Modeling the Effects of Promotional Efforts on Aggregate Pharmaceutical Demand: What We Know and Challenges for the Future - Innovation and Marketing in the Pharmaceutical Industry - page 594

detailing effectiveness is larger for more innovative drugs and drugs that treat a
broader range of ailments. Murphy et al. ( 1992 ) fi nd that promotional effectiveness
is moderated by therapeutic novelty. Hahn et al. ( 1994 ) fi nd that pharmaceutical
promotion has a stronger effect on trial rate when the new drug is of higher quality
and when the corresponding therapeutic market is growing. However, Kremer et al.
( 2008 ) fi nd no signifi cant moderating effects of product characteristics.
Ad ( 3 ) Moderating effects of market characteristics. The fi fth column in Table 20.1
indicates that studies look at a wide range of therapeutic categories, and promo-
tional effectiveness might differ by therapeutic category. There is evidence that
therapeutic category affects diagnosis uncertainty (Joseph and Mantrala 2009 ), and
this might cause physicians to utilize wider information sources such as detailing
and journal advertising for some prescription decisions. The level of pharmaceutical
expenditure also differs across categories, and this could lead to different degrees of
saturation across categories (Vakratsas and Kolsarici 2008 ). Gatignon et al. ( 1990 )
fi nd that market growth is a strong and positive moderator of detailing effectiveness.
These fi ndings are confi rmed by Kremer et al. ( 2008 ) who fi nd signifi cant differ-
ences in responsiveness among therapeutic categories, and conclude that there are
interactions between the types of marketing instrument.
Another market characteristic that might infl uence sales responsiveness is the
country under study. Differences in (self) regulation on, for example, the use of
promotional instruments, cultural differences, and different saturation levels might
cause heterogeneity in promotional effectiveness across countries (Chintagunta and
Desiraju 2005 ). However, Kremer et al. ( 2008 ) do not fi nd systematic evidence of
such effects in their meta-analysis.
Ad ( 4 ) Moderating effects of model and data characteristics. The third column of
Table 20.1 indicates that slightly more than half of the studies accommodate endo-
geneity. Studies that do not assume that promotional expenditures are exogenous
would potentially result in a correlation between promotional expenditures and the
error term and create a bias in the promotional effect estimates. If promotional
expenditures are set as a percentage of sales, this will generate endogeneity, which,
when ignored, will lead to a positive bias (Shugan 2004 ), which means that the
promotional effects are overestimated. Indeed, Kremer et al. ( 2008 ) fi nd that studies
that accommodate endogeneity report promotional elasticities that are signifi cantly
lower than studies that do not.
The fourth column of Table 20.1 indicates whether heterogeneity in promotional
effectiveness across brands or categories is accounted for. We have already discussed
that promotional effectiveness is highly heterogeneous, across instruments (see Ad
(1) above) and across therapeutic categories (see the previous subsection). Leefl ang
and Wieringa ( 2010 ) establish that promotional effectiveness also differs between
brands within a category. Hence studies that pool brands, categories, or instruments
deserve special attention, because the results of the study might be affected by the
pooling decision. As an example of this phenomenon De Laat et al. ( 2002 ) investi-
gated the effects of promotional investments on prescribing behavior by pooling
across 140 brands in 11 categories. Leefl ang and Wieringa ( 2010 ), using the same