y? B-ioa SRAC No. 512 November i990 Southern g3; Regional ‘ l Aquaculture Center Analysis of a Local Retail Market for Catfish and Crawfish The Texas Agricultural Experiment Station Charles J. Arntzen, Director The Texas University System legeStation, Texas [Blank Page in Orwnfl Bakiletin] ' \ 5w. {v ‘é’ ~ r . m r " ' 2* .-. 1 ‘ n’ 1 \l T‘: , J ll; _ r ~: 1 '51; . ,. \\ » - hi,‘- |< ‘i n Analysis of a Local Retail Market for Catfish and Crawfish Oral Capps, Ir. F and Johannes Adrianus Lambregtsl lRespectively, professor and graduate research assistant, Department of Agrictiltural Economics, Texas A&M University, College Station. Keywords: catfish, crawfish, market analysis. ‘a Contents Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Introduction . . ." . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Nature of Scanner Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Data Source . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Conceptual Framework for the Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Data Description . . . . . . . . . . ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Individual UPCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Customer Counts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Advertisement Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Statistical Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11 Descriptive Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11 Econometric Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 Own-Price Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20 Cross-Price Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20 Advertisement Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 Seasonality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 Summary, Conclusions, and Implications for Further Research . . . . . . . . . . . . .24 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28 Literature Cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28 Appendix — Graphs of Prices of Catfish and Crawfish Products . . . . . . . . . . . . .30 D 5pm Eggjg- HOLE Executive Summary Aquaculture is a major industry in several southern states, where key aquacultural products include catfish and crawfish. Production levels during the last decade have grown substantially in the aquacultural industry. Little work, however, has been done to assist industry planners in developing effective marketing programs. This lack of knowledge is a formidable barrier to market expansion and stability. The Southern Regional Aquaculture Center (SRAC) identified the lack of market information as a critical industry need. In 1988, in recognition of this need, the SRAC commissioned a project involving researchers from Alabama, Arkansas, Louisiana, Mississippi, South Carolina, and Texas. The overall goal of this project was to gather information to facilitate the expansion of markets for catfish and crawfish produced in the southern region. In this light, this research herein reported deals with analyses of catfish and crawfish products at the retail grocer level. Specifically, the objectives were twofold: (1) to evaluate marketable product forms of catfish and crawfish in supermarkets through the use of scanner data and (2) to estimate retail demand relationships for catfish and crawfish. The time frame of the study is the period January 1987 to November . 988. The source of data is a retail food firm in Houston, Texas. Consequently, this analysis lies within the boun- daries of the traditional catfish and crawfish markets. According to McGee et al. (1989), the south-central region of the United States consumes approximately 45 percent of all catfish produced. Work with scanner data is not a trivial task. Much careful and organized computation is necessary to con- duct analyses successfully using scanner data. This study rests on weekly point-of-sale purchases of catfish and crawfish products. The items correspond to either fresh or convenience (processed) products. The convenience catfish products are Mrs. Paul's Catfish Strips, Hormel Catfish Fillets, and Hormel Catfish Bob- ber Snacks. The fresh catfish products are fresh whole catfish, fresh farm-raised catfish fillets (the aquacultural product in this analysis), and fresh ocean catfish fillets. Similarly, the single convenience crawfish product is Cajun Cook Crawfish Etouffe with Rice. The two fresh crawfish products are fresh cooked crawfish and frozen cooked crawfish meat. The weekly observations (97 in all) began on Wednesday and ended on Tuesday to conform to retail ‘food firm sales and advertising pat- _ ems. Importantly, the retail food firm in this study caters to relatively high-income customers. Customer counts per week for this firm ranged from 577,428 to 861,844 over the time frame analyzed. Advertisement space (in terms of square centimeters) of the respective products varied considerably from week to week. The principal finfish and shellfish products in terms of print space and frequency of advertisement were catfish and shrimp, respectively. Catfish and crawfish received roughly 27 and 1 square centimeters, respectively, of print space on average. The share of finfish advertisement space for catfish was 26 percent, whereas the share of shellfish advertisement space for crawfish was about 0.3 percent. In terms of frequency, advertisements for catfish occurred 48 out of 97 weeks, whereas advertisements for crawfish occurred only once. The availability of fresh catfish relative to fresh crawfish may affect advertisement frequency. Fresh catfishproducts constituted roughly 90 per- cent of all catfish sales during our study. Fresh catfish products generated about $13,570 in sales per week for this retail food firm. The principal fresh catfish products are fresh farm-raised catfish fillets and fresh whole catfish. Convenience catfish products gave rise to roughly $1,500 in sales per week, about 10 percent of all catfish sales. The key convenience catfish products in terms of dollar sales were Hormel Catfish Bobber Snacks and Mrs. Paul's Catfish Strips. For the retail firm, convenience crawfish products constituted $789 week- ly in sales, approximately 84 percent of all crawfish sales. Fresh crawfish products constituted the remain- ing 16 percent. The principal fresh crawfish product was fresh cooked crawfish meat. With few exceptions, purchases of catfish and craw- fish products varied tremendously on a weekly basis. The purpose of this study’ s econometric analysis, per- haps the cornerstone of the project, is to develop models to explain such variation in product movement. The econometric models correspond to demand relationships at the retail level. The dependent variable in the respective demand relationships is purchases per 1,000 customers. The respective exogenous variables are (1) own-price; (2) prices of competing products; (3) advertisement variables; and (4) seasonality. Emphasis is on price and advertisement elasticities. Price elas- ticities refer to percentage changes in purchases caused by unit percentage changes in prices; similarly, adver- tising elasticities refer to percentage changes in pur- chases caused by unit percentage changes in advertising. Elasticities reveal the sensitivity of pur- chases to price changes and / or to promotion efforts. Generally, for both crawfish and catfish products, the explanatory power of the econometric models is on the order of 50 to 70 percent. The econometric models are satisfactory, especially with the relatively large amount of variation to be explained on a week-to-week basis. All own-price elasticities are negative and, except for fresh ocean catfish fillets, fresh crawfish, and fresh cooked crawfish, are statistically significant. The respective elasticities are in the elastic range for all catfish products except for the aggregate convenience catfish. The own-price elasticities for the individual convenience catfish products range from -2.723 to -13.652, and for fresh catfish, the range is from -1.295 to -6.046. The own-price elasticity for fresh ‘crawfish is -0.835, and the own-price elasticity for Cajun Cook Crawfish Etouffe with Rice is -0.812. The demand for fresh cooked crawfish is price elastic. The magnitude of this elasticity is -2.682. In sum, sample evidence exists to indicate that own-price exerts a notable influence on purchases, holding all other factors constant. . For fresh catfish products, only 6 of 24 cross-price elasticities are statistically different from zero. For con- venience catfish products, 5 of 18 cross-price elasticities are statistically different from zero. For the prepared entree Cajun Cook Crawfish Etouffe with Rice, shellfish is the only statistically sig- nificant cross-price variable. The price of beef, the price of finfish, and the price of shellfish influence purchases of fresh crawfish. Prices of competing products, how- ever, do not bear greatly on purchases of fresh cooked crawfish. Own-advertisement elasticities are positive and statistically significant for fresh farm-raised catfish fil- lets and the aggregate of all fresh catfish products. The respective own-advertisement elasticities for these products are 0.058 and 0.109, much smaller in mag- nitude than the corresponding own-price elasticities. Own-advertisement effects are not significant for craw- fish products. With few exceptions, virtually no linear association exists between product price and product exposure (advertisement space). For fresh farm-raised catfish fil- lets, a significant, albeit relatively small, negative as- sociation is evident. Importantly, few cross- advertisement effects are significantly different from zero. That is, advertisement exposure for finfish, shellfish, and the aggregate of beef, pork, poultry, lamb, and veal only slightly affects purchases of crawfish and catfish. Because demand for individual catfish products in the retail firm studied is elastic, incentive to lower prices exists. Such a strategy results in increases in total revenue. This strategy is particularly important because of the general insignificance of cross-product prices. For fresh farm-raised catfish fillets, a strategy to increase advertisement exposure may be worthwhile to boost demand. However, strategies to alter advertisement » exposure for various products to increase demand for other catfish and crawfish products appear not to be worthwhile. Seasonality is a key determinant in purchases of all catfish products, except for whole catfish. For crawfish, on the other hand, seasonality is a key factor only in purchases of fresh cooked crawfish or fresh crawfish. Although scanner data have been available for several years to marketers, such data represent a new form of information to the aquacultural sector. This study constitutes a pilot test of the use of scanner data to investigate the demand for catfish and crawfish products for a local market. Despite the apparent success in analyzing retail demand relationships with scanner data for catfish and crawfish products, concern lies with generalizing the results to regional or national levels. Scanner data from supermarkets in a particular location represent a "con- trolled" experimental situation. The community- specific results may not allow defensible, broad nationwide or regional inferences. Because of this potential limitation, the results of local analyses (such as this study) should not be used on a stand-alone basis. Although this analysis is limited geographically to the Houston area, the methodology can be replicated in other regions. "Q I Introduction Aquaculture is a major industry in several southern states, where key aquacultural products include catfish and crawfish. Production levels of the aquacultural sector have grown substantially during the last decade. Except for the establishment of The Catfish Institute, however, little work has been done to assist industry planners in developing effective marketing programs. This lack of knowledge is a formidable barrier to market expansion and stability. The Southern Regional Aquaculture Center (SRAC) identified the lack of market information as a critical industry need. In 1988, in recognition of this need, SRAC commissioned a project involving researchers from Alabama, Arkansas, Louisiana, Mississippi, South Carolina, and Texas (Hatch, 1988). The overall goal of this project was to gather information to facilitate the expansion of markets for catfish and crawfish produced in the southern region. National surveys of 3,600 consumers, 1,800 retail grocery managers, and 1,800 restaurant managers were conducted from April 1988 to June 1988. The principal aims of these surveys were threefold: (1) to identify socio-economic effects on household consumption pat- terns of catfish and crawfish (McGee et al., 1989); (2) to identify factors associated with the handling of catfish and crawfish by retail grocery outlets; and (3) to identify factors affecting the handling of catfish and crawfish by full-service restaurants. Recently, the fastest growing segment of the seafood industry is the retail food sector, particularly grocery stores. In 1987, seafood sales were $17.8 billion, 5.7 percent of total grocery store sales ($313 billion). Catfish is the leading aquacultural product in the United States. Sales of catfish were nearly $150 million in 1988 (Engle et al., 1988). This research analyzes catfish and crawfish products at the retail grocer level. Specifi- cally, the objectives are twofold: (1) to evaluate marketable product forms of catfish and crawfish sold in supermarkets through the use of scanner data; and (2) to estimate retail demand relationships for catfish and crawfish. The source of data for the analyses in this study is from a retail food firm in Houston, Texas. This analysis lies within the boundaries of the traditional catfish and crawfish markets. According to McGee et al. (1989), the south-central region of the United States consumes approximately 45 percent of all catfish. The time frame analyzed is the period from January 1987 to November 1988. This research directly benefits not only the SRAC but also food retailers, especially given the recent proliferation of seafood delicatessens. Our work complements the national surveys com- missioned by the SRAC. The focus, however, is only on a local up-scale market in Houston. Previous studies that deal with influences on retail grocery demand for catfish or crawfish are few. Engle et al. (1988) conducted a survey to profile prices and quantities sold of the most important seafood products at the retail grocer level in a 13-county area in east-central Alabama and west- central Georgia. No attempt, however, was made to estimate retail demand functions for the specific types of seafood products. Raulerson and Trotter (1973) con- ducted a market experiment in six Atlanta grocery stores to determine the demand for commercially raised catfish during a 2-month period in 1972. Price elas- ticities for catfish ranged from -1.23 to -8.93. Our analysis builds on the work of Engle et al. (1988) as well as on the work of Raulerson and Trotter (1973). Emphasis is on estimating price elasticities and advertisement elasticities for individual catfish and crawfish products at the retail grocer level. The analysis in this research is similar to the work by Capps (1989) that examined retail demand relationships for steak, ground beef, roast beef, chicken, pork chops, ham, and pork loin. Nature of Scanner Data This research rests on the collection, organization, and use of scanner data from a retail food firm in Houston. Data on a weekly basis are from the period January 1987 to November 1988 (97 weeks). Scanner data constitute a readily available source of product- specific information. Such data not only permit analysis of demand for disaggregate commodities but also rep- resent current market conditions. Traditional demand analysis has generally depended upon aggregate annual, quarterly, or month- ly time-series data of purchases and prices. These data often do not represent current market conditions and typically are too general for product-specific decision making. Time-series data from conventional secondary sources, in short, typically lack disaggregate product and price detail. Panels and surveys provide more detailed data for specific products as well as socio- demographic infonnation but are expensive methods of data collection. Generally, a key limitation of panels or surveys is the lack of price information. Prices must be imputed from reported quantity and expenditure figures. Analysts may question the use of such imputa- tions, particularly the estimation of cross-sectional demand functions (Cox and Wohlgenant, 1986). Another key limitation of the use of surveys (not neces- sarily panels) is the lack of time continuity. To illustrate, the United States Department of Agriculture sponsors the Nationwide Food Consumption Survey (NFCS). Since its inception in 1936, this survey takes place only once approximately every 10 years (e.g., 1965-66, 1977- 78, 1987-88). Scanner data, on the other hand, constitute a readily available, current, and timely source of product-specific information. To quote Tomek (1985), "existing secon- dary data seem especially inadequate for studying product demand in retail markets, and fundamental work needs to be done to obtain relevant data" (pp. 913-914). "The data associated with computerized checkout systems in grocery stores could become an important source of information for studying retail demand" (p. 913). Scanner data are not without limitations, however. Because of problems of data integrity and of too much detail creating "data overload," empirical practitioners have been less than enthusiastic about the value of scanner data in market research. Each week as few as 10 to 20 supermarkets will generate the equivalent amount of data as would a panel of 10,000 households. Conse- quently, considerable resources are necessary to reduce the mass of data to useful summary figures for demand analysis purposes. Despite the volume of price, quantity, and expen- diture information, scanner data, at least from retail food firms, typically lack the dimension of consumer socio-demographic data. To circumvent this problem, several firms currently issue customer identification cards (e.g., HEB food stores, personal communication) from which these firms obtain socio-demographic in- formation essential to the derivation of income elas- ticities. For demand analyses based on scanner data from supermarkets, the common experimental unit is the individual food firm (aggregation over consumers), not the individual consumer. This aggregation problem is not necessarily negligible. If the food firm caters to a more or less homogeneous group of consumers, how- ever, this aggregation problem is of little consequence. Despite the sheer volume of information, scanner data files need to be augmented with information per- taining to advertising or promotional activities. Competitors’ actions are also important but are ex- tremely difficult to anticipate, measure, and evaluate. Analogously, it is difficult to represent nonprice effects (merchandising schemes, coupons, services, cleanli- ness, product selection, and reputation for fresh meat or produce). Consequently, the all-other-things-held- constant assumption may fail with the use of scanner data. Food stores supplying the data for meat, poultry, and fish items as well as for produce must have the equipment to generate labels enabling the products to be electronically scanned. This equipment is expensive, sensitive, and may not always produce scannable labels. Because of the inability of particular food stores to scan such items, some scanner data for meat, poultry, fish, or produce may not be available or reliable. In regard to data integrity, food industry observer Richard E. Shulman makes this point: . . caveat about scanning data: It's not accurate. It is representative. Don't expect the scanner to capture 100 percent of all sales. There are dozens of reasons that sales are "lost": bad symbols, poorly trained checkers, etc. The impor- tant thing to understand is that most sales will be cap- tured and the resulting data can be acted upon" (National Grocers Association Technology Newsletter, 1985). Lesser and Smith (1986) point out that scanner data misrepresent item movement (quantity purchased) if the scanning file is not rigorously maintained, if the items cannot be or are not scanned, or if Universal Product Codes (UPCs) are not entered manually. Fur- » therrnore, scanner data may not provide accurate infor- mation if stock shrink accounts for a substantial portion of the movement of a product. Because stock shrink generally contributes approximately 1 to 2 percent of supermarket sales, this factor should not be a major issue for the vast range of products. The integrity of the data is therefore a function of the level of discipline of the retail firm in capturing accurate information. Along this line, Lesser and Smith (1986) conducted a study to evaluate the accuracy of scanner data. Their results suggested that "substantial error is possible when examining individual items on a weekly basis. This factor should be considered when using scanner data" (p. 71). Scanner data from supermarkets in a particular location (for this analysis Houston) presumably repre- ~ sent a "controlled" experimental situation. The com- munity-specific results, however, may not contribute to defensible, broad regional or nationwide inferences. Because of this potential limitation, the results of local analyses should be used not on a stand-alone basis but as supporting evidence in conjunction with a research approach designed to conduct demand analyses with scanner data on a national or regional basis. Data Source The source of data for the analyses in this study is a retail food firm in Houston. The time frame is from January 1987 to November 1988 (Table 1). Weekly ob- servations began on Wednesday and ended on Tuesday to conform to store sales and advertising patterns. The number of supermarkets in operation by this firm over this time interval was 43. Importantly, the retail food firm in this study caters to relatively high-income cus- tomers, roughly 40 percent of whom have annual in- comes in excess of $60,000 per household. The number of finfish and shellfish species sold in this retail firm over the period January 1987 to Novem- ber 1988 was 448. Of the 448 species, 6 were catfish species, and 3 were crawfish species (Table 2). Scanner data are available on a daily basis. Ag- gregation of daily information into weekly information Table 1. Documentation of the weeks for the scanner project, January 1987 to November 1988. 1987 1988 Week Week Week Week ending ending ending ending Week no. date Week no. date Week no. date Week no. date 1 113 27 714 52 105 75 614 2 1 20 28 721 53 1 12 76 621 3 127 29 728 54 1 19 77 628 4 203 30 804 55 1 26 78 705 5 210 31 811 56 .202 79 712 6 217 32 818 57 209 8O 719 7 224 33 825 58 216 81 726 8 303 34 901 59 223 82 802 9 310 35 908 60 301 83 809 10 317 36 915 61 308 84 816 1 1 324 37 922 62 315 85 823 1 2 331 38 929 63 322 86 830 1 3 407 39 1006 64 329 87 906 1 4 414 4O 1013 65 405 88 913 15 421 41 1020 66 412 89 920 16 428 42 1027 67 419 90 927 1 7 505 43 1 103 68 426 91 1004 18 512 44 1117 69 502 92 1011 19 519 45 1117 70 509 93 1018 20 526 46 1 124 71 516 94 1025 21 602 47 1201 72 523 95 1 101 22 609 48 1208 73 530 96 1 108 23 616 49 1215 74 607 97 1115 24 623 50 1222 25 630 51 1229 26 707 Table 2. Universal product codes (UPCs) for catfish and crawfish products. UPC Description Number of observations 1 116211186 Mrs. Paul's Catfish Strips” 87 3760015151 Hormel Catfish Fillets’ 97 376004221 4 Hormel Catfish Bobber Snacks’ 46 20607400000 Fresh whole catfishb 97 20608100000 Fresh farm-raised catfish filletsb 97 20614000000 Fresh ocean catfish filletsb 97 1830012021 Cajun Cook Crawfish Etouffe with Rice’ 97 20608000000 Fresh cooked crawfishb s4 20613600000 Frozen cooked crawfish meatb 39 “Convenience (processed) item — prepared entree. Fresh item. makes computations more manageable. This weekly information also allows for better representation of su- permarket operations. To clarify, price changes are usually initiated once per week, and merchandising activities such as newspaper advertisements and dis- plays are also usually done weekly (Carmen and Figuroa, 1986). This study is based on point-of-sale purchases. For each product, movement (in pounds) and price (in cents/ pound) are reported by week. For commodity aggregates (fresh catfish, fresh crawfish), convenience (processed) catfish, and convenience (processed) craw- fish, the quantities of the various items correspond to the sum of the respective quantities of the relevant UPCs. Implicit prices of the commodity aggregates are weighted averages of all individual UPC prices. The weighting mechanism is the ratio of the sum of all sales over the UPCs to the sum of all quantities. Quality affects may result from such commodity aggregation (Houthakker, 1952; Cox and Wohlgenant, 1986). When distinct items are aggregated into com- modity groups, variations occur in the implicit prices. Furthermore, the weighted average prices change with the quantities of the component goods consumed. Al- though the use of implicit prices potentially limits the analysis, given that the aquacultural products in ques- tion are relatively homogeneous, quality effects at- tributable to commodity aggregation are assumed to be negligible. Emphasis in this study is on demand relationships at the firm level in lieu of the store level. The prices for each UPC are the same across the supermarkets studied, and sales of meat items at the stores are reasonably similar. Hence, data from all supermarkets in the finn are aggregated to form 97 weekly time-series observa- tions. Conceptual Framework for the Analysis Holdren (1960, pp. 117-123) provides the concep- tual framework for this analysis. Attention is on multi- product retail demand functions. According to Holdren (1960, p. 123)," the multiple product retail demand func- tion can be characterized by qi = fi (p1, p2, . . ., p“, a1, a2, . . ., am). (1) where the q's represent quantity variables expressed in appropriate units, the p's represent price variables, and the a's represent attributes of the retailer's non-price offer variation." Advertising, sales promotion activities, hours open, and customer services are concrete ex- amples of non-price offer variation. Seasonality also may affect the quantity variables, all other things held constant (Carmen and Figueroa, 1986). Because they are proxies for tastes and preferences of the collection of consumers who frequent retail stores, the socio- demographic influences in retail demand functions must be considered as well. In light of the previous discussion, the specification for the demand models in this study is as follows: Q“ = KP“, Pit, SEASON, ADV“, ADVJ-t) (2) where Qtt is purchases per 1,000 customers (in pounds [fresh items] or in units (convenience [processed] items) of catfish or crawfish itemi in week t, t = 1, . . ., 97; Ptt is price of catfish or crawfish product i in week t (cents/ pound); Pt-t corresponds to a vector of prices of competing products (j refers to the set of competing products) in week t (cents/ pound); SEASON cor- responds to a set of monthly binary variables to measure seasonality; ADVtt corresponds to the amount of print space given for catfish or crawfish product i in the weekly advertisement flier (square centimeters); and ADV-t corresponds to the amount of print space given for the set of competing products in the weekly advertisement flier (square centimeters). Data corresponding to purchases are converted to a per customer basis. Consequently, the dependent variables reflect purchases per 1,000 customers. Be- cause of unavailability of information, the model specification excludes competitors’ prices and advertis- ing as well as socio-demographic variables. The variables Ptt and Pt-t capture own-price and cross-price effects. Own-price effects are hypothesized to be negative. Cross-price effects may be negative or positive to reflect substitutable or complementary relationships among the commodities in question. For disaggregate analyses, the identification of appropriate substitutes or complements a prion’ is a difficult task. Cheng and Capps (1988) suggest that the demand for finfish and shellfish depends upon poultry, pork, and beef prices. In this study, such prices correspond to weighted average prices of poultry, pork, and beef products. Weighted average prices of finfish and shellfish products are also included in the model specification. Specifically, in the demand relationships for catfish products, the weighted average price of fin- fish items, excluding catfish, is used as a regressor as well as the weighted average price of shellfish items. Similarly, in the demand relationships for crawfish, the weighted average price of shellfish items, excluding crawfish, is used as a regressor along with the weighted average price of finfish items. Because data are only from a single firm, some may argue from the following rationale that price elasticities are not estimable: (1) consumers can respond to price changes by shopping at different stores within a market area, and (2) no information in this study is available on their purchases at other stores or on the prices charged at other stores. According to the Food Marketing In- stitute, however, only 27 percent of shoppers compare prices 'from store to store (Cox and Foster, 1985). Con- sequently, it is- possible to estimate price elasticities. Additionally, multicollinearity between competitor's prices and in-store prices may be too strong to allow for measurement of the separate effects of the variables (Funk, et al., 1977). Therefore, in this study, the omission of competitors’ prices may not be a limiting factor in estimating in-store price elasticities. Local newspaper advertising is the only advertis- ing mode considered in our study. Although television, radio, and in-store displays are used by the chain, these forms are primarily oriented toward creating a favorable corporate image (personal communication with the retail firm). Newspaper advertising (the week- ly advertisement flier of the firm), on the other hand, is geared primarily to promoting specific products. The basic format and design of the newspaper advertise- ments used by the chain were the same throughout the period. Therefore no measure of "creative aspects" of advertising is necessary. In this study, advertising data refer to the amount of print space devoted to each item, measured in square centimeters. This study allows the examination of own- and cross-advertisement effects. All other things held con- stant, own-advertisement effects are hypothesized to be positive, whereas cross-advertisement effects are hypothesized to be negative. The respective set of ad- vertisement variables used in the retail demand relationships corresponds precisely to the set of price variables previously discussed. Competitors’ advertis- ing is excluded because of resource constraints. Data Description This section deals with three components: (1) data for individual UPCs, (2) documentation of customer counts by week, and (3) documentation of advertise- ment space for catfish and crawfish products. Pulling together price/ quantity information on individual UPCs, customer counts, and advertisement space was an exacting task. Individual UPCs Price and quantity information are not necessarily available for all UPCs for all 97 weeks (Table 2). For example, Mrs. Paul's Catfish Strips, Hormel Catfish Bobber Snacks, fresh cooked crawfish, and frozen cooked crawfish meat were available at week 1 of the analysis but eventually were discontinued by the retail firm. The various catfish and crawfish items correspond to either fresh or convenience (processed) products. The three convenience catfish products and corresponding UPCs are prepared entrees (the numbers in parentheses are the actual UPCs): (1) Mrs. Paul's Catfish Strips (1116211186), (2) Hormel Catfish Fillets (3760015151), and (3) Hormel Catfish Bobber Snacks (3760042214). The three fresh catfish products are (1) fresh whole catfish (2060740000O), (2) fresh farm-raised catfish fillets (20608100000), and (3) fresh ocean catfish fillets (206140000O0). The single convenience crawfish product is Cajun Cook Crawfish Etouffe with Rice (1830012021); the two fresh crawfish products are (1) fresh cooked crawfish (20608000000) and (2) frozen cooked crawfish meat (20613600000). This analysis also considers aggregate products, namely fresh catfish, convenience (processed) catfish (prepared entrees), and fresh crawfish. Customer Counts Figure 1 plots customer counts, which per week for the retail firm under study ranged from 577,428 to 861,844 over the time frame. The average customer count was 724,070. Advertisement Space Information on customer counts and advertisement space must be augmented to the price and quantity in- fonnation of the individual UPCs. That is, data pertain- ing to advertisement space and customer counts are not automatically part of the scanner data pertaining to the individual UPCs collected at the point of sale. Thousands Advertisement space (in terms of square cen- timeters) for the respective aquacultural products varied considerably from week to week (Figs. 2 and 3). Descriptive statistics of advertisement variables are ex- hibited in Table 3. On the basis of print space, catfish averaged almost 27 square centimeters. In comparison, crawfish received slightly more than 1 square cen- timeter of print space on average. Advertisements of catfish occurred 48 times over the 97 week span, while advertisements of crawfish occurred only once. The combination of all remaining shellfish products received roughy 57 square centimeters of print space on average; the frequency of such advertisements is 70 of 97 weeks. Additionally, the combination of all remain- ing finfish products received an average of nearly 92 square centimeters of print space. The frequency of such advertisements is 77 out of 97 weeks. The share of shellfish advertisements for crawfish is roughly 0.3 per- cent, whereas the share of finfish advertisement space for catfish is 26.0 percent. The principal shellfish product in terms of print space and advertisement fre- quency is shrimp; the principal finfish product in terms of print space and advertisement frequency is catfish. 900 l 800 roo- soo . WW4 lllllllllllllllllllllllllllllqlllllllllvllllllllllllllllllllllllllllllllllllllllllllllllllllll VVeek Figure 1. Customer counts. 120 100 8O 6O 4O 2O 3OO 250 200 150 i100 J 5O Square Centimeters I I I I §l§§§§§§§+lllII§I§I+I§IIIIIIIIIllIIIIIQIIIIIIIIIIIJIIIIl1ll%l|l|lll_.§ Week I Figure 2. Advertisement space for crawfish. Square Centimeters ,_ l ll L ll ll lllllll L lll l II ll_l_l l__IllI _l§_l|llIlIl ll I I LLLLLLLLLLLLLLLII ' “" T I ' “I T T I“ “ “ “'T““ ' “"'T"“ Week Figure 3. Advertisement space for catfish. Il- 4-‘ .- Table 3. Advertisement space“ for aggregate finfish and shellfish products. Species Mean St. Dev. Sharee Min. Max. Frequency N Crawfish 1 .142 1 1.255 0.0028 0 1 10.85 1 97 Catfish 26.699 53.853 0.2603 0 291 .58 48 97 Shellfishb 56.676 73.988 NA 0 442.02 70 97 Scallops 3.837 8.464 0.1 1 76 0 45.00 24 97 Lobster 1 .719 10.501 0.0258 0 94.20 5 97 Clams 0.820 3.592 0.0287 0 21.75 6 97 Crab 1 .633 7.996 0.0349 0 54.37 6 97 Shrimp 46.330 72.474 0.7014 0 412.02 64 97 v Mussels 0.398 2.021 0.0132 0 13.00 4 97 Oysters 1 .936 5.059 0.0724 0 30.60 15 97 Finfishc 91 .594 93.904 NA 0 399.78 77 97 Shark 6.61 1 18.396 0.0641 0 955.55 25 97 Grouper 1 .528 7.326 0.0125 0 50.70 6 97 Perch 3.993 21 .723 0.0214 0 198.38 8 97 Scrod 1 .540 5.526 0.0209 0 44.40 1 1 97 Swordfish 2.535 16.822 0.0231 0 164.00 9 97 Mackerel 0.147 1 .451 0.0015 0 1 4.30 1 97 Redfish 0.477 3.992 0.0065 0 38.64 2 97 Orange roughy 4.897 22.048 0.0524 0 201 .69 15 97 Rockfish 2.902 9.330 0.0287 0 77.76 17 97 Tuna 22.019 49.982 0.1276 0 247.86 24 97 Mahi-Mahi 0.71 3 2.864 0.0074 0 15.96 6 97 Pollock 4.579 12.130 0.0549 0 94.40 25 97 Salmon 20.836 55.163 0.1355 0 291 .20 24 97 Flounder 2.070 _6.439 0.0226 0 39.00 13 97 Snapper 0.396 2.236 0.01 32 0 1 3.50 3 97 Bluefish 1 .202 7.108 0.0119 0 55.25 4 97 Oreo dory 4.959 23.160 0.0356 A 0 193.60 1 1 97 Turbot 0.423 4.166 0.0016 0 41 .04 1 97 Trout 1 .886 7.284 0.0266 0 45.60 8 97 Halibut 2.595 1 1 .308 0.0228 0 105.00 1 4 97 Whitefish 5.276 14.017 0.0479 0 96.60 21 97 Other productsd 996.10 312.062 340.04 1 875.37 97 97 ‘In square centimeters. bAll shellfish products except crawfish. ‘All finfish products except catfish. dBeef, pork, poultry, lamb, and veal. eShare of either finfish or shellfish advertisement space. 10 Finally, advertisements of beef, pork, poultry, lamb, and veal received roughly 1,000 square centimeters of print space per week. At least one of these products was advertised every week. Statistical Procedures This section deals with two components: (1) descriptive statistics of catfish and crawfish products; and (2) a layout of the econometric analysis. Descriptive Statistics Detailed descriptive statistics of purchases and prices for the catfish and crawfish products are ex- hibited in Tables 4 and 5. Descriptive statistics cor- respond to the mean, median, standard deviation, min- imum, and maximum. The mean and median relate to measures of central tendency, the standard deviation corresponds to a measure of dispersion, and the mini- mum and maximum define the range of the data. In terms of item movement, the key fresh catfish product is fresh farm-raised catfish fillets; the least important fresh catfish product is fresh ocean catfish fillets. The major prepared entrees are Honnel Catfish Fillets and Mrs. Paul's Catfish Strips. The principal crawfish product was the prepared entree Cajun Cook Crawfish Etouffe with Rice. The major fresh crawfish product was fresh cooked crawfish. In terms of price, the most expensive catfish items, on average, were catfish fillets, either farm-raised or ocean catfish. The least expensive were Hormel Catfish Fillets and Bobber Snacks. The Table 4. Desaiptive statistics of purchases of catfish and crawfish species. Standard UPC code N Mean Median deviation Minimum Maximum Purchases Catfish 1116211186“ 87 131.56 149 57.78 1 223 ' 3760015151’ 97 219.94 216 56.46 96 543 3760042214“ 46 84.23 87 47.67 1 274 Convenience catfish’ 97 377.89 388 150.58 1 44 1030 20607400000b 97 804.76 597 560.84 228 2523 206081000001’ 97 2826.2 2365 1113.84 1276 6751 20614000000” 97 12.70 s 32.47 0 276 Fresh catfishb 97 3643.66 3408 1 103.43 1545 7282 Crawfish 1830012021’ 97 274.25 274 72.71 129 504 20600000000” s4 17.20 16 12.42 0 s7 20613600000” a9 2.84 a 1.9a 0 7 Fresh crawfish 97 16.04 15 13.73 0 57 Purchases per 1,000 customers Catfish 1116211186“ 87 0.1898 0.2092 0.0915 0.0012 0.3693 3760015151“ 97 0.3099 0.2999 0.1005 0.1264 0.8584 3760042214“ 46 0.1284 0.1246 0.0774 0.0013 0.4331 Convenience catfish‘ 97 0.541 1 0.5325 0.2573 0.1921 1 .6283 20607400000b 97 1.1279 0.7857 0.8317 0.3002 3.6609 206081000001’ 97 3.9326 3.2800 1.5685 1.6805 8.8272 20614000000b 97 0.0172 0.0062 0.0451 0 0.3940 Fresh catfishb 97 5.0778 4.5341 1 .5935 2.0348 9.5215 Crawfish 1830012021’ 97 0.3766 0.3753 0.0848 0.1699 0.6591 206080000001’ 84 0.0248 0.0218 0.0190 0 0.0954 206136000001’ 39 0.0042 0.0043 0.0030 0 0.0102 Fresh crawfish 97 _ 0.0232 0.0191 0.0201 0 0.095 a Purchases in terms of units. b Purchases in terms of pounds. 11 ' Price in terms of cents/ unit. Price in terms of cents/ pound. Table 5. Descriptive statistics of prices of catfish and crawfish species as well as prices of meat, poultry, and fish. Standard UPC code N Mean Median deviation Minimum Maximum Catfish 1116211186‘ 87 333.00 333.00 0.00 333.00 333.00 376(D15151' 97 258.13 2591B 11.64 217.00 268.00 3760042214‘ 46 255.54 259.00 8.42 217.00 259.00 Convenience catfish‘ 97 279.71 280.26 1 1 .55 240.98 298.38 20607400000” 97 264.12 269.00 30.32 199.00 299.00 206081000001’ 97 421.40 4291K) 38.57 344.00 469.00 206140000001’ 97 556.89 499.00 87.08 459.00 798.00 Fresh catfishb 97 382.41 395.11 38.30 295.56 440.66 Crawfish 1830012021‘ 97 300.02 289.00 32.72 267.00 369.00 20608000000” s4 44.3.52 449.00 8.97 429.00 449.00 20613600000b 39 1138.64 999.00 179.46 999.00 1398.00 Fresh crawfish 77 484.58 449.00 58.82 429.00 671.10 Meat, poultry, and fish prices PPORK 97 290.84 293.19 27.43 204.09 363.60 PPOULT 97 169.50 172.75 27.74 91.68 218.76 PBEEF 97 246.70 253.61 24.36 196.37 282.55 PSHELLC 97 611.07 605.94 107.71 410.56 800.63 PFINd 97 489.32 493.95 34.87 382.28 567.92 ° Weighted average price of fresh shellfish species, excluding crawfish. d Weighted average price of fresh finfish species, excluding catfish. most expensive crawfish item was frozen cooked craw- fish meat, and the least expensive crawfish item was Cajun Cook Crawfish Etouffe with Rice. Fresh catfish and crawfish products were more costly on a per unit basis than were convenience counterparts. Weighted average prices of fresh pork, poultry, and beef at this firm respeCtively averaged $2.90, $1.69, and $2.46 per pound. These prices were lower than those for fresh catfish ($3.82 per pound) and fresh crawfish ($4.84 per pound). Weighted average prices for finfish and shellfish at the retail firm were $4.89 and $6.11 per pound on average, respectively. Average dollar sales and average budget shares per week for catfish and crawfish products are exhibited in Table 6. Budget shares represent the proportion of sales attributable to individual products. Catfish products contributed $15,061 weekly in sales, while crawfish products constituted $937 weekly in sales at this retail firm. Fresh catfish products constituted roughly 9O per- cent of all catfish sales, and produced $13,571 in sales per week. The principal fresh catfish products were fresh farm-raised catfish fillets and fresh whole catfish. Convenience catfish products generated roughly $1 ,500 12 in sales per week, about 10 percent of all catfish sales. The key convenience catfish products in terms of dollar sales were Hormel Catfish Bobber Snacks and Mrs. Paul's Catfish Strips. This set of characteristics of catfish sales is similar to the national study of McGee et al. (1989). In the national study, the most preferred product forms of catfish were fresh fillets followed by fresh whole- dressed fish. The product fonn least preferred was prepared entrees. Convenience crawfish products constituted $789 weekly in sales, approximately 84 percent of all craw- fish sales. Fresh crawfish products constituted the remaining 16 percent. The principal fresh crawfish product was fresh cooked crawfish meat. Graphs corresponding to movement (purchases) over time for each of the catfish and crawfish products are exhibited in Figures 4 - 15. Graphs corresponding to prices over time for each of the catfish and crawfish products are exhibited in the Appendix. The graphs summarize more clearly the variability in item move- ment and in price on a week-to-week basis. With few exceptions, movement varied tremendously per week. Table 6. Average dollar sales and average budget shares per week for catfish and crawfish products. Category Average dollar sales/week Average budget share All catfish products 15,061 Convenience catfish 1,488 0.0988 UPC 1116211186 602 0.0400 UPC 3760015151 653 0.0434 UPC 3760042214 230 0.0153 Fresh catfish 13,571 0.9011 UPC 20607400000 2,020 0.1341 UPC 20608100000 11,523 0.7651 UPC 20614000000 27 0.0018 All crawfish products 937 Convenience crawfish 789 0.8417 UPC 1830012021 Fresh crawfish 148 0.1582 UPC 20608000000 36 0.0386 UPC 20613600000 1 12 0.1196 Pounds (Thousands) 0 lllllllll:lllllllll%lllllllllillllllllIiillllllll}!ll!lIlllwllilllllillllllllwlllllllllilllll Week Figure 4. Purchases of fresh catfish. 13 Units 1200 1000 - 800 r 600' 400 " 200 r o lllllllllllllllllllllllllllllllllllllllllllllllllllllillllllllliIllllllllllllllllllllllllllllll Week] Figure 5. Purchases of convenience (processed) catfish. Variability in prices was not as dramatic as was variability in item movement. Econometric Analysis The purpose of the econometric analysis, the cornerstone of this project, is to develop models to explain the variation in item movement. The functional form chosen for the analysis of any set of demand relationships is open to empiricism. The study rests on the use of the linear functional form. The interpretation of parameter estimates as elasticities is convenient with the double logarithmic functional form. However, this functional representation was not used because of potential zero observations, especially for the advertise- ment variables. Emphasis is on price and advertisement elasticities. Price elasticities refer to percentage changes in purchases caused by unit percentage changes in prices; similarly, advertisement elasticities refer to per- centage changes in purchases caused by unit percent- age changes in advertising. Elasticities are often of primary interest not only to agricultural economists but also to food retailers. Knowledge of price elasticities 14 allows retailers to deal with shortage or surplus situa- tions to minimize price volatility. Advertising elas- ticities reveal the sensitivity of purchases to advertisement efforts. Under the assumption that supply is perfectly elas- tic in this local market, a seemingly unrelated regression (SUR) procedure is workable. Random exogenous fac- tors such as general level of economic activity, competitors’ actions, or prices of nonmeat items within the retail firm may affect purchases of the respective catfish and crawfish products apart from the specified predetermined variables. Consequently, the distur- bance terms of the equations may be contemporaneous- ly correlated. Given that the exogenous variables are not the same in each relationship, gains in estimation ef- ficiency can be expected with the SUR procedure rela- tive to the use of ordinary least squares (Fomby et al., 1984). However, because of differences in the available number of observations for each particular product (see Table 2), the empirical results rest on the use of single- w equation estimation techniques —- either ordinary least squares or generalized least squares. E Pounds 60 50 4O 20"- lllllllllllllllllll|lllllll|lllllllllll lug“! l () I l l T :_llllllllllllllllllllllllllll}ll|\§LL§L L§§§L L T Week Figure 6. Purchases of fresh crawfish. Empirical Results This section concerns the econometric demand analyses for the various catfish and crawfish products. The econometric model corresponds to demand relationships at the retail level. The dependent variable in the respective demand relationships is units of move- ment per 1,000 customers. The purpose of the econometric analysis is to identify and assess factors affecting purchases per 1,000 customers. The respective exogenous (independent) variables are (1) own-price, (2) prices of competing products, (3) advertisement variables, and (4) seasonality (monthly dummy vari- ables). For example, in the econometric model for fresh catfish, price variables corresponding to fresh catfish, convenience (processed) catfish, other finfish, shellfish, and beef, pork, poultry are included. As well, advertis- ing variables corresponding to catfish, other finfish, shellfish, and the combination of beef, pork, poultry, veal, and lamb are included. 15 Ordinary least squares (OLS) or generalized least squares (GL5) regression results for the econometric models are exhibited in Tables 7-10. The list and description of variable names is given in Table 11. Be- cause of a lack of variation in the price of Mrs. Paul's Catfish Strips, we could not estimate the demand relationship for this product (see Figure A. 4). Similarly, because of a lack of variation in the advertisement of crawfish in the form of frozen cooked crawfish meat, the demand relationship for this product could not be estimated. For crawfish products, the adjusted coefficients of determination (R) range from 0.4413 (Cajun Cook Crawfish Etouffe with Rice) to 0.5465 (fresh crawfish). For catfish products, the R measures ranged from 0.5693 (fresh ocean catfish fillets) to 0.9239 (Hormel Catfish Bobber Snacks). For both crawfish and catfish products, the explanatory power of the econometric models generally was on the order of 50 to 70 percent. In all cases, the amount of variation explained by the models was statistically significant. On the basis of Table 7. OLS or GLS regression results for commodity aggregates of catfish. Variable Parameter estimate t-statistic Fresh catfish“ INT ERCEPT 13.10832‘ 3.543 PCATFRES 002865‘ -10.028 PCATCONV -0.00623 _-0.709 PFIN 0.00625‘ 1 .892 PSHELL -0.00036 -0.390 PBEEF 0.00614 1.209 PPORK -0.00196 -0.420 PPOULT -0.00162 -0.402 ADFIN -0.00109 -0.911 ADSHELL -0.00012 -0.088 ADVAOM -0.00006 -0.168 ADCAT 0.01037‘ 4.903 SEASON 1.999“ AD] R-SQ 0.6762 Convenience catfishb INTERCEPT 1 .37236‘ 2.273 PCATFRES -0.00142‘ -3.170 PCATCONV -0.00055 -0.365 PFIN -0.00138‘ -2.646 PSHELL 0.00101‘ 6.678 PBEEF -0.00145‘ -1.898 PPORK 0.00065 0.951 PPOULT -0.00097 -1 .603 ADFIN -0.00007 -0.403 ADSHELL -0.00019 -0.927 ADVAOM 0.0001 1‘ 2.233 ADCAT -0.00074‘ -2.279 SEASON 6.219“ DURBIN-WATSON DW 1.773 AD] R-SQ 0.6971 ‘ Statistically significant at the 0.05 level. a GLS estimates to circumvent serial correlation problems. b OLS estimates. c F-statistic. Table 8. OLS or GLS regression results for individual fresh catfish products. Variable Parameter estimate t-statistic Fresh ocean catfish fillets‘ INTERCEPT 0.39352‘ 2.706 P206140 -0.00004 -0.722 PCATCONV -0.00058‘ -1 .860 PFIN 0.00019‘ 1 .665 PSHELL 0.00004 1.034 PBEEF -0.00011 -0.648 PPORK -0.00042‘ -2.551 PPOULT -0.00020 -1 .457 ADFIN -0.000005 -0.132 ADSHELL 0.000006 0.118 ADVAOM -0.00001 -1.070 ADCAT -0.00001 0.212 SEASON 6.473“ AD] R-SQ 0.5693 Fresh farm-raised catfish filletsb INT ERCEPT 15.97354‘ 4.280 P206081 -0.03100‘ -11.405 PCATCONV -0.01200 -1.309 PFIN 0.00415 1.361 PSHELL -0.00135 -1.472 PBEEF 0.00695 1.487 PPORK -0.00071 —0.173 PPOULT 0.00072 0.201 ADFIN -0.00040 -0.368 ADSHELL 0.00026 0.213 ADVAOM 0.00013 0.417 ADCAT 0.00859‘ 4.356 SEASON 1.989“ DURBIN-WATSON DW 2.254 AD] R-SQ 0.7077 Fresh whole catfishb INT ERCEPT 9.98400‘ 4.513 P206074 -0.02582‘ -11.503 PCATCONV -0.00387 -0.704 PFIN 0.00088 0.468 PSHELL -0.00208‘ -3.710 PBEEF 0.00395 1.423 PPORK -0.00450‘ -1.772 PPOULT 0.00235 1.076 ADFIN -0.00026 -0.388 ADSHELL -0.00004 -0.058 ADVAOM -0.00036‘ -1.918 ADCAT 0.00065 0.571 SEASON 0.681‘ DURBIN-WATSON DW 2.309 AD] R-SQ 0.6214 ‘ Statistically significant at the 0.05 level. ' GLS estimates to circumvent serial correlation problems. b OLS estimates. t c F-statistic. 17 Table 9. OLS or GLS regression results for individual convenience (processed) catfish products (prepared entrees). Variable Parameter estimate t-statistic Hormel Catfish Bobber Snacks‘ INTERCEPT 1 88256" 5.524 P3764 000686‘ -6.201 PCATFRES -0.00020 -1.061 PFIN -0.00005 -0.325 PSHELL 0.0001 3 1 .383 PBEEF -0.00047 -1 .395 PPORK 0.00048 1.271 PPOULT -0.00023 -0.741 ADFIN 0.00004 0.606 ADSHELL -0.00042" -1 .71 4 ADVAOM 0.00002 0.800 ADCAT 0.00038 0.925 SEASON 5.479“ AD] R-SQ 0.9239 Hormel Catfish Filletsb INTERCEPT 1 32190‘ 4.650 P3761 -0.00327* -3.710 PCATFRES —0.00034 -1 .623 PFIN -0.00019 -0.535 PSHELL -0.00030 -1.273 PBEEF 0.0001 4" 1 .91 3 PPORK 0.00004 0.138 PPOULT -0.00018 —0.666 ADFIN -0.00001 -0.147 ADSHELL -0.00002 -0.246 ADVAOM 0.00001 0.792 ADCAT -0.00011 -0.817 SEASON 3.763“ DURBIN-WATSON DW 1.563 AD] R-SQ 0.5789 “ Statistically significant at the 0.05 level. a GLS estimates to circumvent serial correlation problems. b OLS estimates. c F-statistic. 18 l,‘ Table 10. OLS regression results for crawfish. Variable Parameter estimate t-statistic Cajun Cook Crawfish Etouffe with Rice INTERCEPT 0.68921‘ 2.473 P183 -0.00102* -2.629 ’~ PCRAFRES -0.00008 -0.601 ' PFIN 0.00028 1 .1 22 PSHELL 000019‘ -1.790 PBEEF 0.00028 0.732 PPORK -0.00030 0.854 PPOULT 0.00026 0.770 ADFIN -0.000006 -0.076 ADSHELL -0.00014 -0.906 ADVAOM -0.00003 -1.477 ADCRAW 0.00060 0.721 SEASON 1 .502’ DURBIN-WATSON DW 1 .769 AD] R-SQ 0.4413 Fresh crawfish INTERCEPT 0.07830 1 .294 PCRAFRES -0.00004 -1 .330 P183 0.00008 0.987 PFIN 000011‘ -1.960 PSHELL 0.00005‘ 2.246 PBEEF 000018‘ -2.230 PPORK -0.00001 -0.196 a PPOULT -0.00004 -0.613 ADFIN 000003‘ -1 .733 ADSHELL -0».00004 -1 .281 ADVAOM 0.000002 0.520 ADCRAW -0.00001 -0.086 SEASON 2.241“ DURBIN-WATSON DW 1.491 AD] R-SQ 0.5465 Fresh cooked crawfish INTERCEPT 0.08321 0.276 P206080 -0.00015 -0.276 P183 0.00009 0.572 PFIN -0.00005 -1 .034 PSHELL 0.00002 1.141 PBEEF -0.00009 -1.197 PPORK 0.000001 0.014 PPOULT 0.00001 0.238 ADFIN -0.00002 -1 .261 ADSHELL -0.00002 -0.637 ADVAOM 0.000001 0.202 ADCRAW -0.00005 -0.305 SEASON 2.413“ DURBIN-WATSON DW 1.331 AD] R-SQ 0.4839 * Statistically significant at the 0.05 level. a a F-statistic. “ "\ Units 250 200 150- 100" “MW O llllllllllllllllllElllllllllllllllllllllllllllllllIllllllllllllllllllllllllllll I Week Figure 7. Purcahses of Mrs. Paul's Catfish Strips (llPC 1116211186). goodness-of-fit, the econometric models were therefore highly satisfactory, especially with the relatively large amount of variation to be explained on a week-to-week basis. The 0.05 level of significance was chosen for the statistical tests. According to Durbin-Watson (DW) tests, serial correlation problems were evident for fresh catfish, fresh ocean catfish fillets, and Hormel Catfish Bobber Snacks. To circumvent these serial correlation problems, a generalized least squares procedure (Cochrane-Orcutt) was used. For the other products, no serial correlation problems were apparent. The DW test statistics for the remaining products ranged from 1.331 to 2.309. On the basis of condition indices and variance decomposition proportions (Belsley et al., 1980), no degrading collinearity problems were evident. Own-Price Effects Consistent with prior expectations, all own-price elasticities were negative, and except for fresh ocean catfish fillets, fresh crawfish, and fresh cooked crawfish, the respective coefficients were statistically significant. The own-price elasticities were in the elastic range for all catfish products except for the aggregate con- 20 venience catfish. Except for fresh cooked crawfish, the own-price elasticities for crawfish were in the inelastic range. As exhibited in Table 12, the own-price elas- ticities for the individual convenience catfish products ranged from -2.723 to -13.652, and for fresh catfish, the range was from -1.295 to -6.046. The price elasticities for fresh catfish were within the range of previous work by Raulerson and Trotter (1 973). The elastic demands at the retail level for catfish were also consistent with elastic demands documented by Kinnucan et al. (1988) at the processor level. The own-price elasticity for fresh craw- fish was -0.835, and the own-price elasticity for Cajun Cook Crawfish Etouffe with Rice was -0.812. The demand for fresh cooked crawfish was elastic. The mag- nitude of this price elasticity was -2 .682. In sum, consid- erable sample evidence exists to indicate that own-price exerts a notable influence on purchases, holding all other factors constant. Cross-Price Effects For fresh catfish products, only 6 of 24 cross-price elasticities were statistically different from zero. Cross- price elasticities may be either positive (indicative of gross substitutability) or negative (indicative of gross ‘ Statistically significant at the 0.05 level. a Table 11. List and description of variable names. Variable name Description PCATFRES Weighted average price of fresh catfish PCATCONV Weighted average price of convenience (processed) catfish (prepared entrees) PBEEF Weighted average price of fresh beef g PPQRK Weighted average price of fresh pork ~ PPQULT Weighted average price of fresh poultry PFIN Weighted average price of fresh finfish Weighted average price of fresh shellfish PSHELL _ ADFIN igiifiiiifiifiiifiifiifiiiifiiiffeh ADSHELL Advertisement space for beef, pork, poultry, lamb, and veal M Advertisement space for catfish Advertisement space for crawfish ADCRAW Monthly dummy variables (M1, ..., M11) to capture seasonality; (M1=1 if SEASON January, 0 otherwise; ..., M11=1 if November, 0 otherwise). Reference month, Dec be . P206140 Pricggf frresh ocean catfish fillets Price of fresh farm-raised catfish fillets Price of fresh whole catfish P3764 Price of Hormel Catfish Bobber Snacks P3761 Price of Hormel Catfish Fillets P183 Price of Cajun Cook Crawfish Etouffe with Rice PCRAFRES Weighted average price of fresh crawfish P206080 Price of fresh cooked crawfish Q Table 12. A summ of the econometric anal ses for catfish and crawfish s ecies. "Y Y P Own- Own-price advertisement b UPC code elasticity’ elasticity“ ADIRSQ Seasonalityc Catfish 3760015151 -2.723 NS 0.5789 3.763‘ 3760042214 -13.652 NS 0.9239 5.479‘ Convenience catfish -0.284 -0.073 0.6971 6.219‘ 20607400000 -6.046 NS 0.6214 0.681 20608100000 -3.321 0.058 0.7707 1.989‘ 20614000000 NS (~1.295) NS 0.5693 6.473‘ Fresh catfish -2.157 0.109 0.6762 1.999‘ Crawfish 1830012021 -0.812 NS 0.4413 1.502 20608000000 NS (~2.682) NS 0.4839 2.413‘ Fresh crawfish NS (~0.835) NS 0.5465 2.241‘ a At sample means. " Adjusted i‘. c F-statistic. " NS refers to the regression coefficient as not statistically different from zero. Units 600 500 400 300 it 100" 200 “WWW “wow, llllllllllllllllllllllllllllwlllllllll:lllllllllllllllllllllllllllllllllllllllllllllllllllllll Week Figure 8. Purchases of Hormel Catfish Fillets (LIPC 3760015151). complementarity). The price of finfish items positively influenced purchases of fresh ocean catfish fillets as well as the aggregate category of fresh catfish. The cross-price elasticities at the sample means were 5.405 and 0.602. Prepared catfish entrees and fresh ocean catfish fillets were gross complements (cross-price elas- ticity of -9.432 at the sample means). Similarly pork and fresh ocean catfish fillets were gross complements (cross-price elasticity of -7.101 at the sample mean). Shellfish and fresh whole catfish were also gross com- plements (cross-price elasticity of -1.126 at the sample means), and pork and fresh whole catfish were gross complements (cross-price elasticity of -1.160 at the sample means). However, for fresh farm-raised catfish fillets, cross-price effects were not statistically different from zero. For convenience catfish products (prepared entrees), 5 of 18 cross-price elasticities were statistically different from zero. The price of fresh catfish, beef, and finfish negatively affected purchases of the aggregate of all convenience catfish products. At the sample means, the cross-price elasticity for fresh catfish and con- venience catfish was -1.003; for beef and convenience 22 catfish, the cross-price elasticity was -0.661; for finfish and convenience catfish, the cross-price elasticity was -1.247. On the other hand, the price of shellfish positive- ly affected purchases of the aggregate of all convenience catfish products. At the sample means, the cross-price elasticity was 1.140. For the two individual prepared entrees, only one statistically significant cross-price ef- fect was evident. Shellfish and Hormel Catfish Fillets were gross substitutes (cross-price elasticity of 0.276 at the sample means). Cross-price effects are not statisti- cally significant for Hormel Catfish Bobber Snacks. For the prepared entree Cajun Cook Crawfish Etouffe with Rice, shellfish was the only statistically significant cross-price variable. The cross-price elas- ticity at the sample means was -0.308, indicative of gross complements. The price of shellfish positively affected purchases of fresh crawfish (cross-price elasticity of 1.316 at the sample means). The prices of beef and of finfish negatively influenced purchases of fresh craw- fish. The cross-price elasticities of beef and finfish were -1.914 and -2.320, respectively, at the sample means. Prices of competing products were, however, not im- U ~" 20o Units 300 250 l 150- 100" 5O lllllllllllllllllll O ' I 1 Week Figure 9. Purchases of Hormel Catfish Bobber Snacks (UPC 3760042214). portant influences on purchases of fresh cooked craw- fish. Advertisement Effects Consistent with prior expectations, own-advertise- ment elasticities were significantly different from zero and were positive for fresh farm-raised catfish fillets and the aggregate of all fresh catfish products. As ex- hibited in Table 12, the respective own-advertisement elasticities for these products were 0.058 and 0.109 at the sample means; consequently the own-advertisement elasticities were much smaller in magnitude than the corresponding own-price elasticities. In contrast with expectations, own-advertisement effects were negative and significantly different from zero for the aggregate category of prepared catfish entrees. Own-advertise- ment effects were not significant for crawfish products. Few cross-advertisement effects were significantly dif- ferent from zero for either catfish or crawfish products. Advertisement exposure for finfish (shellfish) negative- ly influenced purchases of fresh crawfish (Hormel Cat- fish Bobber Snacks), whereas the aggregate advertisement of beef, pork, poultry, lamb, and veal 23 positively influenced purchases of convenience catfish and negatively influenced purchases of fresh whole catfish. Pairwise correlation coefficients between own- price and own-advertisement effects for catfish and crawfish products were, except for fresh farm-raised catfish fillets and the aggregate category of prepared catfish entrees, not statistically different from zero. .In these cases, a significant albeit relatively small negative association (correlation coefficients of -0.2024 and - 0.2400, respectively) existed between own-price and own-advertisement variables. For the remaining products, no statistical association was evident between product price and product exposure (advertisement space). Seasonality Seasonality was a major determinant of purchases of all catfish products except for fresh whole catfish. All other things held constant, purchases of the aggregate of fresh catfish products were significantly higher in January, March, April, July, August, and October. For fresh farm-raised catfish fillets, purchases were sig- Pounds 3000 I 2500 " T 2000 I 1500 ' 1000 500 ii = l. lllln 0 llllllllllllllllllllLllllllllllllllllllillllllllillllllllllllllllllllilllllllllllllllllllllllll I Week Figure 10. Purchases of fresh whole catfish (UPC 20607400000). nificantly higher from January to May as well as from July to October relative to the other months. Fresh ocean catfish fillets purchases, however, were significantly lower in all months relative to December. For the ag- gregate of convenience catfish products as well as for Hormel Catfish Fillets, purchases were significantly higher from January to May and from September to November relative to other months. For Hormel Cat- fish Bobber Snacks, purchases were significantly higher in January and May and significantly lower in February, September, October, and November relative to other months. Seasonality for crawfish was a key factor only in purchases of fresh cooked crawfish or fresh crawfish. Seasonality was not a statistically significant deter- minant of the purchase of Cajun Cook Crawfish Etouffe with Rice. For fresh cooked crawfish and for fresh craw- fish, purchases were significantly higher from January to June relative to other months. 24 Summary, Conclusions, and Implications for Further Research Although scanner data have been available for several years to marketers, such data represent a new form of infomtation to the aquacultural sector. This study constitutes a pilot test of the use of scanner data to investigate the demand for catfish and crawfish products for a local market (retail food firm) in Hous- ton. The time frame for this analysis was the period January 1987 to November 1988. This study rests on analyses of seven individual catfish and crawfish products as well as commodity aggregates (fresh and convenience catfish and craw- fish). Although work with scanner data was exacting, requiring much computational effort, useful descrip- tive statistics and graphs of prices and purchases can be generated. With additional effort, information on cus- tomer counts and advertising can be obtained for this firm. Extreme caution is in order, however, in the or- ganization of scanner data for analysis. Pounds(Thousands) 7 ‘N e ~ ' H 5 ‘ 4__ 3 2 r ' 1 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII I I I I I I I I I We e k Figure 11. Purchases of fresh farm-raised catfish fillets (UPC 29608100000). '\ Pounds 3 O O 2 5 O " 2 0 O - 1 5O - ,1 O0 - so ~ 1\ 1 wTfT?fvfTJ6\i#?#1g&T1+;;4T&;§n¢§rTwT1¢L+¢é;Iéé IIIIIII _ G ' -\ () I I I T I I I I We e k Figure 12. Purchases of fresh ocean catfish fillets (UPC 20614000000). 25 Units 600 500 " 400 - 300 r 200 - 100" llllllllllllllllllllllllllgilllllllllllllllllllllllllllIll!lllllIll!lllllllllllllllllliillll I I Week Figure 13. Purchases of Cajun Cook Crawfish Etouffe with Rice (UPC 1830012021). The cornerstone of this analysis was the specifica- tion and estimation of econometric models to analyze purchases of catfish and crawfish products on a per 1,000 customer basis. The‘ purpose was to identify and assess key factors to allow producers, processors, and distributors to anticipate consumer behavior in retail markets, improve planning, and provide better service to consumers. With few exceptions, the models adequately cap- tured significant variation in purchase patterns. Generally, the key exogenous variables in this analysis were own-price, own-advertising, and seasonality. In particular, the purchase patterns of the products in question were highly sensitive to price changes and moderately sensitive to the effects of advertising. Fu- ture research to explore potential seasonality in prices and advertising elasticities merits attention. Overall, the research lends encouragement to the possibility of using scanner data in market research. Because demand for individual catfish products in this retail firm was elastic, incentive to lower prices may exist. Such a strategy results in increases in total revenue. Assuming that costs do not change, this 26 strategy is particularly important because of the general insignificance of cross-product prices. Own-advertisement effects were important only for fresh farm-raised catfish fillets and the aggregate of all fresh catfish products. Own-advertisement elas- ticities for these products were positive but very inelas- tic. Nevertheless, a strategy to increase advertisement exposure for fresh farm-raised catfish may be worthwhile to boost demand, subject to the costs of advertising. Without cost information, however, it is impossible to discern whether a strategy to reduce own- price is preferable to a strategy to increase exposure. Such a determination depends upon the costs of the respective strategies. Own-advertisement effects were not significant for crawfish products or for individual convenience catfish products. Few cross-advertisement effects were significantly different from zero. Conse- quently, except possibly for fresh catfish fillets, strategies to alter advertisement exposure of various products to increase product demand are probably not worthwhile. Despite the apparent success in analyzing retail w demand relationships with scanner data, concern lies with generalizing the results to regional or national U Pounds 20- y 10S x O lllllllllllllllllllllllllllllllllllllll lgi,_l_l l_lllllllllllll|llllllllllllllllll I I l T 1 I t 1 Week Figure 14. Purchases of fresh cooked crawfish (UPC 20608000000). levels. Scanner data from supermarkets in a particular location represent a "controlled" experimental situa- tion. The community-specific results may not allow defensible, broad nationwide or regional inferences. Because of this potential limitation, the results of local analyses (such as this study) should not be used on a stand-alone basis. Although this analysis was limited to the Houston area, the methodology can be replicated in other geographic regions, particularly the south-central United States, the traditional market area for catfish and crawfish. Though much recent empirical and theoretical work exists on demand and market analyses, reliable estimates of demand parameters for aquacultural products in general and catfish and crawfish products A in particular are few. Scanner data can be the most 27 detailed and definitive source of retail food industry statistics available to researchers. User of scanner data can expand demand and market analyses. Although the use of scanner data is in the embryonic stage of development, it promises fresh insights for market research. In the next decade, analysts will concentrate on scanner data assembly, management, and analysis (Branson et al., 1987). Con- ceivably, with proper management, scanner data may well be the ultimate data source of demand and market analyses at the retail level. This particular pilot study sheds light on the potential utility of scanner data in market research. Pounds Week Figure 15. Purchases of frozen cooked crawfish meatHIPC 206013600000). Acknowledgments We appreciate the Southern Regional Aquaculture Center (SRAC) for funding this project, which was sup- ported in part by a grant from the United States Depart- ment of Agriculture, Number 87-CRSR-2-3218, supported jointly by the CSRS and the Cooperative Extension Service. Special recognition is due to C.G. Shepherd of the SRAC. We wish to thank SRAC par- ticipantsLynn Dellenbarger, Upton Hatch, James (Bud) Dillard, Carole Engle, Henry Kinnucan, and Robert Pomeroy for reviews of drafts of this bulletin. As well, we wish to thank Wade Griffin, Dick Edwards, Dave Bessler, Don Farris, and Mike Mazzocco for review comments. Any remaining errors or omissions are the sole responsibility of the authors. We give special com- mendations to Nila Reece for programming expertise, data management, and computational assistance. Im- portantly, we wish to thank Randall's Food Markets, Inc., for data procurement. Finally, credit is due to Natalie South for generating the graphs used in this bulletin as well as for diligence in typing the manuscript. Literature Cited Belsley, D.A., E. Kuh, and R.E. Welsch. 1980. Regression diagnostics: identifying influential data and sour- ces of collinearity. New York, New York: John Wiley and Sons. Branson, R.E., et al. 1987. Data sources for demand analysis, in R. Rauniker and C.L. Huang eds. Food demand analysis: problem, issues and empirical evidence. Iowa State Press. Capps, Jr., O. 1989. Utilizing scanner data to estimate retail demand functions for meat products. American Journal of Agricultural Economics, 71, 3:750-60. Carmen, H.F., and E.F. Figueroa. 1986. An analysis of factors associated with weekly food store sales variation. Agribusiness, 2:375-90. Cheng, H.T., and O. Capps, Jr. 1988. Demand analysis of fresh frozen finfish and shellfish in the United States. American Journal of Agricultural Economics, 70, 3533-542. U '\ Cox, C., and R. Foster. 1985. What's ahead for the U.S. food processing industry? Discussion. American Journal of Agricultural Economics, 67:1155-7. Cox, T.L., and M. Wohlgenant. 1986. Prices and quality effects in cross-sectional demand analysis. American Journal of Agricultural Economics, 68, 5:908-19. ‘Engle, C.R., Upton Hatch, and Scott M. Swinton. 1988. Factors affecting retail grocery demand for seafood products in east-central Alabama and west-central Georgia. Journal of the Alabama Academy of Science, 59, 1:1-16. Engle, C., O. Capps, Jr., L. Dellenbarger, J. Dillard, U. Hatch, H. Kinnucan, and R. Pomeroy. The U.S. market for farm-raised catfish: overview of con- sumer, supermarket, and restaurant surveys. Texas A&M University, Department of Agricultural Economics, College Station, unpublished manuscript. Fomby, T.B., R.C. Hill, and S.R. Johnson. 1984. Ad- vanced econometric methods. New York, New York: Springer-Verlag. Funk, T.F., K.D. Meilke, and H.B. Huff. 1977. Effects of retail pricing and advertising on fresh beef sales. American Journal of Agricultural Economics, 59:533-37. Hatch, L.U. 1988. National survey of U.S. fish consump- tion. Presented to the Aquaculture International Congress and Exposition, Vancouver, British Columbia, Canada. 29 Holdren, B.R. 1960. The structure of a retail market and the market behavior of retail units. Englewood Cliffs, New Jersey: Prentice-Hall, Inc. Houthakker, H.S. 1952. Compensated change in quan- tities and qualities consumed. Review of Economic Studies, 191155-64. Kinnucan, H., S. Sindelar, D.W., and U. Hatch. 1988. Processor demand and price-markup functions for catfish: a disaggregated analysis with implications for the off-flavor problem. Southern Journal of Agricultural Economics, 20,2:81-91. Lesser, W.G., and J. Smith. 1986. The accuracy of super- market s¢anning data: an initial investigation. Jour- nal of Food Distribution Research, 17:69-74. McGee, W.M., L.E. Dellenbarger, and J.G. Dillard. 1989. Demographic and attitudinal characteristics of cat- fish consumers. Southern Regional Aquaculture Center Publication 508, Technical Bulletin 168. National Grocers Association Technology Newsletter, November 1985. Raulerson, R.C., and W.K. Trotter. 1973. Demand for farm-raised channel catfish in supermarkets. U.S. Department of Agriculture, Marketing Research Report Number 993, Washington, D.C. Tomek, W.G. 1985. Limits on price analysis. American Journal of Agricultural Economics, 67:905-15. Appendix Graphs of Prices of Catfish and Crawfish Products 3O 500 400 ' 300 200 100 350 300 250 200 150 100 5O Figure A.1. Price of fresh catfish. Ce nts/ Pou nd I lllllll’!l%lllllllll%lllllllll%lllllllll%lllllllll}lllllllll%lllllllll{Illllllllpllllllllpllll Week Figure A.2. Price of convenience (processed) catfish. Cents/Unit lllllllll%lllllllll%llllll111F111!llll%lllllilll}lllllllll%lllllllll%lllllllllwlllllllqlllll Week 31 Figure A.3. Price offresk crawfish. Ce n ts/ Pou nd 700 600 - 500 " I 400 I 300 I 20o 100" O IllIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII Week Figure A.4. Price ofMrs. Paul's Catfish Strips (llPC 1116211186). Ce nts/ U n it 350 I 300 250 I 200 150- 100 I 5O O llllllllIIIllllilllglllllllllgllllIllllglllllllllglIII]I1IIIIIIIIIIIIIIIIIIIIIIIIIIII Week 32 Figure A.5. Price of Hormel Catfish Fillets (UPC 3760015151). Cents/Unit 300 250 200 - 150- 100" O lllllllll%lllllllll%lllllllllgllllllllwlllllllllwllllllll:lllLJlllJPlllllllwlllllllllglllll Week Figure A.6. Price of Hormel Catfish Bobber Snacks (llPC 3760042214). C e n t s / U n i t 300 250 I 200 I 150 100 I 5O Figure A.7. Price of fresh whole catfish (UPC 20607400000). C e n t s / Po u n d 350 30o . I 250 I 200 I 150 I 10o 5O O IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII]Illll1}]!IIlIlll{IIIIIIIIIIIIIIIIIIIIIIIII Week Figure A.8. Price of fresh farm-raised catfish fillets (UPC 20608100000). C e n t s/ Pou n d 500 400 I 300 I 200 100 O 0 IIIIIIIIIIIIIIIIIJIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII Week 34 Figure A.9. Price of fresh ocean catfish fillets (UPC 20614000000). Cents/Pound "“~ 1000 800- "\ 600" 400" 200* O IIIIIIIIIILIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIILIIIIIIIIIIIIII11111IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII I I I I I I I I I Week -""\ Figure A.10. Price of Cajun Cook Crawfish Etouffe with Rice (UPC 1830012021). Cents/Unit 400 300 200" 100" »'\ O lllllllllllIlllHIILLILILJJIIIIlllllllllllllllllllllllllllllllllllllllllllllllilllillllllllllll I I I I I I I I I Week 35 Figure A.11. Price of fresh cooked crawfish (UPC 20608000000). C e n t s / Po u n d I 300 I 200 I 100 O IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII Week Figure A.12. Price of frozen cooked crawfish meat (UPC 20613600000). C e n ts / Po u n d 1600 1400" 1200 .,--==-:=-s-=--=--:===== I 800 600 I 400 I 200 0 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII Week 36 P ‘~4- [Blank Page in Original Bulletin] l [Blank Page in Orifinfl ' ‘ . .. _ ( .v..~ w’ ‘ . h’; * ‘s ._ , 1A >._ ‘\ / » ‘i ‘a [Blank Page in Original Bulletin] ' Mention of a trademark or a proprietary product does not constitute a guarantee or a warranty of the product by The Texas Agricultural Experiment Station and does not imply its approval to the exclusion of other products that also may be suitable. All programs and information of the Texas Agricultural Experiment Station are available to everyone without regard to race, color, religion, sex, age, handicap, or national origin. 8-1663 1.6M—1 1-90 1a