ECONOMETRIC MODEL FOR PREDICTING INDEXES-USING ARMA, VAR AND VECM APPROACH

Abstract

A financial market is a place where purchase and sale of financial products ranging from derivatives, stocks, bonds, currencies, and commodities are conducted. These financial markets offer individuals, corporations, and organisations a place to invest, borrow, and lend money, manage risk, and make money. In a market for financial instruments, both buyers and sellers come together to exchange financial securities at a price determined by supply and demand in the market. The choice is an exchange, where trading occurs through a centralised platform, or an over-the-counter (OTC) market, in which deals occur directly between buyers and sellers, can be used to organise the market. An index is a statistical indicator of how well a specific area of the financial market has performed. The index measures the overall performance of a collection of stocks or other financial instruments that are either comparable in character to one another or that are part of the same sector or industry. By monitoring the price alterations of the underlying assets over time, the value of an index is determined. Stock market indices, which assess a set of equities that are traded on a stock exchange, are the most often used indexes. Indices act as benchmarks for evaluating how well portfolios, mutual funds, and other financial instruments perform. The index may be used by investors to compare the returns on their investments to the returns on the whole market. Indices act as benchmarks for evaluating how well portfolios, The financial market in India is a system that allows people, businesses, and organisations to purchase and sell financial items such bonds, derivatives, stocks, commodities, and currencies. Primary market and secondary market are the two general divisions of the Indian financial market, which is governed by Securities and Exchange Board of India (SEBI). Companies that want to raise money from the public issue new securities, such as bonds, debentures, and shares, on the main market. On the other hand, investors exchange these assets on the secondary market. The National Stock Exchange (NSE) and the Bombay Stock Exchange (BSE) are the two main stock exchanges in India. The BSE Sensex, which monitors the performance of 30 sizable, well-known firms listed on the Bombay Stock Exchange, is the most popular index in India. The performance of the top 50 firms listed on the National Stock Exchange is tracked by the NSE's own index, the Nifty 50. These indices are valued according to the market capitalization of the firms which make up the 1

index. The index functions as a benchmark for the performance of the index-companies and reflects the market's general mood. The objectives of this report are: - • To determine the factor that can affect the Indexes in question namely Nifty 50, S&P 500 and Nikkei 225 using already existing literature published. • To make all the data take as stationary so that time series analysis can be done. • To create a suitable model using Vector Auto Regressive Model (VAR) in order to create a predictive model and the gain an equation for the following indexes • Selecting suitable lag parameter so that optimal model can be created and performing various test in order to make sure all the assumptions of the time series analysis and regression are met. • Performing various test in order to establish that the assumptions of these models are met like Heteroskedasticity, normality etc. • Creating a ARMA model for the same parameter and evaluating which of the 2 models is more optimal. The Secondary Data was collected from various Yahoo finance and government websites. On the basis of the information gained the model is created using various parameters found and supported by literature review. The models are created and suitable lag are selected which are used to create the model equation. The models were tested for in order to confirms that none of the assumptions of the models are violated. The model created in VAR though had a great value of Square but violated a condition of Cointegration which led us to using VECM model in order to correct the VAR model. ARMA model on the other had good R square value which was lower than VAR but had no violation of any assumption. The interpretation of this study is based on the assumption that the information taken is correct and the faults of the methods used are minimized but not completely removed.