An Empirical Analysis Of Meat Demand System Models In Peninsular Malaysia

The trend of meat consumption in Malaysia has shown a significant growth since early 1970’s. Total amount of meat per capita consumed has grown 57.3% from 1990 to 2003. Poultry and pork have been the dominant meats consumed in Peninsular Malaysia. However, the highest growths were chicken and beef, which have increased 99.15% and 57% from 1990 to 2003 respectively, while the consumption of pork per capita declined 22.3 %. Among the factors cited for this meat consumption growth were the consequences of economic development and associated changes in dietary patterns, taste and preferences in the country. Other reports indicated that the greater part of these changes had been due to consumer income growth and price effects besides high population growth. These spectacular changes make the answers to meat demand estimates in Malaysia more complex. Thus, empirical analysis employing the systems approach is used in this study. This study is motivated by two aspects. First, various authors have estimated demand of meat in Malaysia in the past and most of these estimations date back to before 1990s. The elasticities obtained from that estimates cannot be used for predictions since many structural changes have occurred in Malaysia since that time. Second, most of studies conducted in Malaysia have focused on estimation of and explanation on structural parameters in using one demand model. Very limited or no attention has been paid to compare among different demand models. Therefore we are interested to know if there are differences in the performance among demand models and choose the appropriate one that could fit in Malaysia data. Finally, we are interested to carriy out a test to identify any structural changes following the discovery of Nipah virus (NV) in Peninsular Malaysia and also to determine if changing the taste and preferences of consumer due to income growth has caused significant changes in the structure of meat demand in Malaysia. The main objective of the study was to select the best functional form specification in empirical meat demand research in Malaysia. This is carried out through the comparison of three different demand system models, which are: Rotterdam Demand System (RT), Linear Almost Ideal Demand System (LAIDS) and Almost Ideal Demand System with Error Correction Model (LAIDS-ECM) by utilizing Peninsular Malaysia’s per capita beef, mutton, chicken, pork and fish consumption and their retailed prices from 1960 to 2003. The second objective is to analyze the impact NV and trend on consumer demand for meat. Two criteria were used to select the model that better explain the observed Peninsular Malaysia beef, mutton, poultry and fish consumption patterns. First, we conducted non-nested tests designed to test the functional forms of the three demand models for using the likelihood ratio test. Second the mean square errors (MSE) for beef, mutton, pork, fish and poultry per capita consumption for each model were compared. All models were subjected the McGuire et al. (1995) system misspecification testing procedure to ensure that they were statistically adequate. Misspecification tests including normality, Functional form, endogeneity, correlations, demand restrictions, joint conditional mean and joint conditional variance tests were conducted. The study found that the LAIDS model is the preferred demand model for Malaysian data compared to the Rotterdam Demand System and LAIDS-ECM. This was confirmed by both non-nested model test and forecast accuracy. We proceeded the estimation of the price and expenditure elesticities for all models in order to see weather they had any differences in the behavior. The expenditure elasticities were not very different among the selected models ranging from 0.486 to 1.248. The highest expenditure elasticity corresponds to fish from the LAIDS-ECM. In addition, the LAIDS and Rotterdam have almost similar expenditure elasticities although the Rotterdam model has more elastic expenditure for beef and fish. For LAIDS and LAIDS-ECM models the gap is higher, for example, the elasticity for chicken from the LAIDS is 0.74, indicating chicken as a normal good, whereas that from the LAIDS-ECM model is 1.056 showing unit elastic demand. This conflict behavior can be explained in short term that the consumption of chicken is elastic as shown the results of the LAIDS-ECM, which represent the short-term changes in consumer’s consumption trend, but in long-run, it seems to be normal goods as LAIDS model predicted. Net of expenditure effects, mutton is more sensitive to own price changes in comparison to any of the other three meat products. Using the cross price elasticities, there is little substitutability between fish and the other three meat products, while beef shows the highest degrees of substitutability with the other products. This finding implies that the price policies for meats are not too sensitive for consumer at least in short term period. We also found that there is structural break in meat demand in Malaysia. The tests of structural change indicated that trend and Nipah Virus did impact Malaysian meat consumption. The result of the study indicates that Malaysian consumers reacted negatively to pork meat, while positively reacted to poultry, beef and mutton during and after crisis. This reaction accounted for the decrease in consumption of pork that resulted in lower expenditure and own-price elasticities for pork after the crisis. A negative trend in pork per capita consumption has occurred, while per capita consumption of chicken and beef had increased. This will give a good prospective for the beef and mutton industry, which have strong demand in future. Thus, the beef industry player has every incentive to cultivate and explore the consumers demand as this offer greater potential for growth and increased in future. Finally, the information presented in this study was intended to complement previous consumer demand researches, providing useful insights about the meat consumption patterns of the growing consumers in the Malaysia. The main contribution of our study on meat demand research formulation is that we compared the three most applied demand models for estimating the effects of income and prices on meat consumption in Peninsular Malaysia. Most previous researches just used only one model to make policy recommendation. We argue that models selection criteria should be considered before implementing the final outcome of demand to evaluate both meat demand and consumer welfare because a comparison of models would be important since different models produce different elasticities which can be used to predict future demand and is critical for guiding domestic production as well as to increase the consumer’s welfare. Therefore our analysis contributes in decision making process for meat policy formulation in Peninsular Malaysia