Asymmetric Price Transmission in the Brazilian Rice Market: A review of methodologies

Jacques Henrique DiasDoctorate degree student at Universidade Estadual de Maringá

Dr. José Luis ParréProfessor at Universidade Estadual de Maringá

Dr. Waldemiro Alcântara da Silva NetoProfessor at Universidade Federal de Goiás

AbstractThe following lecture is intended to present and brieflydiscuss some arguments on the price asymmetry literature byapplying the most used methodologies to the case of ricemarket in the Brazilian economy. We aim to assess theeconometric models by the aftermath of each of them to thatspecific case. The results using the Error Correction Model(ECT) indicated that there is some asymmetry but it is notpossible to infer that this asymmetry remains to the long run.Previous methodologies also showed that upward anddownward price movements between these markets occurredin a different level and/or speed, the variables aresignificantly different for both rising and falling price phases,which demonstrates that the more sophisticated models canindeed forecast results accurately, shedding light on theoutput of prior models.

Background

By common sense we can perceive that somemarkets are not completely integrated, withupward and downward price movementsbetween these markets occurring in differentspeed, magnitude or both. Price surges and fallsmay not be equally transmitted throughout aproduction chain and this has been a matter ofconcern for many researchers especially in theagricultural economics.

Noteworthy is the fact that economic theory hasdifferent ways to explain APT, mostly, marketpower and other costs inherent to asymmetrieson price transmission. First works dealt withthese cases as exceptions (Tweeten and Quance1969, Woffram 1971, Houck 1977, Ward 1982).

A seminal work by Peltzman 2000 demonstratedthat the asymmetry in price transmission (APT)could be a rule for the great part of the 282products assessed in his work.

The literature in this issue is becoming vast,accompanied by the development of neweconometrical methodologies applied to APTworks (Von Cramon-Taubadel and Loy 1996,Goodwin and Holt 1999, Abdulai 2002).

Main Objectives

This research aims to:1. Tackle the econometric models by showingthe results of these models to the specific case ofrice in the Brazilian economy from the pricetaken by the producers in one node of the chainto price paid by consumers at the end node ofthis product chain;2. Apply the ECT methodology to data in orderto verify long run equilibria among the series.

Footnote #1

The aftermath of the different models to assessthe data might be unclear, but the scope here ismerely to use the framework and present whatkind of output emerge from the developmentmade by the literature on price asymmetriesrather than discussing the details of the pureeconometrics outlined in the footnotes of eachmethod.

Data

We used monthly data retrieved from basicallytwo sources: Agrolink for the price (R$/kg)received by the producers in Rio Grande do SulState and Ibge for the price paid by consumers inthe metropolitan area of São Paulo*, the rangegoes from April, 2004 to June, 2016 whichresults in 147 observations.

*Up to this presentation.

Graph

01

23

$/kg

Jan-04 Jan-06 Jan-08 Jan-10 Jan-12 Jan-14 Jan-16Month

consumer producer

Producer / ConsumerPrice per kg

size: 48,804

vars: 43 20 Nov 2016 21:29

obs: 147

Contains data from base.dta

. describe

Table 1: First Model – Perfect Symmetry Hypothesis

Where:a = Consumer pricel = Producer price

Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

Table 7

VECM – Vector Error Correction Model

Steps (pre-estimation procedures): 1. Causality;2. Unit Root;3. DFA (Augmented Dickey Fuller);

1. Model Selection: Trend and/or Constant Term4. Coefficients.

Causality

• Granger Causality Test*:

As expected, from all previous work assessed,the price paid by consumer is determined by theprice received by producers based on betterresults for the chi squared*. Tests wereperformed with 2, 5 and 12 lags experimentally.(Brandão, 1985 apud Aguiar, 2004).

**For 2 and 12 lags we cannot reject the possibility of the consumer price determining the producer price

Stationarity assumed.

Unit Root

Dickey Fuller test and Phillips-Perron gave thesame results, but for PP-Test we used 5 lags. (Aguiar, 2004)

ADF - Regressions

Three different regression were performed to ascertain a better goodness of fit, the model withtrend and constant term was choosed.

Coefficient

Only the model with trend and constant termshowed a negative L1 coefficient.

Johansen cointegration testThe series are cointegrated in order I(1), so, a VEC model is proposed (Taubadel & Loy, 1996).

Model:

Results

1. The results imply in a faster adjustment forrising price phase, generally 2 periods, and aslower adjustment in falling prices phase, morethan 4 lagged values were significant;2. The series converge;

3. An alternative model with cumulative valuesfor the dummy of price difference was notsignificant;4. The number of zeros in this model is a matterof concern;5. Positive sign of the coefficients for risingphase.

Producer Price Risinglags Coefficient Standard Error P>|z|

_ce1 -0.777264 0.0947465 0

L1

1 -2.228368 0.2659797 0

2 -1.118815 0.2797338 0

3 0.0801676 0.2433794 0.742

4 -0.0564704 0.1776471 0.751

Producer Price Fallinglags Coefficient Standard Error P>|z|

1 -1.293027 0.2090468 0

2 -0.9329264 0.2086226 0

3 -0.5686962 0.1821531 0.002

4 -0.4489684 0.150468 0.003

Autocorrelation

Thank You!

Contact:jacqueshenrique@yahoo.com.br

jhenriquesdias@uel.br

Obs: Stata Code and Database are avaiable for those who wants to help me

References• Abdulai, Awudu. "Using threshold cointegration to estimate asymmetric price transmission in the Swiss

pork market." Applied Economics 34, no. 6 (2002): 679-687.• Aguiar, Danilo RD. "Impacto dos custos de comercialização nas margens produtor-varejo de arroz e de

feijão em minas gerais1." Teoria e Evidência Econômica, Passo Fundo 12, no. 22 (2004): 51-76.• Brandão, Antonio Salazar P. "Moeda e preços relativos: evidência empírica." Brazilian Review of

Econometrics 5, no. 2 (1985): 33-80.• Cramon‐Taubadel, Stephan, and Jens‐Peter Loy. "Price asymmetry in the international wheat market:

Comment." Canadian Journal of Agricultural Economics/Revue canadienne d'agroeconomie 44, no. 3 (1996): 311-317.

• Goodwin, Barry K., and Matthew T. Holt. "Price transmission and asymmetric adjustment in the US beef sector." American Journal of Agricultural Economics 81, no. 3 (1999): 630-637.

• Houck, James P. "An approach to specifying and estimating nonreversible functions." American Journalof Agricultural Economics 59, no. 3 (1977): 570-572.

• Peltzman, Sam. "Prices rise faster than they fall." Journal of political economy 108, no. 3 (2000): 466-502.

• Tweeten, Luther G., and C. Leroy Quance. "Positivistic measures of aggregate supply elasticities: some new approaches." American Journal of Agricultural Economics 51, no. 2 (1969): 342-352.

• Ward, Ronald W. "Asymmetry in retail, wholesale, and shipping point pricing for fresh vegetables." American journal of agricultural economics 64, no. 2 (1982): 205-212.

• Wolffram, Rudolf. "Positivistic measures of aggregate supply elasticities: some new approaches—some critical notes." American Journal of Agricultural Economics 53, no. 2 (1971): 356-359.