Statistical Analysis Of Factors Affecting The Pricing Of Bank Card Receipts

Acquiring market as one part of bank card business,it has been playing a more and more important role. The notice on perfecting the mechanism of bankcard fees pricing has been issued jointly by the national development and reform commission and the people’s bank of China ,on March 18, 2016. The policy changes the single market pricing rules,it no longer takes the government guidance pricing mechanism,but the market independent pricing mechanism. Fee rate is determined by acquiring institutions and special merchants .Competition between market subjects becomes more fierce. In the context of both opportunities and challenges, acquiring institutions how to develop the correct competition strategy and reasonable pricing mechanism,will become the key to success in the competition, it is about market position and development direction. So, formulating scientific single pricing mechanism to identify and quantify factors that influence the pricing for acquiring institutions is urgently needed. Based on the above background, this paper takes bankcards pricing factors as the research object, tries hard to quantify factors influence degree, in order to provide a reference for acquiring market pricing.But acquiring market pricing is a complicated system engineering,it involves the interests of the participating subjects and concerns about the development of bank card acceptance environment and financial industry. Therefore,we need integrate a variety of factors ,such as cost, demand, competition, to reflact multiple relationships in the network bank card comprehensively. However, the existing research literatures are qualitative analysis or model derivation under strict conditions, lack of comprehensive and quantitative analysis of the single pricing.Through combing the literatures and experts’ consultings, this article establishs the influencing index system of bank card single pricing.This index system contains 5 dimensions, 23 indicators.And then, based on the empirical data, it uses the AHP and machine learning method to calculate weight of the various influence factors respectively.The main research content is as follows: 1. The background analysis of bank card industry. Basic theory knowledge of bank card industry were introduced in order to explicit single institution status in the bank card industry and close relations with other participants. This paper shows all kinds of prices in the process of POS transaction and the allocation of interests,that builds a theoretical basis for establishing the influencing index system of bank card single pricing. Through the introduction of business policy background and present situation of the single market,we clear about the acquiring market characteristics. 2. Factors of acquiring market pricing is analyzed by AHP. The author reviews literature and consults expert to establish the influencing index system of bank card single pricing ,and then uses AHP to calculate the weight of the various factors and sort factors by importance.3. Factors of acquiring market pricing is analyzed by machine learning methods. Because of the AHP is based on the expert scoring,that has a strong subjective color,so we need use machine learning method to verify this weight sorting. First ,choose four machine learning methods to classify the empirical data, and choose the optimal method through the misjudgment rate of training set and test set. Second, use the optimal method to order the independent variables .Third compare the results of two methods .There are three main conclusions: 1. AHP tells us that card group pricing makes the most important influence, followed by acquiring institutions, costs,.Exchange fee and network transfer service poundages combination weighting is 0.2472 samely,the POS cost, developing customer cost and POS service cost are also ranked top.2.Machine learning method tells us that merchants bankcard transaction rate has the greatest influence, followed by merchants industry type. The accept will of bankcard and the number of transactions also perform well, e But their influence degree is equivalent to half of the amount of card transactions.3.It is concluded that the results of two methods are basically identical. In addition to the per capita GDP ,all the other indicators sorting results are the same.