CONSTRUCTION OF AN ISLAMIC INDEX USING THE MEDIAN CRITERION OF QUANTITATIVE FILTERS: APPLICATION TO THE AFRICAN STOCK EXCHANGE

Many techniques, combining qualitative and quantitative filters, have been proposed in the literature to construct an Islamic index on financial stock exchange. These are often based on fixed thresholds to define quantitative filters. However, as companies do not have the same characteristics from one sector to another, it is necessary to have heterogeneous thresholds for filtering purposes. Our contribution, here, is to propose a methodology which integrates this fact by using the median (statistic) criterion which is much more robust than the average criterion (vis-à-vis the presence of extreme values) but also the median statistic is consistent with the Wassatiya principle (50% of the way) than the 33% fraction used by many studies.


INTRODUCTION
Islamic finance is based on the principles of "sharia". It has established itself as a financing alternative to the conventional system. Its evolution has led to the creation of new index.
A stock exchange index is defined as an indicator that measures the evolution over time of the performance of a stock market, a sector of activity or a sampled portfolio, Topsacalian (2000). In other words, it is the weighted or unweighted average of the prices of the panels of stocks, grouped on the basis of their zone, their country and / or their sector and / or their capitalization. This definition is broad enough to leave the door open to the creation of multiple stock market indices.
The strong demand for Islamic index has accelerated the launch of other Islamic index by other operators with more or less extensive coverage of geographic areas outside Africa.
The specificity of Africa in Islamic finance is mainly due to the importance of its Muslim community. It turns out that there are wealthy Muslims in Africa who respect Islamic injunctions in financial transactions (Koita & Diaw, 2014). However, there are no "Sharia" compliant clues. This African characteristic constitutes one of the major symbolic arguments in favor of the development of African-style Islamic finance. Thus, the creation of an Islamic sector stock exchange index would be one of the necessary adjustments to strengthen its competitiveness in this sector.
The Islamic index most covered in articles are American, European and Asian indices. In our research, we only encountered three studies in Africa. Koita and Diaw (2014) investigated the possibility of investing in BRVM using the filtering method of the Islamic FTSE Global Shariah Equity Index. Mbengue (2017) proposed an index from the stock exchanges of Ghana, Nigeria and the BRVM using the S&P methodological guide. Ndiaye (2019) studied the possibility of attracting the investment of Gulf countries in the financial centers of Ghana, Nigeria, Morocco by applying the filtering of the DJIMI index.
Our study proposes to use the statistic of median as a filtering method. To the best of our knowledge, this is a first paper that uses the median statistic. Indeed, companies do not have the same characteristics from one sector to another, it is necessary to have heterogeneous thresholds for the purposes of filtering. Our contribution, here, is to propose a methodology which integrates this fact by using the median statistic criterion which is much more robust than the average/mean statistic but also the median is in coherence with the principle of "wassatiya" (golden mean) than the fraction of 33 % used by many studies.
In view of the initiatives taken by the Central Bank of West African States (CBWAS) to adapt banking regulations in order to take charge of the specificities of Islamic finance, the efforts made by the Islamic Bank of Development (IBD), applying the Islamic filtering in the creation of a sector stock market index is appropriate.
In the rest of this research, we will first present the data and the new methodology. Finally, we will present the findings and the comparison with another the existing methodology: DJIMI.

DATA AND METHODOLOGY TO CONSTRUCT A FINANCIAL INDEX Data
In the selection of "Sharia-compatible" shares, we use as a pool of values the 1442 companies making up the 17 stock exchanges (South Africa, Botswana, BRVM, Egypt, Ghana, Keyna, Malawi, Morocco, Mauritius, Namibia, Nigeria, Uganda, Rwanda, Tanzania, Tunisia, Zambia and Zimbabwe).
To calculate the ratios, it was necessary to take into account the annual standardized balance sheets over seven financial years (2009 to 2015), the average monthly stock market price over 84 months of study as well as the number of shares of each of the listed companies.
The number of shares multiplied by the average monthly price (from 01/01/2009 to 12/31/2015) is used to calculate the market capitalization. This gives us three annual values, one for each of the three ratios. A company integrates the Islamic index over the seven years as long as it scrupulously respects each of the filters associated with the three ratios.

Methodology
For a share to be included in this index, it must pass the qualitative and quantitative filters. Indeed, the introduction of the qualitative filter consists in carrying out a sectoral exclusion.
The quantitative filter completes the qualitative filter for a better selection of share and it consists in sorting the companies to keep only those whose financial structure is able to meet the requirements expressed in the form of ratios.
To remain in compliance with the principles of "Sharia", "Sharia" committees impose a number of criteria in terms of liquidity, claims and debt on share traded on financial markets. This filter is also expressed as a ratio.
Among the methods for selecting stocks for a stock market index in Islamic finance, the most commonly accepted criteria, within the meaning of El Khamlichi (2012), are those retained by the "sharia" committee of DJIMI (Debt Ratio, Liquidity Ratio and Debt Ratio).
Unlike previous works, we have chosen to conduct research that better reflects the specificities and needs of Africa in terms of stock markets. Indeed, the ratios applied on international stock exchange differ from one index to another, from one country to another or from one continent to another.
After the qualitative filter, we proceed to a sector classification of listed companies. Then, we take two approaches. The first is to calculate the median statistic over all the companies that have passed the qualitative filter. The second is to calculate the median statistic by business sector that has passed the first filter. Finally, for both approaches, an asset is selected if the value of its ratio is less than or equal to the median statistic of the ratios of selected companies (either all or by sector). We use the median statistic for three main reasons: • A statistical reason: it is more robust than the average in the face of extreme values. • A religious reason: the median statistic is a principle of Islam which is the golden mean or "wassatiya" to define the threshold of acceptance / rejection. • A financial reason: using the median statistic allows the portfolio to be diversified. statistic, the thresholds varying from 10% to 70%. Compared to the sectoral median statistic, the threshold for ratios varying from 33% to 70%.

FINDINGS AND ANALYSIS The Result of The Extra-Financial Filter
The initial goal was to carry out a complete analysis of 37 stock exchanges, due to lack of data, 17 stock markets were studied.
The original purpose was to perform a comprehensive analysis of the thirty-seven (37) stock exchanges in Africa, but the research limit confined us to seventeen (17) stock markets. Of these, three (Mauritius (SEM), Uganda (USE) and Rwanda (RSE)) were automatically eliminated due to a lack of data. Unfortunately, all the companies that make up the fourteen (14) remaining exchanges that passed the first test could not be analyzed in the second screening. This is due to the unavailability of financial statements and also to the absence of detailed items (financial result) from the companies' financial statements.
Of the 1442 companies listed on the 17 stock exchanges, only 418 have passed the first test where 29% of companies are potentially compatible with "Sharia". This is the most restrictive filter in this study. The 71% are made up of companies whose activities are related to finance, insurance and / or leisure as being the most "illicit" sectors from the point of view of compatibility with the requirements of "Sharia". Although this configuration is not surprising, it can be explained by the very nature of their activities and therefore by their instinctive propensity to expose themselves to "riba".    Vol. 8, No. 1;2021 or 51%. It is followed by the service sector with 16 companies, or 27% of the 59 selected, followed by the finance sector with 15% or 9 companies. The telecommunications and transport sectors accumulate only 7%, i.e., 04 and 01 companies respectively. For the sectoral median filter, the industrial sector records 42% of the companies which passed both tests. It is followed by the service sector with 35% of companies among the 57 retained by the second filter. The finance sector occupies the third position with 14%. The remaining sectors (agriculture, telecommunications, transport and others) account for only 9% of companies.

Finding of the Quantitative Filter (Or Financial)
The ones that failed the second quantitative filters represent approximately 22.2% for the DJIMI, 21.77% for the general median and 21.9% for the sectoral median, or respectively 320; 314 and 316 companies. In terms of geographical distribution for the three methods (respectively DJIMI, general median and sectoral median), South Africa dominates with 47%, 48% and 49% of the companies among those which have passed the qualitative and quantitative filters. The second place is occupied by North Africa with 28%, 32% and 33%, of the lot of values retained. West Africa is third with 19%, 14% 14% of companies that have passed the filters. Southern Africa records 6%, 3% and 2%. On the other hand, East Africa does not register any company for the DJIMI filter. Nevertheless, it 3% and 2% by general and sectoral medians.

CONCLUSION
Our paper aimed at applying a new filtering methodology in the process of creating an Islamic index in order to make the Africa zone more competitive in attracting Muslim and foreign capital. The study has been carried out taking into account a pool of stock exchanges from 16 countries and a regional stock exchange (BRVM) in Africa.
We have applied the qualitative and quantitative filters. For the quantitative filter, we have adopted three methods: the DJIMI method, filtering by general median statistic and filtering by sectoral median statistic. This approach produced a number of results.
The qualitative filtering has led to the exclusion of 71% of the stocks making up the 17 African financial markets covered by the study. Quantitative filtering has resulted in the exclusion of 22.2% for the DJIMI, 21.77% for the general median and 21.9% for the sectoral median of listed companies. Those who have passed this quantitative filter respectively represent 14.21%, 15.82% and 15.3% of the companies subjected to this test. South Africa leads the way in terms of the geographic distribution of businesses. The second place is occupied by North Africa. Next comes West Africa. Finally, Southern Africa and East Africa each of these zones counts a company and this, for the three methods.
The study has proven that the African stock exchange is a market where Muslims could invest in at least 53, 59 and 57 listed companies during the study period (2009)(2010)(2011)(2012)(2013)(2014)(2015) with respectively the methods of DJIM, from the statistic of median in general and sectoral case.
In addition, we can retain that the financial index with the sectoral median (statistic) method constitutes the most diversified portfolio compared to the others. On the other hand, the general median statistic counts more listed values.
Finally, beyond the choice of filtering by the median (general or sectoral), it is necessary to rethink the relevant filtering process to take into account the specifics of African companies in terms of Islamic financial investment. The interest of applying a filter in the process of creating an Islamic stock index in Africa's financial center does not only have a symbolic dimension aimed at sending a positive signal to African or foreign Muslim investors; but to prove that this index once available, would have financial arguments of its own.

Copyrights
Copyright for this article is retained by the author(s), with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (https://creativecommons.org/licenses/by/4.0).