Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 Stock Return Predictability through Financial Ratios (Financial Sectors of KSE) Anas Aftab1 and Kiran Naseer2 The aim of the current study is to examine the predictability of the return of common stock of the financial sector of KSE (Karachi Stock Exchange) through financial ratios. By applying fixed effect model, findings of the study of 31 financial companies listed in Karachi Stock Exchange from 2003-2012 is showing that dividend yield, earning yield and earnings per share has significant relationship with stock returns .while return on equity and equity to asset ratio is insignificant. Keywords: Stock Return, Financial ratios, Listed Financial Companies, Karachi Stock Exchange. 1. Introduction Financial statements are imperative to analyse the financial position and performance of firms. Financial managers design financial statement in such a way that can reflect almost all aspects of the enterprise. These statements consist of both types of information’s i.e. Long run information and short run information. And there are many studies which scrutinize these information for financial decisions. One important approach to evaluate common stocks is based on analysis of financial statements. Financial statement analysis has conventionally been seen as piece of the basic analysis required for common stock or equity valuation. Albeit financial reports information is used in some researches to predict firms’ prospect monetary performance, such as earnings and growth[24].Where as other researches explain the impact of financial reports data on share price[15]. Financial ratios are designed keeping in view the level of information available in financial statements. Recent studies have shown that accounting ratios can be used to evaluate the future performance of firms. Investors can be facilitated through financial ratios analysis to make their investment decisions and predicting firm’s stock returns. The predictability of stock return can contribute to attaining the highest return with the lowest risk, so it has become foremost themes for worldwide investors. In previous centuries it was quite difficult to predict stock returns. This was difficult because the level of market efficiency was not clear. In contrast, several research studies documented the predictability of stock return based on different predictors. Researchers try to find out most accurate variables for predicting stock prices. Financial ratios also became the hot spots of various stock return researches, as indicated in the work of [17][5][13][20]. Some researchers were tending towards financial and some were towards profitability ratios i.e. Book to market ratio, price to earnings ratio, dividend yield. Meanwhile some researchers have used cash flow ratios like price to cash flow ratio, cash burn ratio, and some focused on macroeconomic variable like interest rate, law and order situation, foreign exchange and inflation rate etc. ______________________________________________________________________ 1 2 1 Anas Aftab ; Kiran Naseer , Facilities & Fleet Officer, British American Tobacco, Pakistan 2 Department of Management sciences, SZABIST Islamabad, Pakistan 1 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 Albeit many studies are carried out to predict the stock returns of the different markets of developed as well as developing countries using financial statement information, but unfortunately financial institutions have been excluded in most of the studies. Because these institutions have special nature of the last few decades Pakistani financial institutions have been headed by structural changes that are different than non financial sector. These changes occurred due to the external environment variation, especially as a ramification of the ample global financial regulations and monetary policies. The current study explores the power of financial ratios, to predict common stock return of financial companies listed on Karachi Stock Exchange. There are ample of studies that have explored the stock return predictability using financial ratios in the developed as well as in the emerging economies. Some studies investigate the non financial sector while few have focused on the financial sector stock return predictability by using financial ratios. This study will help investors to identify that which ratio has more explanatory power for common stocks. Moreover, According to my finest information, it is the only study in Pakistan which uses the set of financial ratios to predict stock returns of financial sector of Karachi Stock Exchange “KSE”. The aim of the study is to explore the strength of financial ratios to predict stock returns of financial sector of KSE. This study is focusing on only few ratios. Moreover, this study is only predicting the stock returns of financial sector. Future studies can test volatility and long run relationship between accounting ratios and stock returns. The research is delimited due to time constraints to just investigate the role of few financial ratios to predict stock returns of (financial sector) listed companies of KSE. II. LITEREATURE REVIEW In literature there is strong evidence of influence of financial ratios in the prediction on stock returns .There are a number of researches which test the market of different developed and underdeveloped countries. The main studies of developed economies [5][6][27][13][14] find the connection among financial ratios and stock returns. In developing economies studies are also conducted to unearth the association between various ratios and stock returns. Most of the stocks do not have their market value same to its intrinsic value. They are either overvalued or undervalued. The difference in prices of common stock is because of number of factors. The speculative nature of market, availability of information to all investors, manipulation, incomplete disclosure of data and non constant performance by firms are key factors in variation in stock prices [19]. Use of inappropriate tools for evaluation by investors is also key factor in the variation in prices of common stocks. Another study was conducted by[12] also used different ratios to find out the intrinsic value of common stocks. Accounting number has the ability to define firm’s value directly. This ability of financial ratios makes it use more prominent because of pedagogy and its practicality. Financial statements are considered as fundamental tool that can be used to forecast return of common stocks. Financial statements are imperative to discuss firm’s conditions. There are number of studies which examine the financial reports from both paradigms i.e. both annual and interim reports like quarterly, semi-annually and monthly). Even though the use of information from 2 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 financial reports is increasing to predict financial performance of the firm [24], but still other researchers are using these information for common stock evaluation [15]. Study of literature shows that, there is strong support of financial ratios in the predictability of common stock future returns. Evidence to support the use of financial ratios can be found in both developing and developed countries. The notion of stock return dates back to 1950 when the dire need for developing mathematical & statistical models spruced up numerous aspects of the business. One of the subjects of these models is the advent of the portfolio theory, proposed by[10]. Later studies on stock returns were derivations from the Markowitz model. An extensive time series literature was written by [22] to use of financial statement and accounting ratios to predict return of common stock. Financial ratios play crucial roles in predicting stock returns. Literature on the predictability of stock returns shows that among the financial ratios aforementioned ratios have a vigorous theoretical background supported by predictive models. Investors expect from their investment returns that can be either in the form of capital gains or dividend yield. On the basis of this phenomenon the exhaustive ratios which are included among the set of variables to predict stock returns are Earning Yield (EY), Dividend yield (DY), Return on equity (ROE), Equity to asset (EA) and Earning per share (EPS). The previous theoretical and empirical literatures indicate that these three financial ratios are more significant and important on stock return predictability that encompasses a wide range of prediction. These ratios consist of specific characteristics. A . Dividend Yield ( DY) and Stock Return Earlier study shows, return of common stock has a positive relationship with a DY “dividend yield” in cross-section & time-series respectively. The dividend yield has projecting influence for a cross-section return of the common stock [2][27][11]. While many studies also delineate the association between stock returns and dividend yield of time series.[16][5][6][27][21].[1] in their studies tries to forecast interest rate and stock returns with the help of predictive power of dividend yield. The result shows that in the short run dividend yield forecasting is more than the long run. The DY “dividend yield” is measured as dividend per share on market price per share. The following formula demonstrates how to calculate dividend yield: Dividend Yield (%) = (Dividend per Share divided by Market rate per share) x 100 B. Return on Equity (ROE) Return on equity is measure how efficiently any company will use its assets to generate earnings. The relationship between the company's profit and the investor's return makes ROE a particularly 3 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 important metric to be scrutinized. Existing literature findings on return on equity power to predict stock return shows mixed results. [18] in his study test the association between return on equity and stock prices of Jordon stock exchange result of his study reveals that return on equity separately does not have power to predict stock prices. While when it is combined with other ratios then it has power to predict stock returns. Return on Equity = (Net Income After Tax/share holders Equity) C.Earnings Yield ( EY) with Stock Return Approximately seventy year before, the ratio of price to earnings (the reverse of earnings yield) was used to predict stock returns of that era[3].Price to earnings ratio is considered as renowned measurers of valuation today. The empirical literatures put the fundamentals of the predictive power of EY on stock return, the association between stock return and the earnings yield is significant, because the earnings yield plays as a risk factor in relation with stock return. Besides, the earnings yield can also show the efficiency of a market.[26] investigates the relationship between earnings yield, market value and common stock return for NYSE. The finding of the study reveals that on average the common stock of high earnings yield firms earn higher risk-adjusted returns than the common stock of low earnings yield firms. A study conducted by [28] unearth that earnings yield has the positive impact with stock returns in stock market of Malaysia. [16][21] argue that earnings yield has independent predictability of stock returns in addition to the dividend yield.[25]use cross sectional stock return predictability of Philippine Stock Exchange finding of the study shows that earning yield has predictive power on the stock returns. Earnings yield will be measured as Earnings Yield (%) = (Earning per Share / Market rate per share) x 100 B. Earnings per Share(EPS) Every organization is intensely interested to maximize its earnings in order to give maximum return to its shareholder. Another objective behind it is to achieve long term growth of business by getting additional funds. Although the researcher has included it in the study, because fluctuations in financial institution's earnings lean to be less severe over time. It is so because of the capacity of financial institutions to protect earnings with the reserves of the loan to lose, than non-financial sector, so the shock of the above mentioned variable has crucial impact on return of common stock. [25] in his study uses different financial ratios to forecast return of the common stock of both financial as well as in the non financial sector. The result of the study unearths that EPS plays a very important role to predict return in both sectors of Philippine stock exchange. [9] 4 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 investigate the importance of earning per share EPS in Taiwan stock exchange and analyze the long run relationship between EPS and predictability of stock returns. The finding of his study reveals that EPS has forecasting power to predict stock returns.[7] in his study test the forecasting power of ratios in TSE “ Turkey Stock Exchang”. Result of the study is contradictory EPS can predict 6% the current year stock return while same EPS has 63% power to predict the one year later stock return In the current study EPS “earning per share” will be calculated as follows; EPS= Net income After Tax/ Number of Outstanding Shares E. Book value of equity to total assets (EA) One of the reasons for financial sector ignorance in the previous studies of common stock return was the difference of leverage structure. In the current study leverage ratio is used to know whether change in leverage will hold information to change stock returns of financial sector. To find out the strength of the capital structure the equity to asset ratio is used. This ratio is used because it captures the financial soundness of financial sector and tells us about their capital sufficiency. In previous studies leverage is used as an important factor to explain the performance of financial sectors of the stock markets.[4] in their study find financial leverage as an important factor to forecast financial institutions’ returns and to find their risk as well. More explicitly, [23] finds a significant positive relationship between capital ratios & return of the common stock for financial companies. He also delineates several important methods used to increase capital ratios; increases in earnings are linked with the biggest stock price increases. In the current study equity to asset is measured as Equity to total asset = Total Equity / Total Asset. In Sum, the existing literature presents a comprehensive explanation about common stock returns predictability and financial ratios. Nonetheless, few findings are contradictory due to several reasons. Some of them are, change in datasets, stock exchange markets and due to difference in the country’s environment. Many researches are conducted around the world about stock return predictability through use of financial ratios but, inadequate work has been done on financial sectors to test the power of these ratios to forecast common stock return[25][7]. The current study concentrates on the gap identified in the literature and test the relationship among financial ratios and stock return of financial sectors in context of Pakistan, by using econometrics tools. 5 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 III THEORETICAL FRAMEWORK Hypotheses Based on the above theoretical evidences from literature, following five hypotheses are formulated A. H1: Dividend yield is associated with stock returns. H2: Return on equity is associated with stock returns. H3: Earning Yield is associated stock returns. H4: Earning per share is associated with stock returns. H5: Book value of equity to asset is associated with stock returns. B. Model In the present study the data which is used is panel data. Panel data has joint effect of times series and cross sectional data. Common effect model and fixed effect model has been used. After applying the HAUSMAN test best model is selected. The equation is as follow: SRit = α0 + α DYit + α ROEit + α EYit + α EPSit + α ETAit + u Where; SR = Stock Returns DY = Dividend yield ROE = Return on equity EY = Earning Yield EPS = Earnings per share ETA = Equity to asset ratio IV METHODOLOGY To test the relationship among financial ratios and stock returns of financial institutions various econometric techniques are applied. Separately, descriptive statistics’ is applied on independent and dependent variables. Correlation is also applied and the assumptions of panel data are also fulfilled. V ANALYSIS OF DATA Descriptive Statistics Table(I) SR DY 0.025 0.104 Mean EY 7.622 EPS 9.344 0.802 20.007 21.044 ETA 9.519 SD 0.201 Min -0.714 .005 -.598 -196.28 -47.58 3.25 Max 1.117 1.515 13.81 200.15 196.52 14.052 310 310 310 310 310 310 N 0.132 ROE 0.226 6 2.513 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 Table 1 delineates the descriptive statistics of stock returns and five independent variables for 31 listed financial companies for period of 2003- 2012. Value of arithmetic mean of stock returns (SR) is .025. In the same way dividend yield has mean score of 0.104 with std. deviation of 0.132 and return on equity has 0.226 averages with deviation of 0.802. Arithmetic means of earning yield, earnings per share and equity to asset ratio are 7.62, 9.34 and9.51 with deviation of 20, 21 and 2.51 respectively. B Correlation among Variables Table (II) SR DY ROE EY EPS SR 1 DY -.182 1 ROE .083 -.062 1 EY .146 -.070 .059 1 EPS .171 -.185 .112 -.018 1 ETA .021 -.319 .006 .010 .161 ETA 1 To find out the association among the abovementioned set of variables, Pearson correlation is used. Correlation finding is very fruitful to discover Multicollinearity among all variables. It is an an important assumption of linear regression. Several studies have cited in their findings that if the correlation among all variables is 0.90 or more, it would be the major reason of Multicollinearity. Table 2displays the relationship amongst all variables and it directs that there is no any issue of multicollinearity. C Linear Regression Model (Common Effect Model) Table(III) Coefficients Standard Error t Stat Intercept .067 .048 1.35 DY -.243 .089 -2.71* ROE .012 .013 0.88 EY .002 .001 2.45* EPS .0013 .001 2.60* ETA -.0044 .004 -0.95 R Square 0.07 Adjusted R Square 0.06 F Statistics 5.09* Observations 310 *Significant at the level of 5% .Heteroskedasiticity Test Ho = There is no Heteroskedasiticity H1 = There is Heteroskedasiticity BP / CW test: 7 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 Chi 2= 0.76 Prob = 0.38 Table 3 shows the analysis of linear regression or common effect model. According to these results three hypothesis are accepted. Dividend yield is showing the negative but significant relationship with dependent variable which is stock price. Means by increasing one unit of dividend yield it will decrease 0.234 unit of stock return and this decrease would be significant. Earning yield and earnings per share are also showing significant relationship with stock returns and this relationship is positive while return on equity and equity to asset ratio is insignificant at level of 5 %. The current model is linear regression model which is also called common effect model. In case of panel data there are some assumptions of common effect model which should be checked. The most important is Heteroskedasiticity. According to the chi square and probability test the value is 0.38 which is insignificant. It leads to reject H1 and shows that there is not any problem of Heteroskedasiticity in current data. Further fixed effect, random effect model and Hausman test for best model is also estimated. In panel data most researches has shown that fixed effect model is appropriate instead of simple linear model or common effect model. D. Results of Fixed Effect Model Table(IV) Coefficients Standard Error t Stat Intercept .117 .293 0.40 DY -.405 .107 -3.78* ROE .001 .014 0.07 EY .002 .005 3.37* EPS .0022 .005 3.59* ETA -.009 .031 -0.29 R Square 0.13 F Statistics 7.90* Observations 310 *Significant at the level of 5%. Hausman Test for best Model Ho = Difference among coefficients are not systematic (Either fixed or random or common or between) H1 = Difference among coefficients are systematic (Fixed Effect) Chi2 = 15.22 Prob = .0095 E Discussion For Panel data estimation especially when all models in which common effect, fixed effect, random effect and between fixed and random model are estimated the important thing is to choose the best among all models. For this purpose Husman test is estimated. Null hypothesis of Husman test is that there is not systematic difference among coefficients which means either fixed effect or random effect; the result will show the same analysis. Alternative hypothesis is that there is systematic difference among models and fixed effect is best choice for analysis. Significant value of chi square is showing that null hypothesis is rejected and alternative accepted and it means fixed effect model is best choice among all models for estimation of this study. 8 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 Table 4 shows the fixed effect model. According to fixed effect model three hypotheses are accepted. As discussed above common effect model this Dividend yield is showing the negative but significant relationship with dependent variable by using fixed effect model as well. Further earning yield and earnings per share are also showing significant relationship with stock returns and this relationship is positive while return on equity and equity to asset ratio is insignificant at level of 5 %. R square is 0.13 shows the explanation of dependent variable by independent variable. In case of panel data this value of R square is acceptable because panel data has combine effect of time series and cross section data. F statistics shows the fitness of model which means over all model is significant for this estimation or not. Significant value of F stat at 5 % level shows that this model is fit for current estimations. VI CONCLUSION Present study explores the relationship among return of common stock and five financial variables(dividend yield, earning yield, return on equity, earning per share and equity to asset ratio)Data of ten years of 31 financial companies (2003-20012) listed in Karachi Stock Exchange is used to apply statistics models. Based on earlier studies five hypotheses have developed. By applying all models for estimation of panel data fixed effect model is selected for estimation through hausman test.Result shows that dividend yield, earning yield and earnings per share have significant relationship with stock returns , while return on equity and equity to asset ratio is insignificant. Albeit the current study is an attempt to provide insight to finance managers and to investors however, current study has few limitations. One of them is that the sample size is not big enough it is small and only financial institutions are selected. The sampling technique which is adopted is convenient and it may be possible that might not be generalize able on the whole set of population .In the developed economies there are studies which have found return of common stock of different sector it may be a crucial point for future studies. More variables can be added like macroeconomic variables with large sample size for different sectors to get the more generalized result. Besides, a comparative study can be conducted to test the predictability of stock returns of financial and non-financial institutions. 9 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 REFERENCES [1] A. Ang, and G. Bekeart, “Stock returns predictability,” The Review of Financial Study, pp.651707,2006. [2] Brennan, Michael, TarunChordia, and A.Subrahmanyam,“Alternative factor specifications, security characteristics and the cross-Section of expected stock returns,”Journal of Financial Economics, vol.49,pp. 345–373,1998. [3] D.Dodd,,and B. Graham, “Security Analysis” The McGraw-Hill Book Company, Inc.1934. [4] E.Brewer,W.E,Jackson,T,Mondschean,“.Risk,regulation,andS&Ldiversificationinto non traditional assets” Journal of Banking and Finance vol. 19, pp.723–744,1996. [5] E.F. Fama, and K. French, “Dividend yields and expected stock return” Journal of Financial Economics.vol.22,pp.3-25,1988. [6] E.Fama, andK. French, “The cross-section of expected stock returns” Journal of Finance, vol.47, pp. 427-465,1992. [7] E.Zeytinoglu, “The impact of market-based ratios on stock returns: the evidence from insurance sector in turkey,Journal Of Finance And Economics,2012. [8] G.Leledakis and S.Cristos, “The stock return predictability of the European financial sector,” Journal of Finance, vol.43,pp. 540-555,2005. [9] H.L.Chang,Y.S. Chen,C.W. Su, and Y.W Chang,"The relationship between stock price and eps: evidence based on taiwan panel data.". Economics Bulletin, vol.30, no.3, pp.1-12, 2008. [10] H.Markowitz,(1952). “Portfolio selection” Journal of Finance, vol.7, pp. 77-91, 1952. [11] H.R.Litzenberger,and K.Ramaswamy, “The effect of personal taxes. and dividends on capital asset prices: Theory and empirical evidence,” Journal of Financial Economics, vol.7,pp.163– 196,1979. [12] J.A. Ohlson, “Financial ratios and the probabilistic prediction of bankrupt” Journal of Accounting Research, vol.18,no.1,pp. 109-13.1991. [13] J.Lewellen, “Predicting returns with financial ratios” Journal of Financial Economics, vol.74, 209-205,2004. [14] J. Pontiff, and L. Schall, “Book-to-market ratios as predictors of market returns” Journal of Financial Economics, vol., 49, pp. 141–160, 1998. [15] J.S. Abarbanell,andB.J. Bushee, “Abnormal returns to a fundamental analysis strategy” The Accounting Review, vol. 73, no. 1,pp. 19-45,1997. [16] J.Y.Campbell, “Valuation ratios and the long-run stock market outlook,” Journal of Portfolio Management, pp. 11-26, 1998. [17] L. Bhandari, “Debt/equity ratio and expected common stock returns: empirical evidence” Journal of Finance, vol.43,pp. 507-528,1988. [18] M.Majid and S.Mukhled, “The relationship between the roa, roe and roi ratios with jordanian insurance public companies”Iinternational Journal of Humanities and Social Science,vol.2,no.11,2012. 10 Proceedings of Annual Paris Economics, Finance and Business Conference 7 - 8 April 2016, Espace Vocation Haussmann, Paris, France ISBN: 978-1-925488-04-3 [19] M.I.Obeidat,“The Internal financial determinants of common stock market price: evidence from abu dhabi securities market. Journal of Economic and Administrative Sciences, vol.25,no.1,pp. 21-46,2008. [20] M.Omran,and A. Ragab,“Linear vs. non-linear relationship between financial ratios and stock returns: Empirical evidence on Egyptian firms” Journal of Accounting and Finance,vol. 3,no.2,2004. [21] O.Lamont, “Earnings and Expected Returns,” Journal of Finance, vol.53, 1563-1587. 1998. [22] P.Brown, “The predictive content of quarterly earnings” The Journal of Business,vol. 41,no.4,pp.488-497,1993. [23] R.Cantor,andR.Johnson,“Bankcapitalratios,assetgrowth,andthestockmarket”Quarterly Review, Federal Reserve Bank of New York ,pp. 10–24,1992. [24] R.Thiagarajan,“Fundamental information analysis” Journal of Accounting Research,vol.31,no.2,pp. 190-215,1993. [25] Reynald and Benderlipe, “Predicting stock returns using financial ratios: evidence from selected philippine companies” Journal of Financial Economics, vol.12, pp. 380–390, 2006. [26] S.Basu,“Therelationshipbetweenearningsyield,marketvalueandreturnforNYSEcommonstocks:f urtherevidence,”JournalofFinancialEconomics,vol.12,pp.129-56,1983. [27] S.P. Kothari, and J. Shanken, “Book-to-market, dividend yield, and expected market returns: a time series analysis”Journal of Financial Economics. Vol. 44, pp. 169–203, 1997. [28] S.T. Lau, T.C. Lee, and T. H. McInish, “Stock returns and beta, firms size, E/P, CF/P, book-tomarket, and sales growth: evidence from Singapore and Malaysia,” JournalofMultinationalFinancialManagement,vol.12pp.207-222, 2002. 11