Neural Network-Ant Colony Optimization Model of Relationship between Accounting Numbers and Stock Prices in the Nigerian Stock Market

Abstract:

This work investigates the value relevance of accounting data in the Nigerian stock market, with a view to determining whether accounting information has the ability to capture data that affect share prices of firms listed on the Nigerian Stock Exchange (NSE). Piece of accounting data is termed value relevant if it is significantly related to the dependent variable, which may be expressed by the stock price.  The methods used for gauging information contents of various accounting numbers were Ordinary Least Squared (OLS), Random Effects Model (REM), and Fixed Effects Model (FEM). The findings show that there is a significant relationship between accounting information and share prices of companies listed on the NSE. Dividends are the most widely used accounting information for investment decisions in Nigeria, followed by earnings and net book value. The study therefore recommends that the firms should improve the quality of earnings as manipulated earnings (of which dividends are sub-sets) have large effects on share prices. The paper also recommends that all companies listed on Nigerian Stock Exchange should prepare Simplified Investor’s Summary Accounts (SISA) with emphases on the most widely used accounting information along the required mandatory detailed financial statements to suit Nigerian peculiarities.