ASYMMETRY IN PRICE TRANSMISSION MECHANISM : THE CASE OF SLOVAK POTATO MARKET

This paper examines price transmission mechanism between farm and retail levels in vertical chain of potatoes. Time series analysis starting with cointegration approach is used to study price linkages between producer and consumer prices in potato market in Slovakia. We test for an existence of structural break in time series data (Gregory Hansen test) in the observed period and allow for an existence of non-linear relationship between prices at various levels of vertical chain by using threshold autoregressive models. We found an evidence of structural break and existence of asymmetry in price transmission along the potato supply chain.


INTRODUCTION
Price transmission from producer to consumer prices is a key characteristic of the supply chain and has received a lot of attention in the literature.There was a significant effort devoted in the empirical literature to quantify of magnitude, speed, and nature of price transmission from producer to consumer prices.Vertical price transmission may be imperfect if price changes at one level are not fully transmitted to another level; if there is a time lag between price adjustments at different levels or if there is an asymmetry in reaction between positive and negative price shocks (Bunte 2006).In agricultural markets we often observe that an increase of producer prices is transmitted more fully and faster to consumer prices while producer price decrease is passed-through the supply chain to consumer prices incompletely and at a lower speed (Vavra and Goodwin 2005).Empirical studies show that asymmetric price transmission is a rule rather than exception (Peltzman 2000).
In the food supply chain, asymmetric adjustment to price shocks can be observed due to the existence of adjustment costs, menu costs and information asymmetries (Ball and Mankiw 1994) Majority of empirical studies confirm asymmetry in food price transmission.Von Cramon-Taubadel (1998) investigated pork prices in Germany and found that the wholesale prices reacted more rapidly to positive shocks than to negative shocks originating at the farm level.Abdulai (2002) showed that increases in producer prices of pork in Switzerland are passed on to retail prices faster than reductions in producer prices.Vavra and Goodwin (2005) studied retail, wholesale and farm level prices of the U.S. beef, chicken and egg markets.Their results indicate significant asymmetries both in terms of speed and magnitude of the adjustment.
On the other hand, Serra and Goodwin (2003) found that asymmetries were not present in the price transmission of highly perishable dairy products in Spain.
Price transmission is closely related to the literature on determination of producer-retail marketing margins.Marketing margin is a difference between the price paid by consumers and that obtained by producers or alternatively marketing margins represent the value of services added to the basic agricultural commodity (Tomek and Robinson 2003).These services include packaging, labelling, processing, transport, information et cetera.The growing size of marketing margins is the source of many political disputes and often results in policy actions aiming at reducing them.
Despite huge economic and political interest in agricultural price transmission in Europe and elsewhere, there has not been any study on the subject conducted in Slovakia.Our paper is the first to evaluate price transmission along the potato supply chain in Slovakia and it is among the first attempts also to shed some light on the economics of determination of food margins at the Slovak potato market.
Potatoes are a key commodity for both producers and consumers in Slovakia.Producer prices of potatoes have a strong seasonal pattern but since 1990 the prices of potatoes are relatively stable.However, production of potatoes has been declining over the last decade.There are various distributional channels used in the marketing of potatoes in Slovakia.Some producers sell potatoes directly to final consumers while other producers of potatoes utilize middlemen (wholesale, retail).Potatoes are also sold directly to processors and restaurants.There are currently four producers groups in active in production and distribution of potatoes in Slovakia.Those producers groups have market share of 17 percent.According to survey conducted by Seifertova et al. (2009) 28 percent of consumers buy potatoes from retailers (hyper and supermarkets), 14 percent of consumers obtain potatoes directly from producers while 30 percent of households still produce their own potatoes.Other consumers use a combination of distributional channels to get potatoes.

Data Description
Monthly price data for Slovakia (from 1999 to 2011) are used to estimate the relationship between producer and retail prices of potatoes.Producer prices were obtained from the ATIS 1 and consumer (retail) prices come from the Statistical Office of the Slovak Republic.ATIS conducts a representative survey among producers of potatoes on area larger than 15 hectares.Weighted average prices are reported.Consumer prices are obtained from the Statistical Office of the Slovak Republic.The Statistical Office of the Slovak Republic reports average weighted prices for the whole territory of Slovakia.A logarithmic transformation of variables is applied, such that results may be interpreted in percentage change terms.

Price development in the Slovak potato market
Price development in the Slovak potato market follow year by year repeated pattern, starting with rise in potato 1 A division of the Agricultural Payment Agency in Slovakia prices during spring months, with later decrease in demand for potatoes in summer months (loss of demand from school canteens) leading to decrease in potato price (Table 2).There were some peculiarities in price development in Slovak potato market during the period 1999 -2011.The reasons for these deviations are following: • The first exemption occurred in 1999; short domestic supplies of potatoes during spring months caused rising imports followed by decrease in potato price to 0.22 EUR per kg.The price of potatoes during this period was permanently lower than the price level achieved in previous years (during spring time).Low potato prices remained until august 1999 when consumers began to lay in for winter stocks.In October 1999 the supply of potatoes increased due imports from Hungary and increased competition led to decrease in prices.Low prices remained also during the year 2000, the annual decrease in potato price reached 30.8 % which reflected the surplus in potato market over the whole supply chain.

Estimation Approach
We apply time-series modelling techniques to evaluate vertical price transmission from producer to consumer prices and vice versa.In this study, linear cointegration, cointegration with structural breaks in the series, and an asymmetric error correction model are employed to quantify an extent, speed and nature of price adjustment within the potato supply chain in Slovakia.The aim of using different approaches was to compare them and choose the best-fitting error correction model.As the first step, we test the stationarity of time series using two unit root tests: the augmented Dickey-Fuller (ADF) test and the Phillips-Perron (PP) test.The number of lags of a dependent variable is determined by the Akaike Information Criterion (AIC).If both time series are not stationary, they are suitable to test for cointegration relationship between them.We employ the Johansen approach to test for cointegration.
The Johansen approach starts with a vector autoregressive model and reformulates it into a vector error correction model: where Z t is a vector of non-stationary variables (producer and consumer prices), A are different matrices of parameters, t is time subscript, k is the number of lags and ε t is the error term assumed to follow i.i.d.process with a zero mean and normally distributed N(0, σ2) error structure.The estimates of Γ i measure the short-run adjustment to changes in the endogenous variables, while Π contains information on the long-run cointegrating relationships between variables in the model.
However, Meyer and von Cramon-Taubadel (2004) noted that the results of tests for price transmission must be interpreted with great caution if there is reason to suspect that there are structural breaks in the price series being investigated.Particularly in transition economies, like Slovakia and other Central and Eastern European countries, there were a lot of changes (economic reforms, significant foreign direct investments in retail sector, EU integration, adoption of CAP, adoption of euro…) that could cause structural break in time series data.From this reason we used a Gregory -Hansen test (1996) that allows for the presence of a onetime endogenously determined structural break in the cointegrating vector.A cointegration procedure, which allowed for structural breaks, was used by Bakucs, Falkowski and Ferto (2012) and others.
We consider four Gregory -Hansen types of models (standard cointegration, cointegration with level shift, cointegration with level shift and trend, and cointegration with regime shift).
where Υ is the dependent variable, Χ is the independent variable, t is time subscript, ε t is the error and k is the break date, φ is a dummy variable (0 for t ≤ k and 1 for t > k).
The above cointegration tests assume symmetric price transmission.In order to capture asymmetric movements in the residuals, Enders and Granger (1998) and Enders and Siklos (2001) propose to use threshold cointegration approach.Assuming the long run relationship between two non-stationary variables X and Y: where μ is the error term.Engle and Granger (1987) show, that cointegration exists if the null hypothesis ρ=0 is rejected in: where ξ is the error term for the residuals.Adjustment of the series of residuals expressed in 1 − t ρµ would be symmetric.To capture the asymmetry in adjustment process, a two-regime threshold cointegration approach should be used: where I t is the Heaviside indicator I t =1 if μ t-1 ≥ τ or I t =0 if μ t-1 < τ.If μ t-1 is bigger than the threshold τ, then adjustment is at the rate ρ 1 .If μ t-1 is smaller than the threshold τ, adjustment is shown in ρ 2 .When ρ 1 =ρ 2 , then the adjustment process is symmetric.If the null hypothesis ρ 1 =ρ 2 =0 is rejected, then X and Y are cointegrated and the following threshold autoregressive model (TAR) is estimated: where ΔY t and ΔX t are dependent and independent variables in their first differences, E is the error correction term, δ represents the speed of adjustment coefficients of ΔY t if Y t-1 is above and below its long-run equilibrium, θ, δ, α and β are coefficients and υ is the error term, t is time subscript and j is the number of lags.Two error correction terms are defined as: Enders and Granger (1998) and Enders and Siklos (2001) proposed also a model for cointegration, known as momentum threshold autoregressive model (MTAR).The term "momentum" describes the rate of acceleration of prices and takes into account steep variations in the residuals; it is especially valuable when the adjustment is believed to exhibit more momentum in one direction than the other.Heaviside Indicator in this case is Threshold error correction models were used for example by Goodwin and Holt (1999) Abdulai (2000Abdulai ( , 2002) ) used both TAR and MTAR models and found out, that the MTAR models fit data better than the others.
To summarize, four asymmetric models are considered in our study.They are threshold autoregression model with threshold value equal to zero (TAR); threshold autoregression model with threshold value estimated (consistent threshold autoregression model -cTAR); momentum threshold autoregression model with threshold value equal to zero (MTAR); and consistent momentum threshold autoregression model with threshold value estimated (cMTAR).A model with the lowest AIC and BIC will be used2 .

Cointegration results
Non-stationary time series can lead to statistically significant results due to purely spurious correlation.We therefore tested for the stationarity of the price series using augmented Dickey Fuller (ADF) and Phillips Perron (PP) tests.The Augmented Dickey-Fuller and Phillips-Perron tests confirmed that all our time series are non-stationary; we stationarized them by taking first differences.The tests indicated that all variables were stationary in first differences.The lags of the dependent variable in the tests were determined by Akaike Information Criterion (AIC).The stationarity tests showed that the original time series are non-stationary and could be used for cointegration analysis.
Johansen cointegration test results (Table 3) indicate that there is no cointegration relationship between producer and consumer prices of potatoes.Johansen test does not reject the null hypothesis of nocointegration.However, when we allow for the existence of structural break (Gregory -Hansen procedure) we find cointegration relationship with structural break.The estimated structural break date is March -July 2008.These structural breaks are related to the impact of turbulence on the global commodity markets and food commodity price peak in 2008.

Threshold cointegration
Threshold cointegration models allow for non-linear relationship between producer and consumer prices and vice versa.The theory does not guide us in the exact model specification and therefore in this paper we used four different threshold models: threshold autoregression model, consistent threshold autoregression model, momentum threshold autoregression model, and consistent momentum threshold autoregression model.We report the results for models with the lowest AIC and BIC.Threshold cointegration tests lead to the same results as Gregory -Hansen test.Estimated models show, that the prices are cointegrated with threshold adjustment (Table 4).From the tests it follows that there is strong evidence of negative asymmetry for producer price of potatoes.This means that in the case of potatoes retailers react faster to the shocks that stretch the margin than to shocks that squeeze the margin.In other words, consumer prices of potatoes react more fully or rapidly to a decrease in producer prices than to an increase in producer prices.
, inventory behaviour of retailers (Balke et al. 1998; Reagan and Weitzman 1982), the nature of government intervention in agricultural commodity markets (Gardner 1975; Kinnucan and Forker 1987), asymmetric information among the firms (Bailey and Brorsen 1989), the market power (Zachariasse and Bunte 2003; Wann and Sexton 1992; Gohin and Guyomard 2000; von Cramon-Taubadel 1998 and others), the interaction between market power and economy of scale (McCorriston et al. 2001; Lloyd et al. 2006), intertemporal optimizing behavior of firms (Azzam 1999), the form or retail demand and farm input supply (Weldegebriel 2004), the share of commodity costs in the cost of final product (Bettendorf and Verboven 2000) and other reasons.
Cointegration with Level Shift and Trend t

Table 2
Descriptive statistics of potatoes price series

Table 1
Production and trade data for potatoes in Slovakia Development of producer and consumer price of potatoes (in EUR per kg).Price development in domestic and foreign potato markets (in EUR / 100 kg) potatoes gradually rose and in April 2006 it was by 140.7 % higher than the year before.In addition, the usual decrease in potato price during summer holidays was milder in 2006 than it was before.In December 2006 potato price reached 0.35 EUR/kg

Table 3
Johansen cointegration test results Source: calculated.