1. Methodology

This section enlightens the

methodology of the study, that is the research design, the population of the

study, the sample size and sampling technique and the sources and methods of

data collection used. Similarly, the techniques for statistical analysis of

data are also described in this part.

The Research Design

This research scrutinizes the presence

of statistical association between accounting information and the share price

of listed banks in Mauritius, South Africa and Botswana mainly to determine the

value relevance of book value of equity, earnings, and dividend. To test the value

relevance of the accounting information, the study relies on the financial

statements and the factbooks published by the stock exchanges.

The population of the study include 4

banks listed on the Stock Exchange of Mauritius (SEM), the 5 banks listed on

the Johannesburg Stock Exchange (JSE) and the 4 banks listed on the Botswana

Stock Exchange (BSE) for a period of ten years (10) ranging from 2007-2017.

Mauritius

South Africa

Botwana

ABC Banking corporation LTD

Standard Bank Group

Investec Bank

Mauritius Commercial Bank

Sasfin Holdings Limited

Barclays Bank

State Bank of Mauritius

Barclays Africa Group

Standard Chartered Bank

NedBank Group

First National Bank

Absa Bank

The main criteria for sample selection

is that it must have been listed on the stock exchange for a period not later

than 2007 so as to ensure that the financial reports and share prices are

accessible for the establishment of a continuous trend of data for the period

under study. Hence, the study sampled ….

banks. This study does not examine companies of other sectors than the banking

sector.

Source of data and variables

This study employs data extracted

from the annual reports and accounts of the sampled banks for the study, that

is, secondary data, similar to previous studies for instance, Che (2007) and

Babalola (2012). The data of major concern includes the component of accounting

information which are obtainable from the financial statements of the listed

banks and the Factbooks published by the

stock exchanges of Mauritius, South Africa and Botswana .

The dependent variable for this study

is the market price of the share of the sampled banks. The book value per share,

the earnings per share and the dividend per share form part of the independent

variables.

Modelling BVS, EPS and DPS as a function of price

The

methodology employed for this investigation derives from association studies.

This technique assesses the value relevance of accounting information through

the statistical association between financial indicators and the firm’s market

value presuming the financial markets are efficient. Ohlson (1995) postulated

that the market value of a firm is a linear function of its earnings and its

net situation. Association studies uses regression analysis involving

regressing the profitability value for a given share in a market on the

accounting numbers for which that value relevance should be appreciated. This study

used the price model which investigates the impact of accounting information on

the market price of shares based on the concepts of Barth et al (1998) and Filip

and Raffournier (2010).

The

regression model is generally expressed as follows:

MVit = ?0 + ?1 X1 +

?2 X2 + …

+ ?n Xn +

?it

Where:

MVit

= the market value of the company;

X =

different financial indicators selected

?it

is the random error

The

basic model derived within the Ohlson (1995) framework is

Pit = ?0 + ?1

BVSit + ?2 EPSit + ?it

The

Ohlson model is modified to accommodate the effect of dividend as follows;

Pit

= ?0 + ?1 BVSit + ?2 EPSit

+ ?3 DPSit + ?it (1)

Where,

Pit

= Share price of company i at the end of year t ;

EPSit

= Earnings per share of company i during period t ;

BVSit

=Book value per share for company i during period t ;

DPSit

= Dividend per share of company i during period t ;

?it = random error;

The

equation (1) expresses the price per share as a function of their accounting

values and the relative earnings. The model allows the assessment of the joint

explanatory power of earnings, book values and dividend. To investigate the

explanatory power of each of the variables separately equation 1 is broken down

into the following equations whereby the share prices are regressed separately

on book values, earnings per share, and dividend per share.

Pit

= ?0 + ?1 BVSit + ?it (2)

Pit

= ?0 + ?2 EPSit + ?it (3)

Pit

= ?0 + ?3 DPSit + ?it (4)

The first

(equation 2) presents the price only as a function of the accounting value of

the share BVS, the second (equation 3) interprets the price as a function of

the earnings per share while the third one (equation 4) interprets the price as

a function of the dividends per share.

The

correlation between the accounting variables and the market prices is assessed

by the regression coefficient R2. The latter clarifies the extent of

the association between the share prices and the accounting variables. It depicts

the value content of the accounting numbers and their ability to reflect the

data transmitted on the financial market which is incorporated in the company’s

market value. The higher R2, the greater the value relevance of the

accounting variables and the better indicator of the market value of the firm. The

accounting information is regarded as value relevant if the obtained adjusted R2

is significantly different from zero, as in similar studies (Beisland, 2008 and

Barth, Beaver & Landsman, 2001).