Valuing Quality Attributes and Country of Origin in the Korean Beef Market
Chanjin Chung, Tracy Boyer,
Abstract
For beef exporters, one of the important questions in the Korean beef market is why Korean consumers are willing to pay almost three times more for domestic Korean beef than they pay for imported beef. To answer this question, we surveyed 1,000 shoppers in Seoul, Korea, and conducted a conjoint analysis on consumers’ willingness to pay (WTP) for country equity of domestic vs. imported beef and quality attributes of marbling, freshness, genetically modified organism (GMO)-free feed ingredients and antibiotic-free production. Among all factors contributing to the price differentials, the most important factor seems to be the country of origin followed by the use of GMO feeds and antibiotics in beef production, marbling grade and freshness. This study finds that Korean consumers value origins of imported beef approximately $14/lb less than the Korean origin. Korean consumers’ valuation of beef quality and country of origin differs by some demographic groups: older vs. younger generations, homemakers vs. non-homemakers and consumers who prefer to purchase packaged beef vs. consumers who prefer to purchase butcher shop beef. Our empirical findings suggest that the top priority for beef exporters who wish to increase sales and value of their beef in the Korean market must be to counter Korean consumers’ strong ethnocentrism by improving the value of their country of origin.
1. Introduction
Since the liberalisation of the imported beef market in 2001, Korea has experienced a steady increase in beef importation resulting in a 53% market share for imports as a portion of total beef consumption in 2003 (US Meat Export Federation, International Market – Korea, http://www.usmef.org/tradelibrary/korea.asp). Major exporting countries to the Korean beef market were the USA, Australia, New Zealand and Canada with respective market shares of beef imports of 68%, 21%, 10% and 1% (US Meat Export Federation). In 2003, Korea was the US cattle producers’ third largest export market following Japan and Mexico. However, like many other countries, Korea closed its market to US beef on 24 December 2003 after the discovery of BSE (bovine spongiform encephalopathy) in the USA. on 1 April 2007, the US and Korean governments concluded negotiations on the US–Korean Free Trade Agreement (FTA), and one of the key FTA negotiations was to allow the USA to resume exports of beef to Korea. In particular, the Korean government agreed to import all US beef products from animals of all ages including deboned beef, bone-in beef, offal and variety meats, and processed beef products while also agreeing to abide by the US BSE inspection scheme (Office of the US Trade Representative, http://www.ustr.gov/assets/Document_Library/Fact_Sheets/2008/asset_upload_file459_14857.pdf). Under this scheme, only 1% of US cattle are inspected each year. This particular agreement triggered several large street protests in June 2008 because protesters believed that the beef import deal was hastily made before the newly elected Korean President’s first trip to Washington and that the deal compromised public health standards in Korea. Nevertheless, US cattle producers see the US–Korean FTA as potentially the biggest and most important bilateral trade agreement in history because Korea currently represents a $1bn market and could grow to be the USA’s top beef export market (National Cattlemen’s Beef Association, http://www.beefusa.org/govesouthkoreanfta.aspx).
One of the interesting and important issues in the Korean beef market is the large price difference between domestic and imported beef. Overall, the prices of imported beef are approximately three times lower than the price of Korean beef at the retail level even after accounting for the 40% tariff on imported beef.2 The reasons for the large retail price difference between domestic Korean and imported beef are unclear especially as both Korean and US beef cattle are grain fed. The source of the price difference between Korean beef and imported beef from Australia and New Zealand may be attributable to basic differences in the actual taste of each origin’s beef due to the animals’ feed. Korean beef is grain-fed beef, which Korean consumers may prefer (Unterschultz et al., 1998), whereas imported beef from Australia and New Zealand is mostly grass fed. Nevertheless, the differential between domestic and imported grain-fed beef remains puzzling.
Consumers’ different perceptions on various beef quality characteristics and their valuation of country of origin must play a part in this price differential. Consumers’ valuation of beef quality relies heavily on their perceptions because, like many other credence goods (Darby and Karni, 1973), quality differences among different country of origin beef may not be easily discernable even after consumption. Asymmetric information problems (Akerlof, 1970) occur in the Korean beef market because beef producers of each country know whether they have used antibiotics, genetically modified organism (GMO) feed ingredients and other appropriate or inappropriate methods in production, but consumers only know producers’ claims about the quality of their products.
A 1995 discrete choice study by Unterschultz et al. (1998) finds that although chefs, both Korean and non-Korean, prefer US beef, purchasing managers for hotels prefer Korean beef. They also find that Australian beef is not significantly preferred or disfavoured by chefs, but is significantly less desirable to hotel managers. However, Unterschultz et al. (1998) did not survey Korean consumers, who, on average, are less knowledgeable and experienced in purchasing beef compared with hotel chefs and purchasing managers.
This paper examines Korean consumers’ preferences for country of origin, quality and price attributes of domestic vs. imported beef in the retail supermarket setting. The primary purpose of this study was to examine how consumers value product origin and quality attributes in the Korean beef market. Specifically, the study measures the value of country of origin of Korean beef, US beef and imported beef from other countries such as Australia, New Zealand and Canada. The study also estimates consumers’ values of quality attributes such as marbling, freshness and GMO free (never fed GMO feed) and antibiotic free (never given antibiotics).3 In addition, this study estimates differences in values among various socioeconomic groups using variables such as income, age, gender, education, travel experience to beef-exporting countries and cooking and shopping preferences.
Choice-based conjoint analysis is used in this study because beef consumption and price data by country of origin and targeted quality attributes are not available due to the lack of labelling of origin in Korean grocery stores and restaurants.4 Therefore, a common belief is that much of the imported beef has been sold as Korean beef. Choice-based analysis is also preferred because one of the key objectives of this study is to estimate the marginal values of product origin and quality factors simultaneously, assuming a retail environment where consumers have a complete knowledge of a product’s origin and quality. To be consistent with this assumption of a transparent retail environment, our survey provides participants with complete information about product quality and origin using printed descriptions and pictures.
2. Previous Research
Recent studies in the marketing and business literature suggest that consumers tend to prefer products from their own country because of an affinity for their home (Sharma et al., 1987; Verlegh and Steenkamp, 1999; Srinivasan and Subhas, 2003). In recent years, marketers have extended the anthropological concept of ethnocentrism to ‘consumer ethnocentrism’, which relates to how individuals’ buying habits are influenced by loyalties towards their own countries and/or antipathy towards other countries. For example, when the US Congress passed the 2002 Farm Bill, it included the mandatory country-of-origin labelling of beef, pork and fish. The legislation intends to promote US products using consumer loyalty towards ‘Made in USA’ products. Along the same lines, various Korean farmer organisations promoted a slogan, ‘Sin-To-Bul-E’, which is translated as ‘the best food products for Korean people are those produced from Korean soil’. The slogan is based on an Asian belief that by nature human beings are created to consume food products grown in nearby local areas. Obviously, Korean farmers wish to exploit ‘consumer ethnocentrism’ for their own marketing purposes.
Several studies in the economics literature have been conducted on the willingness to pay (WTP) for beef based on its origin and attributes. These studies were principally motivated by recent issues of food safety and BSE or Mad Cow disease. A contingent valuation survey of Japanese consumers by McCluskey et al. (2005a) finds that women and those with concerns about food safety have increased WTP for BSE-tested beef. Loureiro and Umberger (2005) conducted a contingent valuation survey of US consumers for certified US meat products under a country-of-origin labeling programme, finding that US consumers believe that they have the safest meat in the world. However, consumers’ WTP for certification is small, and consumers may also value attributes other than country of origin. Umberger et al. (2002) look at US consumer preferences for Argentine grass-fed beef and US grain-fed beef to find that 23% of their sample is willing to pay $1.36 more per pound for Argentine grass-fed beef over American corn-fed beef.
Discrete choice experiments and conjoint choice methods have been used in a variety of settings to test for consumers’ choice preference on specific attributes of food products, beef in particular. Grunert (1997) analyses consumer’s evaluation of beef quality in four countries: France, Germany, Spain and the UK. Results of Grunert’s conjoint analysis indicate that important quality attributes in all four countries are that the beef tastes good and is tender, juicy, fresh, lean, healthy and nutritious. Among these attributes, the most important and concrete attribute on which consumers base their quality evaluation is fat content (lean). Information about country of origin, breeding and feeding has no effect on quality perception. McCluskey et al. (2005b) use a discrete choice experiment to examine the importance of health benefits of grass-fed beef among US consumers. They find that price, fat and omega-3 fatty acids influence choice most, in that order. Tonsor et al. (2005) test for national differences in preferences on hormone free, GMO free, country of origin and farm source verification in beef markets of Germany, France and the UK. Unlike many previous studies in the discrete choice experiment literature, the Tonsor et al. study allows for the actual purchase of these goods, which avoids potential hypothetical bias problems when no money is exchanged (a similar non-hypothetical experiment was run by Lusk and Schroeder, 2004).5Tonsor et al. (2005) find heterogeneous preferences within and across consumers in these three countries for genetically modified beef and hormone-free beef. Perhaps because the survey occurred prior to the discovery of BSE in the USA, they find no significant difference in WTP for domestic beef over American beef. Another US study by Mennecke et al. (2006) finds that the US state region is the most important characteristic to consumers, followed by animal breed, traceability, type of feed and beef quality. They also find that women are more likely to be concerned about food safety and health. Carlsson et al. (2007) estimate Swedish consumers’ WTP for beef quality attributes such as labelling requirement, GMO feed, outdoor feeding and transportation to slaughter. Their analysis finds that Swedish consumers are willing to pay a price premium for the use of mobile abattoirs over long-distance transportation of live farm animals to slaughtering houses. The results were consistent with and without an opt-out alternative. Recently, Schnettler et al. (2009) also estimate consumer’s WTP for country of origin and information regarding animal welfare with alternative price options in the Chilean beef market. They find that country of origin and information regarding animal welfare are more important factors affecting beef purchase than price.
3. Conceptual Models
Researchers have used a number of choice experiments to elicit values of quality attributes of non-market goods. In the areas of marketing and environmental economics, many studies have used either contingent valuation techniques or conjoint analysis techniques for this purpose. Proponents of conjoint analysis and discrete choice experiments argue that they can measure the WTP for multiple attributes simultaneously, and their statistical designs can allow for the reduction in collinearity among the variables (Holmes and Adamowicz, 2005). They further argue that the conjoint techniques can estimate the incremental benefits that consumers derive from various individual attributes of a non-market good (DeShazo and Fermo, 2002). In our study, we use a conjoint analysis model to examine the marginal WTP for different attributes of beef in the Korean market, including country of origin, marbling, freshness and products free of GMO feed and antibiotics. Respondents are asked to choose one bundle (or treatment combination) of attributes at a given price. The implicit value of changes in one attribute holding all else constant is then computed, such as the marginal value of each quality attribute.
Following Lancaster’s hedonic theory, we use a random utility model to represent utility of quality characteristics and origin of beef in the Korean beef market. The random utility model for individual i choosing alternative j is denoted by:
(1)where Xijβ represents the deterministic part of the utility, whereas the error term εij represents the stochastic portion of the utility. The vector, Xij, includes quality characteristics and the country of origin and β is a corresponding parameter vector. Then, the probability that individual i chooses jth alternative from the choice set Si is:
(2)Here, the choice set Si includes different countries of origin, and different levels of marbling and freshness, and the use of GMO feed ingredients and antibiotics.
Equation (1) can be estimated using a multinomial logit model (MNL) based on the assumption that the error term, εij, is independently and identically distributed with the extreme value distribution. For the MNL, we also assume that the scale factor is one because the assumption of identically distributed error term allows the scale factor to be fixed, and it is typically unidentifiable within any particular dataset (Lusk et al., 2003).
The MNL can be extended to the mixed multinomial logit model (MMNL) considering respondent heterogeneity. The MNL assumes that all individuals share the same parameters for all attributes, which indicates that individuals have the same preferences for quality attributes and country of origin for beef in the Korean market. The MNL is also derived based on the independence of irrelevant alternatives (IIA) assumption, which results from the assumption of independently distributed errors across alternatives. The MMNL allows the parameters of attributes to vary across population and relaxes the IIA assumption (Train, 1998; see also Layton and Brown, 2000; Revelt and Train, 1998; Lusk and Schroeder, 2004; Campbell, 2007 for further details of the application of MMNL).
A simple MMNL cannot explain causes of any observed heterogeneity in individual preferences because each parameter value varies stochastically around its mean. one way to address this limitation is to consider differences in preferences due to demographic and other characteristics of respondents. However, in discrete-choice experiment models, researchers cannot directly include these characteristics in econometric models because they remain the same across all alternatives, whereas only attributes of the goods under evaluation vary across alternatives. Therefore, individuals’ characteristics must interact either with the choice attributes or with the alternative-specific constants (Hanley et al., 2001). In our study, we consider various interaction terms between choice attributes and individual demographic characteristics such as age, income, gender, education, travel experience and package preference.
4. Survey Design and Data
Data were collected through an in-person consumer survey targeted toward Korean beef consumers. The survey was conducted in January and February 2007. Eleven agricultural economics students (undergraduate and graduate students) who speak Korean served as interviewers and conducted surveys with 1,000 grocery shoppers using randomised choice sets and attribute level designs for each respondent. The interviewers were sent to small and large grocery stores in Seoul, Korea, and a convenience sample of survey participants was solicited at store entrances. Using choice experiments (often referred to as conjoint analysis), respondents were asked what type of beef they would buy among the orthogonally designed choice sets (Kuhfeld et al., 1994). Each choice set included alternative types of beef with various levels of price, quality and country-of-origin attributes.
Table 1 summarises attributes and levels of each attribute included in the choice experiment design. The attribute price has 11 levels ranging $10–35 per pound.6 Five levels of marbling grade are considered following the current grading scale in Korea, and three levels are considered for freshness. The survey also includes three quality attributes: chilled or frozen beef, free of antibiotics or not and free of GMO feed ingredients or not in beef production. Finally, three different countries-of-origin are considered: Korean, US and Other (Australia, New Zealand and Canada). Therefore, with the number of attributes and attribute levels in Table 1, a full factorial design results in 3,960 scenarios or beef attribute combinations. For each survey, choice sets were constructed by randomly assigning scenarios as options to each choice set without repetition. Random assignment of profiles from a full factorial design has been shown to work efficiently in estimating values (Lusk and Norwood, 2005). Before the survey, interviewers were trained in survey administration and beef attributes described in the survey so that they could sufficiently explain differences in the levels of each attribute to the survey participants. Pictures and descriptions of each attribute were also provided to interviewers so that they could present them to survey participants, and an example of a choice set is illustrated in Figure 1.7
Attributes and levels in choice experiment design
| Attributes | Levels |
|---|---|
| Price ($/lb) | 10, 12.5, 15, 17.5, 20, 22.5, 25, 27.5, 30, 32.5, 35 |
| Marbling | Extra premium, premium, A, B, C |
| Freshness | High, medium, low |
| Chilled (vs. frozen) | Yes, no |
| Free of antibiotics | Yes, no |
| Free of GMO feed ingredients | Yes, no |
| Country-of-origin | Korean, US, other exporting countries |
An example of choice set used in choice experiment
Table 2 presents summary statistics of the demographic characteristics of survey participants. Of 1,000 participants, 6.3% were in their twenties, 14.8% were in their thirties, 61.4% were in their forties and fifties and 17.5% were over 60 years of age. The gender ratio was heavily skewed towards women; 16.3% were male and 83.7% were female, which is not surprising because most grocery shoppers in Korea are traditionally female.8 Almost 60% of participants had some college level education, a college degree or a postcollege graduate degree. The majority of participants, 59.9%, were full-time homemakers. Low-income (less than or equal to $24,000 annually), medium-income ($24,001–60,000 annually) and high-income (>$60,000 annually) households made up 13.5%, 57.6% and 28.9% of the sample, respectively. Survey participants were also asked if they had travelled to foreign countries at least once or had lived in foreign countries for more than a one-month period. These questions were asked to see if travelling or residing in a foreign country influenced a consumer’s preference on beef from that country. Approximately 60% of the participants had travel experience in foreign countries, but only 10% had lived for more than a one-month period in foreign countries. In particular, 9.5% had traveled to the USA, whereas 6.1% responded that they had an experience of living in the USA. Two questions were asked about consumers’ beef purchasing and cooking preferences. Consumers were asked how likely they were to purchase packaged beef rather than purchasing beef from a specialist butcher. Approximately 50% (1 on a 1- to 5-point scale) responded that they strongly prefer packaged beef to butcher shop beef, whereas only 14.6% (5 on a 1- to 5-point scale) responded that they were most likely to buy beef from a butcher. When consumers were asked how much they enjoy cooking on a 1- to 5-point scale (1 = least enjoyable activity and 5 = most enjoyable activity), about 57% responded either 4 or 5.
Summary statistics on demographic characteristics of survey participants
| Variable | % |
|---|---|
| Age | |
| Twenties | 6.3 |
| Thirties | 14.8 |
| Forties | 26.9 |
| Fifties | 34.5 |
| Over 60 years | 17.5 |
| Gender | |
| Male | 16.3 |
| Female | 83.7 |
| Education | |
| Less than high school | 4.8 |
| High school | 35.6 |
| Some college/college | 52.7 |
| Graduate school | 6.9 |
| Occupation | |
| Professional job | 9.3 |
| Full-time homemaker | 59.9 |
| Office worker | 8.2 |
| Public official | 2.4 |
| Own business | 8.5 |
| Unemployed | 11.7 |
| Annual income ($) | |
| ≤24,000 | 13.5 |
| 24,001–60,000 | 57.6 |
| 60,001–96,000 | 20.4 |
| ≥96,000 | 8.5 |
| CTravel | |
| USA | 9.5 |
| Canada | 1.8 |
| Australia | 1.7 |
| New Zealand | 0.8 |
| Other countries | 45.5 |
| None | 40.7 |
| CStay | |
| USA | 6.1 |
| Canada | 0.8 |
| Australia | 0.3 |
| New Zealand | 0.6 |
| Other countries | 2.8 |
| None | 89.4 |
| Prefpac | |
| 1 | 47.8 |
| 2 | 9.5 |
| 3 | 9.7 |
| 4 | 18.4 |
| 5 | 14.6 |
| Cooklike | |
| 1 | 15.4 |
| 2 | 13 |
| 3 | 14.8 |
| 4 | 27.2 |
| 5 | 29.6 |
5. Empirical Models
The deterministic part of individual i’s utility of choosing option j, Vij = Xijβ, in equation (1), can be specified as:
(3)where Price refers to the retail price of beef ($/lb); Marbling1 to Marbling4 are dummy variables that represent the grade of marbling: Marbling1 = 1 if beef is extra premium grade, 0 otherwise; Marbling2 = 1 if beef is premium grade, 0 otherwise; Marbling3 = 1 if beef is grade A, 0 otherwise; Marbling4 = 1 if beef is grade B, 0 otherwise. Grade C of marbling has been dropped to avoid the perfect collinearity problem. Freshness1 and Freshness2 are dummy variables that represent the extent of beef freshness with three different levels: Freshness1 = 1 if beef is highly fresh, 0 otherwise; Freshness2 = 1 if beef is moderately fresh, 0 otherwise. The dummy variable for the lowest freshness is used as a base. Chilled, Antibiotics and GMO are dummy variables that represent the following: Chilled = 1 if beef is freshly chilled (instead of frozen), 0 otherwise; NonAntibiotic = 1 if beef is produced without feeding antibiotics to cattle, 0 otherwise; NonGMO = 1 if beef is produced without feeding GMO feed ingredients to cattle. Finally, three country-of-origin variables for beef are considered in this study: Korea, the USA and other. Origin_US and Origin_Other are dummy variables that represent beef from the USA and other countries such as Australia, New Zealand and Canada with Korean beef as a base: Origin_US = 1 if beef is imported from the USA, 0 otherwise; Origin_Other = 1 if beef is imported from other countries (stated above), 0 otherwise. Obviously, when the status quo option is chosen, the utility level in equation (3) is set to be zero.
Marginal rates of substitution among any attributes, xi, can be calculated as the ratio of the coefficients. WTP or marginal value of an attribute can be estimated by dividing the estimated coefficient for the attribute by the coefficient of payment vehicle (e.g. price). For example, when βk is the estimated coefficient for the Price attribute as in equation (3), WTP for jth attribute is:
(4)To facilitate statistical inferences for the WTP estimates, standard deviations are estimated using a parametric bootstrapping procedure. Following Krinsky and Robb (1986), 1,000 observations were randomly drawn from multivariate normal distributions using estimates and variance–covariance matrices from estimation methods discussed previously.
6. Estimation and Results
The regression results from the MNL and MMNL are reported in Table 3.9 All coefficients are statistically significant at the 5% level and show the expected signs. The negative sign for the price coefficient indicates that the level of consumers’ utility decreases as the price of beef increases. The positive signs for marbling coefficients imply that Korean consumers have greater preferences for well-marbled beef. The regression results also indicate that Korean consumers prefer fresh, chilled (rather than frozen), antibiotic-free and GMO-free beef. The negative signs of the parameter estimates for the country of origin suggest that Korean consumers value Korean beef more than imported beef, even if all controlled quality attributes are the same. Overall coefficients from the MNL are slightly smaller than those from the MMNL.
Marginal effects from random utility models: Multinomial logit model (MNL) vs. mixed multinomial logit model (MMNL)
| Independent variables | MNL | MMNL |
|---|---|---|
| ||
| Price | −0.17 (0.043) | −0.17 (0.05) |
| Marbling1 | 0.79 (0.06) | 0.92 (0.09) |
| Marbling2 | 0.64 (0.06) | 0.76 (0.08) |
| Marbling3 | 0.64 (0.06) | 0.75 (0.09) |
| Marbling4 | 0.30 (0.07) | 0.36 (0.08) |
| Freshness1 | 0.76 (0.05) | 0.95 (0.07) |
| Freshness2 | 0.38 (0.05) | 0.47 (0.06) |
| Chilled | 0.20 (0.04) | 0.27 (0.05) |
| NonAntibiotic | 0.85 (0.04) | 1.26 (0.09) |
| NonGMO | 0.79 (0.04) | 1.18 (0.09) |
| Origin_US | −2.20 (0.06) | −2.44 (0.21) |
| Origin_Other | −2.05 (0.05) | −2.39 (0.21) |
| Log Likelihood | −7,293 | −7,250 |
| Number of observations | 10,000 | 10,000 |
A likelihood test result is reported in Table 4. The likelihood test of the MNL vs MMNL model rejects the null hypotheses of a fixed coefficient vector because χ2-statistic = 86 and χ2-critical value = 23.68. Based on estimated coefficients in Table 3, WTP is calculated for each quality attribute and country of origin using equation (4) and is reported in Table 5. Overall, WTP estimates from MMNL are bigger than those from MNL. The standard deviations of WTPs are estimated using the Krinsky–Robb bootstrapping technique, and all WTP estimates are significant at the 5% level. Consumers were willing to pay more for better marbling grades, and the WTP difference between the base grade (grade C) and higher grades (Extra Premium, Premium, grade A and grade B) ranges from $1.81/lb to $4.77/lb from MNL, whereas the differential ranges from $2.15/lb to $5.54/lb from MMNL. The results also suggest that consumers were willing to pay more for fresh beef. WTPs for high and medium beef freshness levels are $4.63/lb and $2.31/lb, respectively, higher than the least-fresh beef from MNL and are $5.70/lb and $2.84/lb, respectively, higher than the least-fresh beef from MMNL. Chilled beef is valued more than frozen beef by $1.19/lb (MNL) and $1.61/lb (MMNL). Antibiotic-free and GMO-free beef are valued more than generic beef by $5.15/lb and $4.77/lb, respectively, from MNL, and $7.54/lb and $7.12/lb from MMNL. Finally, the effect of country-of-origin seems to be the most important factor in Table 5. MNL estimates show that the value of country of origin of Korean beef is $13.35/lb more than the value of country of origin of US beef, whereas it would be $12.41/lb more than the value of country of origin of imported beef from Canada, Australia and New Zealand. These WTP estimates become a bit larger when we use the MMNL. Consumers’ value of Korean origin is $14.63 more than the value of US origin and $14.38 more than their value of country of origin from other imported beef. The relative value of country of origin of US beef seems slightly lower than the value of the ‘other’ origin imported beef in the Korean market. This result may be because the survey was conducted while US beef was still banned from the Korean market and many Korean consumers had negative perceptions of the ‘Made in USA’ label due to the ongoing US–Korean FTA negotiations.
Likelihood ratio tests: MNL vs. MMNL
| Models | χ2-statistic | No. of restrictions | χ2-critical value at 5% | Test for null hypothesis |
|---|---|---|---|---|
| MNL vs. MMNL | 86 | 14 | 23.68 | Rejected |
Comparison of willingness to pay for MNL and MMNL ($/lb)
| Independent variables | MNL | MMNL |
|---|---|---|
| ||
| Marbling1 | 4.77 (1.99) | 5.54 (2.05) |
| Marbling2 | 3.88 (1.56) | 4.57 (1.71) |
| Marbling3 | 3.86 (1.57) | 4.49 (1.63) |
| Marbling4 | 1.81 (0.72) | 2.15 (0.77) |
| Freshness1 | 4.63 (2.01) | 5.70 (2.43) |
| Freshness2 | 2.31 (0.98) | 2.84 (1.02) |
| Chilled | 1.19 (0.52) | 1.61 (0.60) |
| NonAntibiotic | 5.15 (2.36) | 7.54 (2.67) |
| NonGMO | 4.77 (2.22) | 7.12 (3.20) |
| Origin_US | −13.35 (6.13) | −14.63 (7.88) |
| Origin_Other | −12.41 (5.67) | −14.38 (7.90) |
One of our objectives in this study is to examine whether WTP (or values) for quality attributes and country of origin differ across demographic groups and groups with different levels of travelling experience to foreign countries, and different levels of cooking and shopping preference. To pursue this objective, we first generated dummy variables for three age groups (Age2030 = 1 if a respondent is in twenties or thirties, otherwise Age2030 = 0; Age4050 = 1 if a respondent is in forties or fifties, otherwise Age4050 = 0; Age60 = 1 if a respondent is over 60, otherwise Age60 = 0); three income groups (Income_L = 1 if a respondent’s annual household income is less than or equal to $24,000, otherwise Income_L = 0; Income_M = 1 if a respondent’s annual household income is more than $24,001 and less than or equal to $60,000, otherwise Income_M = 0; Income_H = 1 if a respondent’s annual household income is more than $60,000, otherwise Income_H = 0) and three travel experience groups (TRST_US = 1 if a respondent traveled to the USA at least once or lived in the USA for more than a one-month period, otherwise TRST_US = 0; TRST_Other = 1 if a respondent travelled to other foreign countries such as Canada, Australia, New Zealand and others at least once or lived in these countries for more than a one-month period, otherwise TRST_Other = 0; TRST_No = 1 if a respondent has no travel or living experience in foreign countries, otherwise TRST_No = 0). We also generated dummy variables for education (Edu_H = 1 if a respondent has at least some college degree including a two-year college degree, otherwise Edu_H = 0); shopping preference (Prefpac = 1 if respondents answered ‘1’ in Table 2, indicating that they strongly prefer packaged beef to butcher shop beef, otherwise Prefpac = 0); cooking preference (Cooklike = 1 if respondents answered over ‘4’ in Table 2, indicating that they enjoy cooking very much, otherwise Cooklike = 0); gender (Gender = 1 if a respondent is male, otherwise Gender = 0) and homemaker (Homemaker = 1 if a respondent’s occupation is full-time homemaker, otherwise Homemaker = 0). Then, we add interaction terms of these dummy variables with independent variables to the original specification in equation (3), estimate MMNL, and calculate WTPs for each quality characteristic and country of origin using equation (4). Again, corresponding standard errors are estimated following the Krinsky and Robb (1986) bootstrapping procedure.
Table 6 shows the results of using demographic characteristics to differentiate values for beef. Of 11 dummy variables included in the model, only three dummy variables (Age60, Prefpac and Homemaker) appear to have statistically significant impacts on WTPs for quality characteristics and country of origin. When people over 60 years of age evaluate the extra premium marbling grade vs. grade C, they tend to value extra premium at $2.15/lb more than people in the age category 20s and 30s. When people over 60 years of age evaluate Korean beef over US beef, they value it at $6.26/lb more than those in their twenties and thirties. When they evaluate Korean beef over other imported beef, they value it at $5.96 more than those in their age category 20s and 30s. People who prefer to purchase beef from grocery stores as a packaged good tend to care more about marbling, freshness, chilled (over frozen), use of antibiotics and GMO feed and country-of-origin than those who buy beef from butcher shops. Unlike the USA, where specialist butchers have become a rare luxury, shopping at a store with packaged beef is considered more up-market in Korea. For example, people who prefer to purchase packaged beef were willing to pay $11.50/lb more for Korean beef over US beef than those who prefer to purchase beef from butcher shops. Full-time homemakers also cared more about beef quality and country-of-origin than those in other occupations. For example, they were willing to pay $1.16/lb and $0.99/lb more for antibiotic-free and GMO feed-free beef, respectively, than those in other occupations. They also would pay $9.58/lb more for Korean beef over US beef than those in other occupations.
WTPs($/lb) with demographic interaction terms from MMNL
Willingness-to-pay estimates in Tables 5 and 6 indicate that among all factors contributing to the price differences the most important factor seems to be the country of origin followed by the use of GMO feeds and antibiotics in beef production, marbling grade and freshness. This finding is consistent with a finding from a recent consumer survey on Korean beef (Rho et al., 2007). In their survey, consumers were asked why they purchased Korean beef even if it was three times more expensive than imported beef. The most popular answer (34.9%) was ‘It is just because it is produced in Korea’. Other reasons included food safety (31.8%), freshness (21.5%) and quality (11.8%). Our finding is also consistent with messages of various advertisements funded by Korean beef producers. The majority of TV and radio commercials funded by the Korean Beef Checkoff Council tend to focus on marketing the ‘Made in Korea’ slogan by appealing to Korean consumers’ ethnocentrism.
7. Conclusions
One of the important questions that US beef producers want to know about the Korean market is why the US beef price is significantly lower than the price of Korean beef (the US beef price is almost one-third of the domestic beef price) even accounting for the 40% tariff on all imported beef. To answer this question, we surveyed 1,000 shoppers in Seoul, Korea, and conducted a conjoint analysis on consumers’ WTP for quality attributes and country of origin. The quality attributes considered in our study include marbling grade, level of freshness, chilled (or frozen) and being free of GMO feed ingredients and antibiotics in the beef production. Among all factors contributing to the price differences the most important factor seems to be the country of origin followed by the use of GMO feeds and antibiotics in beef production, marbling grade and freshness. Our study finds that Korean consumers value origins of exporting countries approximately $14/lb less than the Korean origin. Korean consumers’ valuation of beef quality and country of origin differs by some demographic groups. Older generations (over 60 years of age) tend to pay more attention to quality, food safety and country-of-origin issues than younger generations (in their twenties and thirties). Consumers who prefer to purchase packaged beef and who are homemakers also show a higher preference for beef of Korean origin and perceive Korean beef as of higher quality and safety than those who prefer to buy butcher shop beef and who are non-homemakers, respectively.
The strong evidence found for the high value of country of origin in the Korean beef market is consistent with many previous studies in the literature (Sharma et al., 1987; Verlegh and Steenkamp, 1999; Umberger et al., 2002; Mennecke et al., 2006). The findings indicate that exporting countries (including the USA) should concentrate their advertising and promotion resources on improving the value of country of origin. In fact, many Korean commodity groups have already used Korean consumers’ high ethnocentrism for their advertising and promotion campaigns. For example, in 2007, the Korean beef check-off programme spent approximately 80% of their $12.3m budget on advertising and promotion programmes (Rho et al., 2007), which were heavily based on ethnocentrism. The programmes have consistently emphasised that Korean beef comes from domestic Korean cattle raised on Korean soil and therefore, must be the best for Korean consumers in terms of taste, freshness and health and food safety. Our empirical findings suggest that the top priority for beef exporters who wish to increase sales and value of their beef in the Korean market must be to counter Korean consumers’ strong ethnocentrism by improving the value of their country of origin, i.e. simply promoting quality characteristics such as marbling may not suffice; they must create consumer loyalty to the ‘Made in USA’ label through various promotion and advertising programmes. However, for a country of origin promotional campaign to work, individual countries would have to succeed in lobbying the Korean government to expedite labeling of the specific country of origin in the retail and restaurant markets. Ironically, the street protests over importation of US beef in June 2008 have led the Korean government to expedite labeling requirements for all beef. Another way of countering Korean consumers’ strong ethnocentrism may be to lower the price of imported beef by reducing the tariff rate (currently 40%) through FTAs and other trade negotiations.
Footnotes
- 2
In July 2007, the price of Korean beef (rib, extra premium) was $33/lb, whereas the price of the same grade Australian beef was $10/lb at the retail level (Korean Agro Fisheries Trade Corporation, 2007). The price of US beef was not available in 2007 due to the import ban. However, before the import ban, it was slightly higher than prices of other imported beef.
- 3
We attempt to separate the value of country of origin from values of other quality attributes such as marbling, freshness, GMO free and antibiotic free through econometric estimation. However, as a reviewer pointed out, the value of country of origin may also represent values of other quality attributes (such as food safety) that are not considered in the survey.
- 4
Labelling of country of origin was voluntary until it was mandated in July 2008. However, the labelling of origin has not been fully implemented except in large supermarkets and restaurants.
- 5
However, the researchers were actually not capable of selling the steak because of the EU ban on hormone-treated beef. This was explained to the participants after the experiment was completed.
- 6
In the survey, prices are for rib in Korean won per Guen: traditionally Korean consumers have purchased meat per Guen which is a weight equivalent to 1.32 pounds (600 g). For readers’ convenience, the prices have been transformed to $/lb (USD in July 2007).
- 7
Descriptions of attributes and other survey instructions were written in Korean for Korean participants. However, key parts of the descriptions were translated into English and listed below for readers.
•Marbling. Marbling grade represents the amount of intramuscular fat inside beef (rib). Five levels of marbling grade we use in the survey are consistent with marbling grades used by processors and retail grocers in Korea. For example, the grade Extra Premium contains the highest amount of intramuscular fat, whereas grade C contains the lowest amount of intramuscular fat among the five grade levels. Pictures of each of these grades are provided for your reference.
•Free of antibiotics. Cattle in Korea, USA, Canada, Australia and New Zealand are routinely administered antibiotics during feeding to reduce the chance of illness and to achieve higher growth and feed conversion rates. The beef (rib) with the label ‘Free of antibiotics’ is guaranteed not to have been administered any antibiotics during production.
•Free of GMO feed ingredients. Cattle are fed corn and soybean meal during production. Some of the corn and soybeans (which produced soybean meal) may have been genetically modified because a significant amount of grains and oilseed meals available for livestock production is genetically modified. The beef (rib) with the label ‘Free of GMO feed ingredients’ is guaranteed not to have been fed any genetically modified grains and oilseed meals during production.
•Country-of-origin. Three different origins are considered for this survey. They are Korea, USA and Other. The country of origin ‘Other’ includes Australia, New Zealand and Canada. - 8
Overall, the distribution of demographic characteristics in our sample is consistent with recent surveys from other studies. Rho et al. (2007) surveyed Korean consumers’ beef purchasing behaviour and the age distribution in their sample was 9.4%, 13.2%, 25.2%, 35% and 17% for age groups of 20s, 30s, 40s, 50s and over 60s, respectively, and the proportion of women was 72.9%. The income distribution was 13.5%, 54.5%, 21.1% and 10.9% for annual income of less than $24,000, $24,001–60,000, $60,001–96,000 and more than $96,000, respectively. Park et al. (2007) surveyed Korean consumers’ pork shopping behaviour and the age distribution was 5%, 28%, 38%, 22% and 7% for age groups of 20s, 30, 40s, 50s and over 60s, respectively. In their sample the proportion of women was 79.3%.
- 9
We estimated MNL and MMNL using proc mdc procedure (SAS Institute, 2002–2003).
원문 : http://onlinelibrary.wiley.com/doi/10.1111/j.1477-9552.2009.00218.x/full
품질 속성 및 국가 평가
한국산 쇠고기 시장 원산지
'축산알기 > 국내실태' 카테고리의 다른 글
| 부산물 유통 실태 : 2016 (0) | 2017.04.14 |
|---|---|
| 수입축산물 유통실태 : 2016 (0) | 2017.04.14 |
| 축산물 유통 SWOT 분석자료 모음 - 3/3 (0) | 2017.04.07 |
| 축산물 유통 SWOT 분석자료 모음 - 2/3 (0) | 2017.04.07 |
| 축산물 유통 SWOT 분석자료 모음 - 1/3 (0) | 2017.04.07 |
댓글