|
CONTENTS
An Experimental Application to Voluntary
Funding of Generic Advertising
Tables
Next Meeting
NEC-63
2005 Next Meeting
October 13-14, 2005
Hyatt Regency
Savannah, GA
Nutrition, Obesity and Commodity Promotion
|
|
Printable PDF
An Experimental Application to
Voluntary Funding of Generic Advertising
by Kent D. Messer, Todd M. Schmit, and Harry M. Kaiser
Cornell University
Generic advertising programs have been a popular tool used by many agricultural
commodity organizations in the United States to enhance market demand,
raise prices, and increase producer net revenue. These programs operate
by assessing producers in an industry and using the collected funds for
generic (non-branded) advertising and promotion of the commodity. Currently,
there are thirteen federal programs and over fifty state programs in existence.
The majority of economic studies evaluating generic advertising programs
have found large benefits for producers relative to costs.
Funding for some commodity programs originally came from voluntary donations
from participants via a voluntary contributions mechanism (VCM). While
initial contributions for the advertising programs using a VCM were typically
high, free-riding and decreased donations eventually became a significant
problem, raising questions of equity and fairness. As a result of these
concerns, essentially all VCMs were abandoned and producers held referenda
on whether to adopt mandatory assessments to fund the advertising programs.
Virtually all programs in operation today are mandatory.
Some individual producers have recently challenged the constitutionality
of mandatory generic advertising programs arguing that being required
to contribute money to generic advertising programs is an infringement
of their rights to free speech. Currently, there are over 70 First Amendment
challenges to generic advertising programs being litigated. To date, there
have been decisions delivered by district and circuit courts on both sides
of the issue, upholding the constitutionality of some of the programs
and ruling others unconstitutional. Because of these challenges, there
is a need to assess whether a new institutional arrangement that maintains
the voluntary spirit of the court findings will achieve the same goals
and benefits of generic advertising.
An alternative funding mechanism that could potentially yield long-term
benefits to producers is the provision point mechanism (PPM) for public
goods. The PPM, which has never been used to fund generic advertising
for agricultural commodities, has two desirable characteristics given
the current legal environment: (1) it is voluntary and thus would not
likely be vulnerable to legal challenges based on freedom of speech, and
(2) it has been shown in both the lab and the field to reduce the incentives
for free-riding and to generate greater total contributions than the VCM
does. The PPM operates by announcing a threshold (or goal) for the fundraising
campaign and soliciting contributions to achieve this threshold. If the
threshold is met or exceeded, the contributions collected are used to
fund the public good; otherwise all of the contributions are returned
and no funding is provided.
Several key questions related to the institutional design of a PPM need
to be explored to find the combination of features and procedures that
could lead to maximum producer welfare. The first question that arises
is what impact producer referenda have on contributions to the advertising
program and, ultimately, on producer surplus. The second question is what
the optimal threshold is for the PPM. The third question is what combination
of institutional features leads to stability of contributions to the advertising
program over time. The fourth question is the impact that effectiveness
of the advertising program has on producer contributions. These questions
are the subject of the research summarized in this paper.
Experimental Design
All experiments were conducted at Cornell University and the subjects
(N=240) were recruited from undergraduate economics courses. The experiments
were carefully designed to simulate the key economic and psychological
elements that influence producer contributions to generic advertising.
The elements included inelastic supply and demand, an incentive compatible
market mechanism, stochastic variation in demand, producer discussion,
and confidential refund-by-requests that enable subjects to receive a
refund of their assessments they paid on their sales.
Each experimental session involved three separate parts, where twenty
subjects assumed the role of producer and seated in front of a private
computer spreadsheet informing them of their costs for producing up to
three units of a fictitious commodity. The first part of the experiment
enabled subjects to become familiar with the experimental platform and
did not involve an advertising program. The second part demonstrated the
benefits of the advertising program and had a mandatory program. Simulating
the potential change that could result if mandatory programs are ruled
unconstitutional, the third part replaced the mandatory program with a
voluntary PPM where the PPM thresholds were varied.
Control of the rate of return was the most critical economic element
to simulate in the experiment. Increases in demand due to advertising
in the next period were based on several benefit-cost ratios – 2:1,
4:1, and 6:1. These rates of return are similar to the rates of return
commonly observed with generic commodity advertising promotion. A particular
benefit-cost ratio was constant throughout an experimental session.
In the third part of the experiment, the advertising program was implemented
and contributions expended only if the PPM threshold was met or exceeded.
The subject participation thresholds used in the experiment were 50%,
70%, and 90%. If the threshold was not met, then all subjects received
a refund of their assessments and there was no advertising program. In
the subsequent round, subjects were given the opportunity to reach the
threshold again.
To test the influence of producer referenda on contribution behavior,
in some of the experimental sessions, subjects were asked to submit confidential
votes on whether they would prefer the PPM with a certain threshold level
or whether they would prefer no advertising program. Referenda were held
prior to the start of a series of rounds for each PPM threshold. In the
other sessions, subjects were not given a choice and were simply informed
that for the next series of rounds the advertising program would be funded
by the PPM with a certain threshold level.
Results
Of primary importance, were subjects’ behavior when PPM thresholds
were varied. An econometric model was developed to determine the relationships
between gross producer surplus on PPM threshold level (PPM), accounting
for advertising contributions (ADV_CONT-1), market demand
(DEMAND), group referendum type (GRPREF), and treatment
round (ROUND). The model was specified to account for the three-level
hierarchical nature of the experimental data, where subject-level information
is nested within experimental groups (or blocks) and observed over rounds
(i.e., repeated measures). For more specifics, see Messer, Schmit, and
Kaiser.
Regression estimates for the models and utilizing
the experimental data calibrated on a benefit-cost ratio of 4:1 are included
in table 1. All estimated parameters were
statistically significant for the producer surplus model at the 0.05 significance
level or less. The statistical significance of the covariance parameter
estimates lends support to the hypothesized three-level hierarchical error
structure. As expected, both demand and final advertising contribution
levels were significantly high given the price impacts from changes in
demand.
PPM thresholds significantly affected producer profits, as did their
effects across referendum groups and program duration (round). A striking
result is that subjects offered significantly higher average contributions
in the referendum sessions than in the non-referendum sessions for all
PPM thresholds. Simulation of the econometric model indicates that with
the exception of the lowest threshold levels, predicted gross producer
surplus in the referendum program were dramatically higher than those
in the non-referendum program. For the referendum program, these relationships
are illustrated in figure 1, which computes predicted gross producer surplus
across thresholds and rounds. This simulation with its hump-shaped curve
illustrates the practical trade-off between high levels of producer contributions
and actually achieving the PPM threshold necessary to implement the program
(and retain these contributions).
Also apparent from the simulation is the answer to
the third question regarding the stability of producer surplus across
rounds. While the sign on ROUND is negative suggesting deterioration
in return levels over time, when interacted with the other variables included
in the model, the net effect is one of improved stabilization for the
referendum program, particularly near the optimal PPM threshold level
of 82% (figure 1). In fact, model simulations
show that the round-by-round changes in gross producer surplus were over
six times larger in the non-referendum programs than in the referendum
programs. Certainly, the higher and more stable gross producer surplus
levels validate the inclusion of voting in PPM-funded programs. That producers
can maximize higher profit levels at relatively higher PPM thresholds
also means that additional funding goes to the generic promotion program,
resulting in larger demand-enhancing impacts.
As mentioned above, control of the rate of return was the most critical
economic element to simulate the experiments. We conducted additional
experiments involving referendum calibrated at return levels both above
(6:1) and below (2:1) the initial experimental settings. While these experiments
do not capture the entire range or reported payoff ratios, we felt that
they provide additional insight into the role of program efficacy on contributions
in a PPM setting. As expected, as benefits from advertising increased,
so did subject contributions. Specifically, the average percentage of
contributions increased from 59% in the case of a 2:1 BCR, to 63% for
the 4:1 BCR, and to 68% for the 6:1 BCR across all threshold levels. The
improved demand enhancing impacts as advertising’s rate of return
increased were also reflected in the average producer surplus levels across
BCRs (3.3, 4.5, and 6.3, respectively).
Supplemental regressions of similar form and specification
to the 4:1 BCR data were conducted on the additional sets of advertising
payoff experiments. Given the changes in contribution behavior, it is
not surprising that as advertising effectiveness decreases, so does the
PPM threshold level that maximizes gross producer surplus. The optimal
PPM threshold dropped from 82% to 68% as the BCR decreased from 4:1 to
2:1 (table 2). Likewise, as effectiveness
improved, the optimal threshold level reached the maximum threshold level
evaluated within the experimental data; i.e., 90%. Expected threshold
achievement at the 6:1 advertising effectiveness level was similar to
that observed in the 4:1 case (77% and 76%, respectively). However, as
effectiveness dropped to 2:1, expected threshold achievement dropped sharply
to less than 50% of the time (table 2).
Conclusions
In light of uncertainties about the constitutionality of mandatory generic
advertising programs for agricultural commodities, it is useful to investigate
alternative voluntary funding mechanisms in case they become needed. Empirical
results from an experimental setting indicate that including producer
referenda as part of the program design positively affects both producer
profits and contribution probabilities. Given how participation in these
referenda strongly affected subjects’ contribution behavior, advertising
programs should encourage these types of institutions that help secure
higher funding levels. In addition, substantially higher program stability
was evident when the program included the referendum and the threshold
was set at or near the level where producer profits are maximized.
These results provide valuable information to commodity organizations
that wish to design promotion programs that may pass constitutional muster
and achieve the largest benefits possible to the producers who fund them.
The range in optimal thresholds highlights the crucial nature of the underlying
advertising performance measure in the experimental set up. An additional
realization is that if commodity programs go to a voluntary PPM type of
program, knowledge on the long-term relative performance of their promotions
programs will be crucial to setting PPM operational parameters. Furthermore,
extending this type of experimental application to producer groups and
commodity organizations is a next logical step in making these types of
institutional designs practical in a real-world setting.
For further details and related research see: Messer, K.D., T.M.
Schmit, and H.M. Kaiser, “Optimal Institutional Mechanisms for Funding
Generic Advertising: An Experimental Analysis” American Journal
of Agricultural Economics (forthcoming).
|
