The New USDAERS Retail Pork Price Series: Evaluating Price Differences and Wholesale to Retail Price

1
The New USDA-ERS Retail Pork Price
SeriesEvaluating Price Differences and
Wholesale to Retail Price Response

• Dawn Thilmany, Oscar Arana and Steve Davies
• Colorado State University
• 2003 WAEA Meetings, Denver CO

2
The Issue at Hand.

• Will richer data allow us to better understand
  the retail and wholesale meat price relationship?
• New retail meat scanner data is USDA-ERS response
  to critique of BLS data
• Controls for featured price promotions often
  associated with meat
• Provides a volume index for the meat and share of
  meat sold through features

3
Project Overview

• Asked by USDA-ERS to explore how new retail meat
  scanner data may be used for empirical work
• Began with pork price series because of
  interesting seasonal dynamics and contrasts among
  cuts
• Limited by the time series that is currently
  available to analyze

4
Paper Overview

• Previous Research
• Wholesale to Retail Price Transmission
• Meat Retail Pricing
• Some Interesting Graphs
• Price, Feature and Volume Relationships
• Initial Empirical Modeling

5
Pork Price Transmission

• Schroeder (1988) investigated the price
  transmission of specific pork cuts
• Different cuts possess different lag lengths.
• Overall, there is evidence that retailers react
  more quickly to increasing wholesale prices than
  decreasing wholesale prices.
• Short-term price transmission for cuts was
  unidirectional from wholesale to retail level
• With the continuing decline in wholesaling
  activity relative to retailer's management of
  input procurement, (due to?) retail
  concentration, the outcomes may change

6
More Recent Work

• Goodwin and Harper (2000) investigated linkages
  among farm, wholesale and retail markets using
  weekly price data
• Farm, wholesale and retail pork prices were
  cointegrated and the transmission of shocks
  appeared to be largely unidirectional from farm
  to wholesale levels.
• Beef sector, Goodwin and Holt (1999)
• Farm, wholesale, and retail beef market shocks
  were also unidirectional, and that retail shocks
  were largely confined to retailers

7
Retailer Behavior

• According to Capps, Jr. et al (1994), retailers
  usually absorb some of the seasonal variation in
  supply and demand
• With increasing competition, retail strategies
  allow rapid adjustment in prices to prevent
  consumers from searching for lower prices and
  switching stores (Weaver et al, 1989).
• New loyalty programs with feature prices

8
New Retail Scanner Data

• Information on standard prices, featured prices
  (loyalty programs), total sales volume index, and
  an index of the share of sales under a featured
  price (142 meat product)
• We will focus on twenty-three pork products
• Only supermarkets with annual sales of 2 million
  or more and that voluntarily provide their
  information are included in the study (USDA-ERS,
  Perry and Southard).
• Commercial sources combine the data from
  retailers to protect confidentiality.
• These data do not include sales from food
  service or restaurants, butcher shops, warehouse
  clubs, or food distributors.

9
A Couple of Points

• Prices among cuts differ
• Not surprising but suggests differing wholesale
  values, shrink and labor
• Ranges suggest fairly stable syndicate prices
• Feature prices, and steepness of discount
  varies across cuts
• Lets examine some relationships graphically

10
Retail, Feature and Wholesale Prices-Center Cut
Pork Chops
11
Retail, Feature and Wholesale Rib Prices
12
Differences across Cuts

• Center cut pork chops show little difference
  between retail feature prices, and stable
  retail-wholesale price margin
• Pork chops are more like bacon, tenderloins and
  pork steaks, less variance in prices and use of
  feature promotional prices
• Ribs shows a more distinct difference between
  retail and feature prices
• Retail promotions, particularly in high demand
  periods (summer), a broader retail strategy
• These promotions influence volumes

13
Sales Volume, Share of Feature Sales and Retail
Prices-Tenderloin
14
Sales Volume, Share of Feature Sales and Retail
Prices-Ham
15
The Role of Features

• Tenderloins exhibit some seasonality
• Small share of sales through promotions, with
  fairly stable volume price levels.
• Ham has very seasonal sales, centering on the
  Christmas and Easter holidays,
• Share of sales through promotions are over
  one-half in several of those peak months.
• A retailing strategy may create more apparent
  retail price variability
• Weakening retail-whole price transmission

16
Methodology

• A pooled cross-section, time series analytical
  approach
• Limited degrees of freedom
• Theoretical and technical reasons to use
  cross-sectional time series models to increase
  the degrees of freedom (Gujarati, 1995).
• Plus, some of the primal cuts at wholesale level
  are the same for a number of retail cuts
• STATA follows Baltagis (1995) treatment of a
  two-stage least-squares extension of the simple
  panel-data random effects and fixed effects
  estimators

17
Hypothesis

• Based on a Granger causality, and following
  Schroeder (1988), the following model can be
  used.
• Rit bo ? bij Rit-j ?cij W it-j at
• Wit do ?dij Rti-j ?fij Wit-j et
• where t refers to month, i refers to individual
  pork cut, R is the retail price of the cut, W is
  the wholesale price, and a and e are random
  errors
• We will test whether there is a significant
  relationship between retail and wholesale prices
  in one or both directions
• As well as retail prices adjusted by promotions

18
Relationships with Just Prices

• Prices lagged by one month were negatively
  related to current retail (bt-1 -0.26) and
  wholesale (dt-1 -0.89) prices
• Prices lagged two months were significant for
  retail (bt-2 -0.22), but not wholesale prices
• A larger cumulative effect for wholesale prices
• Significant, persistent neg. causality of
  retail on wholesale prices, but wholesale pos.
  shocks retail prices at a greater level of
  magnitude
• Bidirectional causality, but wholesale model
  lacks explanatory power
• Retail prices are shocked by previous wholesale
  prices, and then recalibrate

19
An Expanded Price Model

• Add volume and feature indices
• Retail model stays much the same, but wholesale
  prices have more gradual influence on prices
• Unexpectedly, both Feature and Overall volume
  positively influence retail price levels
• Wholesale price model changes so that lagged
  values have positive influence, but retail prices
  still move counter to wholesale changes
• Only volume index has an influence (negative)
• Still remains a weak model

20
Adjusted Retail Prices

• No lagged, own price effects
• Lagged retail neg. shock and lagged wholesale
  pos. shock adjusted price
• No effect from feature price levels
• Both the volume index and feature volume neg.
  shock adj. prices
• Same lagged wholesale retail shocks
• Again, past wholesale price movements have pos.
  shocks, while retails are neg.

21
Conclusions

• Wholesale and retail prices have bidirectional
  influences
• Retail may overreact to wholesale shocks, and
  then gradually decline
• Wholesale model underspecified?
• Volume of sales promotions shock retail and
  adj. retail differently
• Differential behavior among cuts?

22
Appendices
23
Wholesale Prices, Base and Extended Models
24
(No Transcript)