Market Performance in the Era of Buy Local: A Welfare Analysis of
Colorado Apples to Assess Winners and Losers
Wenjing Hu, Dawn Thilmany McFadden, Yuko Onozaka and Dustin L. Pendell2
Colorado State University and UiS Business School, University of Stavanger, Norway
.
CSU Working Paper
The authors wish to acknowledge funding support from USDA-CSREES NRI Project
#2008-35400-18693 and the Colorado Ag Experiment Station.
Market Performance in the Era of Buy Local:
A Welfare Analysis of Colorado Apples to Assess Winners and Losers
Abstract
The widespread emergence of state-based and other local labeling programs is one signal
of the general acceptance that consumers value such indicators. Subsequent discussions
of the benefits, tradeoffs or market performance implications of such marketing strategies
are now prevalent. This study focuses on the fresh apple market, and assesses the market
dynamics and welfare effects in the presence of differentiated markets for locally grown
products within direct markets as an alternative to shipping points for conventional
retailers using an equilibrium displacement model. The markets are segmented for
demand and supply; first by estimation of stated consumer valuations of regional-origin
labeling (demand side), and then, by estimated differential costs and intermediate prices
in marketing channels (supply side). The results show that, in the long run, consumers
would shift their demand toward local apples and the increased share of locally produced
apples would be marketed directly rather than through conventional supply chains.
Although suppliers who continue marketing through shipping points would lose in the
short run, all local suppliers would gain in the long run. These results suggest market
incentives will encourage integration of local produce into more retail food supply chains.
Key Words: local food promotion, marketing channels, welfare analysis, apple markets,
equilibrium displacement model
2
1. Introduction
There is a growing public interest in localized food systems since localization
activities are perceived to impact several key issues including: improving environmental
outcomes, more localized control of food-security to address anti-corporate sentiment,
and supporting community-based economic development strategies (Martinez et al.,
2010). In recent years, the local foods movement was ranked as a top story by several
relevant media outlets, including the Packer and The National Restaurant Association’s
annual What’s Hot list (Galbraith, 2011). Subsequently, there are an increasing array of
policies and programs targeted to support the development of local food systems (a listing
of federal programs are now compiled at the U.S. Department of Agriculture’s Know
your Farmer, Know your Food website), but critics note that such programs are being put
forth without adequate evaluation of how local foods will affect market performance and
the welfare of key stakeholders.
Within markets, the interest and promotion surrounding local foods has
significantly increased the demand for local produce (Adelaja et al., 1990; Carpio and
Isengildina-Massa, 2009; Costanigro et al., 2011). The structure of markets has also been
affected, as demand has stimulated a proliferation of localized, direct marketing supply
chains linking growers directly to consumers, such as farmers’ markets and community
supported agriculture organizations, farm stands and on-farm sales. According to the
2007 Census of Agriculture, direct sales of agricultural products for human consumption
amounted to $1.2 billion in 2007, a 117.79% increase in sales from 1997, but still only
0.4% of total agricultural sales; moreover, many believe those numbers are significantly
below the volume of sales now occurring in direct, and locally-focused markets.
1
One driving motivation for the locally-focused policies and programs relates to
potential gains to producers, consumers and local markets from having more food choices
available. There are some studies that have investigated the impacts of a more localized
food system on market players. However, most of these studies were based on surveys of
markets and consumers and data from sales/financial reports. (Brown and Miller, 2008;
Darby et al., 2008; Myers, 2004). More recently, studies have evaluated the impact of
local promotions using more theoretically-based economic models (e.g., Carpio and
Isengildina-Massa, 2010). However, with a focus primarily on consumers’ responses to
promotions of local or state-based labels, such studies may neglect supply side
implications related to the restructuring of supply chains. This study contributes to the
literature by filling this gap using a system-wide economic approach which examines
how consumers’ response and the restructuring of the supply chain affect the welfare of
consumers and suppliers.
In particular, we aim to answer the following questions regarding the impacts of
the introduction of local labeling: How do consumers and suppliers respond to the local
labeling efforts? What would the new structure of demand and supply markets look like
given such responses? What will be the supply chain-wide economic impacts of a local
labeling program? Does the potential for targeted, differentiated pricing strategies
encourage some producers to redirect their supply to segmented marketing channels to
optimize potential profits? To answer these questions, the structural and performance
dynamics of fresh apple markets is investigated, while accounting for demand shocks due
to the introduction of local labeling by segmenting markets by product origin. The
marketing channels on the supply side are segmented to include direct, short supply
2
chains, in addition to the more conventional grower-shipping point-terminal market-retail
chain. Such delineation allows one to determine how more localized systems influence
the dynamics of market and evaluate the performance of markets in the face of increasing
consumer demand. The results from this study will also contribute to the literature by
providing insights on how strong consumer responses to local produce offerings (albeit
among a relatively small set of buyers) may affect market dynamics. By allowing for
segregated markets, akin to what occurs more formally with organic produce, both the
conceptual and empirical framework provides a method to analyze welfare effects within
a more differentiated food system.
The analysis of such delineated market system requires a great deal of detailed
information for each segment of the market. Consequently, this study focuses on a
generalized U.S. and Colorado fresh apple markets. Focusing on fresh apple market has
several advantages. Apples are generalizable to many different countries and regions, as
it is one produce category in which cultivation is possible within diverse geographic
locations. Thus, more localized supply is plausible in terms of growing conditions,
although economic feasibility of such production is a question explored in this study. In
case of U.S., local and regional production areas still remain in most states, but generally,
with reduced volume since most areas did not fare well in the face of competition from
global trade partners and dominant U.S. production areas in the late 1990’s. The state of
Colorado exhibits these typical trends—high consumption of fresh apples, dominant
presence of apples imported from outside of Colorado, shrunken but increasing
production of state apple production, the presence of direct marketing channels, and state
3
program to promote Colorado grown fresh produce. Thus, Colorado can arguably be
considered as a “representative” case of local apple market.
The unique characteristics of apples are also considered in the local vs. non-local
context. Apples are not highly perishable, so that the local harvest can be stretched across
a longer season with some post-harvest handling. On the other hand, supply response is
constrained by the long lead time for apple orchard establishment, thus, consideration of
short- and long-run market adjustment is important. It is also worth noting that fresh
produce is a category gaining particularly high attention from consumers, producers,
government agencies, and media in the context of food miles and the local food
movement (Thilmany McFadden and Low, 2012). Thus, impacts from shifting to more
regionalized market in fresh produce category is of great interest to many stakeholders.
2. Background and Previous Research
U.S. Apple Market
U.S. apple production was 9,515 million pounds in 2008, and was commercially valued at
$2,206 million in revenue. The U.S. was the third largest among global producing
countries after China and EU, accounting for 6.4% of world apple production. In 2008,
34% of U.S. produced apples were processed to make juice, applesauce, cider, and other
processed products, but the majority of U.S. production (66%) entered the fresh market.
In 2008, apples ranked first in fresh fruit consumption, at 16.2 pounds per capita (U.S.
Department of Agriculture: Economic Research Service, 2012). Apples are grown in all
50 U.S. states, however, Washington is the largest fresh apple supplying state, accounted
for 72% of domestic supplies (U.S. Department of Agriculture: Economic Research
4
Service, 2012). The apple industry has been shrinking in most states since the late 1990s
under the pressure of global competition. Colorado’s production of apples for 2008 was
reported at 18 million pounds, including the supply of apples that were directed to fresh
market sales (10 million pounds). In short, production could no longer meet the demand
of over 5 million Coloradans (81 million pounds if one just considers fresh apples), but
current supplies of fresh apples are mostly marketed in-state to fill 12% to 13% of current
fresh demand.
As is the case with most product categories, the conventional supply chain for U.S.
fresh apples includes growers, packers, shippers, processors, brokers, and retailers.
However, because apples are not as perishable as some other crops, they can be marketed
over a longer season using controlled atmosphere (CA) storage. There are still fairly
competitive market conditions as packers and marketers sell their apples to retailers on
the spot market (Lynch and Coleman, 2010), while big retailers like Wal-Mart and
Costco use contracts with large suppliers to ensure year-round supplies. According to
Gomez (2010) and Pirog and Tyndall (1999), the production in main consumption areas
such as Iowa was mainly consumed within state through direct sales such as farmers’
markets, while the apples produced in main production regions such as Washington and
New York were mainly exported or shipped to other states through packers and shippers.
Still, this market structure’s flexibility has allowed for the apple market to revert to more
localized sourcing conditions for short seasons in a number of places in the country. Thus,
the interaction between Colorado and the U.S. apple market is a representative case study
to examine effects of a localized marketing strategy.
Impacts of Local Promotion
5
Local promotion programs have not only increased consumers’ demand for local produce,
but also stimulated the spread of more localized supply chains. There are a few veins of
literature that examine the impacts of farmers’ markets and CSAs on producers,
consumers, and communities (Brown and Miller, 2008; Darby et al., 2008; Henneberry et
al., 2008; Myers, 2004) using surveys of markets and consumers and data from
sales/financial reports. Carpio and Isengildina-Massa (2010) were one of the first to go a
step further in exploring the welfare implications of any potential shifts to more locally
marketed foods using a quantitative model based on economic theory. We follow their
lead in estimating these same effects, but for a more specific food category, apples, and
further delineating changes in the structure of marketing channels.
Preference for local attributes
There is a significant amount of work that has been conducted to investigate
consumers’ demand for local produce. According to Adams and Salois’ (2010) review of
literature on consumers’ perceptions and willingness to pay (WTP) for local and organic
food characteristics, studies started to find that consumers were willing to pay a higher
premium for local than for organic attributes as early as the late 1990s. In the early
2000’s, Loureiro and Hine (2002) and Mabiso et al. (2005) found consumers were willing
to pay more for local potatoes and apples labeled “U.S.A. Grown”, respectively. In later
studies, Darby et al. (2008) and Hinson and Bruchhaus (2005) found that the majority of
respondents were willing to pay a premium for locally produced strawberries.
In a more broadly framed study, Carpio and Isengildina-Massa (2009) found that
95% of South Carolina consumers prefer state-grown produce to all types of domestic
produce, and 78% and 30% of consumers chose state-grown produce at a 5% and 50%
6
premium level, respectively. The resulting estimated mean WTP premium for South
Carolina grown produce was 27.5%.
More specific to this study, Onozaka and Thilmany McFadden (2011) found that
U.S. consumers had a strong positive preference for local apples averaging a WTP of
$0.22/lb or $0.24/lb premia with respect to domestic apples, depending on the marketing
channel where they were purchased. In a complementary project, Costanigro et al. (2011)
found the average premium for the local apple attribute was 101% in a store experiment.
Equilibrium displacement model
An equilibrium displacement model is a commonly used method to analyze the
impacts of exogenous shocks such as measuring the performance of food programs and
policies (e.g. Brester et al., 2004; Carpio and Isengildina-Massa, 2010; Lusk and
Anderson, 2004; Thompson et al., 2005). Lusk and Anderson (2004) and Brester et al.
(2004) examined the welfare effects of the Country-of Origin Labeling (COOL) program
in the livestock industry. Thompson et al. (2005) used an EDM to assess the direct and
distributional effects of state-financed quality control and regional origin assurance
programs in Bavaria so they were required to separate the market into two regions. Using
an EDM, it is relatively easy to deal with substitution among markets and segmented
markets by examining multiple stages in the supply chain as well as multiple markets
(Burrell et al., 2006), but this study is one of the first to say a producer redirecting some
of their production to direct sales, which represents a new marketing channel.
Although EDMs have been widely used in assessing the effects of food programs
and policies, there have been a very small number of examinations of the performance of
local promotion programs. One example is Carpio and Isengildina-Massa (2010)’s
7
evaluation of the potential economic impacts of a regional promotional campaign in
South Carolina. They disaggregated the markets into two segments: locally labeled and
marketed and mass marketed with no designation of production source.
Carpio and Isengildina-Massa (2010) provided a good starting point of reference
to evaluate the effects of local promotion programs; however, there are some limitations
in their research. First, the parameters used in their study were mostly aggregate demand
and supply parameters, such as demand and supply elasticities for fruits and vegetables,
which made it difficult to estimate the parameters accurately and lowered the precision of
the estimation of welfare changes. Second, they only differentiated on the demand side
and did not segment the markets in the supply side (thereby assuming the costs of
marketing two different ways would not vary). However, the demand shock due to the
local campaign will not only affect the performance of the local market, but also change
the structure of local marketing channels. Thus, this study will further their research in a
number of ways: by investigating a specific market, the fresh apple market in Colorado,
and segmenting by marketing channels on the supply side, and with more differentiation
between short- and long-run scenarios.
3. Methodology
An equilibrium displacement model (EDM) is developed for Colorado fresh apples and
used to analyze the impacts of local labeling. This study will segment markets by
estimation of increased consumer valuation due to regional-origin labeling with quality
control on the demand side as well as realizing differential costs and prices in marketing
channels on the supply side. Several assumptions for the apple market and marketing
8
channels are made in order to implement this EDM: all locally labeled apples are
consumed in Colorado; the effects of the local marketing initiatives are concentrated in
Colorado; Colorado apples and domestic apples are weakly separable; apples marketed
through different market channels are weakly separable; Colorado supplies of Colorado
produced fresh apples can be marketed directly in local markets or marketed through
shipping points in or outside the state; and, domestic apples (from other regions of the
U.S.) can only be marketed through shipping points and cannot enter direct markets.1
The economy is separated into two regions: Colorado (Region A) and the rest of
the United States (Region B) which has supply and demand relationships with the
Colorado apple sector. The EDM framework specifies demand and supply equations for
each region, market clearing conditions and price relationships (price margins), yielding a
total of 13 linear equations. The structural model can be described as follows:
Region A (Colorado)
Demand:
(1)
𝐷𝐴𝑙 = 𝐷𝐴𝑙 (𝑃𝐴𝑙 , 𝑃𝐴𝑑 , ∝𝑙 )
(2)
𝐷𝐴𝑑 = 𝐷𝐴𝑑 (𝑃𝐴𝑙 , 𝑃𝐴𝑑 , ∝𝑑 )
Supply:
(3)
𝑆𝐴𝐹 = 𝑆𝐴𝐹 (𝑃𝐴𝐹 , 𝑃𝐴𝑆 )
(4)
𝑆𝐴𝑆 = 𝑆𝐴𝑆 (𝑃𝐴𝐹 , 𝑃𝐴𝑆 )
(5)
𝑆𝐴𝑆𝐴 = 𝑆𝐴𝑆𝐴 (𝑃𝐴𝐹 , 𝑃𝐴𝑆 , 𝑃𝐴𝑙 , 𝑃𝐴𝑑 , 𝑃𝐵𝑑 )
Region B (rest of the United States)
Demand:
1
More complete details are available in Hu (2012)
9
𝐷𝐵𝑑 = 𝐷𝐵𝑑 (𝑃𝐵𝑑 )
(6)
Supply:
𝑆𝐵𝑆 = 𝑆𝐵𝑆 (𝑃𝐵𝑆 )
(7)
Market-Clearing Conditions
(8)
𝐷𝐴𝑙 = 𝑆𝐴𝐹 + 𝑆𝐴𝑆𝐴
(9)
𝐷𝐴𝑙 + 𝐷𝐴𝑑 + 𝐷𝐵𝑑 = 𝑆𝐴𝐹 + 𝑆𝐴𝑠 + 𝑆𝐵𝑠
Price Relationships
(10)
𝑙
𝑃𝐴𝑑 (1 + 𝑡𝐴−𝐶𝑂
) = 𝑃𝐴𝑙
(11)
𝑑
𝑃𝐵𝑑 (1 + 𝑡𝐴−𝐶𝑂,𝑈𝑆
) = 𝑃𝐴𝑑
(12)
𝐹
) = 𝑃𝐴𝐹
𝑃𝐴𝑆 (1 + 𝑡𝐴−𝐶𝑂
(13)
𝑆
𝑃𝐵𝑆 (1 + 𝑡𝐴−𝐶𝑂,𝑈𝑆
) = 𝑃𝐴𝑆
The price parameters 𝑃𝑖𝑘 , k=l, d, F, S; i=A,B are derived from different sources to
construct the demand and supply equations where l and d denotes local and domestic,
respectively; F and S denotes apples marketed through direct markets and through
shipping points, respectively; A and B denotes Colorado and the rest of the United States,
respectively. ∝𝑘 represents demand shifters for differentiated apples related to attributes
that consumers perceive through the local labeling efforts. 𝐷𝑖𝑘 and 𝑆𝑖𝑘 represent the
𝑙
demand and supply for apples through channel k in region i, respectively. 𝑡𝐴−𝐶𝑂
is the
𝑑
market margin between farmers’ market and retail market in Colorado. 𝑡𝐴−𝐶𝑂,𝑈𝑆
is the
𝐹
market margin between Colorado and domestic retail markets. 𝑡𝐴−𝐶𝑂
is the supply price
difference between selling in direct markets and selling to shipping points in Colorado (if,
10
empirically, such a difference is found to exist in spatial analysis of market prices). This
is confirmed in Gomez (2010) that the producers’ share of the retail dollar was 80% with
direct marketing in New York, while the shares of Washington and New York suppliers
𝑆
for conventional supply chain are 35% and 47-60%, respectively. 𝑡𝐴−𝐶𝑂,𝑈𝑆
is the supply
price difference between Colorado and domestic shipping points. All variables and
parameters for the structural model are defined in Appendix A.
It is important to note that equation (5) shows that the supply for Colorado apples
marketed through shipping points and shipped back to Colorado (SASA) is a function of
direct market price (𝑃𝐴𝐹 ), Colorado shipping point price (𝑃𝐴𝑆 ), farmers’ market price for
local apples (𝑃𝐴𝑙 ), domestic apple retail price in Colorado (𝑃𝐴𝑑 ), and domestic apple retail
price in the rest of the United States. (𝑃𝐵𝑑 ). 𝑃𝐴𝐹 and 𝑃𝐴𝑆 determine whether to market direct
or through shipping points. Then processors need to decide whether it is economical to
ship apples back to Colorado or to other states (since we assume they will not be
differentiated as Colorado grown once they are in wholesale channels). The share of
apples marketed through different channels ultimately depends on the farmers’ market
price of local apples and the prevailing domestic apple retail price.
In equilibrium, the demand for local apples in Colorado equals the supply of
Colorado apples in direct markets plus the supply of Colorado apples through shipping
points to Colorado, which means 𝐷𝐴𝑙 = 𝑆𝐴𝐹 + 𝑆𝐴𝑆𝐴 (equation (8)). In the equilibrium, total
apple demands equals total apple supply (since we assume no imports or exports of fresh
apples and all apples produced in a harvest year are consumed in the same year and no
storage will enter the market of next year) (equation (9)).
11
Totally differentiating equations (1)-(13) yields the EDM of the apple market. The
EDM can be found in Appendix B (see equations B.1-B.13).
All variables and parameters for the EDM are defined in Appendix A. Specifically, 𝛾 =
𝑑𝑙𝑛(∝𝑙 ) represents the increase in consumer demand for local apples resulting from local
labeling efforts.
4. Model Inputs
Prices and quantities
The Colorado farmers’ market apple price is used to represent the “retail” price of local
apples and the fresh Red Delicious apple retail price in Southcentral retail market (which
includes Colorado) is used to represent the conventional retail price of domestic apples in
Colorado. The Red Delicious apple is chosen due to its market share dominance in the
apple market and the consistent availability of price data. The Red Delicious apple price
in the Northwest retail market is used to represent the retail price of domestic apples in
the rest of the U.S. based on the results from complementary retail market relationship
analysis (Hu, 2012). Although the Southwest market was a market leader in the price
formation of all other retail markets, it is not chosen due to the limited production of Red
Delicious apples in that region. The Northwest retail market significantly affected the
price formation process of other retail markets except the Southeast market. More
importantly, the Northwest region is the main production area for fresh apples in the U.S.
The direct market price is represented by the Colorado farmers’ market apple
price. Due to the lack of Colorado shipping point price data, the Colorado shipping point
price is estimated using the Washington shipping point price for Red Delicious apples
12
and chosen based on its dominant market position (Hu, 2012), the relatively short
distance to Colorado compared to other major production regions, and the consistent
availability of data in that series.
Weekly shipping point price, retail price and domestic truck rate report data for
fresh Red Delicious apples are collected from USDA’s Agricultural Marketing Service
(U.S. Department of Agriculture: Agricultural Marketing Service, 2012). Weekly
Colorado farmers’ market prices are collected from Colorado State University
Extension’s Fresh Produce and Meat Market Reports (2012). Weekly Midwest, Rocky
Mountain, East Coast, and West Coast on-highway diesel fuel prices are collected from
the U.S. Energy Information Administration (2012) to adjust the shipping point price to
account for direct transportation costs. The seasonal Colorado farmers’ market price only
covered a period from August to October, 2011, so the farmers’ market price is only used
as base price for estimation of price relationship parameters to populate the EDM. Based
on estimates outlined in Table 1 using data described in Appendix A, the production and
consumption levels for the rest of the country are estimated as 6,235 (SBs) and 4,851
million pounds (DBd), respectively.
Since there are no reported data on the volume of directly marketed apples in
Colorado, the amount of directly marketed apples is estimated based on the proportion of
directly marketed fruits in Colorado (see Hu, 2012 for more details on estimation).
Similarly, the consumption of local apples in Colorado is calculated using the proportion
of local food sales through all channels within the U.S. The resulting estimates of locally
13
marketed demand and supply are given in Table 1. 2 The estimated production and
consumption levels of apples used in the simulation of EDM are shown in Table 2.
Price premium & market margin
The price premium between local and domestic apple prices is estimated based on the
price difference between farmers’ markets and retail markets:
(14)
𝑙
𝑡𝐴−𝐶𝑂
=
𝑃𝐴𝑙
𝑃𝐴𝑑
3.35
− 1 = 1.61 − 1 = 1.08.
Table 3 presents the market margins estimated between the Colorado retail market and
domestic retail market, the Colorado shipping point (if it were to exist) (𝑃𝐴𝑆 ) and the
domestic shipping point (𝑃𝐵𝑆 ). Given those estimates, the resulting supply price difference
between selling in Colorado farmers’ markets (𝑃𝐴𝐹 ) and selling to Colorado shipping
𝐹
points (𝑃𝐴𝑆 ) is estimated as 𝑡𝐴−𝐶𝑂
= 0.89. Upper and lower bounds for the marketing
margin are only estimated for the short-run scenarios. In the long run, all transaction costs
and other factors that affect the price difference can be adjusted, thus the price difference
is assumed to be constant and the average value of the lower and upper bound is used.
More details of the estimation are available in Hu (2012). The estimated market shares
and price margins are shown in Table 3.Demand elasticities
The demand for apples in Colorado that differentiate local and domestic origins was first
estimated in a 2008 national survey data from Onozaka et al. (2008) which utilized to
2
Although the Census of Agriculture is available for 2007, the production and
consumption of fresh apples are estimated for 2008 and the proportions of direct
marketed apples and consumption of local apples are assumed to be stable from 2007
(when Census data was available) to 2008.
14
estimate the WTP and market share for local and domestic apples in Colorado. 3 The
estimation shows that the demand for fresh local apples in Colorado is a kinked demand
curve due to segmentation of consumers in the market (Figure 1). The horizontal axis is
the market share of consumers who would purchase local apples under a specific price.
As we can see in Figure 1, the kinked point is at price level of $1.80/lb and is close to the
prevailing retail price of conventional Washington Gala apples given in the survey’s
choice experiment ($1.89/lb). If the local apple price is lower than the kinked point (i.e.,
the price of local apples is lower than the price of domestic apples), almost all consumers
would buy local apples. However, if local apple price is higher than the kinked point (i.e.,
the price of local apples is higher than the price of domestic apples), only a segment of
the consumers would buy local apples.
Figure 2 suggests that about ten percent of consumers are relatively insensitive to
price changes, indicating they are relatively loyal local apple consumers. This is
consistent with the results given in Onozaka, Nurse and Thilmany McFadden (2011) that
11% of the sample in 2008 survey shop at direct markets where locally grown products
are sold. It is interesting to note that the premium suggested by the 11% representative
consumer in the 2008 survey data ($1.00/lb) is close to the premium estimated by actual
prices paid given historical price data ($0.81/lb).
Given the kinked demand that is revealed, we infer that this EDM analysis may
need to be handled differently than what is commonly done, which inherently serves as a
form of sensitivity analysis. The two separate EDMs will be estimated, one shocked at
3
Details of the data and estimation results can be found in Onozaka and Thilmany
McFadden (2011).
15
the median consumer level (median market buyers, or MMB) and one at an 11% market
share level of direct market (current market buyers, or CMB). Subsequently, a premium
of WTP for local apples with respect to domestic apples is set as $0.20/lb and $1.00/lb.
Based on the 2008 mean price in Southcentral and Northwest retail markets, the base
prices of domestic apples in Colorado and other states are set as $1.61/lb and $1.56/lb,
respectively. 4 The own price elasticities are derived through the formula:
(15)
𝜀ℎ𝑖𝑖 =
∆𝑀𝑎𝑟𝑘𝑒𝑡 𝑆ℎ𝑎𝑟𝑒 𝑖
∆𝑊𝑇𝑃𝑖
𝑊𝑇𝑃𝑖
∗ 𝑀𝑎𝑟𝑘𝑒𝑡 𝑆ℎ𝑎𝑟𝑒 𝑖,
while the cross-price elasticities are estimated using:
(16)
ij
εh =
∆Market Share i
∆WTPj
WTPj
* Market Share i .
The distribution of estimated individual-level WTP among Colorado consumers
compared against respondents from the remainder of the U.S. were not found to be
significantly different in the 2008 survey data, thus, domestic own price elasticities in
Colorado and the rest of the country are assumed to be the same.
Demand shocks
The WTP for local apples with respect to domestic apples is based on the premium
estimated by Onozaka and Thilmany McFadden (2011) at $0.20/lb for MMB and
4
The premium suggested by the median representative consumer premium of WTP in
2008 survey data ($0.20/lb) is different from the premium estimated through historical
prices ($0.81/lb). This may reflect the specific situation of Colorado, where supply of
local apples could not meet the demand for local apples which drives up the price of local
apples and above median consumers’ willingness to pay, so that supplies go to the most
fervent local consumers.
16
$1.00/lb for CMB. The introduction of the local label is assumed to be an exogenous
shock (𝛾) because the production source was not known by consumers before local
labeling programs were established and promoted (and the gain from that new labeling
effort was the justification many states made for the development of state programs).
Demand shocks are calculated based on the following formulas:
(17)
𝑃𝑟𝑒𝑚𝑖𝑢𝑚
𝛾 = − 𝑑𝑜𝑚𝑒𝑠𝑡𝑖𝑐 𝑝𝑟𝑖𝑐𝑒 ∗ 𝜀𝐴𝑙𝑙 = 0.019
Median Market Buyers (MMB)
and
(18)
𝑃𝑟𝑒𝑚𝑖𝑢𝑚
𝛾 = − 𝑑𝑜𝑚𝑒𝑠𝑡𝑖𝑐 𝑝𝑟𝑖𝑐𝑒 ∗ 𝜀𝐴𝑙𝑙 = 0.372
Current Market Buyers (CMB).
Supply elasticities
On the supply side, elasticities are estimated in two steps following Alston, Norton, and
Pardey (1995). First, decomposed supply elasticities for Colorado (𝛽𝐴 ) and other states
(𝛽𝐵 ), together with the supply elasticity of apples produced and consumed in Colorado
and marketed through shipping points (𝛽𝐴𝑆𝐴 ), are estimated based on Armington (1969)
and following Alston, Norton, and Pardey’s formula (1995). Second, further decomposed
supply elasticities are estimated using the aggregate elasticities estimated in the last step.
The second step is mainly based on Armington’s formula (1969) using parameters
defined in Tables 1-4 and Table A.1:
𝑖𝑗
(19)
𝐷𝑖 (𝜌
𝛽𝐴 = 𝜔𝐴𝐴
𝑖 𝛽𝐴 + 𝜏) − 𝛿𝑖𝑗 𝜏
(20)
𝐷𝑖 (𝜌
𝛽𝐵 = 𝑤𝐵𝐵
𝑖 𝛽𝐵 + 𝜏) − 𝛿𝑖𝑗 𝜏
𝑖𝑗
i=S,F,SA
i=S,F.
Both short- and long-run scenarios are considered in this study. Based on the
available data in the 2008 survey, the demand parameters in both the short- and long-run
models are assumed to be identical. In contrast, the supply elasticities are different for the
17
short and long run. For the long-run EDM, the aggregate own-price supply elasticity (β)
is chosen to be 1.0 based on previous literature and economic theory (Carpio and
Isengildina-Massa, 2010; Chavas and Cox, 1995). In the short run, when producers have
a limited ability to react to changes in demand by changing their supply, the aggregate
own-price supply elasticity is chosen to be 0.44 based on Chavas and Cox’s estimation.
The elasticity of transformation (𝜏) is included to ensure local marketed apples and apples
marketed through shipping points are treated as substitutes delineated by marketing
channel (following Carpio and Isengildina-Massa (2010) which is -0.5 in short run and 1.8 in long run. δij is the Kronecker delta (𝛿𝑖𝑗 = 1 when i=j; 𝛿𝑖𝑗 = 0 when i≠j). The
expansion elasticity of directly marketed apples (𝜌𝐹 ) is assumed to be 0.5 in short run
and 1.0 in long run. Apples marketed through shipping points have expansion elasticities
that are recovered from equations (21) and (22) (James and Alston, 2002):
(21)
(22)
𝐷𝐹
𝐷𝐹
𝜌𝑆 = (1 − 𝜌𝐹 ∗ 𝜔𝐴𝐴
)/(1 − 𝜔𝐴𝐴
)
𝑆𝐹
𝑆𝐹
𝜌𝑆𝐴 = (1 − 𝜌𝐹 ∗ 𝜔𝐴𝐴𝑙
)/(1 − 𝜔𝐴𝐴𝑙
).
5. EDM results
The price, quantity and changes in producer welfare due to new local labeling efforts and
promotions are presented in Table 4. Two scenarios were considered. The first scenario
was assuming a “fixed supply,” which analyzed the effects in the very short run when
suppliers could not react to the increase in consumer demand. In this scenario, the
increase in producer welfare was only due to the price change. The second scenario
allowed for “elastic supply”, which analyzed the effects in a relatively long run when
suppliers could react to the shocks in consumer demand. In this scenario, both the prices
18
and quantities adjusted to demand shifts. Within each scenario, simulations based on
MMB and CMB are presented.
Local labeling increases consumers’ WTP for local apples relative to domestic
apples in Colorado, and subsequently, the supply price for direct markets increases (by
67.67% in the short run, and 128.82% (MMC) and 115.33% (CMB) in the long run)
compared to the supply price for shipping points, which decreased by 21.33% in short run
and increased by 39.82% (MMB) and 26.33% (CMB) in the long run. Subsequently, in
Colorado, demand for both local and domestic apples increases in the long run (by 12.07%
(MMB) and 10.78% (CMB) for local apples, 874.14% (MMB) and 63.48% (CMB) for
domestic apples. The increase is larger for domestic apples partially due to the currently
low production capacity of Colorado, which cannot meet the state demand.
On the supply side, the increase in the direct market price compared to the
shipping point price lead the Colorado supply for direct markets to increase by 230.97%
(MMB) and 206.84% (CMB) relative to the supply for shipping points, which decreases
in the long run by 62.33% (MMB) and 65.17% (CMB). In terms of producer welfare,
Colorado suppliers for direct markets gain ($0.58 million) while Colorado suppliers for
shipping points lose ($1.37 million to $1.52 million) in the short run. In the long run,
however, both suppliers will gain, but the suppliers for direct markets will gain more
(over $2 million) than the suppliers for shipping points (over $1 million).
The big loss in the short run is due to the high premium between supply price for
direct market and shipping point which is based on historical price differences in these
channels, and explains why many remaining Colorado orchards now focus on locallyfocused market strategies. Additionally, the high cross price demand elasticity between
19
local apples and domestic apples derived from 2008 survey data contributes to the loss.
These market forces drive the shipping point price much lower, but the suppliers could
not reduce supply due to “fixed supply” in the very short run. Because of the lower level
of direct sales of apples in Colorado compared with supplies for shipping points, the gain
of suppliers for direct markets could not offset the loss of suppliers for shipping points,
but those losses may just be the persistence of broader competitive challenges small
production states have faced in the apple industry. Thus, there is a big loss in Colorado in
short run. The big change in prices and quantities are due to the high premium between
supply price for direct market and shipping point and the high cross price demand
elasticity between local apples and shipping points.
Overall, there is little difference in the results for lower bound parameters and
upper bound parameters partly due to their association with short run supply parameters.
In the EDM, the suppliers are assumed not to be able to react to the changes in demand in
the short run. However, this does not imply that the results are not sensitive to supply
prices and price differences. On the other hand, most of the changes based on CMB are
similar to the changes based on MMB, except the changes in Colorado demand for
domestic apples. This coincides with the fact that CMB consumers are more loyal to local
apples so their inelastic demand yields a larger relative shock to the demand for local
apples. Again, much of the shock resonates from the significant difference in the
elasticity of demand for domestic apples with respect to local apple price using CMB
(inelastic) and MMB (significantly elastic).
Sensitivity analysis
20
Sensitivity analysis was conducted to test the robustness of the estimated welfare
measures with respect to changes in the values of the parameters. Based on the approach
proposed by Zhao et al. (2000), truncated normal distributions are assigned to the twelve
parameters (𝜀𝐴𝑙𝑙 , 𝜀𝐴𝑙𝑑 , 𝜀𝐴𝑑𝑙 , 𝜀𝐴𝑑𝑑 , ∆WTP, 𝛽 , 𝜏, ρ F , 𝑡𝑙 , 𝑡𝑑 , 𝑡𝐹 , 𝑡𝑆 ) which specify the possible values
of each parameter and the corresponding probabilities. All other parameters are functions
of these twelve parameters. There are 100,000 simulated observations and 35,201 are left
for the short run and 59,933 for the long run after discarding the observations that violate
the substitution restriction (see Hu (2012) for complete results). The 95% confidence
intervals of the changes in producer welfare in the brackets in Table 4 show that suppliers
for direct markets would always gain from local promotion efforts. Based on median
consumers’ WTP, suppliers for shipping points would always lose in the short run due to
the shift of demand toward local apples, and subsequently, the supply shift toward direct
markets. Whether the gain by suppliers for direct markets could always offset the loss by
suppliers for shipping points in the long run depends on market conditions. However, in
the long run, suppliers for shipping points would more likely to gain. Most of the changes
in producer welfare simulated based on the variables and parameters in Table 1-4 lie in
the 95% interval.
The sensitivity to individual parameter analysis shows that the changes in
Colorado total producer welfare are sensitive to all parameters and the intervals show that
the influence of all parameters are statistically different from zero5. The changes are most
sensitive to the cross price demand elasticity between local apples and domestic apples
(𝜀𝐴𝑙𝑑 , 𝜀𝐴𝑑𝑙 ), aggregate own price supply elasticity (𝛽), elasticity of transformation between
5
More complete details are available in Hu (2012).
21
marketing through direct markets and through shipping points ( τ ), and the price
difference between supply price for direct markets and for shipping points (𝑡𝑓 ). The
results suggest that factors connected with the substitution between local apples and
domestic apples dominate the changes in producer welfare.
6. Policy Implications and Conclusions
This study assesses the potential gains and losses under the policies and programs that
support the revitalization of local and regional food systems by examining the impacts of
changes in the Colorado apple market: a sector that has seen active relocalization as a
response to its fading market share in the face of national and international competition.
By integrating changes in the prices and demanded quantities of Colorado labeled apples
relative to domestically produced apples, as well as changes in prices and supplied
quantities of directly marketed apples relative to more conventionally marketed apples
through major shipping points, the welfare changes for both consumers and producers
were explored.
The results show that, in the long run, consumers will shift their demand toward
local apples due to increased labeling efforts, promotion and access to markets that
implement local labeling (building on findings from earlier consumer analysis by
Onozaka and McFadden, 2011). Increases in Colorado’s production of apples would be
expected given higher prices, and also, new supplies would be marketed more directly
relative to the volume marketed through shipping points due to those demand shocks. In
short, the implications for Colorado producers are mixed depending on their market
orientation. For this reason, our findings motivate great interest in what the 2012 Ag
22
Census will show in terms of sales growth and number of producers re-investing in
specialty crop production and using more direct marketing channels.
These results are interesting for several reasons. First, local labeling programs,
together or in addition to an increasing number of direct markets, seem to be the
mainstays of the emerging local foods movement (Martinez, et al, 2010). As a
complement to work done by Carpio and Isengildina-Massa (2010), it shows there may
be long term gains to producers as they strategically position some of their produce to
more localized markets, thereby justifying resources invested in local promotion
programs by many state Departments of Agriculture and mapping tools that create
awareness of producers (e.g., MarketMaker and LocalHarvest).
Second, the widespread acceptance of an emerging local food sector working
alongside the conventional supply chains, in contrast to continued polarization that
suggests a zero-sum game where there will be losers to offset the winners of local food
gains, is precarious. We show there may be a short-run downside for the majority of
producers. However, these results may be sensitive to the elasticity of transformation
between marketing through direct markets and through shipping points (τ), and the price
difference between supply price for direct markets and for shipping points (𝑡𝑓 ), so more
empirical research on these economic forces is needed. For example, using price wedges
to represent marketing costs will cause the gain to producers to be overstated due to new
labeling efforts and promotional campaigns; we are currently constrained to refine such
estimates by the availability of data. However, this study’s extrapolation is more careful
than previous studies that assumed levels from the broader literature, since most of these
parameters were derived from market-based data. Furthermore, empirical evidence from
23
local orchards in Colorado suggest that local markets are the highest margin outlets, even
when marketing costs are considered, but locally-focused apple producers do concede
that some of the price premia is lost to additional marketing labor and expenditures.
Third, there are likely winners and losers in programs to support local foods, as is
the case with many policies. When markets start to fragment, welfare analysis should be
decomposed by those who are buying or selling through different channels in order to
assess the winners and losers, as well as including information for policymakers to assess
the net welfare change and whether prevailing market forces (i.e., strong consumer
interest in product assurances) inform how public choice would guide their decisions. In
the long run, consumers may “vote with their dollars” and shift to the market they feel
suits them best, and in response, suppliers could realize more perfect price discrimination
if they shift some of their production to other supply chains, which can be Pareto optimal.
In essence, suppliers will create more refined pricing strategies with regard to different
consumer groups, and redevelop marketing channels to improve on the efficiency losses
inherent in more fragmented distribution systems (Martinez et al, 2010).
One notable limitation of the study is that we only consider the fresh apple market,
even though most state marketing programs for local foods cover the full array of food
products. Although limiting our scope to a single market was necessitated by the
complexity of our model, this may also explain why these welfare changes are much
smaller than those reported by Carpio and Isengildina-Massa. Previous studies had to use
parameter estimates that were more simplified in terms of market channels (not allowing
directly marketed produce to vary from products offered in more conventional markets),
which is a limitation given that WTP does seem to vary by where consumers shop
24
(Onozaka et al.,
2011). This study is also constrained by the overly simplified
assumption that global markets do not matter. Based on the available data, this is
necessary at this point. However, given the importance of the international markets, with
increasing volume and value of international trade of fresh produce, as well as the
relevance of global food market in the context of local food movement, the extension of
the current research to include the international market seems warranted.
Despite several limitations of this research, the results and modeling efforts are
readily applicable to other products and regions where parameters can vary based on the
characteristics of the food product and the dynamics of the market, such as level of
imports, number of areas that have a suitable climate for production, lands and facilities
to produce the food, and level of processing needed. Moreover, this study provides useful
insights to those who seek to analyze local food markets, or more generally, any other
relocalized sectors using an EDM on the key factors that matter most such as the
sensitivity of consumers’ demand for a relocalized product to the price of a conventional
product, the transformation between marketing methods associated with relocalization,
and the costs of switching marketing channels.
25
References
Adams, D.C., Salois, M.J., 2010. Local versus organic: A turn in consumer preferences
and willingness-to-pay. Renew Agr Food Syst 25, 331-341.
Adelaja, A.O., Brumfield, R.G., Lininger, K., 1990. Product Differentiation and State
Promotion of Farm Produce: An Analysis of the Jersey Fresh Tomato. Journal of Food
Distribution Research 90, 73-85.
Alston, J.M., Norton, G.W., Pardey, P.G., 1995. Science Under Scarcity: Principles and
Practice for Agricultural Research Evaluation and Priority Setting. Cornell University
Press, Ithaca, NY.
Armington, P.S., 1969. Theory of Demand for Products Distinguished by Place of
Production. Int Monet Fund S Pap 16, 159-178.
Brester, G.W., Marsh, J.M., Atwood, J.A., 2004. Distributional impacts of country-oforigin labeling in the US meat industry. J Agr Resour Econ 29, 206-227.
Brown, C., Miller, S., 2008. The Impacts of Local Markets: A Review of Research on
Farmers Markets and Community Supported Agriculture (CSA). Am J Agr Econ 90,
1296-1302.
Burrell, A., Gijsbers, G., Kosse, A., Nahon, D., Requillart, V., Zee, F.V.d., 2006.
Assessment of Research Methodologies and Data Constraints: Preparatory Economic
Analysis of The Value-adding Process Within Integrated Supply Chains in Food and
Agriculture. European Commission.
Carpio, C.E., Isengildina-Massa, O., 2009. Consumer Willingness to Pay for Locally
Grown Products: The Case of South Carolina. Agribusiness 25, 412-426.
Carpio, C.E., Isengildina-Massa, O., 2010. To Fund or Not to Fund: Assessment of the
Potential Impact of a Regional Promotion Campaign. J Agr Resour Econ 35, 245-260.
Chavas, J.P., Cox, T.L., 1995. On Nonparametric Supply Response Analysis. Am J Agr
Econ 77, 80-92.
Colorado State University Extension, 2012. Fresh Produce and Meat Market Reports.
Fort Collins, Colorado. Online. Available at
http://www.extension.colostate.edu/boulder/ag/abm.shtml#prices. [Accessed February
2012].
Costanigro, M., McFadden, D.T., Kroll, S., Nurse, G., 2011. An In-Store Valuation of
Local and Organic Apples: The Role of Social Desirability. Agribusiness 27, 465-477.
Darby, K., Batte, M.T., Ernst, S., Roe, B., 2008. Decomposing local: a conjoint analysis
of locally produced foods. Am J Agr Econ 90, 476-486.
26
Galbraith, D., 2011. Top 10 Produce Industry Stories of 2011. The Packer. Online.
Available at http://www.thepacker.com [Accessed February 2012].
Gomez, M., 2010. Moving Local Foods from Farm to Consumers: Lessons from NYS
Apples. Research & Policy Brief Series, Cornell University.
Henneberry, S.R., Agustini, H.N., Taylor, M., Mutondo, J.E., Whitacre, B., Roberts,
B.W., 2008. The Economic Impacts of Direct Produce Marketing: A Case Study of
Oklahoma’s Farmers’ Markets, Southern Agricultural Economics Association Annual
Meeting.
Hinson, R.A., Bruchhaus, M.N., 2005. Louisiana Strawberries: Consumer Preferences
and Retailer Advertising. Journal of Food Distribution Research 36, 86-90.
Hu, W., 2012. Evaluating Performance and Welfare Dynamics of Local Food Systems:
The Case of Local Markets in the Colorado Apple Industry, Department of Agricultural
and Resource Economics. Colorado, Fort Collins. Available at
http://digitool.library.colostate.edu///exlibris/dtl/d3_1/apache_media/L2V4bGlicmlzL2R0
bC9kM18xL2FwYWNoZV9tZWRpYS8xNzE1MzE=.pdf.
James, J.S., Alston, J.M., 2002. Taxes and quality: A market-level analysis. Aust J Agr
Resour Ec 46, 417-445.
Loureiro, M.L., Hine, S., 2002. Discovering Niche Markets: A Comparison of Consumer
Willingness to Pay for Local (Colorado-Grown), Organic, and GMO-free Products.
Journal of Agricultural and Applied Economics 34, 477-487.
Lusk, J.L., Anderson, J.D., 2004. Effects of country-of-origin labeling on meat producers
and consumers. J Agr Resour Econ 29, 185-205.
Lynch, B., Coleman, J., 2010. Apples: Industry & Trade Summary. United States
International Trade Commission. Commissioners: Aranoff, S.,Pearson, D. R.,Tanner
Okun, D.,R. Lane, C.,A. Williamson, I.,A. Pinkert, D.,A. Rogowsky, R.,LaneyCummings, K., Washington, DC.
Mabiso, A., Sterns, J., L., H., Wysocki, A., 2005. Estimating Consumers’ WillingnessTo-Pay for Country-Of-Origin Labels in Fresh Apples and Tomatoes: A Double-Hurdle
Probit Analysis of American Data Using Factor Scores, American Agricultural
Economics Association Annual Meeting.
Martinez, S., Hand, M., Pra, D.M., Pollack.S., Ralston.K., Smith, T., Vogel, S., Clark, S.,
Lohr, L., Low, S., Newman., C., 2010. Local Food Systems: Concepts, Impacts, and.
Issues, in: Agriculture, U.S.D.o., ERS (Eds.).
Myers, G.S., 2004. Howard County Farmers’ Market Economic Impact Study 2004.
Report. Howard Co. (MD) Economic Development Auth. Agricultural Marketing
Program.
27
Onozaka, Y., McFadden, D.T., 2011. Does Local Labeling Complement or Compete with
Other Sustainable Labels? A Conjoint Analysis of Direct and Joint Values for Fresh
Produce Claim. Am J Agr Econ 93, 689-702.
Onozaka, Y., Nurse, G., McFadden, D.T., 2011. Defining Sustainable Food Market
Segments: Do Motivations and Values Vary by Shopping Locale? Am J Agr Econ 93,
583-589.
Pirog, R., Tyndall, J., 1999. Comparing Apples to Apples: An Iowa Perspective on
Apples and Local Food Systems. Leopold Center for Sustainable Agriculture, Iowa State
University, Ames.
Thompson, S.R., Anders, S., Herrmann, R., 2005. Regional-Origin Labeling with Quality
Control: An Economic Analysis, 2005 International Congress.
U.S. Department of Agriculture: Agricultural Marketing Service, 2012. Washington, DC:
USDA/AMS. Online. Available at http://www.marketnews.usda.gov/portal/fv.
[Accessed April 2012.].
U.S. Department of Agriculture: Economic Research Service, 2012. U.S. Apple Statistics.
Washington, DC: USDA/ERS. Online. Available at
http://www.ers.usda.gov/publications/FTS/2008/09SEP/FTS334.pdf [Accessed February
2012].
U.S. Energy Information Administration, 2012. Gasoline and Diesel Fuel Update. EIA.
Online. Available at http://www.eia.gov/petroleum/gasdiesel/ [Accessed February 2012].
United States Census Bureau, 2012. Online. Available at http://www.census.gov/.
[Accessed January 2012.].
United States Department of Agriculture: National Agricultural Statistics Service, 2008.
2007 Census of Agriculture. Washington, DC: USDA/NASS. Online. Available at
http://www.agcensus.usda.gov/Publications/2007/index.php. [Accessed April 2011.].
Zhao, X.Y., Griffiths, W.E., Griffith, G.R., Mullen, J.D., 2000. Probability distributions
for economic surplus changes: the case of technical change in the Australian wool
industry. Aust J Agr Resour Ec 44, 83-106.
28
Table 1. Supply and Demand Quantities Calculation
Variables
Supply
Values of Variables
Calculation
U.S. utilized fresh
apple (S)
6243.9 million
pounds
Production of the rest
of the country (SBS)
Colorado utilized fresh
apple (SA)
10 million pounds
6233.9 million pounds
(0.16% of U.S.sup)
(99.84% of U.S. sup)
Delta directly sold ag
products proportion
3.267%
Weights of Delta fruit
production w.r.t.
Colorado fruit production
38.164%
Mesa directly sold ag
products proportion
7.723%
Weights of Mesa fruit
production w.r.t. 4
counties total fruit
production
Montezuma directly sold
ag products proportion
43.912%
Weights of Montezuma
fruit production w.r.t. 4
counties fruit production
Montrose directly sold ag
products proportion
3.790%
Weights of Montrose fruit
production w.r.t. 4
counties fruit production
3.674%
Colorado shipping point marketed
apples (SAS):
9.47 million pounds
(94.73% of CO total supply)
Colorado directly marketed
apples (SAF):
0.53 million pounds
(5.27% of CO total supply)
(39.94% of local apple consumption)
4 counties weighted
proportion of direct
fruit sale: 5.27%
1.166%
A
Colorado consumption of local
apple through shipping points
(SASA) : 0.79 million pounds
(60.06% of local apple
consumption)
0.901%
29
Table 1. Supply and Demand Quantities Calculation, Continued
Variables
Values of Variables
Calculation
Demand
U.S. population
304,797,761
U.S. apple
consumption (D):
4846.28 million
pounds
U.S. per capita
consumption of fresh
apples
15.9 pounds
Colorado population
4,939,456
Colorado apple
consumption (DA):
78.54 million pounds
(1.62% of U.S. total
consumption)
U.S. local food sales
through all channels
U.S. all food sales
5000 million dollars
297220.491million
dollars
The rest of the country apple
consumption (DBd):
4767.75 million pounds
(98.38% of U.S. total consumption)
Colorado consumption of local
apples (DAl):
1.32 million pounds
(1.68% of CO total consumption)
Colorado consumption of
domestic apples (DAd):
77.22 million pounds
(98.32% of CO total
consumption)
Proportion of local
food sales :
1.68%
30
Table 2. Prices and Quantities Used for the EDM
Variablesa,b
Directly Marketed
(i=F)
Shipping Point
Marketed (i=S)
Demand
Farmers’ market price (𝑃𝐴𝑙 ) ($/lb)
3.35
Southcentral retail price (𝑃𝐴𝑑 ) ($/lb)
1.61
(𝑃𝐵𝑑 )
1.56
Northwest retail price
($/lb)
CO aggregate demand (DA) (mil.lbs.)
REST aggregate demand
(DBd)
CO consumption of local apple
78.54
(mil.lbs.)
(DAl)
CO consumption of domestic apple
(mil.lbs)
(DAd)
4767.75
1.32
77.22
Supply
Supply price (𝑃𝐵𝑖 ) ($/lb)
0.68
Supply price (𝑃𝐴𝑖 ) ($/lb)
3.35
(farmers’ market price)
1.61
(real supply price)
CO aggregate supply (SA) (mil.lbs.)
REST aggregate supply
CO directly marketed
(SBS)
(SAF)
(mil.lbs.)
10
6233.90
(mil.lbs)
0.53
(SAS)
9.47
CO shipping point marketed
(mil.lbs)
0.71 (LR)
0.75U 0.68L(SR)
(SASA)
CO SP marketed local consumed
0.79
(mil.lbs)
Note:
a. l stands for local apples and d stands for domestic apples. F stands for supply through direct markets and S
stands for supply through shipping points. SA stands for consumption of local apples through shipping
points. L stands for lower band and U stands for upper band.
b. All prices are deflated by 1998 January CPI (United States Census Bureau, 2012).
31
Table 3. Market Shares and Price Margins Used for the EDM
Parametersa
Directly Marketed
(i=F)
Shipping Point Marketed
(i=S)
Demand Market Shares
𝐷𝑑
𝑤𝐴𝑇
0.0159
𝐷𝑙
𝑤𝐴𝑇
0.0003
𝐷𝑑
𝑤𝐵𝑇
𝐷𝑖
𝑤𝐴𝐴
𝐷𝑆𝐴
𝑤𝐴𝐴
𝐷𝑖
𝑤𝐵𝐵
𝐷𝑙
𝑤𝐴𝐴
𝐷𝑑
𝑤𝐴𝐴
𝐷𝑖
𝑤𝐴𝐴𝑙
0.9838
0.0067
0.9933
0.0101
--
1
0.0168
0.9832
0.4015
0.5985(SA)
0.0001
0.0015
Supply Market Shares
𝑆𝑖
𝑤𝐴𝑇
𝑆𝐹
𝑤𝐴𝐴
𝑆𝑆𝐴
𝑤𝐴𝐴
𝑆𝑆
𝑤𝐴𝐴
𝑆𝐹𝑙
𝑤𝐴𝐴
𝑆𝑆𝐴𝑙
𝑤𝐴𝐴
𝑆𝑖
𝑤𝐵𝑇
0.0530
0.0790
0.9470
0.4015
0.5985
--
0.9984
Price Margins
𝑙
𝑡𝐴−𝐶𝑂
𝑑
𝑡𝐴−𝐶𝑂,𝑈𝑆
𝐹
𝑡𝐴−𝐶𝑂
𝑆
𝑡𝐴−𝐶𝑂,𝑈𝑆
1.08
0.22(LR) 0.27U 0.17L (SR)
0.89
0.25
Note: l stands for local apples and d stands for domestic apples. F stands for supply through direct markets and
S stands for supply through shipping points. SA stands for consumption of local apples through shipping points.
L stands for lower band and U stands for upper band.
32
Table 4. Demand and Supply Elasticities Used for the EDM
Parametersa
Directly Marketed
(i=F)
Shipping Point Marketed
(i=S)
Demand Elasticities
Colorado&Colorado(𝜀𝐴𝑙𝑙 )
-0.16 (MMB) -0.60 (CMB)
Colorado&Domestic(𝜀𝐴𝑙𝑑 )
5.31 (MMB)
Domestic&Domestic(𝜀𝐴𝑑𝑑 )
Domestic&Colorado(𝜀𝐴𝑑𝑙 )
Domestic&Domestic(𝜀𝐵𝑑𝑑 )
-0.35 (MMB) -0.35 (CMB)
7.73 (MMB) 0.56 (CMB)
Demand shock (𝛾)
0.019 (MMB) 0.372 (CMB)
2.84 (CMB)
-0.05 (MMB) -0.05 (CMB)
Supply Elasticities
Colorado grown(𝛽𝐴𝐹𝑖 )
1.79 (LR) 0.50 (SR)
Colorado grown(𝛽𝐴𝑆𝑖 )
-0.79 (LR) -0.06 (SR)
Colorado grown (𝛽𝐴𝑆𝐴,𝑖 )
-0.01 (LR)
Colorado grown (𝛽𝐴𝑆𝐴,𝑙 )
Colorado grown (𝛽𝐴𝑆𝐴,𝑑 )
Other States grown(𝛽𝐵𝑆𝑖 )
Aggregate own price elasticity
of supply(𝛽)
Elasticity of transformation(𝜏)
Expansion Elasticity(𝜌𝑖 )
0
(SR)
-0.01 (LR) 0
(SR)
1.01 (LR) 0.44 (SR)
-0.01 (LR) 0
(SR)
-0.01(LR) 0 (SR)
-0.01(LR) 0 (SR)
--
1.00 (LR) 0.44 (SR)
1.00 (LR) 0.44 (SR)
-1.80 (LR) -0.50 (SR)
1.00 (LR) 0.50 (SR)
1.00 (LR) 1.00 (SR)
Note:
a.
l stands for local apples and d stands for domestic apples. F stands for supply through direct markets and S
stands for supply through shipping points. SA stands for consumption of local apples through shipping
points. L stands for lower band and U stands for upper band.
33
Table 5. Price, Quantity, and Producer Welfare Changes
Variablesa
Fixed Supplyd
(MMBb)
lowere 𝑡𝐴𝑑 =0.17
upper 𝑡𝐴𝑑 =0.27
 = 0.019
Fixed Supply
(CMBc)
lower 𝑡𝐴𝑑 =0.17
upper 𝑡𝐴𝑑 =0.27
 = 0.372
Elastic Supply
(MMB)
Elastic Supply
(CMB)
 = 0.019
 = 0.372
%∆𝐷𝐴𝑙
-0.10
-0.10
0.01
0.02
12.07
10.78
%∆𝐷𝐴𝑑
849.29
849.03
62.46
62.46
874.14
63.48
%∆𝐷𝐵𝑑
-6.40
-6.15
-0.38
-0.13
0.83
0.24
%∆𝑆𝐴𝐹
0
0
0
0
230.97
206.84
%∆𝑆𝐴𝑆
0
0
0
0
-62.33
-65.17
%∆𝑆𝐴𝑆𝐴
0
0
0
0
-2.19
-2.25
%∆𝑆𝐵𝑆
0
0
0
0
14.82
1.33
%∆𝑃𝐴𝑙
110.02
109.99
120.34
120.34
113.33
125.14
%∆𝑃𝐴𝑑
2.92
2.92
12.33
12.34
5.33
17.14
%∆𝑃𝐵𝑑
-13.72
-23.71
-4.64
-14.62
-16.67
-4.86
%∆𝑃𝐴𝐹
67.67
67.67
67.67
67.67
128.82
115.33
%∆𝑃𝐴𝑆
-21.33
-21.33
-21.33
-21.33
39.82
26.33
%∆𝑃𝐵𝑆
-46.33
-46.33
-46.33
-46.33
14.82
1.33
34
Table 5. Price, Quantity, and Producer Welfare Changes, Continued
Variables
∆𝑃𝑆𝐴𝐹 (mil.$)f
∆𝑃𝑆𝐴𝑆 (mil.$)
∆𝑃𝑆𝐴 (mil.$)
Fixed Supply
(MMB)
lower 𝑡𝐴𝑑 =0.17
upper 𝑡𝐴𝑑 =0.27
Fixed Supply
(CMB)
lower 𝑡𝐴𝑑 =0.17
upper 𝑡𝐴𝑑 =0.27
Elastic Supply
(MMB)
Elastic Supply
(CMB)
0.58
0.58
0.58
0.58
2.37
2.00
(0.11, 1.15)g
(0.20, 2.03)
(0.38, 0.66)
(0.31,0.74)
(1.42, 34.54)
(1.98, 4.61)
-1.37
-1.52
-1.37
-1.52
1.84
1.19
(-3.68, 0.67)
(-8.08, 0.37)
(-2.28, 0.40)
(-2.11, -0.53)
(-90.04, 2.67)
(-0.12, 0.39)
-0.79
-0.94
-0.79
-0.94
4.21
3.19
(-4.17, 1.12)
(-4.19, 1.11)
(-4.16, 1.10)
(-4.14,1.11)
(-245.78, 14.56 )
(-23.15, 12.78)
Note:
a. All simulations are based on 2008 average prices and quantities that are deflated by 1998 January CPI (United States Census Bureau, 2012).
b. The shock to demand for local apples due to origin labeling efforts was estimated to be 0.019 using MMB.
c. The shock to demand for local apples due to origin labeling efforts was estimated to be 0.372 using CMB.
d. In the “fixed supply” scenario, a perfectly inelastic supply situation, suppliers cannot react to the changes in demand by changing the
quantities supplied. In the “elastic supply” scenario, suppliers can adjust their supply to the changes in the demand.
e. In the “lower 𝑡𝐴𝑑 ” scenario in “fixed supply”, lower bound estimated 𝑡𝐴𝑑 is used in the simulation, while upper bound estimated 𝑡𝐴𝑑 is used in
the simulation for “upper 𝑡𝐴𝑑 ” scenario.
f. ∆𝑃𝑆𝐴𝐹 is changes in producer welfare of Colorado suppliers for direct markets, ∆𝑃𝑆𝐴𝑆 is changes in producer welfare of Colorado suppliers for
shipping points, and ∆𝑃𝑆𝐴 is the total changes in producer welfare of Colorado suppliers.
The values in brackets are 95% probability intervals of producer welfare changes.
35
5
Local Price ($/lb)
4
3
2
1
0
0.2
0.4
0.6
Market Share of Buy Local (%)
0.8
1
Figure 1 Demand for Local Apples
Note: The dashed line is the real demand curve based on 2008 survey data, while the solid line is a trend line.
Price Premium of Local vs Domestic ($/lb)
2.0
1.5
1.0
0.5
0
0
0.2
0.4
0.6
Market Share of Buy Local (%)
0.8
1
Figure 2 Price Premium of Local Apples vs Market Share
36
Appendix A
Table A.1 Variables and Parameters Definition
Variables
Definition
Demand
𝑃𝑖𝑘
Retail (farmers’ market) price for apple k in region i
𝐷𝑖𝑘
Supply
The demand for apple k in region i
𝑃𝑖𝑘
The supply price for apples marketed through channel k in region i
𝑆𝑖𝑘
The supply of apples through channel k in region i
Transaction costs ratio
𝑙
𝑡𝐴−𝐶𝑂
Market margin between farmers’ market and retail market in Colorado
𝑑
𝑡𝐴−𝐶𝑂,𝑈𝑆
𝐹
𝑡𝐴−𝐶𝑂
Market margin between Colorado and domestic retail markets
𝑆
𝑡𝐴−𝐶𝑂,𝑈𝑆
Weights
𝑤𝑖𝑖𝐷𝑘
𝐷𝑘
𝑤𝑖𝑖𝑙
𝐷𝑘
𝑤𝑖𝑇
𝑤𝑖𝑖𝑆𝑘
𝑆𝑘𝑙
𝑤𝐴𝐴
𝑆𝑘
𝑤𝑖𝑇
The supply price difference between selling in direct markets and selling to
shipping points in Colorado
The supply price difference between Colorado and domestic shipping points
Region i share of demand for apple k (or apples marketed through channel k)
with respect to Region i total demand
Region i share of demand for local apples marketed through channel k with
respect to region i total demand for local apples
Region i share of demand for apple k with respect to U.S. total demand
Region i share of supply for apples marketed through channel k with respect to
Region i total supply
Region i share of supply for apples marketed through channel k with respect to
region i total supply of local apples
Region i share of supply for apples marketed through channel k with respect to
U.S. total supply
Elasticities
𝜀𝑖𝑘𝑘
Apple k own price demand elasticity in region i
𝜀𝑖𝑘ℎ
Apple k cross price demand elasticity with respect to apple h price in region i
𝛽𝑖𝑘𝑘
𝛽𝑖𝑘ℎ
Apples marketed through channel k own price supply elasticity in region i
𝛽𝐴𝑆𝐴𝑘
Apples marketed through channel k cross price supply elasticity with respect to
price marketed through channel h in region i
Colorado shipping point marketed local apple cross price supply elasticity with
respect to price of apple k (or price of apples marketed through channel k )
37
Table A.1 Variables and Parameters Definition, Continued
Variables
Other Parameters
𝛾
𝛼𝑘
𝜌𝑘
𝜌𝑆𝐴
𝜗
𝜏
𝜀
𝛽
Definition
Premium
Demand shock to local apples: 𝛾 = − domestic price ∗ 𝜀𝐴𝑙𝑙
Expenditure elasticity of apple k
Expansion elasticity of apples marketed through channel k
Expansion elasticity of shipping point marketed local apples
Elasticity of substitution
Elasticity of transformation
Aggregate own price elasticities of demand
Aggregate own price elasticities of supply
38
Appendix B: Equilibrium Displacement Model for the Colorado Apple Market
(B.1)
(B.2)
𝑑𝑙𝑛(𝐷𝐴𝑙 ) = 𝜀𝐴𝑙𝑙 𝑑𝑙𝑛(𝑃𝐴𝑙 )+𝜀𝐴𝑙𝑑 𝑑𝑙𝑛(𝑃𝐴𝑑 ) + 𝛾
𝑑𝑙𝑛(𝐷𝐴𝑑 ) = 𝜀𝐴𝑑𝑙 𝑑𝑙𝑛(𝑃𝐴𝑙 )+𝜀𝐴𝑑𝑑 𝑑𝑙𝑛(𝑃𝐴𝑑 ) −
𝐷𝑙
𝑤𝐴𝐴
𝐷𝑑
𝑤𝐴𝐴
𝛾
(B.3)
𝑑𝑙𝑛(𝑆𝐴𝐹 ) = 𝛽𝐴𝐹𝐹 𝑑𝑙𝑛(𝑃𝐴𝐹 )+𝛽𝐴𝐹𝑆 𝑑𝑙𝑛(𝑃𝐴𝑆 )
(B.4)
𝑑𝑙𝑛(𝑆𝐴𝑆 ) = 𝛽𝐴𝑆𝐹 𝑑𝑙𝑛(𝑃𝐴𝐹 )+𝛽𝐴𝑆𝑆 𝑑𝑙𝑛(𝑃𝐴𝑆 )
(B.5)
𝑑𝑙𝑛(𝑆𝐴𝑆𝐴 ) = 𝛽𝐴𝑆𝐴,𝐹 𝑑𝑙𝑛(𝑃𝐴𝐹 )+𝛽𝐴𝑆𝐴,𝑆 𝑑𝑙𝑛(𝑃𝐴𝑆 )+𝛽𝐴𝑆𝐴,𝑙 𝑑𝑙𝑛(𝑃𝐴𝑙 )+𝛽𝐴𝑆𝐴,𝑑 𝑑𝑙𝑛(𝑃𝐴𝑑 )
+𝛽𝐴𝑆𝐴,𝑑 𝑑𝑙𝑛(𝑃𝐵𝑑 )
(B.6)
𝑑𝑙𝑛(𝐷𝐵𝑑 ) = 𝜀𝐵𝑑𝑑 𝑑𝑙𝑛(𝑃𝐵𝑑 )
(B.7)
𝑑𝑙𝑛(𝑆𝐵𝑆 ) = 𝛽𝐵𝑆𝑆 𝑑𝑙𝑛(𝑃𝐵𝑆 )
(B.8)
𝑆,𝑆𝐴𝑙
𝑆𝐹𝑙
𝑑𝑙𝑛(𝐷𝐴𝑙 ) = 𝑤𝐴𝐴
𝑑𝑙𝑛(𝑆𝐴𝐹 ) + 𝑤𝐴𝐴
𝑑𝑙𝑛(𝑆𝐴𝑆𝐴 )
(B.9)
(B.10)
(B.11)
(B.12)
(B.13)
𝐷𝑙
𝐷𝑑
𝐷𝑑
𝑆𝑆
𝜔𝐴𝑇
𝑑𝑙𝑛(𝐷𝐴𝑙 ) + 𝜔𝐴𝑇
𝑑𝑙𝑛(𝐷𝐴𝑑 ) + 𝜔𝐵𝑇
𝑑𝑙𝑛(𝐷𝐵𝑑 ) = 𝜔𝐴𝑇
𝑑𝑙𝑛(𝑆𝐴𝑠 )
𝑆𝑆
𝑆𝐹
+𝜔𝐵𝑇
𝑑𝑙𝑛(𝑆𝐵𝑠 ) + 𝜔𝐴𝑇
𝑑𝑙𝑛(𝑆𝐴𝐹 )
𝑙
𝑑𝑙𝑛(𝑃𝐴𝑙 ) = 𝑑𝑙𝑛(𝑃𝐴𝑑 ) + 𝑡𝐴−𝐶𝑂
𝑑
𝑑𝑙𝑛(𝑃𝐴𝑑 ) = 𝑑𝑙𝑛(𝑃𝐵𝑑 ) + 𝑡𝐴−𝐶𝑂,𝑈𝑆
𝐹
𝑑𝑙𝑛(𝑃𝐴𝐹 ) = 𝑑𝑙𝑛(𝑃𝐴𝑆 ) + 𝑡𝐴−𝐶𝑂
𝑆
𝑑𝑙𝑛(𝑃𝐴𝑆 ) = 𝑑𝑙𝑛(𝑃𝐵𝑆 ) + 𝑡𝐴−𝐶𝑂,𝑈𝑆
Data for the Colorado Apple Market (see Table 2)
The production of apples for fresh use in Colorado in 2008 was 10 million pounds (SA), while the
production for fresh use for the United States was 6,245 million pounds (S) (United States
Department of Agriculture: National Agricultural Statistics Service, 2008). The U.S. per capita
consumption of fresh apples in 2008 was 16.2 pounds(United States Department of Agriculture:
National Agricultural Statistics Service, 2008), while the population of Colorado and the United
States was 4.94 million and 304.37 million in 2008, respectively (United States Census Bureau,
2012). Thus, the estimated consumption of fresh apples for Colorado and the U.S. in 2008 was 80
million pounds (DA) and 4,931 million pounds (D), respectively.
39