Got Produce? Generic-Promotion Expert Enters Debate With Some Shocking Analysis
Jim Prevor’s Perishable Pundit, August 11, 2009
As part of our extensive discussion of the proposed generic promotion program for produce, we have often called on the proponents to enlist qualified experts to review the costs and benefits of such a program.
Now they have announced that Taylor Farms was kind enough to donate some money so that the Produce for Better Health Foundation could ask Harvey Kaiser, a Professor at Cornell University who specializes in the economics of generic promotion, to answer some questions that have been raised about the commodity promotion board.
PBH summarized Professor Kaiser’s findings in three points, two of which we found somewhat shocking:
• The median average for these types of programs is a 10-fold return on investment to the producer, or whoever pays.
• The bulk of the cost would be borne by the consumer.
• While there would be a small increase in production due to enhanced consumer demand, the net result is a sustainable price increase due to the promotion program.
The question of who ultimately bares the cost is complex and it is entirely plausible that consumers might bare the costs of the program — although there is no guarantee and we question PBH’s ascribing such a definitive statement to the professor.
The shock, though, comes in the first and last point that PBH highlights from the Professor’s findings:
To say that the “median average return” for “these types of programs” is a “10-fold return” and the return will go to the “producer, or whoever pays” sounds very odd.
First, we are not aware of much research at all on “these types of programs” — that is to say multi-product in a multiplicity of formats such as fresh, frozen, canned and juice — so we wonder where this claim comes from. Second, we don’t see how the decision of where to collect the money, say at grower level or first-handler level, could possibly change who gets the benefit of the program. Third, how is it possible that the world is filled with programs promoting commodities and industries — California Iceberg Lettuce, Washington Apples, California Tree Fruit, PromoFlor, the Western Australian produce industry — and these producers are earning “10-fold return” and yet they keep refusing to fund the programs? Are producers really so dense?
This is all certainly problematic but the stunner is the claim that the program will cause only a “small increase in production” and, in fact, will result in a “sustainable price increase.” This is shocking because the whole justification for PBH’s involvement in this matter is that it will increase consumption. Since a substantial increase in consumption is impossible without a substantial increase in production, it seems highly unlikely that consumption will boom at the same time prices are supposed to be rising.
If this assessment is accurate, then PBH ought to just say: “So sorry, we made a mistake. This isn’t in our domain.”
The whole assessment by PBH of what Professor Kaiser believes is so odd that we were anxious to read the professor’s assessment ourselves.
Fortunately, PBH has made Professor Kaiser’s piece, titled Background Brief on Checkoff Programs, available. The brief consists of an introduction and nine questions, presumably posed by PBH, and the professor’s answers. We’ve added our own assessment underneath the professor’s responses. You can see it in blue.
Background Brief on Checkoff Programs
Harry M. Kaiserr
Many agricultural commodities produced in the United States have collective marketing programs aimed at increasing overall market demand (both domestic and foreign) and enhancing producer revenues. These programs, which are sometimes referred to as “checkoff programs,” are funded through assessments on firms in the industry.
Currently, there are 17 federal programs and numerous state check-off programs in existence (over 50 in California alone). The budgets for these programs total about $1 billion annually in the agricultural sector. The assessments are usually mandatory for all firms after a majority of producers approve the checkoff in a referendum. The revenue raised by checkoff programs is invested in a variety of marketing (and sometimes research) activities, and varies by commodity. The main activity used by the majority of checkoff programs is generic media advertising. Popular examples of media campaigns include: Got milk?; Beef: It’s what’s for dinner; The incredible edible egg; Pork: the other white meat; Milk mustache; and the Dancing Raisins.
In an attempt to increase consumption of fruits and vegetables, the national fruit and vegetable industry is currently considering implementing a checkoff program for promotion and research. The proposed checkoff program would raise approximately $30 million through a 0.046% assessment collected from first handlers and importers, and would be used to promote fruit and vegetable consumption through marketing and educational programs.
The proposed fruit and vegetable checkoff has generated a lot of debate. It has also generated some important questions and issues. In this brief, I address some of the most important questions that have been raised by affected parties. As Director of the Cornell Commodity Promotion and Research Program, and Professor of Marketing, I have conducted independent research on generic advertising programs for over 23 years and have published 100’s of research articles on various dimensions of the economics of these programs. It is with this background and expertise that I bring to hopefully inform the debate.
Question: Who bears the burden of the cost of the program, growers, first-handlers or consumers?
Answer: Economists refer to this issue as the “incidence” of the assessment. Regardless of whether the grower or the first handler pays, who bears more of the incidence depends upon which group is most sensitive to price changes. The group that is the least sensitive to price changes will bear more of the burden. For most food products in the U.S., consumers tend to be very insensitive to price changes because food items typically compose a small percentage of their budget. First handlers and growers of fruit and vegetables are probably more sensitive to price changes than are consumers. So, in general, if the entire industry pays, regardless of where the assessment occurs, the incidence of the assessment will likely be spread across the industry, with most of the assessment borne by the consumer.
PUNDIT: As they say, the answers you get depend on the questions you ask, and here, Professor Kaiser addresses the incidence of cost when businesspeople are typically concerned with the incidence of risk. In a field such as produce, where margins are thin at every stage of the distribution chain, business executives hope and expect that all their investments will ultimately be covered by the consumer. But sometimes they are not.
The relevant question for many is who gets stuck with the bill if markets do not adjust to cover these expenses? Clearly, the answer is the guy who pays the expense. So, for example, if retailers, in pursuit of a “value” reputation, refuse to increase the FOB price they will pay for a given item, it is someone before the retailer in the supply chain who has to pay the bill.
Professor Kaiser is, of course, correct that input costs tend to be allocated to those who are least concerned with paying more, but there is no guarantee that one will be able to pass on costs of any type, and the incidence of risk thus remains with the one paying the bill. For many, this is a risk they would prefer not to take. Many first handlers would like to see the incidence of risk spread over various sectors of the supply chain rather than concentrated with them.
Question: Won’t first handlers simply force growers to pay the assessment?
Answer: Not likely for reasons alluded to above. First handlers will likely push much of the assessment forward to parties buying their produce. Consumers are not very sensitive to price changes, and if the full assessment, 0.046%, shows up in the form of a higher consumer price, there should be minimal negative impact on demand.
PUNDIT: There is a little confusion here, and it is probably caused by Professor Kaiser’s work being so heavily focused on dairy. In the produce industry, first handlers often handle produce on a commission basis. The contracts typically explicitly call for first handlers to deduct all expenses from grower returns. So these first handlers will bill back for USDA inspections, delivery costs, warehousing, and, most assuredly, for any fees or assessments.
Now, first handlers selling on a commission basis will also certainly try to get their growers a higher price that will compensate for any expenses. It is very common for a handler to ask his grower “What return do you need to come out?” If he knows the grower needs $5, the handler will certainly aim to achieve that, but, of course, markets are unforgiving and it is not always possible.
This whole issue is especially problematic because it leads to a division between voting authority and paying the bill. In most generic promotion efforts, the one who pays gets the vote. Under this system we have in produce, very often the first handler will get the vote but the grower will actually pay the bill: A kind of taxation without representation.
Question: Do these programs really increase consumption?
Answer: There have been hundreds of evaluation studies that have addressed this question, and the overwhelming conclusion is yes they do. Kaiser summarized the results of 21 studies that estimated “promotion elasticities” for a representative set of state and national generic promotion programs.
A promotion elasticity measures the percentage change in consumer demand given a 1% change in promotion expenditures. All of 21 studies found positive and statistically significant generic promotion elasticities. The median and average elasticities from these studies are 0.045 and 0.096, respectively, i.e., a 1% increase in generic promotion expenditures results in a 0.045% and 0.096% increase in demand for the commodity when holding all other demand determinants constant.
The spread in promotion elasticities in these 21 studies range from a low of 0.005 to a high of 0.428. While statistically different from zero, it is clear that the typical impact of these programs on commodity demand is quite small. Indeed, in the majority of these studies, demand factors such as price, income, and population demographics have been found to have a larger impact on demand than generic promotion. Generic promotion, however, is one the few demand factors that the industry can control.
The prime reason why these programs have small impacts on demand is that the level of generic promotion is quite small. All of the mandatory generic advertising checkoffs are smaller, in some cases much smaller, than 1% of the price received by producers. Hence, it is not surprising that generic advertising has a small, but positive impact on food demand.
PUNDIT: We’re not sure why the professor decided to omit footnotes, thus making going back to his full research difficult. What we can say is that several things about this point need much more exposition: First, it is not clear how one can measure an increase in demand to the third decimal place on a product with fixed supply. Second, the research does not seem to actually identify the impact of the promotion; instead, it seems to use promotion as a kind of residual.
So the researchers identify an increase in consumption, then deduct for things they know could influence consumption, say an increase in population and then whatever is left, they declare that to be due to the promotion. Yet this claim is unsupported. It is more a measure of our ignorance than a measure of the effectiveness of promotion.
Question: Does the increase in consumption due to generic promotion translate into increased profits for growers and first handlers?
Answer: This question gets at the heart of a potential paradox of checkoff programs known as “rent dissipation,” i.e., effective generic promotion results in a higher demand and market price, which eventually causes an increase in output by growers and first handlers, which causes market price to fall.
In theory, it is possible that the expanded output could result in growers and first handlers not being any better off even with effective promotions because of this effect. However, while theoretically possible, virtually all actual evaluation studies of existing industry-sponsored generic promotion programs show that this is not the case. For example, Kinnucan examined this issue for generic catfish advertising, and Kaiser has looked at this issue for numerous promotion programs for various commodities and these studies have shown only a small increase in output due to the price enhancement of promotion. The net result has been a sustainable price increase due to promotion programs.
PUNDIT: In our piece, Got Produce? The Rent-Dissipation Hypothesis And The Issue Of Cui Bono, we specifically looked at the generic catfish advertising and drew from it the conclusion that Professor Henry W. Kinnucan of Auburn University drew:
With few production alternatives existing for catfish ponds and equipment, asset fixity operates as a natural deterrent to entry or expansion, causing a relatively inelastic supply response at the farm level. Furthermore, demand for catfish at the wholesale level is only slightly elastic and is probably inelastic at the farm level. This combination of elasticities, coupled with the magnitude of the demand shift as represented by the advertising elasticity, results in sufficient rents from increased advertising to more than offset incremental costs over any reasonable time horizon. Thus, the notion that producers are no better off with the promotion program than without it is not supported by our analysis.
Our findings are generalizable only to the extent that other industries have characteristics similar to those of the catfish industry. Asset fixity, which accounts for the sluggish supply response for catfish, may exist in other industries, especially those involving perennials such as almonds, raisins, walnuts, oranges, etc. This may not be the case for vegetables and some row crops, where inputs are less specialized and production lags are shorter. Then, too, farm-raised catfish is a relatively new product; this increases the likelihood that consumers will respond to catfish advertising. Clearly, the rent-dissipation hypothesis needs to be tested over a wider range of commodities before we can be confident that cooperative advertising ventures can indeed generate sustainable benefits for producers in the face of uncontrolled supply response in competitive markets.
In other words, products that are difficult or expensive to increase production of tend to benefit from higher prices due to stimulated demand from generic promotion, whereas products for which production is easily and inexpensively increased do not realize higher prices due to the supply response.
We find Professor Kaiser’s explanation that actual evaluation studies of existing generic promotion efforts rarely show such an effect unpersuasive because the Professor does not account for the “establishment and survival bias” in this comparison.
Survival bias is often illustrated with mutual fund advertising. A quick glance at the ads in the business section of the local paper shows that virtually every mutual fund seems to have had exceptional performance — yet actively managed mutual funds underperform the indexes. What accounts for this discrepancy? It is called “survivor bias” — the mutual fund companies know consumers don’t like to invest in funds with losing track records, so they are constantly closing or merging out of existence poorly performing funds and leaving only the winners extant.
The consequence of this is that if you study all existing mutual funds, you will get a performance record far exceeding that of all mutual funds that ever existed.
In the same way, Professor Kaiser’s studies are all on programs that actually exist. But the very establishment of a program is a selective act. So, in the produce industry for example, Northwest pears, an expensive crop to plant with a long lead time, has a generic promotion program, and California avocadoes, once again an expensive crop to plant with a long lead time and one with real restrictions on availability of land and water, both have generic promotion programs. In contrast the California Iceberg Lettuce industry voted to end its program. Leafy greens, a row crop relatively easy and inexpensive to expand production, has only the small and voluntary Leafy Greens Council.
When Dr. Kaiser points to research on existing programs, he is looking at programs that passed through many stages of approval and maintenance and comparing them to a proposed program that has not gone through such a process. That is inherently an asymmetric comparison.
Think about this kind of comparison in your own company or organization. Five projects are proposed… as they go through the process, only one meets the Return on Investment criteria and is implemented and is highly successful. You cannot deduce anything about the four projects that didn’t make it from the one project that cleared the vetting process.
Equally we can’t deduce anything about the rent dissipation process on produce commodities that do not have a commodity promotion program from commodities that have gone through the process and chose to establish and sustain a commodity promotion program.
Question: Are these programs profitable for the participants that pay for them?
Answer: This is the bottom line and most important question that first handlers and growers should ask. Here, the evidence from evaluation studies is overwhelming that the benefits of these programs exceed the cost. Economists typically measure the benefits of these programs as the incremental net revenue resulting from the increase in demand and market price due to generic promotion, while the cost is generally measured as either the cost of the promotion, or total cost of the checkoff program.
Table 1 (below) lists 14 studies that have evaluated individual generic promotion programs for fruits and vegetables in the United States:
Table 1. Estimated Benefit-Cost Ratios for Fruits, Vegetables, and Other Generic Promotion Programs.
Fruits and Vegetables
Alston et al.
California Table Grapes
Alston et al.
California Dried Plums
Erickson et al.
Carter et al.
Carman and Craft
Capps et al.
Florida Orange Juice
Richards and Patterson
Costo et al.
Ward and Forker
Ferguson et al.
Van Sickle and Evans
Median — 10.0 Average — 16.0
Crespi and Sexton (2005)
Schmit et al (1997)
Williams et al. (2004)
Florida Orange Juice
All Dairy Products
Davis et al (2000)
Kaiser and Schmit (1998)
Murray et al. (2001)
Median — 5.7 Average — 6.3
Of these 14 studies, the median average benefit-cost ratio (BCR) was 10.0, indicating the average benefits were ten times larger than the costs, and none of these studies had a BCR that was below 1.0. The average BCR was 16.0, indicating average benefits were 16.0 times larger than the costs. Also noted in Table 1 are other commodity promotion benefit-cost ratios, with a median average BCR of 5.7 and an average BCR of 6.3. Indeed, a more thorough perusal of the literature reveals very few studies that have measured a BCR that was less than 1.0 for any checkoff program.
This is actually quite interesting and tells us that academics in this field have chosen a biased measurement of success. Read the key line:
“Economists typically measure the benefits of these programs as the incremental net revenue resulting from the increase in demand and market price due to generic promotion, while the cost is generally measured as either the cost of the promotion, or total cost of the checkoff program.”
Note that Professor Kaiser is taking the increase in revenue and comparing that revenue increase to the cost of the promotion or of the checkoff. In other words, he is saying that economists in this field choose to ignore the cost of production!
More realistically, if we want to measure the profitability of a promotion or program, we should take the increase in revenue caused by increased demand or higher prices, deduct not only the cost of the promotion, but also deduct any costs involved in expanding production or purchasing more supplies and deduct the costs involved in selling these added items.
In other words, the proper “benefit” to look at is the additional net profit, not net revenue, realized by an entity and contrast that with the costs of the program.
There are two other troubling matters in the way Professor Kaiser defines profitability. First, he seems to be speaking of profitability for the industry as opposed to for the entity that is actually paying the assessment.
In an industry such as produce with many family farms, it is reasonably common to find a farmer with 500 acres who is not really looking to buy more land. If the industry becomes more profitable by increasing production and consumption, this farmer will be paying to allow other farmers to expand production. If prices remain fixed, but an additional expense is imposed, our family farmer’s profit will go down.
The second issue is that investments are typically measured against a cost of funds or hurdle rate. Lots of investments that would be profitable are not made because the profitability is less than some alternative investment or less than the cost of funds. Even accepting generic advertising is profitable, would product-specific generic advertising be more profitable than the proposed multi-product/multi-format (canned, frozen, fresh, 100% juice) generic promotion order?
Finally what is the time lag? Time is money. How long must one invest before one realizes a return?
Question: Why are the estimated benefit-cost ratios (BCRs) for these programs so large?
Answer: The reason estimated BCRs are so large for generic promotion programs is not because benefits are large in an absolute sense, but rather they are large relative to costs. Because the costs of these programs are so tiny in relation to industry revenue (almost always well under 1%), the upshot of this is it does not take much of an increase in demand or in price to produce a high BCR.
Kinnucan and Zheng presented an interesting illustration to highlight this point, using 11 federal programs encompassing over 80% of all checkoff program revenue. In their study, the authors calculated how much of a farm price increase would be necessary to yield a BCR equal to 1.0 (benefits = costs) for the 11 checkoff programs. The average increase in price to yield a BCR of 1.0 was a mere 0.94%. In other words, if the checkoff program increases price of the commodity by 0.94%, the program is breakeven in terms of costs and benefits. If the farm price increase due to generic promotion is 5%, Kinnucan and Zheng found an implied BCR of 8.2. To the extent that generic promotion could increase the farm price by 5%, which appears very plausible, the potential BCR for a $30 million fruit/vegetable campaign would be 8.2, delivering $246 million in returns to the producer.
Professor Kaiser is highly esteemed, yet we are perplexed at on what basis he determines that it is “very plausible” that a $30 million total budget — not a $30 million dollar campaign — could possibly increase produce prices by 5%. If this did happen, Professor Kaiser’s estimate of the benefit to the industry is wildly low.
The Produce Marketing Association had Battelle do an Economic Reach and impact of the Fresh Produce and floral industry study. It includes floral but excludes frozen and canned, and it came up with a sales number at the production source — that is before processing, marketing and distribution of over $33 billion dollars. So, using the PMA number, a 5% increase in price at the production level would be over $1.6 billion. PBH says an assessment rate of 0.046% will realize $30 million… that means that they are saying the industry size is $65 billion, which means a 5% increase in price would represent a $3.25 billion increase in value. These seem bizarrely outsized returns to expect from a $30 million-a-year investment.
Question: Would the proposed fruits and vegetables promotion and research checkoff program have as high of BCR as those listed in Table 1?
Answer: While it is impossible to predict with absolute certainty what the BCR for a future program would be, based on the past performance of generic promotion programs for individual fruits and vegetables, it is highly likely that the benefits of the future program would be substantially higher than the costs. Based on the median BCR from Table 1 of 10.0, this would mean that the $30 million per year investment in generic promotion of fruits and vegetables would return $300 million in additional net revenue to the producer.
One of the questions related to this whole enterprise is the degree to which experience with individual commodity promotion can be extrapolated to a national campaign covering many fruits and vegetables sold in many forms. Most efforts to increase sales are modest. Promote one snack fruit and one typically depresses the sales of other snack fruits. Many shopping lists just say “fruit” and so a great deal on pears can affect apple sales.
For a total industry program to be a success, the additional sales have to come from elsewhere in the store. So an increase in pear sales has to either lead consumers to spend more in the supermarket or that pear sale has to reduce, say, chocolate chip cookie consumption.
This whole area of study is known as Beggar-thy-Neighbor Advertising, and there is some indication in this UC Davis paper that the produce industry could benefit by closer cooperation between individual commodity promotion groups.
What there is not, however, is any real evidence that we can project the success of this much harder task — taking sales from chocolate chip cookies — based on the much easier task of getting someone to buy a pear rather than an apple.
Question: Would this mean that every fruit and vegetable commodity would receive the same BCR?
Answer: No. Indeed it is unlikely that a generic promotion program that promotes a broad category of fruits and vegetables would return benefits of the promotion equally among commodities. That is, if the promotional campaign for such a program was effective, then it is likely that some commodities would fare better than others.
Unfortunately, there have not been a lot of studies that have looked at this type of distributional question. One study by Schmit et al. examined the impact of generic milk advertising on whole, 1%, 2%, and skim milk products, as well as generic cheese advertising on American, processed, and other cheeses. The authors found that generic milk advertising had virtually identical positive effects on the demand for fluid milk products. However, while generic cheese advertising had a positive effect on processed cheese, it had a substantially larger positive impact on American cheese demand. So it is likely that the demand impacts from a generic fruit and vegetable promotion campaign will be different for some commodities.
And this difference in “demand impacts,” combined with a difference in “supply response,” is very problematic.
In the dairy industry, the milk can be used to make cheese, butter, various types of milk, and the “demand impacts” don’t matter because it all comes from the same cow.
In produce we face a real situation in which the supply response will be so different on different commodities that it is quite possible that, for years and years, leafy green growers could be subsidizing pear growers.
What if the industry segments that have high barriers to entry vote for this plan and the industry segments with low barriers to entry vote against –are we prepared to just shove this down their throats?
Question: Most checkoff programs have been commodity-specific. Would a promotion program with the broad category of all fruits and vegetables be as effective?
Answer: There have not been many examples of generic promotion programs that are as broad as the fruit and vegetable category combined. However, a recent study by Global Insight, Inc., evaluated the combined impact of all of the USDA’s Foreign Agricultural Service generic export promotion programs on U.S. exports. The results of this study found a BCR of about 5 for all programs combined. That is, a $1 investment in all generic export promotion returned $5 in export revenue.
We’re not really sure the relevance of this data. First of all the funds are provided by USDA specifically because they don’t believe that the producers would pay for the export promotion if USDA didn’t ante up. Nobody in the industry would object if USDA or HHS wanted to promote produce consumption for, say, reasons of public health. Second, the USDA programs are not mandatory. So the hurdle is much lower. It is a serious thing to compel a man to fund a program when he would prefer not to. Third, there is no basis to extrapolate from these export programs to domestic programs.
Finally, this “BCR” is a wacky number. We have to look at profit increases, not sales increases.
We certainly thank Professor Harry Kaiser for helping the industry think through such a crucial issue.