Prasad,

Thanks for visiting our blog and for your question. Bouman and Jacobsen (2002) tested whether the Halloween effect was sector-specific and found no evidence to support such a hypothesis; it was present in all sectors of the market without significant differences. Jacobsen and Visaltanachoti, in "The Halloween Effect in US Sectors," (2006) found that all US stock market sectors and 48 out of 49 industries performed better during the November-April period than in the May-October period between 1926 and 2006. In more than two-thirds of all sectors and industries, the return difference was statistically significant. To your point, they did find performance differences across sectors and industries. However, the strongest Halloween effect was found in the clothes, construction, machinery and steel sectors. The weakest Halloween effect was found in the food, consumer, oil, and utilities sectors.

-Dave