Having estimated the propensity sco

Having estimated the propensity score for each firm from the logistic model, we applied a one-to-one nearest neighbor matching without replacement to match RFID-adopting firms to control firms. More specifically, we matched each RFID-adopting firm in the base year (t – 2) to a control firm in the same industry (four-digit SIC) with the nearest propensity score for that year. As a result, 146 sample firms were successfully matched with a control firm. The post-match model in Table 1 presents the matching procedure is successful. After applying propensity score matching, we found that the coefficient of most variables experiences a sharp decrease in magnitude, and become statistical insignificant between adopter and matched non-adopter. We found the coefficient of firm size is significant because RFID-adopting firms are mainly large firm.