Transcript Cattle prices 1990 - 2005

Lkassbi et al. (2008) Transmission of Price Trends
from Consumers to Producers
Why this is an interesting paper?
1 – Quite short
2 – Has to do with Canada and market concentration
3 - Light on the description of a rather complex mathematical concepts
4 - Illustrates some of the practical complexities of this type of analysis
Objective of the Paper
Examine the relationship between farm and retail food prices for several
agricultural products. (supply management)
Market power being exercised by the marketing chain, specifically retailers.
Terminology to define in simple terms:
Purchasing Power Parity
Alternative to using the exchange rate to measure the buying power of income
Based on a basket of goods, and the $US cost in two countries
Price flexibility
how price responds to a quantity change
supply or demand side of a market
Non-stationary versus stationary series
mean and variance do not change over time
Integration and Co-integration of time series
(1) Integration – 1 series
Differencing the series (once or more) produces a stationary series
(1  L) X t  X t  X t 1  X
where
L  the difference operator
(2) Co-integration – 2 series
some combination of two series produces a stationary series
E.G. subtracting one series from the other
=> they share some common element
Economic Model and Method
(1) Derive equations that connect retail prices to farm prices
Assuming a competitive market equilibrium
(2) Use data to determine:
Are results consistent with a competitive market ?
(3) If NOT => evidence that there might be an abuse of market power
Step 1: Derive Equations
Pr  Arf  F  Arw  W  Arz  Z
and
Pf  A ff  F  A fw  W  A fz  Z
where
Pr  retail price
Pf  farm price
F  farm sup ply
W  farm inputs
Z  dem andshifters
All variablesare expressedin logarithms
coefficients (A) are the price flexibilities
Step 2: Testing for Market Power
Competitive market => MVP of input equal to price of input
Farm AND Retail
If
MVP > P(input) => evidence of market power
Assuming competitive markets, the two equations can combined
Pr  B f  Pf  Bw  W  Bz  Z
(Equation 2)
coefficients (B) are a function of the price flexibilities
NOT competitive, the equation is a bit different
Pr  B f  Pf  Bw  W  Bz  Z
 Bc  C
(Equation 3)
Includes other variables (C) that influence retail price:
stocks of the farm products and farm input prices.
Practical (Statistical) Considerations
If the variables in (3) are “stationary”, then a standard F-test can be used
Ho: BC = 0
If the variables are “integrated” then:
Test if (2) OR (3) are co-integrated.
If (2) is co-integrated without the variable C, then BC = 0
=> evidence for a competitive market.
The Symmetry Test
The Concept:
When the retail price increases, the retail level demand for a farm product
should change in much the same way as when there is a change in the farm
price.
There is symmetry.
Arf   S f  A fz
(4)
Where S f is the cost share of the farm productto produce the retail product
=> Price flexibilities at the retail and farm level should be related
Change in a shift variable (e.g. bad weather) should be reflected in how
the change in the farm price affects the retail price.
Constant Returns to Scale
Producing the retail (food) product when using the farm product as an input
E.g. Are there constant returns to using corn to produce corn flakes?
Increasing returns => potential market power
The Test:
Price flexibility at the retail and farm level should be the same, given an
exogenous shift (shock) at the farm level
(bumper crop -- Z)
Arz  Afz
(5)
Data Used in the Analysis - Canada
(4) supply managed commodities: eggs, milk, chicken and turkey
Monthly data on farm and retail prices (CANSIM, Statistics Canada)
Quantity data are, based on per capita commercial disappearance data
1997 – 2005 for 98 observations
Results:
Data were not stationary (constant mean & variance)
Had to test for co-integration and Constant Returns
Co-integration relationships found for retail and farm prices for all products
Eggs: a single co-integration in eggs consistent with no market power.
Other three commodities have multiple co-integration relationships, consistent
with market power.
CRTS was rejected in all cases.
This result, by itself implies market power for all four commodities.
Conclusion
(1)
Conclude that their results are suggestive of market power for these
four commodities
(2)
Temper their conclusions by recognizing that the data they used
might not be the best
(3)
Other more sophisticated statistical methods could have been used