Abstract

The present research work formulates the probability distributions from Markov models for exploring the patterns of cashless transactions through debit/credit cards with reference to State Bank of India (SBI). The transitions of number of monthly E-transactions (in millions) based on three states namely decrement, remain-Same and increment are considered for classification. The assumed states are obtained based on the classifications of finite differences among the monthly sequence of transitions. While classifying the states, a standard notion of process capability using one sigma control limits of the averages is considered. The real time data of State Bank India (SBI) debit transactions (in millions) is procured from the web sources of Reserve Bank of India (RBI) for the period of 112 months (approximately 9.5 years). This study comprises of two major parts, primary part deals with (i) formulation of transition probability matrix (TPM), (ii) derivation of discrete probability mass functions (pmf) of the said three states, and (iii) finding the explicit mathematical relations of various statistical characteristics from the developed pmfs. Whereas the second part deals with (i) the computation of statistical measures using the derived mathematical relations, (ii) carrying out the numerical data analysis and (iii) understanding the model behaviors with real time data on monthly bank transactions. This study has a significant scope of extending the decision support systems on visualizing the bank’s transaction track records. Portfolio management wing of the banking sector can make use of these studies for extracting the hidden parameters of vital decision making by finding the key performance indicators on E-transactions.