Generic advertising campaigns promote the general qualities of a product to current and potential customers leading to an improvement in the demand of all firms who market that product. Recent national generic advertising campaigns include the ones for milk ("Got Milk?''), florists ("Think Flowers.''), the National Basketball Association ("I Love this Game.'') and beef ("Beef. It's What for Dinner.''). Also, at a local or regional level, it is a common practice for shopping plazas, malls, car dealer associations and local industry groups to conduct such campaigns.
The focus of this dissertation is to understand, through analytical modelling and an experimental economics perspective, the best way to organize such an advertising campaign. We investigate two types of mechanisms currently under use by managers- Voluntary Contribution Mechanisms (VCMs) and Mandated Contribution Mechanisms (MCMs). In the former, industry members can decide if they want to participate and if so, how much they wish to contribute to the campaign. In MCMs, the government stipulates a payment rule by legislation and all industry members must behave in accordance with this rule. Non-compliance is illegal. The key strategic problem in VCMs is ensuring participation by all firms whereas in MCMs the concern is that firms will not truthfully reveal privately held information.
We study two types of VCMs- the Simple VCM and the Provision Point VCM. The Simple VCM is the one used in practice and here the advertising budget is set as the sum of contributions of industry members. However, we show that this can never lead to an industry welfare maximizing advertising budget. In order to overcome this, we propose the Provision Point VCM where the campaign is conducted only if contributions exceed a pre-determined threshold.
The experimental investigation in this dissertation focusses on evaluating the relative performance of the Simple and two Provision Point VCMs under different conditions. The key criterion for evaluation is efficiency- which is defined as the ratio of the actual contributions divided by the maximum possible contributions (the latter is also referred to as the Pareto Optimum). We study the short-term and long term impact of face-to-face communication on contributions to the advertising campaign and if contributions will differ when payoff information of participants is incomplete. A key contribution of this dissertation is using actual managers involved in a generic advertising campaign as participants. The important findings from the forty four economic experiments conducted are summarized below:
{1.} A low level of efficiency was found in the Simple VCM case (37.30\%) when compared to the Provision Point VCMs (42.19\% in the low case and 98.44\% in the high case).
{2.} When the provision point was set at the Pareto Optimum, 98.44\% efficiency was obtained and the provision percentage (i.e. the fraction of occasions on which the provision point was equalled or exceeded) was 93.75\%. This is an important result since previous studies had predicted a drop in the provision percentage.
{3.} Communication led to gains in efficiency in all cases. An extremely weak long-term effect was found in the Simple VCM case and a strong long-term effect was found in the Provision Point VCM (Low) case.
{4.} No differences were observed in the level of contributions between the complete and incomplete information cases for both Simple and Provision Point VCMs. This is a surprising result which needs closer attention in the future.
In the analytical section of the dissertation, we develop a very general model for n>2 firms and for a general class of advertising response functions.
The key problem in Simple VCMs is obtaining complete participation in the advertising campaign. Our equilibrium results suggests that this will never occur and based on the industry structure, we obtain either "free-rider'' or "cheap-rider'' equilibria. A general result obtained here is that the equilibrium advertising budget will always be strictly less than the Pareto Optimum when advertising is feasible.
We first study the case when the Provision Point is set at the Pareto Optimum. In this mechanism, equilibrium results suggest that optimal advertising is always a feasible outcome. In certain cases, zero advertising emerges as an equilibrium. This is, however, Pareto Inferior to the optimal advertising budget and, hence, will not be predicted to occur. Next, we establish that in all other Provision Point VCMs, optimal advertising cannot occur as a Nash equilibrium.
Finally, in the case of MCMs, the key strategic problem relates to ensuring that all industry members truthfully reveal information that is privately held. We design a mechanism that ensures that for all industry members, truthful information revelation is the dominant strategy. Further, we show that the cost shares implied by this mechanism have an intuitive explanation as the product of an equal split of the optimal advertising budget and a correction factor that accounts for group size, industry power and advertising response effects.