The relation between the knowledge of reservation prices and the degree of success of bundling strategies

International Journal of Business Strategy, Jan, 2008 by Jose G. Aguilar-Barcelo, Marco A. Pinto-Ramos, Natanael Ramirez-Angulo

A low statistical significance level indicates that it is less likely that the sample comes from a population in which the two variables are unrelated. As usual, the breakpoint for statistical decisions is a significance level of 5% (p-value of .05), i.e., 95% of confidence is necessary to reject the null hypothesis of independence.

At a glance, the data in Table 2 do not show any strong pattern of association. Nevertheless, it should be noted that is common, according to the empirical evidence, that those bundles for which a medium level of competition exists also show a moderate level of Lerner index. Furthermore, we must highlight that the firm is almost never able to set a price far above the cost when the bundles face strong competition in its neighborhood, ceteris paribus. Nonetheless, although it is true that the values on the secondary diagonal of the crosstab are not low (consistent with neoclassic microeconomic theory) the relationship shown in this Table has a p-value of .30 for the [chi square] test, and therefore the possibility that both variables are independent cannot be ruled out.

Information in Table 3 neither allows ruling out the possibility that the variables are independent since they have a p-value of .77 (23% of confidence) according to the [chi square] test. On the one hand, a high efficiency component would allow reaching higher levels of Lerner index if we keep the level of competition constant and at a moderate level; on the other hand, with an aggressive competition, these "savings" could be used to compete on prices, thus reducing the Lerner index levels. It is not clear which effect might dominate.

The [chi square] test for the relationship shown in Table 4 is significant at 2%, which indicate that it is very unlikely to obtain a sample that would yield such a different values -or greater- from the estimated values, assuming independent variables. Thus, the hypothesis of independence between these two variables can be rejected. The foregoing would be evidence in support of that the capacity of the firms to set a price above cost in their bundles have a relationship with their expectation about how the consumers value and relate the products comprising them.

Assuming that this expectation is true, we would have, also, that low Lerner indexes correspond each other with high asymmetries in the reservation prices of consumers and vice versa, therefore the firms would be taking advantage of the decrease in variability and dispersion in reservation prices resulting from grouping products, in comparison with that of individual products, offering prices closer to cost in the former case; these results are not fully supported by the theory. This way, with a bundling strategy the firm would motivate some consumers to buy the bundle even presenting a very significant low reservation prices for some of the individual products (Adams & Yellen, 1976). Nonetheless, for the case of high reservation prices, the firm might be maintaining, in turn, high Lerner indexes for its individual products in an attempt to capture the consumer surplus. Though supported by the theory of bundling, this conjecture is only an assumption, since, unfortunately, there is no way of knowing the Lerner index of single products in the case when the commercial offer of the sample firms did not include bundle sales. In any case, the Lerner index of the individual products was not included in the data analysis.