Regression analysis results

Market return as control variable

	Competitor 1	Competitor 2
	P-value		P-value
MSFT	-0.0361	0.2818	-0.0297	0.5095

Table 12 – Abridged Table 7. Estimated values of and P-value less than 0.1 denotes 90% significance level

The only case when the values of and are negative is for Microsoft and its competitors – Apple and Google. However, the results are not significant even at 90% confidence level, therefore no arbitrage profits can be safely earned.

Industry return as control variable

	Competitor 1	Competitor 2
	P-value		P-value
AAPL	-0.1225	0.0000	-0.1480	0.0000
MSFT	-0.5681	0.0000	0.0561	0.2091

Table 13 – Abridged Table 8. Estimated values of and P-value less than 0.01 denotes 99% significance level

If the regression controls for the industry instead, Apple and Microsoft show significant opposite effects from competitors’ stocks. 1% increase in Google and Hewlett-Packard stocks causes -0.1225% and -0.148% decrease in Apple’s share price. In addition, 1% increase in Apple’s share causes a negative change of -0.5681% in Microsoft stock price. The results are significant at 99.9% confidence level; therefore arbitrage profits can be earned.

To do so, programs calculate intraday price change for Microsoft and Hewlett Packard. A positive price change triggers a short sell transaction of Apple stock, while a negative price change triggers a buy transaction. Similarly, programs calculate intraday price change for Apple stock and buy/sell Microsoft stock.

Influence on Competitors

	Competitor 1	Competitor 2
	P-value		P-value
AAPL	-0.7736	0.0000	-0.9780	0.0000
MSFT	-0.0485	0.1191	-0.0708	0.3635
NCR	-0.0117	0.8934	0.0427	0.1433

Table 14 – Abridged Table 9. Estimated values of and P-value less than 0.01 denotes 99% significance level

This time, only Apple’s significant news shows significant opposite effect on its competitors. However, the magnitude of the effect is striking. A 1% increase in share price for Apple causes -0.7736% and -0.978% decrease in Google and Hewlett Packard stocks.

Arbitrage profits can be earned from this information. If programs detect significant good news for Apple (or significant increase in its share price), they can short sell Google and Hewlett-Packard stocks.

Conclusion

Overall, the paper succeeded in identifying companies which have Situation X-type events more commonly than Situation Y. It also partially managed to explain the competitors’ stock price reactions, with limitations. Arbitrage profits can be earned for some companies if:

1) news articles are released during the trading hours

2) the duration of price adjustment to direct news is different to the duration of adjustment to competitors’ news (beneficial, but not necessary)

3) technology industry shows low returns on the day (discussed below)

4) transaction costs are low (Apple’s significant news effect on Google and Hewlett Packard is high enough to disregard transaction costs)

5) relevant software is created, which analyses competitors’ news and intraday stock returns

Hypothesis Check

Date: 2015-12-17; view: 709