Seppo Suominen Essays on cultural economics

2.5Robust checking

The robustness of the findings above is checked in multiple ways. The estimation is made with alternative WOM measures and different sample sizes. Since there are two alternative variables (Previous week’s attendance in Helsinki (HKIADM1) and TOP10) published in the newspaper NYT that measure WOM, the other is used to check the robustness of estimation. The results are in table 2-5 below. Since the WOM-variables (HKIADM1 and TOP 10) are correlated (-0.832) these can not be used simultaneously.

In Helsinki top 10 listing the movie with the biggest previous week’s admissions is numbered as 1, the movie with the second biggest admission is numbered as 2, and so on up to 10. Hence TOP10 variable should get a negative coefficient.. Each movie has a different intercept in the fixed effects model (not shown). The number of screens, the time variable (Weeks since released) and the critical reviews get similar results than in the table 2-4. However, the price variable is not significant and the seasonal variable (All Screens) is significant in the random effects model. Since these variables are moderately positively correlated (0.252) the results indicate that during high season (other than summertime) either the average prices are higher or more probably firm attenders go to see a firm during the weekend when the prices in general are higher than during the weekdays. Durign the low season (summertime, see figure 1 in appendix) film attenders might prefer more weekday evenings than during the high season and the admission tickets on average are cheaper.

Following the idea of Basuroy, Desai and Talukdar (2006) a third alternative variable for WOM is used: a cumulative number of screens since its release, however excluding the week in question. The results are shown below in table 2-6. The results are in line with the previous results in which the WOM is TOP10.

In the appendix some further results are shown with the sample excluding the first week but consisting only films that have been critically reviewed and the critics published in the newspaper. The sample size is 201. Otherwise the results are in line with the previous except that results indicate that the attendance is unit elastic with respect to price and critical reviews regardless of the WOM variable used. However, the fixed effects model favoured by the Hausman test shows no significance for the critical reviews variable. The full sample (n=1060) resuts are also shown in the appendix. Regardless of the WOM variable used the critical reviews have a positive and significant effect on film admission. The dummy variable for the films not reviewed (NOT CR) is with this sample always significant and negative indicating that any review published in the newspaper from the lowest (“waste of time”) to the highest (“superior”) has a positive impact on attendance.

The price variable is not significant with the full sample while the variable is significant in the smallest sample (n = 201) including only the admission starting from the second week and with all films reviewed in the newspaper. Since the full sample includes admission starting from the first week, it can argued that the latecomers that do not go to see a film during the first week are more price sensitive than those that go to see a film during the first week. In the smallest sample the critics review variable parameter in the random effects model is approximately equal to one indicating that if a firm has been critically reviewed the critics has a powerful effect on admission figures.

In all samples (n = 1060, n = 520 or n = 201) the number of screens and the time variable (weeks since released) are always significant and the parameter estimates are reasonable. If the first week is excluded (n = 520 or n = 201) the absolute value of parameter estimate for the time variable seems to be lower than it is in the full sample indicating that after the second week the admission figures dimish faster than they do in the first two weeks.

In the fixed effects model the critical review variable is not significant if the first week is excluded and even though there is some variation in the variable within each firm the variation is mostly captured in the individual constant variables. Therefore the fixed effects model is not suitable for studying the effects of critical reviews on the film admission.

Table 2: Estimation results, all movies with previous admission in Helsinki, n = 520

Test statistics for the classical model indicate that conventional regression analysis (OLS) without group dummy variables is not suitable for explaining weekly movie admissions. The t-statistics for critical reviews variable that illustrates significance is misleading due to misspecified model.

With the Finnish data, movie admission is inelastic with respect to number of screens. The screen variable does not take into account the number of actual seats in the hall. Blockbusters with a vast admission are shown in larger auditoriums and with more daily showings than arts movies. Increasing the number of screens is not as flexible as increasing daily showings if the movie turns out be a blockbuster. If the number of screens is still increased, these are probably with lower number of actual seats and therefore the relative admission increase is lower, and that might explain the inelasticity.

Two important hypotheses were imposed. Positive critical reviews should have a positive impact on movie attendance, and the results indicate that this is true. The other hypothesis proposes that word-of-mouth should have a positive impact on attendance which is also verified.