Australasian Agribusiness Review - Vol.14 - 2006

Paper 15

ISSN 1442-6951

---

 

The FMD Outbreak in the Taiwan Pig Industry and the Demand for Beef Imports into Taiwan

Hui-Shung (Christie) Chang, Chung-Jen Hsia and Garry Griffith*

 

Abstract

In March 1997, a FMD epidemic broke out in the Taiwan pig industry and within four months some 40 per cent of the pig population was lost to the disease. The demand for pork fell substantially following the outbreak due to food safety concerns by consumers, and this raises the question of whether there were any related impacts in the demand for other meats. The objectives of this study were to determine econometrically the relative impact of the FMD outbreak on the demand for imported beef from Australia, New Zealand and the United States, and to draw marketing implications for beef suppliers to Taiwan. The study found that overall the FMD outbreak had little impact on beef imports although beef demand had changed slightly for New Zealand and United States over a six-month period. This result suggests that there might be a missed opportunity for beef suppliers, given that food scares are generally short-term in nature and a quick response immediately after the outbreak is required to take advantage of the situation.

Key words: competitiveness, consumer response, import demand, meat demand, FMD.

* Dr Hui-Shung (Christie) Chang is a Senior Lecturer in the School of Economics, University of New England, Armidale, Australia; Dr Chung-Jen Hsia is a Research Fellow in the Agricultural Marketing Division, Council of Agriculture, Taipei, Taiwan; and Dr Garry Griffith is a Principal Research Scientist in the NSW Department of Primary Industries, Beef Industry Centre, Armidale, Australia and an Adjunct Professor in the School of Economics, University of New England, Armidale, Australia.

The contact details for Chang are: (61) (2) 67732855 by phone, (61) (2) 67733596 by fax, and hchang@une.edu.au by e-mail.

1. Introduction

In March 1997, a foot and mouth disease (FMD) epidemic broke out in the Taiwan pig industry and within four months some 40 per cent of the pig population was wiped out (Shieh, 1997). The demand for pork fell substantially following the outbreak due to food safety concerns. Because pork consumption accounts for almost one third of per capita meat consumption in Taiwan, averaging 40 kg per person per year, a significant reduction in pork demand could be expected to lead to an increase in consumption of other meats, such as beef. Further, because more than 90 per cent of beef consumed in Taiwan is imported, mainly from Australia, New Zealand and the United States, a possible move by consumers towards beef would seem to present a marketing opportunity for all beef suppliers to Taiwan. The objective of this study is to examine the impact of the FMD outbreak, as well as price and other demand shifters, on the demand for imported beef in the Taiwan market.

This paper differs from other studies on the link between food scares and meat demand (eg FAO, 2002, Dahlgran and Fairchild, 2002; Tsai, 1999; Burton and Young, 1996; Wessells et al., 1995; van Ravenswaay and Hoehn, 1991) in that it focuses on the impact of a food scare in pigs on the demand for imported beef and the implications for overseas suppliers, rather than on other meat sectors in the domestic market. The reason for this focus is our interest in using this paper as a case study for assessing the willingness and readiness of major beef suppliers to Taiwan in responding to market changes.

The paper begins with an overview of the FMD outbreak in March 1997 and a brief description of the beef market in Taiwan, including trends in beef consumption and beef imports. In the following section, the empirical model examining factors affecting beef import demand is developed and explained. The modelling results are then discussed, followed by implications for the beef exporters to Taiwan and concluding remarks.

2. The FMD Outbreak in the Taiwan Pig Industry

The FMD epidemic in the Taiwan pig industry broke out in 1997, almost 70 years after the last eradication in 1930 (FAPRI, 2000). The first suspected case was noted on 14 March 1997 on a pig farm in Hsinchu prefecture, 60 km south of Taipei. The case was confirmed on 19 March by the Taiwan Animal Health Research Institute and verified on 25 March by the FMD World Reference Laboratory in Pirbright, United Kingdom. The disease spread quickly throughout the island but was contained within two months by slaughtering of infected pigs and blanket vaccination. During the epidemic, 1,300 farms were affected in March, followed by 3,864 farms in April, 975 farms in May, five farms in June and three in July. By the end of July, a total of 6,147 pig farms had contracted the disease, about one quarter of all pig farms in Taiwan (Shieh, 1997; FAO, 2002). Among the 4.66 million pigs found on the infected farms, about 185,000 died of the disease. The remainder, including 1.01 million pigs showing clinical signs of FMD, were destroyed and disposed of by burying, incineration, rendering or burning. When the epidemic was over in July, about 40 per cent of the total pig inventory was lost.

The pig price dropped dramatically, from $NT (New Taiwan dollars) 55 to $NT 17/kg immediately after the announcement of the FMD outbreak by the Council of Agriculture on 20 March, and remained at about one quarter to one half of the original price for one and a half months (Shieh, 1997). In late May, the pig price returned to $NT 40/kg temporarily but soon fell back to $NT 30/kg and stayed low for quite some time. Since the cost of production was estimated to be around $NT 40/kg, the loss to producers was devastating. Monthly pork retail prices between January 1996 and December 1998 are shown in Figure 1, with the decline in April 1997 clearly evident.

There were two reasons for the significant fall in pig prices. Firstly, about 40 per cent of pigmeat is exported. Japan, the biggest export market, accounted for more than 90 per cent of total Taiwan pigmeat exports at the time (Ho, 1997). The outbreak of FMD meant that all pigmeat exports to Japan (about 27 million tonnes per year) were banned and had to be diverted to the domestic market, resulting in a huge glut and consequent price fall (Shu, 1998). Secondly, immediately following the outbreak, consumers were very concerned about food safety and the effect on their personal health. Despite assurances from the government, consumers were sceptical and the demand for pork fell. The combined effect of an increase in supply (due to export diversion) and a decrease in demand contributed to the significant fall in pig prices. The food scare, and hence the reduction in demand, was alleged to have lasted from six to 12 months (Yang et al., 1999).  

The pig industry is a major livestock sector in Taiwan, accounting for 21.1 per cent of the total value of agricultural production in 1996. In 1997, the figure dropped to 11.8 per cent, and then rebounded partially to 13.0, 15.7 and 14.3 per cent in 1998, 1999 and 2000, respectively (Taiwan Provincial Government, 2001). The economic impacts of the FMD outbreak on the pig and related industries were substantial. The total costs were estimated to be about $US6.6 billion, including the direct costs of control measures (stamping out and mass vaccination) and compensation to farmers[1], and the indirect costs to agricultural and related industries (Yang et al., 1999; FAO, 2002). The loss of jobs in the pig industry and related industries was estimated to be around 65,000. Export revenues fell from $US 1.6 billion in 1996 to $US 234 million in 1997, while Gross Domestic Product (GDP) was reduced by 0.64% (an equivalent of $US 14.3 billion) (Ho, 1997; FAO, 2002).

3. Meat Consumption in Taiwan

Pork has been the major meat item for the Taiwanese during the past three decades. Since 1995, annual per capita consumption has been about 40 kg (Table 1), accounting for about one third of total consumption of pork, beef, lamb, poultry and seafood. It is clear from Table 1 that annual per capita beef consumption in Taiwan (3.29 kg in 2001) is very low, compared with other meats. This is because beef has traditionally not been included in the Taiwanese diet. Beef consumption has increased slightly since the mid-1970s as a result of market liberalisation, economic growth and changing consumer lifestyles (Hsu, 1997). However, beef consumption in Taiwan is likely to remain low (Chang and Hsia, 2000).

Table 1. Annual per capita meat and fish consumption and shares in Taiwan in selected years

	Beef/Veal		Pork		Mutton/Lamb		Poultry		Fish/Seafood		Total
	Kg	%	Kg	%	Kg	%	Kg	%	Kg	%	Kg
1965	0.39	0.92	12.07	28.57	0.06	0.14	1.99	4.71	27.74	65.66	42.25
1975	0.94	1.50	17.51	28.00	0.17	0.27	8.36	13.37	35.56	56.86	62.54
1985	1.66	1.79	34.23	36.92	0.39	0.42	18.02	19.44	38.41	41.43	92.71
1995	3.16	2.76	39.76	34.76	1.04	0.91	32.09	28.05	38.34	33.52	114.39
1996	2.81	2.48	40.62	35.81	1.10	0.97	30.49	26.88	38.42	33.87	113.44
1997	3.44	2.88	39.05	32.66	1.30	1.09	33.49	28.01	42.30	35.37	119.58
1998	3.29	2.80	41.13	34.98	1.25	1.06	32.49	27.63	39.41	33.52	117.57
1999	3.75	3.09	38.76	31.91	1.36	1.12	33.89	27.90	43.72	35.99	121.48
2000	3.31	2.81	39.50	33.53	1.28	1.09	33.37	28.33	40.34	34.24	117.80
2001	3.29	2.57	41.13	32.13	1.25	0.98	32.94	25.73	49.41	38.60	128.02

Source: Taiwan Agricultural Yearbook, Taiwan Provincial Government.

Consumer perceptions and relatively high beef prices are given as the main reasons for the slow growth in beef consumption (Asian Market Intelligence, 1994). Another reason is that Taiwan is a mature market for animal protein and it is difficult for beef to gain a substantial foothold (CIE, 1995). That is, per capita meat consumption of about 120kg (see Table 1) is similar to developed countries such as Australia and Japan. Therefore, an increase in beef consumption would have to come at the expense of other meats. In general terms, market share can be increased by providing a better product, lowering price, improving access and product promotion. However, sudden changes in the environment may also create marketing opportunities. The FMD outbreak in March 1997 would seem to have provided such an opportunity whereby consumers could have been lured towards eating more beef. Indeed, Tsai (1999) found that following the FMD outbreak, consumers had substituted beef and chicken for pork. Table 2 shows the changes in monthly per capita meat consumption before and after the FMD outbreak.

Table 2. Monthly per capita meat consumption (in kg), 1996.1-1997.12

	Pork	Beef	Chicken
1996 Jan	4.364	0.268	2.418
Feb	3.941	0.167	2.038
Mar	4.294	0.189	2.303
Apr	3.405	0.227	2.436
May	3.712	0.198	2.704
Jun	2.216	0.169	2.641
Jul	3.891	0.185	2.864
Aug	4.140	0.228	2.916
Sep	3.98	0.166	2.866
Oct	4.056	0.207	2.892
Nov	4.349	0.144	2.751
Dec	4.195	0.111	3.399
1997 Jan	4.263	0.190	2.963
Feb	3.266	0.119	2.160
Mar	2.796	0.218	2.58
Apr	3.78	0.224	2.795
May	2.979	0.297	3.225
Jun	3.088	0.355	3.298
Jul	3.518	0.307	3.210
Aug	4.052	0.266	3.112
Sep	4.359	0.247	3.053
Oct	4.756	0.190	3.393
Nov	4.281	0.196	3.222
Dec	4.244	0.218	3.011

Source: Tsai (1999).

4. Beef Imports into Taiwan

Beef production in Taiwan is very low and over 90 per cent of beef consumed is imported. Domestically-produced beef comes mainly from culled dairy cattle and draft yellow cattle. Following the beef market liberalisation in 1975, beef imports into Taiwan have increased from 156 tonnes in mid-1970 to 56,904 tonnes in 2001 (Council of Agriculture, 2001). The ratio of total beef imports to total beef consumption has increased from 20 per cent in 1974 to more than 90 per cent since 1985 (Taiwan Provincial Government, 2001).   

Australia has been the largest supplier to Taiwan, followed by New Zealand and the United States. Based on frozen beef imports during 1989-2001, market shares, in volume terms, for Australia, New Zealand and the United States were 63, 21 and 16 per cent, respectively. The average unit import values (CIF) were $US 2.71/kg, $US 3.84/kg and $US 5.27/kg, respectively (see Table 3). It is clear that Australian beef was relatively cheaper than New Zealand beef, which, in turn, was cheaper than United States beef, over this period. It is also clear that the market share, import quantity and price of Australian beef in Taiwan were more stable than those for New Zealand and the United States. The coefficients of variation (CoV) in Table 3 for prices were 0.15, 0.25 and 0.23 for Australian, New Zealand and United States beef, and for quantities imported were 0.24, 0.52 and 0.52, respectively.

Table 3. Descriptive statistics of the monthly import data, 1989.1-2001.12

Variablea	Mean	St. Deviation	Minimum	Maximum	CoV
QSUSA	0.16	0.07	0.04	0.34	0.42
QSAUS	0.63	0.11	0.39	0.84	0.18
QSNZ	0.21	0.08	0.05	0.41	0.39
QUSA	632.30	326.66	114.70	1591.80	0.52
QAUS	2478.30	600.00	947.21	4356.10	0.24
QNZ	878.41	457.72	74.19	2210.20	0.52
PUSA	5.27	1.20	3.02	8.07	0.23
PAUS	2.71	0.42	1.95	3.40	0.15
PNZ	3.84	0.95	2.41	7.71	0.25
QTOTAL	3989.00	980.44	1432.10	7067.30	0.25

Notes: ^a The variables are defined as follows: QS, Q and P are the volume share, import quantity (in tonnes) and unit import price (in $US/kg) for beef from the United States, Australia or New Zeal and, respectively; and Q_TOTAL is the total beef imports (in tonnes) from the three major suppliers.

The price differentials reflect transport differentials and quality differentials among the different sources of supply. Quality differentials of frozen beef imports are attributable mainly to differences in production methods (grass-fed versus grain-fed), degree of processing (eg whole carcass, boxed beef, bone-in versus boneless) and retail cuts (eg sirloin steak versus shin shank) (Chang and Hsia, 2000). Generally speaking, United States beef is considered of higher quality, comprising grain-fed beef, and is classified as "special grade" for tariff purposes (Hwang, 1993). By comparison, Australian and New Zealand beef and domestically produced beef are classified as "general grade". The quality of beef determines where the products are sold. For example, United States and New Zealand beef tends to go to the Western-style restaurants and modern supermarkets. On the other hand, Australian beef (comprising mostly shin shank and other lower value cuts) goes mainly to the Chinese restaurants and the local wet markets (Chen and Hsiao, 1995; CIE, 1995; NZMPB, 1997).

Total frozen beef imports from January 1989 to December 2001 are shown in Figure 2. Imports from Australia, United States and New Zealand are shown in Figures 3-5, respectively. Note that immediately after the FMD outbreak in March 1997, there appeared to be increases in Australian beef imports for three months (April, May and June), which were then followed by four consecutive falls from July to October. By comparison, United States beef imports seemed to fall immediately in April, increased in May, June and July, fell again in August, increased again in September and fell again in October and November. For New Zealand beef imports, there was an initial fall in April, followed by two consecutive increases in May and June, which were followed by four consecutive falls from July to October. So, did the FMD outbreak have an impact on beef imports? An econometric demand model was developed to help shed light on this seemingly ambiguous situation.

 

5. Estimation of Beef Import Demand

In this section, the empirical model used to examine the impact of FMD on beef import demand in Taiwan is described, along with the estimation procedures. The system of import demand equations for the three major beef suppliers to Taiwan are as follows:

(1)                 Q_it = α_i + d_i Q_it-1 + b_i2 P_{AUS, t} /P_USA,t  + b_i3 P_{NZ, t} /P_USA,t  + ¡_i Qtotal_t  + _{θi FMD} + ∑_j f_{ij FMDt-j} + ∑_k j_ik D_k + ε_it,         i = Australia, New Zealand and the United States,

where

Q_it = the volume of frozen beef imports from source i at time t; and t = 1 (January 1989), 2 (February 1989), ..., 156 (December 2001);

Q_it-1 = the volume of frozen beef imports from source i at time t-1;

Qtotal_t = the volume of total frozen beef imports at time t;

P_{AUS /} P_{USA =} the ratio of unit import value of frozen beef imports from Australia

(P_AUS) to the unit import value of the United States (P_USA);

P_{NZ /} P_{USA =} the ratio of unit import value of frozen beef imports from New Zealand (P_NZ) to the unit import value of the United States (P_USA);

FMD = a dummy variable used to capture the impact on beef imports of the FMD outbreak in March 1997, where FMD = 1 for March 1997 and FMD = 0 elsewhere;

_{FMDt-j = FMD lagged j periods; j = 1,2, …, 6;}

D_k = a set of monthly dummy variables using January as the reference point, where k  = 2 (February), 3 (March),..., 12 (December); and

                e_it = white noise error term which may be contemporaneously correlated.

Equation (1) states that the demand for beef imports from each supply source depends on the relative prices of imported beef from each source, the volume of total imports, the FMD outbreak, and seasonality. The current model was selected after a variety of specifications were experimented with in the preliminary analysis, including the Almost Ideal Demand System (AIDS) of Deaton and Muellbauer (1980). Despite desirable theoretical properties associated with a formal systems approach, results from the preliminary analysis indicated that based on the available data the model presented in Equation (1) was the most appropriate for examining the beef market in Taiwan. That is, the model yields more reasonable results in terms of goodness-of-fit, signs and magnitude of the estimated coefficients, compared with the AIDS model and other specifications.

The underlying assumptions are that (i) imported beef is separable from domestically produced beef and other meats; (ii) imported beef from different sources is differentiable by source of origin; and (iii) prices are exogenous. The separability assumption is a necessary and sufficient condition for examining the demand for beef imports as a group under a multi-stage utility maximisation process discussed in Deaton and Muellbauer (1980, pp. 127-134)[2].

Treating agricultural commodities from different sources of supply as differentiable is consistent to what was first proposed by Armington (1969). The price exogeneity assumption is justified on the basis that total beef imports into Taiwan are small relative to total beef exports from each of the supply sources, either individually or jointly. In 2000, the volume of beef exported to Taiwan accounted for only 2.8 per cent of total beef exports by Australia, United States and New Zealand (ABARE, 2001). Therefore, changes in beef import demand in Taiwan are not expected to have an impact on individual prices. A similar approach was employed by Kim et al. (2004) where beef imports from Australia, New Zealand, the United States and Canada were modelled as a demand sub-system based on similar assumptions.

In addition, it is hypothesized that the response of beef imports to the FMD outbreak is not immediate but may be delayed or dissipated over a period of time because of uncertainties surrounding the severity of, and consumers’ reaction to, the outbreak, as well as the lead time in getting new shipments in from overseas. As such, possible distributed lagged responses were incorporated in the model. After experimenting with different lag structures, a linear structure with a six-month lag was found to best represent the demand response.[3]

A lagged dependent variable is also included in the model based on a partial adjustment assumption. Thus the quantity of beef imports from each supplier is assumed to be unable to fully adjust to new market conditions within one month, due to factors such as supply contracts, ordering and shipping lags, and other types of inertia in the beef marketing and distribution system in Taiwan. The presence of a lagged dependent variable suggests that adjustments may be separated into short-run and long-run responses.

The assumption that the error terms in Equation (1) may be contemporaneously correlated means that there may be efficiency gains if the three equations are estimated as a system. However, the system results based on Seemingly Unrelated Regression (SUR) indicated that the cross-equation correlations between the equation errors were weak. This means that there is little efficiency gain from the systems estimation. Indeed, when the results from SUR and Ordinary Least Squares (OLS) for the United States and New Zealand equations were compared, they were almost identical. Therefore, the import demand equations were estimated individually using OLS procedures, except for Australia. The latter was estimated using the iterative Cochrane-Orcutt procedure because of second-order autocorrelation (AR(2)) that was detected in the OLS estimates. The estimated coefficients of autocorrelation were 0.69 and 0.19 and statistically significant at the five per cent level. AR(2) residuals means that random effects on the quantity imported in this period are correlated with the random effects in the preceding two periods. All econometric work was based on Shazam (Version 8.0) (White, 1999).

The data used for the empirical analysis are monthly data from January 1989 to December 2001 on volumes and values of frozen beef imports from Australia, the United States and New Zealand. Data on imports are obtained from Monthly Statistics of the Republic of China, published by the Directorate-General of Budget, Accounting and Statistics (2001). There are a total of 156 observations. Descriptive statistics for the data are presented above in Table 3.

6. Results

The estimated results for the three import demand equations are presented in Table 4. Firstly, the estimated regression equations explain between 72 and 87 per cent of the variations in individual import demands. Secondly, all of the estimated price coefficients have the expected signs. That is, estimated own-price coefficients are negative while estimated cross-price coefficients are positive. These results are consistent with demand theory and Armington’s proposition that imports from different sources are potential substitutes. However, not all the estimated price coefficients are statistically significant at the ten per cent level. This means that prices may not be the most important factor in explaining individual import demands. Thirdly, the quantity of total imports plays a very significant role in determining imports from particular sources. Fourthly, the lagged dependent variable is significant and correctly signed for the New Zealand and United States equations, but negative and significant for the Australian equation. The latter result, along with the AR(2) correction required in this equation, may be evidence of misspecification error or it may have something to do with the marketing strategy employed by the Australian suppliers. Fifthly, there are strong seasonal variations in the demand for Australian and New Zealand beef (although these are not reported for space reasons).

Table 4. Estimated results of import demand equations, 1989.1-2001.12

	Australiaa	New Zealand	United States
Qt-1	-0.07 (-1.76)b	0.47 (8.52)	0.53 (10.43)
Paus/usa	-771.81 (-2.12)	172.21 (0.64)	47.64 (0.20)
Pnz/usa	100.57 (0.50)	-221.95 (-1.33)	291.77 (2.05)
Pusac	671.24 (3.50)	49.74 (0.23)	-339.41 (-1.13)
Qtotal	0.60 (22.99)	0.21 (8.73)	0.18 (8.73)
FMD	-63.31 (-0.29)	193.18 (0.91)	188.35 (1.04)
FMD-1	221.93 (0.85)	-150.73 (-0.71)	40.00 (0.22)
FMD-2	258.69 (0.90)	-98.88 (-0.46)	126.81 (0.69)
FMD-3	-180.31 (-0.59)	396.91 (1.80)	-195.84 (-1.04)
FMD-4	-251.54 (-0.85)	-269.82 (-1.23)	87.05 (0.48)
FMD-5	126.28 (0.49)	-248.79 (-1.16)	-221.60 (-1.21)
FMD-6	-108.49 (-0.49)	-316.78 (-1.49)	494.37 (2.72)
Constant	776.90 (2.81)	-290.45 (-1.81	-633.54 (-4.65)
Adjusted R2	0.87	0.80	0.72

^{Notes: a} The Australian equation is estimated using the iterative Cochrane-Orcutt procedure

   to correct for AR(2).

^b The figures in parentheses are t-ratios.

^c These price coefficients are recovered from the homogeneity conditions implied by using relative prices, and their t-ratios are calculated accordingly based on the usual formula.

Finally, the impact of the March 1997 FMD outbreak on imported beef demand is almost nil. It can be seen from Table 4 that all of the individual coefficients associated with the FMD variables are statistically insignificant at the five per cent level of significance, with only two exceptions. These are the increase in the demand for New Zealand beef three months, and the increase in the demand for United States beef six months, after the outbreak. These lagged responses may have something to do with the time required to deliver new shipments to Taiwan. The FMD coefficients also were tested jointly as a group based on likelihood ratio tests. The calculated c² statistics are 7.84, 10.04 and 12.12 for Australian, New Zealand and United States equations, respectively. They are smaller than the tabulated value of 14.07 with seven degrees of freedom at the five per cent level of significance. So although demand might have changed (some increases and some decreases) over the six-month period, overall, the FMD outbreak had very little impact on the demand for beef from individual suppliers. Again, after factoring out the impacts of other variables, the changes in import demands are not as apparent or profound as those that appear in Figures 2-5. The differing results imply that evaluation of the impact of the FMD outbreak, using data for the adjacent months only, may be quite misleading.     

Table 5. Estimated demand elasticities

	Australian   price	New Zealand price	United States price	Total       imports
Australia imports	-0.16**a	0.03	1.43**	0.96**
New Zealand imports	0.10	-0.18	0.30	0.97**
United States imports	0.04	0.33**	-2.83	1.14**

Notes: ^a ** indicates that the estimated elasticities are statistically significant at the five per cent level.  

Estimated demand elasticities are presented in Table 5. These elasticities are defined as the per cent change in quantity demanded with respect to a one per cent increase in either prices or total imports. One finding is that while the demand for beef from Australia and New Zealand are inelastic with respect to their own-prices, the demand for United States beef is own-price elastic. This means a one per cent increase in the own-price would result in 0.16 and 0.18 per cent decreases, respectively, in Australian and New Zealand beef exports to Taiwan, but a 2.83 per cent decrease in United States beef imports. Another finding is that Australian beef is a strong substitute for United States beef, with an estimated cross-price elasticity of 1.43. This means that a one per cent increase in the price of United States beef will result in a 1.43 per cent increase in the quantity demanded of Australian beef. United States beef is also a substitute for New Zealand beef, with a significant cross-elasticity value of 0.33. However, there is no indication that New Zealand beef is a substitute for either Australian beef or United States beef. For United States and New Zealand beef, the results suggest that there are significant partial adjustment effects that mean that the long-run responses are approximately twice those of the short-run reported in Table 5.

Some of these results need to be considered with caution since not all the estimated elasticities are statistically significant at the five per cent level. Nevertheless, the differing results suggest that the beef market in Taiwan is complex and more research, perhaps of a qualitative nature, on marketing strategies is needed to explain the differences.

The estimated coefficients associated with total imports are positive for all three demand equations and statistically significant at the one per cent level. This result implies that the demand for beef from all suppliers increases with an increase in total import demand. However, the distribution of the increase is not even across suppliers. For a one per cent increase in total beef imports, the quantity demanded of Australian beef will increase by 0.96 per cent, of New Zealand beef by 0.97 per cent and of United States beef by 1.14 per cent. The overall response to the total imports seems to suggest that although Australian and New Zealand demands may be increasing in absolute terms, the rate of increase is slightly less than the overall growth. This may eventually lead to deterioration in their market shares and competitive positions in the Taiwan market. One explanation suggested in Chang and Hsia (2000), which is particularly true for Australia, was that Australian beef exported to Taiwan tended to be of lower quality, compared with New Zealand and United States beef. This might have worked well in the past when consumers were less aware of quality differentials in beef and lower quality beef suited traditional ways of cooking beef in Taiwan. However, as consumers become more affluent and better informed, demand for higher quality products will increase.    

Seasonality also appears to exert a strong influence on beef import demands. In particular, the modelling results show that at the five per cent level of significance, Australian demand is statistically significantly lower from February through to August and statistically significantly higher in November and December, when compared with the reference month of January. In contrast, the demand for New Zealand beef is statistically significantly higher in March and statistically significantly lower in September, October and November, when compared with January. Further, the demand for United States beef does not show any seasonality at all because all the coefficients associated with seasonality dummies are statistically insignificant. Since total beef imports to Taiwan were found to be significantly lower in February, September and December than in January in the preliminary analysis, the seasonal patterns found here seem to suggest that when total beef import demand is at its lowest as in February, the reduced demand appears to be absorbed mostly by Australia. One possible reason for this coincidence is the fact that Australian beef is sold predominantly in the wet markets that are patronised mostly by households and small businesses. During the festival months (such as the Chinese New Year which falls between late January and mid-February and the Autumn Festival in September or October), when households buy more pork, chicken and fish for ceremonial purposes, demand for beef in the wet markets is therefore substantially reduced. By comparison, United States and New Zealand beef is directed to the restaurant trade, where demand is less affected by traditional festivities and cultural events. Seasonality was also found to be a significant factor for affecting household demand for meat and seafood in Japan (Johnson et al., 1998).

When comparing seasonality among suppliers, the lack of seasonality in the import demand for United States beef may be due to the nature of grain-fed beef supply, which is less affected by weather and season, thus allowing more consistent quantities throughout the year. Grass-fed beef from Australia and New Zealand, on the other hand, can be significantly affected by weather, and hence production tends to be more seasonal.

Note that the seasonality patterns indicated by the estimated results do not entirely agree with the patterns illustrated in Figures 3-5. This is because the patterns seen in those figures are a combined effect of a number of factors, including seasonality, prices, total imports and FMD. As such, the coefficient associated with the monthly dummy variable is the net effect of seasonality after the impacts of all other factors have been accounted for. Pure seasonal patterns in total and individual imports are shown in Figure 6 using monthly averages. It seems clear that seasonality is most apparent in total and Australian beef imports, and that these two patterns are highly correlated.  

7. Marketing Implications

This analysis shows that the impact of the FMD outbreak on beef consumption was marginal and only temporary, and dissipated within six months. This result is consistent with previous findings (eg Dalhlgran and Fairchild, 2002) that “consumers tend to have short attention spans when it comes to food safety issues and in the absence of frequent message repetition, the demand alterations soon disappear”. However, Burton and Young (1996) showed that the BSE scares in Great Britain in the early 1990s, in addition to its short-term effect, had resulted in a permanent loss of market share for beef over a period of three to four years after the BSE incident.

Taken together, these results seem to suggest that there was a missed opportunity for beef suppliers to Taiwan. Given that food scares are generally short-term in nature in terms of their impact on consumption patterns and that there exists the possibility for a permanent switch, a quick response and a more aggressive marketing campaign was required from beef suppliers, either individually or jointly, immediately after the outbreak. This may have resulted in increased sales, at least in the short term, and an improved market position in the longer term. It is understood that Meat and Livestock Australia spent around $A5.5 million to promote Australian beef in Japan soon after the BSE incident there in September 1999. However, according to Stewart (2000), no such effort was directed at Taiwan when the island was hit by the FMD outbreak in pigs in 1997. 

Further, it also appears that the minimal and short-term effect of food scares on demand experienced so far does not necessarily suggest that consumers will continue to be forgetful and forgiving in the future. Research has shown that consumers are becoming more demanding and vigilant when it comes to food safety as they become more aware and educated of the issues. This means that next time around, any food scare may have a more permanent effect on consumers’ preferences and consumption behaviour. The implication is that monitoring of, and responding to, market changes will become even more important for food industries to achieve a competitive advantage in their market.

Finally, Australia and New Zealand appear to be in direct competition with each other, more so than with the United States. This observation is made based on the different and opposite seasonal patterns observed between Australia and New Zealand and the fact that the estimated own-price and total imports elasticities for both countries are very similar. This implies that some purchasing patterns exist that seem to play Australia and New Zealand against each other. United States beef, on the other hand, seems to be in a unique market position, suggesting that the United States may have been successful in differentiating its beef products in the Taiwan market. This latter result is similar to what was reported in Kim, Kim and Veeman (2004).

8. Conclusions

In March 1997, a FMD epidemic broke out in the Taiwan pig industry. In the following six months, a large proportion of the pig population was wiped out and the demand for pork fell substantially. The objective of this study was to determine the impact of the FMD outbreak, and other factors, on the demand for imported beef in the Taiwan market. Demand functions for beef imports from Australia, the United States and New Zealand were estimated based on monthly data from January 1989 to December 2001.

The major finding of this study is that the FMD outbreak had little impact on the demand for imported beef regardless of the source of origin. The analysis also found that there was strong seasonality in beef imports, but that the patterns of seasonality differed among suppliers. The study also found that Australia is in direct competition with New Zealand in the Taiwan beef market, but less so with the United States.

The marginal impact of the outbreak on the import demands from individual suppliers may suggest that there was a missed opportunity for beef suppliers, especially Australia, given its proximity to Taiwan. Given that food scares are generally short-lived in terms of their impact on consumption behaviour, a quick response and a more aggressive marketing campaign from the beef suppliers immediately after the outbreak might have resulted in increased sales, at least in the short term. However, in the future the effects of food scares may be longer-lasting. There is evidence that consumers are becoming more vigilant, less forgetful and less forgiving, so that another large-scale food scare could lead to a permanent change in consumption patterns. Marketers who are not willing or equipped to respond more swiftly next time around may lose substantial sales, as well as market position.

9. References

ABARE (Australian Bureau of Agricultural and Resource Economics) (2001 and previous issues). Australian Commodity Statistics 2001. Canberra: ABARE.

Armington, P.S. (1969). A theory of demand for products distinguished by place of production. International Monetary Fund Staff Paper, 16, 159-176.

Asian Market Intelligence (1994). A Report on Consumer Usage of and Attitudes Toward Meat in Taiwan. Seoul: Asian Market Intelligence.

Burton, M. and Young, T. (1996). The impact of BSE on the demand for beef and other meats in Great Britain. Applied Economics, 28, 687-693.

Chang, H.S. and Hsia, C.J. (2000). Beef import market share in Taiwan: implications for Australia. Australian Journal of Agricultural and Resource Economics, 44, 217-321.

Chen, C.W. and Hsiao, W.C. (1995). The structure of Taiwanese demand for meats. Agriculture and Economics, 16, 111 - 141.

CIE (Centre for International Economics) (1995). Taiwan Meat Market Strategy. Meat Research Corporation Project M.315A. Canberra.

Council of Agriculture (2001 and previous issues). Agricultural Trade Statistics of Republic of China. Taipei, ROC.

Dalhgran, R.A. and Fairchild, D.G. (2002). The demand impacts of chicken contamination publicity – a case study. Agribusiness, 18, 459-474.

Davis, G. (1997). Product aggregation bias as a specification error in demand systems. American Journal of Agricultural Economics, 79, 100-109.

Deaton, A. and Muellbauer, J. (1980). Economics and Consumer Behaviour. Cambridge: Cambridge University Press.

Directorate-General of Budget, Accounting and Statistics (2001 and previous issues). Monthly Statistics of the Republic of China. Taipei: Executive Yuan.

FAS (Food Agricultural Service) (2000). Impact of food-and-mouth disease on meat trade. International Agricultural Trade Report. USDA. www.fas.usda.gov/dlp2/highlights/2000/FMDimpact.html.

FAO (Food and Agriculture Organisation) (2002). Animal diseases: implications for international meat trade. Intergovernmental Group on Meat and Dairy Products, 19^th Session, 27-29 August 2002, Rome. www.fao.org/DOCREP/MEETING/004/Y6975E.HTM.  

FAPRI (2000). Animal disease outbreaks and their impacts on trade. www.fapri.org/bulletin/Oct2000/AnimalDiseaseOutbreakAndThierImpcatsOnTrade.htm.

 Ho, J.Y. (1997). The pig industry and the outbreak of FMD. Industrial Investigation and Technology, 122, 2-16.

Hwang, T.-C. (1993). Meat expenditure pattern, forecast, and impact from liberalisation in Taiwan. Journal of Agricultural Economics, 54, 117-164.

Johnson, A.J., Durham, C.A. and Wessells, C.R. (1998). Seasonality in Japanese household demand for meat and seafood. Agribusiness, 14, 337-351.

Kim, R.B., Kim, Y.J. and Veeman, M. (2004). Adjustment in the South Korean beef import market under beef import liberalisation policies. Agribusiness, 20, 201-216.

Nayga, R. and Capps, O. (1994). Tests of weak separability in disaggregated meat products. American Journal of Agricultural Economics, 76, 800-808.

NZMPB (New Zealand Meat Producers Board) (1997). Taiwan Market Report. Wellington, New Zealand.

Shieh, H.K. (1997). Short communication: the FMD situation in Taiwan. Journal of Chinese Society of Veterinary Science, 23, 395-402.

Shu, Y.J. (1998). A study on the future development of swine industrial under the impact of the foot and mouth disease in Taiwan. Journal of Chiayi Institute of Technology, 59, 69-80.

Stewart, J. (2000). Clean image may not save Australia's beef industry. Australian Broadcasting Corporation. http://www.abc.net.au/am/s114236.htm.

Taiwan Provincial Government (2001 and previous issues). Taiwan Agricultural Yearbook. Department of Agriculture and Forestry, Taiwan, ROC.

Tsai, J. (1999). The impact of foot and mouth disease epidemic on the consumption of meat in Taiwan. Taiwanese Agricultural Economics Review, 4, 205-221.

van Ravenswaay, E.O. and Hoehn, J.P. (1991). The impact of health risk information on food demand: a case study of alar and apples. In J.A. Casewell (ed.), Economics of Food Safety (pp. 155-174), New York: Elsevier.

Wessells, C.R., Miller, C.J. and Brooks, P.M. (1995). Toxic algae contamination and the demand for shellfish: a case study of demand for mussels in Montreal. Marine Research Economics, 10, 143-159.

White, K. (1996), Shazam User's Reference Manual. New York: McGraw-Hill.

Yang, P.J., Ju, R.P., Jong, W.B. and Song, H.T. (1999). The 1997 FMD outbreak in Taiwan and its economic impacts. Animal Husbandry Fortnightly, 61, 25-29.