed.

Table 3:Skewness/Kurtosis Test
Multicollinearity Test: Gujarati (2003) mentioned that the VIF will be 1 if there is no collinearity between two independent variables.And if the VIF >10 then the chances of multicollinearity are very high.In the below table, the VIF is 8.49 which is less than 10 and it indicates that there is no risk of multicollinearity for these sample variables.

Table 4:Multicollinearity Test
Heteroskedasticity Test: Gujarati (2003) stated that heteroscedasticity is present if the p value is less than a significance level of 0.05.However, according to the below table result in this study heteroscedasticity is absent as p value is 0.2165 which is greater than significance level of 0.05.

Table 5:Breusch-Pagan test
Panel Data Regression Analysis:There are several estimation methods in panel data.The most general and frequently used panel data methods are Pooled OLS, Fixed effects, and Random effects model.All the three models are being estimated for comparison reasons.In terms of determining the appropriate model all three model results will be examined to make statistical decisions.First of all, the P value will be considered for all the models then Adj.R-squared followed by F statistics and P value of F statistics.
Results of Model Estimation:The below table illustrates the estimation for all the three models.The three models are Pooled, Fixed, and Random models and the model provides the result differently while considering the dependent variable such as ROA and ROE.However, the results reveal that the Adj.Rsquared for ROA is -0.009 for Pooled OLS model while 0.03 is for both fixed and random model.This indicates that the macroeconomic indicators contribute about 0.9% by considering pooled OLS while 3% and 3% contribute by considering fixed effects and random effects model respectively to the profitability as measured by ROA.In terms of p value of F-statistics, the random model and fixed model show the significance which indicate that these models are appropriate for this data as ROA dependent variable, on the other hand, pooled OLS shows insignificance value that suggest inappropriate models for this data.Moving to coefficient's significance, the result of all three models show that in pooled OLS models none of them macroeconomic variables are significant while in random and fixed effects models have at least one variable significant.This also indicates that random and fixed effects models are more appropriate compared to pooled OLS models.
In terms of ROE result as dependent variable, the adj.R-squared of ROE is -0.01 for pooled model, 0.03 for fixed effects model, and 0.03 for random effects model.This signifies that the macroeconomic indicators explain approximately 1% , 3%, and 3% in pooled OLS, fixed effects, and random effects models respectively of the variation of the bank's profitability as measured by ROE.In terms of P value of coefficient for all the models, both fixed effects model and random effects model have at least one variable significance at 10% level, while in pooled OLS model does not have significance variable.This result indicates either fixed effects model or random effects model is appropriate for ROE.In terms of p value of F-statistics, the result shows all the models are insignificant however, random effect model and fixed effects model result are much better than pooled OLS model.Therefore, the result suggests choosing between fixed effects and random effects models.To compare and evaluate between fixed effects model and random effects model, the Hausman test will be conducted.

Table 6:Model Estimation Results Summary
Hausman Test:The Hausman test was developed by Hausman (1978) and it has a null hypothesis that random-effect specification procedures have a consistent coefficient.The results of the test are outlined below.

Table 7:Results summary of Hausman Test
The results show that the value of the Prob>chi2 is 0.99 and 1.000 for ROA and ROE respectively which is greater than the significance level of 0.05.Therefore, the result indicates that the null hypothesis would be more appropriate.The null hypothesis of the Hausman test means that the random-effects model is more appropriate.Thus, the panel data regression with random-effects model is preferred in this study.
Random Effects Model Results:As per Hausman test, the random effects model results displayed below which were analyzed by the econometric software STATA.

Table 8:Results of Random effects Model Empirical Result
From the result table, the adj.r-squared is 0.03 and 0.03 for ROA and ROE as dependent variables respectively which indicates that 3% variation in ROA and ROE were explained by independent variables.Further, the F-statistics were included to see whether all the coefficients in the model are different from zero.If the prob.F statistic <0.05 which indicates the model is appropriate.As a result, from the result table, it is shown that the prob.(F) is 0.01 with ROA which is less than 0.05 and suggest that the model is appropriate while with ROE the prob.(F) is 0.12 which is greater than the significance level of 0.05 and suggests that the model is not appropriate with ROE.However, by comparing all the models, a random model is more appropriate compared to other models.Moreover, the result discovered that 1 unit increase of GDP decrease ROA by 2.03, 1 unit increase in Inflation increase the ROA by 0.40, 1 unit increase in interest rate increases ROA by 0.85, 1 unit increase in unemployment rate increases ROA by 2.42, and 1 unit increase in exchange rate increases ROA by 12.05.if all the factors are held constant ROA will decrease by 15.89.While, 1 unit increase of GDP decrease ROE by 18.32, 1 unit increase in Inflation increase ROE by 2.96, 1 unit increase in Interest rate increase ROE by 1.82, 1 unit increase in unemployment rate increase ROE by 18.54, and 1 unit increase in Exchange rate increase ROE by 77.16.if all the factors are held constant ROE will decrease by 101.97.
In terms of interest rate, the results found insignificant positive relation with ROA and ROE.This suggests that the interest rate did not impact the UK bank's profitability.This outcome is consistent with Vong and Chan (2008), and Mamatzakis and Remoundos (2003).On the other hand, Anwar and Herwony (2006), Suffian and Habibullah (2010), Alper and Anbar (2011), Zhang and Dong (2011), Kanwal andNadeem (2013), andObillo (2015) found significant relation with profitability.
Moving to the unemployment rate, the result indicates insignificant impact on profitability measured by ROA and ROE but positive relationship with profitability.This finding is supported by Singh andSharma (2016), Gordijenkaite (2015), and Dewi et al. (2019).On the opposite, Ab. Bari (2018) and Messai and Jouini (2013) found significant impact of unemployment on profitability on their studies.
In terms of exchange rate, the finding of this study indicates an insignificant positive effect on both profitability measurement ratios of ROA and ROE.This indicates that exchange rate fluctuations did not influence the bank profitability in the UK.This result is supported by Manyo et al. (2016), andKiganda (2014).On the other hand, the study by Abebe (2006), Babazadeh and Farrokhnejad (2012), Getachew (2016), Casey et al. (2014), and Maigua and Mouni (2016) found significant impact on profitability.
The study investigated the effect of macroeconomic factors on profitability of the 23 listed commercial banks in the London Stock Exchange for a 5-year period from 2015-2019.The study specifically sought to determine, examine, evaluate and establish the effect of real GDP, interest rate, inflation, exchange rate, and unemployment rate on bank profitability.The effect of macroeconomic factors on bank profitability was examined using panel regression analysis.The regression results show that macroeconomic factors have an insignificant effect on bank profitability in the UK except GDP which show the negative significant impact on profitability.

Conclusion
The present study aimed at investigating the impact of macro-economic indicators on profitability of commercial banks in the UK.The macroeconomic variables that were considered in this study included real GDP, real interest rates, inflation, exchange rate (US dollar/ GBP), and unemployment rate.As per the first objectives of the study to determine the concept of the macroeconomic indicators and profitability.This research has fulfilled the first objectives by determining the concept of macro-indicators and profitability.The study discussed leading and lagging indicators.Jeang (2010) enunciated that lagging indicators only start showing some change through them when the economy has predominantly started following a certain pattern of change or direction; the lagging indicators are more precise than the leading indicators, they portray the change of the macro-economy in a better way.Lagging indicators are included such as GDP, inflation, exchange rate, and unemployment rate.On the other hand, Lee (2010) articulated that a leading indicator is any type of macro-economic factor that changes in a direction before the rest of the economy begins to go in a particular direction.Breeden (2013) enunciated that the leading indicators help the participants in the market to observe the future trend of the market and lets policymakers predict significant changes in the economy based on which the policymakers can decide on their policies.Leading indicators included such as Interest rate.Further, this research discussed the importance of profitability, measurement of profitability.The profitability could be measured such as ROA, ROE, and NIM.However, this study has conducted ROA and ROE as profitability measurement.
In terms of second objectives, the study has reviewed the effect of macroeconomic indicators on profitability through previous research.However, the findings of the studies were mixed.In majority of researches GDP was found as a significant variable such as (Bourke (1989); Molyneux and Thornton (1992); Athanasoglou et al. (2008); Bashir (2003); Zang and Dong ( 2011)).Rest of the macro-indicators were found to have positive relation with profitability except a couple of studies such as Zeitun (2012), Otuori (2013), Suffian andChong (2008), andSaksonova andSolovjova (2011) were found to have negative relation with profitability.
Moving to the third objective of developing the methodology by using appropriate models for the sampled data.This research has estimated all the panel data regression models to justify the goodness of fit the data.Also run a couple of diagnostic tests to check the data.Further this research has conducted a Hausman test to choose between random and fixed effects models.Finally, the random effects model was applied to data which was obtained from 23 banks financial statements from 2015-2019.Through the random effects model, the study explored the effects of macro-indicators on profitability of commercial banks in the UK.
Finally, the study concludes that GDP has a significant negative effect on profitability of commercial banks in the UK measured by ROA and ROE.Which indicates a negative contribution to profitability.However, rest of the sampled macro indicators do not have influence on bank profitability.This indicates that the macro-economic indicators do not impact on profitability of commercial banks in the UK.
The study was faced by a number of limitations.First, only 23 commercial banks are listed in the London Stock Exchange(LSE) against a total of 156 banks in the banking sector in the UK.Secondly, this study made use of ROA and ROE as measures of financial performance.There are other measures of financial performance including NIM (Net Income Margin) among others.Lastly, the study was specific to the UK and therefore suffers from the limitations of country specific studies thus cannot be generalized to banks in other countries other than the UK.For future research, this study can be extended to cover longer time periods.Unbalanced panel data can be used to incorporate the banks which are recently established.Quarterly data can be analyzed to reveal more precise results.Other econometric techniques can be applied to verify the relationship.More macroeconomic factors such as CPI, imports, exports, tax rates and income level can be focused on.In addition to domestic banks, foreign banks may be included in the sample.Furthermore, listed banks in different countries or a group of countries can be evaluated.