The Analysis of World Events and Stock Prices

Abstract

The research reported below has two major objectives. The first is to examine in a scientific fashion certain broad relations between world events and movements in stock prices. The second is to illustrate and suggest specific applications of some techniques for measuring meaning that have not previously been used to any great extent in business research. events were defined as occurring in this study when a fiveto eightcolumn headline occurred in the New York Times. The quantitative attributes of headlines-the size, the ready availability of a historical series, their reliability, and other factors-make them an appropriate source for this study. The headlines were classified by untrained observers into twenty categories of meaning. They were also classified on a seven-point good-bad scale. Many characteristics of the time series of headline classifications and dates seem of interest. For example, the longer the string of consecutive world event headlines in the preceding days, the more likely it is for world event headlines to occur on the following day. Another interesting resuit is that bad headlines tend to follow bad headlines and good headlines tend to be followed by good headlines. There are also some noteworthy seasonal properties to the headline series. The most unequivocal pattern of influence reported below is that large changes are substantially more likely following world events than on randomly selected days. For clusters of world events, such as those occurring on crises, the proportion of large changes is substantially greater than the proportion following relatively isolated world events. Contrary to popular belief, the particular category into which a world event falls does not appear to add much additional information concerning future price movements. There is a strong tendency for large price changes on the first and second day following world events to show the same direction of change. This tendency for large changes, in the average, to be followed by continuations is characteristic of all large changes, even those that do not follow world events. On Days 2-5 following extremely bad world events, there is a tendency for rises to occur. On those occasions, the market appears to be overreacting to bad news. The tendency to overreact to extremely bad news emerges in a separate study of market movements covering presidential illnesses, a fairly homogeneous class of events, during the past fifty years. The relation between world events and the stock market has attracted considerable attention within the invest* This report is based on the author's Ph.D. dissertation, World Events and Stock Prices: A Study of Large New York Times Headlines and Subsequent Movements in the Standard & Poor's Composite Index. It is a pleasure to acknowledge with gratitude the many helpful suggestions of Prof s. James Lorie and Harry Roberts, Dr. M. F. M. Osborne, and Mr. Winfield Smith, which were incorporated n this report.