Linkages between news and dividend decisions using neural networks

Dividends are considered to be one of the most important management decisions from the point of views of investors, however in the area of corporate finance it still remains a controversial issue [1]. Despite the existent literature on dividends and stock market relationships, little focus is found on media events and its impact on Indian stock market outcomes. This article uses neural network theory to examine the quantitative relation between news, dividend payouts and its determinants. Further the research focuses on understanding whether different sectors and industries relate to stock market outcomes post event impact. Using a sample of 1362 companies listed on the National Stock Exchange of India Ltd between the periods of 2001–2015, we find that variables of net sales, profit after tax, cash balances, total reserves, previous year dividends, net block of assets and forex revenue earnings and expenses are significant in determining the payout decision across the nine sectors examined. The results from random forest regression score above the other neural models used in the study. Lastly, inflation, gold prices, US federal interest rates, ECB interest rates and dollar exchange rates carry significant influence on the determinants of dividend payout decision and eventually the payout.