CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
PROGRESS REPORTS
Project title: Measuring and understanding the pattern of margins between farm and retail prices for
Calilfornia specialty crops to increase growers returns
Dec 2011
Data Compilation
As indicated in the last report, our goal in this project was to complete data collection and be ready for data analysis.
In this period of the report, we completed the data compilation and began data analysis. As outlined in the previous
report, in addition to commodity price information, it is also important to acquire other industry related information
specific to commodity, including domestic production, marketing methods, and international trade. In light of this,
our data consist of three broad categories including prices, production, and trade. Price data are vast. Price data
include detailed crop specific fruit prices, along with vegetable prices which can be used as comparison. These data
are collected from various sources including Bureau of Labor Statistics, National Agricultural Statistics Service and
Economic Research Service. Price data include prices at different marketing stages (farmgate and retail), and at
different time intervals (annual and monthly). We collected the time series data from as far back in time as is
available. These data goes back as far as 1913, but for some commodities only begin in 1996 (data time period
depends on the specific crop). Production data are mostly collected from USDA sources. Given California is the
major producer for fruits and vegetables, which are the commodities of our particular interest, the geographical region
for production data is limited to California. Production data include production and acreage of major fruit crops and
some representative vegetables. Trade data are collected from Foreign Agricultural Service, USDA. From the
Service’s Global Agricultural Trade System, U.S. import and export quantities and values are collected for most
major fruits including peaches, apples, nectarines, strawberries, kiwis, avocadoes, oranges for the period of 19892009. More detailed information is provided in the appendix.
Data Analysis
To summarize our output (thus far) on descriptive data analysis, we present two figures (in the appendix) which
describe the historical movement of aggregate price indices. The first figure presents annual Consumer Price Indexes
(CPI) for urban consumers for general food and two different food bundles, fresh fruit and fresh vegetables for the
period of 1950-2009 with the index values based on the prices of 1982-84. All three CPIs increased together at a
relatively low rate until the early 1980s. However, since then all three indexes increases at considerably different
rates, showing that fresh produce prices rose far more than general food prices. Compared to the price of 1982, the
CPIs for fresh fruit and vegetables rose 3.5 times and 3.1 times in 2008, respectively, while the general food index
rose 2.3 times. The second figure depicts farmgate and retail spreads for fresh fruit prices over the same period. Up to
the early 1980s, the producer price index (PPI) (or farmgate prices) is slightly higher than the consumer price index
(CPI). However, after this period, the CPI rose very rapidly while there was little change in PPI. As a result, the gap
between CPI and PPI widens. These two graphs show the central point of our project—for the past two decades, fresh
fruit retail prices have risen more than any other food groups. On the other hand, producer received prices have either
decreased or changed little.
SCBGP
Revised 12/08/2009
Attachment 4
Page 1 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
We are currently investigating historical prices for specific crops in relation to different marketing stages (farmgate
and retail), crop seasons (based on domestic crop seasons) and the importance of international trade. Our crop
specific analysis includes all major fruits. However, price analysis at different marketing stages is limited to the crops
where data are available.
Historical Consumer Price Index for Urban Consumers
350
250
150
50
-501950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
General food
350
Fresh fruit
Fresh vegetables
Historical Consumer Price Index and Produer Price Index
for Fresh Fruit
250
CPI for fresh fruit
PPI for fresh fruit
150
50
-501950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
After preliminary investigation on historical trend of data, our data analysis is based on the two levels of hypotheses.
The first is to verify whether the price gap between retail and producer levels are the relevant price gap for domestic
producers. That is, CPI is calculated based on the prices of in season as well as off season while PPI is based on only
in season prices. Thus, CPI relevant to our study (with the focus on farm income implications) would be the CPI for
in season prices. This investigation is done using monthly prices. The second set of hypotheses is developed to
understand why the prices have evolved as they are. Those hypotheses include: a. differential market power that has
changed over time, b. differential changes in degree of marketing services, c. differential changes in the prices of
marketing services for which some products are more intensive (including waste), d. change in the elasticity of
demand for fresh produce relative to other items so prices have changed given some market power, e. differential
changes in the farm price because of costs of production evolving differently (because of input prices), f. differential
changes in costs regionally (California relative to Iowa?) where different products are grown in different ratios in
different regions, g. differential changes in technical change by crop, and h. differential changes in product
SCBGP
Revised 12/08/2009
Attachment 4
Page 2 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
characteristics that are costly to supply. We also investigate the possibility of the availability of offseason crops and
quality adjusted (i.e., fresher or other improved characteristics) products.
Provide a comparison of actual accomplishments with the goals established for the reporting period.
As was outlined in the previous report, our project started late due to the administrative delay in funding transfer.
However, we steadily achieved the ojectives of our workplan. Our work schedule indicates that by this term we
would be finishing data collection and conducting preliminary descriptive data analysis, where the data summary will
be prepared at various levels, by broad level of crops, by specific crop, and by specific marketing channel. Given
this, our progress in this project closely follows the workplan schedule.
Present the significant contributions and role of project partners in the project.
I have been working with Dr. Daniel Sumner closely in each stage of work. Dr. Sumner provided his insight in
developing hypotheses and other stages of project.
Clearly convey progress toward achieving outcomes by illustrating baseline data that has been gathered to date and
showing the progress toward achieving set targets.
For data collection, as outlined above, we have collected a large amount of data from many different sources. They are compiled
to be ready for data analysis. Data analysis work is already under way. Preliminary data analysis is almost completed. Further
data analysis specific to the hypotheses mentioned above are being formulated. Our targets are specified in the workplan and we
are closely following the workplan schedule.
Apr 2011
In this period, our major activities can be summarized broadly in two areas, including additional data
collection and data analysis. More detailed description of activities conducted in each area is provided below.
On-going Data Analysis: This on-going analysis is conducted using annual time series of product price indexes. We
have analyzed the time trend of price indexes for fresh fruit as an aggregate compared to food or vegetable price
indexes. The time series indexes are mostly developed from data reported by the Bureau of Labor Statistics (BLS).
Our investigation indicates that the rate of increase in the historical price of the aggregate of fresh fruit has been
comparable to other food prices up to the early 1980s. However, since then, fresh fruit prices have risen much more
rapidly than general food prices. On the other hand, the producer price index (representing wholesale prices) of fresh
fruit has increase much less, resulting in the falling share of farmgate price in retail value for fresh fruit.
In the second stage of our analysis, we have expanded our investigation into specific crops. In this crop
specific study, our analysis uses more detailed data for each crop. The main purpose of this second stage analysis is
to investigate the farm margin at the crop specific level using more detailed price data. So far, we have relied on
annual price data published by the BLS. However, BLS generates annual price data by aggregating the monthly
figures using the simple averaging method. This means that the prices, regardless they are collected during the season
or off-season, are treated with the same weight in the annual price calculation.
SCBGP
Revised 12/08/2009
Attachment 4
Page 3 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
These simple average annual prices may be biased for two main reasons. First, we have witnessed an increase
in imports of fresh fruit in the last two decades. Therefore, products are available in the market during the traditional
“off-season” and off-season market prices are tied to import prices. Second, introduction of early varieties and plant
hybrid technology extended the season of domestic production. The annual fruit prices data published by BLS reflect
prices of the product supplied from domestic as well as import sources. Moreover, the annual price data may be
biased upward relative to the information from two or three decades ago because the price of early season crops,
which is generally higher than the mid-season price, is treated with the same weight in annual price calculation, even
though production in the early season may be in small quantity. In addition, even if the weights were correctly
assigned in the annual averages, the comparison with historical data might be misleading because the nature of the
products has changed. Simply stated “early-season” fruit is not really the same product as a “mid-season” fruit on
either the buyer or seller side of the market and the “early-season” price is likely to be higher. That means, the
aggregate price could rise, because early season fruit became a larger share, even if the price of neither type of fruit
increased. We will continue to investigate this data issue. Given that this study focuses on the farm share in retail
value in the fruit sector, our objective in the next step of analysis is to investigate the issues surrounding farm share.
We next construct more relevant price data and analyze them.
Construction of annual prices which reflect domestic supplies during the season: Our objective is to construct price
data which account for various supply flows during the season. To do so, we began with monthly price data. These
monthly price data are the most disaggregated data available from public sources on a consistent basis. We also
collected shipment information from Agricultural Marketing Service of USDA (we collected the data as much as
possible from Agricultural Marketing News). Then, using each month’s share of shipment during the season as a
weight, we constructed weighted season prices for the period of 1980-2010 (1980 was the earliest year for which
monthly retail prices were available from BLS). When shipment data are not available, we use the simple average of
the months which are defined as harvest months by UC-Davis crop specialists. When it is hard to define the season
due to wide geographic production areas, we used the monthly import and export data to define the season.
We have also begun to investigate in detail potential anomalies in the individual commodity price indexes
because of changes in product definitions, oddly timed base periods and changes in products covered in aggregations.
Analysis: We initially applied the weighted price method for five commodities, including fresh apples, fresh
strawberries, navel oranges, fresh peaches, and fresh grapes. We have summarized the initial results in graphic form
and presented the results for three commodities. We selected these three for presentation because each represents
unique characteristics in the market. Fresh apples in the U.S. market are mostly supplied from domestic sources and
strawberries have a season which is long relative to other fruits while table grapes are one of the fruits which
represent a large import share. Currently, about half of U.S. annual consumption is supplied from foreign sources.
While the prices of apples and strawberries were weighted using shipment data, shipment data for table grapes
separately were not available and thus prices of grapes were calculated as a simple season average. Figures A-C
present retail and farm prices for these three crops. The first and second graphs in each figure describe nominal and
deflated prices at the retail and grower level. For all these three crops, real grower prices had not declined as was
previously believed. Even though they fluctuated to some extent, the real grower price for fresh apples has been
steady over the three decades. On the other hand, the real price for fresh strawberries has declined about 20-25
percent (judging from the linear trend line). However, given the strawberry yield has more than doubled over the
SCBGP
Revised 12/08/2009
Attachment 4
Page 4 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
same period, a decline of real price per unit by 20 percent is relatively a small decline. The real price for fresh grapes
shows steady trend with wide fluctuation.
Figure D presents the farm share in retail value for each of three commodities considered. The farm share for
fresh apples has fluctuated at around 17 percent, whereas the farm share for strawberries has clearly trended
downward over time. The farm share for fresh grapes has fluctuated widely with no clear trend.
Figure A. Fresh Apples: Historical Weighted Average Prices in Current and Real Value
Fresh apples: Grower prices and retail prices in real
value ($/lb, 2000=100), weighted by monthly shipment
of domestic production
Fresh apples: Grower prices and retail prices in
current value ($/lb), weighted by monthly shipment of
domestic production
retail
price
1.40
grower
price
0.35
retail
price
1.20
grower
price 0.40
0.30
1.00
0.35
1.20
1.00
0.80
0.30
0.80
0.25
0.60
0.20
0.15
0.20
0.60
0.25
0.40
0.15
0.40
Retail price
0.10
0.20
0.20
Grower price
0.05
0.00
0.00
1980
1985
1990
1995
2000
0.10
Grower price
0.05
0.00
1980
0.00
Retail price
1985
1990
1995
2000
2005
Figure B. Fresh Strawberries: Historical Weighted Average Prices in Current and Real Value
SCBGP
Revised 12/08/2009
Attachment 4
Page 5 of 15
2005
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
Fresh strawberries: Grower prices and retail
prices in real value ($/lb, 2000=100), weighted
monthly shipment of domestic production
grower
retail by
price
Fresh strawberries: Grower prices and retail
prices in current value ($/lb), weighted by
monthly shipment of domestic production
grower
retail
price 0.90
price
3.00
0.80
2.00
0.80
2.50
price0.90
2.50
0.70
0.70
2.00
0.60
1.50
0.60
0.50
0.50
1.50
Retail price
1.00
Grower price
0.50
0.00
1985
1990
1995
2000
0.30
Retail price
0.30
0.20
0.50
0.20
Grower price
0.10
0.10
0.00
0.00
1980
0.40
1.00
0.40
0.00
1980
2005
1985
1990
1995
2000
2005
Figure C. Fresh Grapes: Historical Weighted Average Prices in Current and Real Value
Table grapes: Season average grower prices
and retail prices in real value ($/lb, 2000=100),
season including Jul, Aug, Sept and Oct grower
retail price
Table grapes: Season average grower prices
and retail prices in current value ($/lb),
season including Jul, Aug, Sept and Oct
grower
retail
price
price
2.50
0.50
2.00
1.50
1.00
1.80
0.40
1.60
0.35
0.40
1.40
0.35
1.20
0.30
1.00
0.25
0.80
1985
1990
1995
2000
0.20
Retail price
0.00
Figure D. Historical Farm Shares
SCBGP
Revised 12/08/2009
Attachment 4
Page 6 of 15
0.05
0.00
1980
2005
0.15
0.10
Grower price
0.20
0.05
0.00
1980
0.25
0.40
0.10
0.00
0.30
0.60
0.15
Retail price
Grower price
price0.45
0.45
0.20
0.50
2.00
1985
1990
1995
2000
2005
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
Share of grower price in retail price: fresh apples,
strawberries, table grapes
0.70
Fresh apples
0.60
Fresh Strawberries
0.50
Table grapes
0.40
0.30
0.20
0.10
0.00
1980
1985
1990
1995
2000
2005
Provide a comparison of actual accomplishments with the goals established for the reporting
period.
Compared to the goals stated in the plan, we had to spend more time and effort in
collecting data. That was due to the lack of already published data because they did not best suit
for our research purpose. This was not anticipated. Nevertheless, we are in good shape in
achieving the goals stated in the workplan. Our work schedule indicates that by this term we
would be finishing data collection and major analysis work is under way. Our current progress
in this project closely follows the workplan schedule.
Present the significant contributions and role of project partners in the project.
BLS is probably one of the most important data sources. USDA collects its own (price)
data at the grower level prices, but for retail prices, BLS is the only source. Probably the most
widely used data by applied researchers in the time series context would be annual time series.
However, it is very important to understand that BLS usually uses the simple average method to
construct annual data, rather than the quantity weighted method. (However, USDA uses the
weighted average method in the construction of annual grower received prices. Shipment data
are supplied from USDA’s Agricultural Marketing Services.) The consequence of this simple
average method used in annual retail price calculation is the prices which are higher than
otherwise. The implication of these higher prices in farm share calculation is serious, leading to
erroneously lower farm shares.
I have been working with Dr. Daniel Sumner closely in each stage of work. Dr. Sumner
provided his insight in every stage of the project including data related issues.
Clearly convey progress toward achieving outcomes by illustrating baseline data that has been
gathered to date and showing the progress toward achieving set targets.
SCBGP
Revised 12/08/2009
Attachment 4
Page 7 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
We gathered more data and continue analysis. It is vitally important to employ data
suitable to the project focus. While annual data are popular among applied researcher and BLS
is one of the most reliable sources, we again demonstrate in this study how important to use
properly executed data. We show that using more relevant data leads to somewhat different time
trends of marketing margins and farm shares.
If a project target has already been achieved, consider amending the outcome measure. This
permits the project staff to “stretch” the goals in order to go beyond what they are already
doing.
We have not yet accomplished the project goal. There is no need to amend the outcome
measure.
Nov 2011
Goals and Objectives: Our goal in this period was to investigate the margin between retail and
farm level prices more closely. In more detail, our objective was to make progress on three
questions: 1) how has the margin between farm and retail prices been divided between farm to
wholesale margin and wholesale to retail margin? 2) What has been the pattern in the margin
across seasons of the year? 3) what have been the correlations between the prices at different
links along the supply chain? and 4) do these patterns differ for different fruits?
Data and Methods: Our analysis thus far had been mostly conducted using monthly and annual
data. To achieve the goals stated above, our effort in this period was devoted into, first,
developing the retail and farm price data that included the least amount of transportation and
marketing inputs. For each of our initial selection of fruits, we designated a representative
producing region in the country and collect both farm and retail prices for that region. Further,
to help us compare farm and retail prices we chose the quality of the products traded at these
marketing channels to be as similar as possible. Thus, we selected the most disaggregated
data—specific to a variety, size of fruit, grade, and container (when available) from Agricultural
Market News, which is published by Agricultural Marketing Services, USDA. Agricultural
Market News reports publish daily, weekly and seasonal price aggregates for the shipping point
(or FOB), terminal (or wholesale) and retail markets. We have examined weekly prices in some
detail. The prices at the shipping point were provided in most detail, classified by product
variety, size, grade, container, and districts where products were originated. Terminal prices also
include these product/container information for each terminal market location. Retail prices
were reported in less detail; they are classified by variety, size (depending on the fruit), and
broadly defined region. Given data collection effort was time-consuming, we initially focused
on recent years.
Preliminary findings –price patterns at the vertically coordinated markets: Our preliminary
investigation can be best summarized graphically. The graphs below depict the weekly prices of
three supplying channels for each of five fruits, for several recent years (price lines are
SCBGP
Revised 12/08/2009
Attachment 4
Page 8 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
discontinuous because prices during the off-season were not available). First, we present the
specifics on data.
For fresh apples, the data period is from Oct. 5, 2007 to Sept. 23, 2011 (no data were available
before 10/05/2007) and the apple variety is red delicious and all prices are converted into per
pound price. Further specific itemized information on product is: for retail market, unit=1 pound,
region=Northwest of U.S.; for terminal market, region=Seattle, grade=WaExFcy,
origin=Washington, size=88s or mid-range package=carton tray pack; for shipping point,
region=Washington state, other specifications are the same as what are reported in the terminal
market.
For peaches, the data period is from 5/19/2007 to 10/1/2011, and the variety chosen is “various
yellow flesh available.” Other product specifics are: for retail, region=Northwest U.S.; for
terminal market, region=Los Angeles, size=42s, package=carton 2 layer tray pack; for shipping
point, region=Central and Southern San Joaquin Valley California, size=40-42s,
“preconditioned”.
For grapes, the data period is from 10/5/2007 to 10/7/2011. The variety is “red/white seedless.”
Other specifics include: for retail, region=Northwest U.S.; for terminal market, region=Los
Angeles, variety=Thompson seedless, size=large, origin=California and imports, package=all
containers; for shipping point, regions=Coachella Valley and Chile imports.
For strawberries, the data period is from 1/6/2007-10/8/2011. Data specifics are: for retail,
region=Northwest of U.S.; for terminal market, region=Los Angeles, size=medium to large,
origin=Oxnard and Salinas-Watsonville, package=flats 12-pt baskets; for shipping point,
region=Oxnard and Salinas-Watsonville, and other specifics are the same as reported in terminal
market.
For oranges, data period is from 1/6/2007 to 6/18/2011. AMS retail data are very sparse and
when reported those are on a per orange basis. Thus, for retail prices, we used BLS monthly
U.S. city average price data for Navel oranges, and used identical weekly prices for each month.
Other data specifics are: for terminal market, region=Los Angeles, variety=Navel with no
subvariety, grade=Shprs1, size=88s, origin=California, package=7/10 bushel cartons; for
shipping point, region=Central&Southern California and Arizon, other specifics are the same as
reported in terminal market.
SCBGP
Revised 12/08/2009
Attachment 4
Page 9 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
Prices of fresh grapes at shipping point, terminal
and retail level
5
0
10/5/2007
10/5/2008
Retail
10/5/2009
10/5/2010
Terminal market
Shipping point
Prices of fresh apples at shipping point, terminal
and retail level
2
1.5
1
0.5
0
10/5/2007
10/5/2008
Shipping Point
10/5/2009
Terminal market
10/5/2010
Retail
Prices of fresh peaches at shipping point, terminal and
retail levels
6
4
2
0
5/12/2007
5/12/2008
Shipping point
5/12/2009
Terminal market
5/12/2010
Retail
SCBGP
Revised 12/08/2009
Attachment 4
Page 10 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
Prices of fresh strawberries at shipping point, terminal
and retail levels
10
5
0
10/5/2007
10/5/2008
Shipping point
10/5/2009
10/5/2010
Terminal market
Retail
Prices of fresh oranges at shipping point, terminal and
retail levels
2
0
Shipping Point
Terminal Market
Retail
A few points stand out from the above graphs.
1) While shipping and terminal prices are closely correlated, the correlation of retail prices
with shipping and terminal prices is much weaker.
2) The FOB and terminal prices are relatively stable from week to week. However, retail
prices show much more weekly fluctuation.
3) Seasonal highs and lows are clear for FOB and terminal prices, but less so for retail
prices.
4) Reported strawberry prices are higher at the terminal level than at the shipping point
level, suggesting that fruit and contract characteristics may not be the same in these
markets.
These findings are significant in the sense that these were not found or expected in the monthly
price analysis.
Provide a comparison of actual accomplishments with the goals established for the reporting
period.
Weeky level data collection conducted during this term was not initially anticipated. However,
after we used monthly (shipment-weighted) level data, we realized that that information was not
adequate in analyzing the margin itself. We needed more disaggregated data and decided to tap
SCBGP
Revised 12/08/2009
Attachment 4
Page 11 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
into AMS’s weekly data. Compiling the AMS data in the form which can be used in our analysis
was time-consuming and tedious. However, this weekly data reveal information which was not
transmitted from the monthly data. This was a small deviation from our work plan. However, in
terms of our overall objective of the project, we are one step closer.
Present the significant contributions and role of project partners in the project.
The findings reported here can be only obtained from the most disaggregated data. These
findings were not even indicated in our previous progress report, which used relatively
disaggregated monthly level data. The findings here are very valuable on analyzing the margins
among vertically coordinated marketing channels. I have been working with Dr. Daniel Sumner
closely in each stage of work. Dr. Sumner provided his insight in every stage of the project
including data related issues.
Clearly convey progress toward achieving outcomes by illustrating baseline data that has been
gathered to date and showing the progress toward achieving set targets.
The most important feature of this study is the use of disaggregated data. This study is one of a
very few which employ disaggregated data being careful to assure product consistency—weekly
level prices for the uniform quality of fruits at all marketing levels from the markets located in
general board regions. Assembly of the data has been time consuming but enables us to consider
marketing margins in the most consistent way. Therefore, our progress is consistent with our
goals and targets.
If a project target has already been achieved, consider amending the outcome measure. This
permits the project staff to “stretch” the goals in order to go beyond what they are already
doing.
We have not yet accomplished the project goal. There is no need to amend the outcome
measure.
Problems and Delays
Note unexpected delays, impediments, and challenges that have been confronted in order to
complete the goals for each project. Explain why these changes took place.
As mentioned earlier, during this term, we collected additional data, which enables us to analyze
further in more detail. This was not anticipated at the beginning of the project. It is common
that as we gain more information about the research topic, additional data needs arise.
Mention the actions that were taken in order to address these delays, impediments, and
challenges.
This did not present any major problem because we were able to obtain new findings which
contribute to achieving our research targets. We are somewhat behind schedule for completion
but the project results will be richer as a result. We have been careful with the project budget, so
we expect to be able to complete the study and provide even richer information from the
research.
SCBGP
Revised 12/08/2009
Attachment 4
Page 12 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
Review measurable outcomes to determine if targets are realistic and attainable. An objective
that is too stringent should be scaled back and identified in the performance report. Keep in
mind that targets may slip due to all kinds of factors, such as employee turn-over and bad
weather.
To maintain our data collection effort, we have focused on five fruit crops. We do not believe
this will undermine the general applicability of our study in any serious way. Data availability
dictated the choice of these five crops. But, they represent a wide spectrum of fruit industry
characteristics--most important, perishability and farm and marketing concentration. Fresh
peaches and strawberries are quite perishable, while apples and oranges are much less perishable.
About half of fresh strawberry supply originates from a few large producer and marketing firms,
and this may indicate different pricing relationships for fresh strawberries compared to peaches
and the other fruits. Therefore, these five major fruits will be able to represent a broad spectrum
of fruits.
Apr 2012
Our main task in this period was to finish preliminary analysis using weekly level price data. In
building a time series model, the most important consideration is the consistency of the data with the
model assumptions. Thus, in order to arrive at a valid model specification, we have to check the
stationarity of the data. If the variables in the regression model are not stationary, the standard
assumptions for asymptotic analysis will not valid (that is, the usual “t-ratios” will not follow a tdistribution). Thus, we cannot validly undertake hypothesis tests about the regression parameters.
Further, if data are not stationary, there is a risk of spurious regressions (if two variables are trending
over time, a regression of one on the other could have a high R-square value even if the two are totally
unrelated). Thus, for any econometric analysis that uses time series data (particularly, for the data
containing the variables which trend similarly over time), investigating the stationarity property of the
data is a necessary procedure. Another test that has to be conducted before the model estimation is the
causality test, which investigates the direction of causality (i.e., this test statistically determines which
variable functions as an independent variable). We conducted unit root tests for our price data and the
causality tests between the shipping point price and wholesale price for each fruit considered. Test
results are summarized below.
Apples: The unit root tests for apple price data indicates that our null hypotheses were rejected
for both shipping point and terminal prices. Thus, we found no unit root, supporting the
stationarity of apple price data. Further, the causality test indicates that shipping point prices
Granger cause terminal prices. The optimal lag of the influence is found to be four.
Strawberries: The unit root tests for strawberry price data indicates that our null hypotheses were
rejected for both shipping point and terminal prices. Thus, we found no unit root and can
conclude that strawberry price data are stationary. Further, the causality test indicates that
SCBGP
Revised 12/08/2009
Attachment 4
Page 13 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
terminal prices Granger cause shipping point prices. The optimal lag of the influence is found to
be three.
Table grapes: The unit root tests for grapes price data indicates that our null hypotheses were
rejected for both shipping point and terminal prices. Thus, we found no unit root and can
conclude that grapes price data are stationary. Further, the causality test indicates that shipping
point prices Granger cause terminal prices. The optimal lag of the influence is found to be three.
Peaches: The unit root tests for peach price data indicates that our null hypotheses were rejected
for both shipping point and terminal prices. Thus, we found no unit root and can conclude that
peach price data are stationary. Further, the causality test indicates that shipping point prices
Granger cause terminal prices. The optimal lag of the influence is found to be four.
Note that shipping point prices Granger-cause terminal prices for all fruits except for
strawberries. That is, as has been commonly assumed or evidenced in the previous literature on
price transmission involving vertical markets, upstream prices influence downstream prices (i.e.,
shipping point price affects the terminal price). Our results indicate that this commonly held
expectation is upheld for all fruits, except strawberries. Our result on strawberry requires further
investigation, and in the next term, we will examine various aspects including the structure of the
strawberry industry to investigate this issue.
We have also conducted a preliminary (very preliminary) analysis of asymmetric price
transmission between shipping point prices and terminal prices. We have found positive
asymmetric price transmission for apples and negative asymmetric price transmission for
peaches and table grapes between shipping point and terminal prices. Given these results are
valid as mean value estimations, to better investigate the process of price transmission, we plan
to conduct quantile analysis. The quantile estimation will provide price transmission information
specific to the magnitude of price change.
Problems and Delays
As mentioned earlier, we had to collect additional data. This was not anticipated in the early
stage of the project. However, it is not uncommon that as we gain more information about the
research topic, additional data needs arise. During this term, we conducted preliminary
estimations on data properties and asymmetric price transmission possibilities. Nevertheless,
further work is needed to refine the model and then re-estimate the system. We will be
requesting a no-cost extension of this project for 6 months, which will revise the completion date
SCBGP
Revised 12/08/2009
Attachment 4
Page 14 of 15
CALIFORNIA DEPARTMENT OF FOOD & AGRICULTURE
SPECIALTY CROP BLOCK GRANT
BI-ANNUAL PROGRESS REPORT
CFDA# 10.170
of this project to December 31, 2012. The no cost extension request is currently being processed
at the UC Davis grant office (before it is sent out to the CDFA).
Future Project Plans
Once the model is fine-tuned, we plan to focus on the estimation of asymmetric price transmission, which
investigates empirical evidence on whether the exogenous price shock at the shipping point price (it
would be the wholesale price, depending on the causality of the price effect) affects the next prices in the
chain (mostly, the wholesale level) by different amounts depending on whether prices move up or down.
Further, to investigate the process of asymmetric transmission depending on the magnitude of the price
change, we will be conducting the quantile estimation. The quantile analysis generates the degree of
asymmetry in transmission specific to the level of price change that is classified using the quantile
system.
SCBGP
Revised 12/08/2009
Attachment 4
Page 15 of 15