TDOC z TA245.7 13-1541 B873 July 1986 NO.1541 ‘i; Implications for Management Decisions THE TEXAS AGRICULTURAL EXPERIMENT STATION Neville P. Clarke, Director The Texas A&M University System College Station, Texas 1E TEXAS AGRICULTURAL EXPERIMENT STATION! Neville P. Clarke, Directorl The Texas A&M University System! College Station, Texas CONTENTS INTRQDUCTIDN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I PRQCEDURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 RICE QUALITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Quality Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Modeling Quality Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 ECDNOMIC QUALITY CDNTRDL MEASURES . . . . . . . . . . . . . . . . . . . 32 LIMITATIQNS QF THE STUDY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 SUMMARY AND CONCLUSIDNS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 APPENDIX A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 APPENDIX B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Rice Quality Factors: Implications for Management Decisions WARREN R. GRANT, Agricultural Economist National Economics Division Economic Research Service U. S. Department of Agriculture M. EDWARD RlST ER, Assistant Professor The Texas Agricultural Experiment Station Texas A&M University {Department of Agricultural Economics} B. WADE BRORSEN, Assistant Professor The Purdue Agricultural Experiment Station Purdue University {Department of Agricultural Economics} a This research was funded jointly by the Texas Rice Research Foundation (Econo-Rice Project), the Texas Agricultural Experiment Station (Project 6507), and the Economic Research Service, USDA. PREFACE This research is an extension of research reported in Texas A ricultural Experiment Station Progress Report 4202 (Brorsen et al. I984). Ad itional markets and years of data have been analyzed. Data from bid/acceptance markets in the western Texas Rice Belt are augmented by American Rice, Incorporated data from throughout the Texas Rice Belt. Analysis of the coo erative data is limited to identifying the different levels of rice quality attri utes. ACKNOWLEDGEMENTS The assistance of Rusty Bergstron, Jacko Garrett, Craig Howell, B.E. Jeter, Billy Knowlton, and Jack Wycoff in providing data for this research is sincerely appreciated. Assistance from Sue Ellen Galvan, Julleen Grant, Raymond Grant, Anna Hallaran, Nina Nobles, Dianne Rister, Ruth Saenz, and Iris Saito in data processing and typing is ratefully acknowledged. Review comments and suggestions from Bill Blaci, Bart Drees, Arthur Gerlow, Melvin Parker, James Stansel, Mo Way, Bill Webb, and Michael Wohlgenant are also appreciated. A special thanks to Kim Trant for editorial assistance. Rice Quality Factors: Implications for Management Decisions INTRODUCTION The quality attributes of a given lot of rice affect the value. Many quality attributes are related to management practices, e.g., insect damage, weed seeds, red rice. Producers need t0 know the value of these qualit attributes when making economic management decisions and in deciding w ether or not to accept an offered price. Researchers could also benefit from this information by knowing where to concentrate research efforts. A 1984 Texas Agricultural Experiment Station progress report (Brorsen et al. 1984) indicated damage caused by rice stink bugs (peck) was a major cause of price discounts in three bid/acceptance markets in the Texas Rice Belt, suggesting a need for further research on stink bu control. This paper includes data from additional markets and years to verify t ese findings. This paper analyzes quality differentials for rough rice prices observed in bid/acceptance markets on the western side of the Texas Rice Belt (i.e., west of Houston). The basic economic question investigated relates to the magnitude of discounts or premiums associated with differences in qualit as implied by historical rice market data. Implications for production-relate decisions are also developed. Cooperative sales data are used to relate the levels and frequencies of occurrence for the respective quality attributes within each county comprising the Texas Rice Belt. DATA United States rough rice is marketed through contracts with mills, bid/acceptance markets, negotiated sales, and cooperative mills on a pooled basis. Six cooperative mills handle over one-half of U.S. rice production (Mullins et al. 1981). Bid/acceptance markets are the second most im ortant marketing channel in Louisiana and Texas, accounting for about one-t ird of the rough rice marketed annually (Mullins et al. 1981). Milling of rough rice produces whole kernels and several by roducts: brokens, brewers, bran, millfeed, and hulls. Whole kernels (e.g., head rice yield) are the most important in terms of revenue. The value of rough rice should be directly related to its expected milling outturn, qualit characteristics, and the general level of rice prices (i.e., sup ly/ emand situation). Among the quality factors that grades are inten ed to measure are red rice, weed seeds, damaged kernels (including peck), off-color, chalk, and off-types of rice kernels. These factors should be negatively related to price, since they are all undesirable characteristics. Graders evaluate each lot of rou h rice before market day in a bid/acceptance market. The expectedgmillin yield and quality characteristics are estimated, and an appropriate U.S. stan ard grade is assigned to each lot 0f rough rice marketed by producers. Applicable U.S. standards associated with the several available grades for rough rice are specified in Table 1. Ins ectors generally adhere to these grade standards; however, across markets wit different inspectors, some inconsistencies may be observed in grading techniques (Brorsen et al. 1984). Rice buyers partici ating in bid/acceptance markets use the available information regarding ua it factors and grades and their own visual inspection of lot samples, as wel as the needs of their mill to determine their respective bid rice for a given lot.‘ The markets are open 1 day a week on an intermittent asis. Buyers make sealed bids which sellers have 24 hours to accept or reject. Lots on which bids are rejected are usually marketed at a later date either through the bid/acceptance market or a privately negotiated sale. Sales records, grade sheets, and confirmed prices were obtained from five bid/acceptance rice markets in Texas for the 1981-82, 1982-83, and 1983-84 marketing years. These records accounted for 24, 26, and 27 percent, respectively, of Texas production during the years surveyed. The markets are located at Alvin, Danbury, Ba Cit , El Campo, and Ganado. Data were available from two other mar ets but were not used because of missing quality values. Each market in the study is located in the western side of the Texas Rice Belt (Figure 1). Onl long grain rice varieties were included in the analysis. To determine t e extent of rice quality across the region, bid/acceptance data are augmented by American Rice, Incorporated (ARI) data for 1982-83 and 1983-84. PROCEDURE Rice qualit factors were analyzed to determine their effects on rough rice prices. The un erlying assumption of the statistical procedure used in this analysis is that goods are valued for their utility-bearing characteristics and prices of goods vary directly with the s ecific amounts of each characteristic the goods contain (Lucas 1975). The o served product prices are, thus, a composite of the value of the product's characteristics? The analytical method discussed in Appendix A was used to derive the discount associated with a one-unit change in a quality variable. These 1 Determination of whether the lot was to be milled as white or parboiled rice was difficult because some buyers were buying for rice mills that use both processes. Only one firm is 100 percent parboiled rice, while several firms are 100 percent white rice. Mill processing was not considered in this analysis. Damaged kernels, especially those damaged by the rice stink bug (peck), are amplified more in parboiled rice than in white rice. Thus, premiums or discounts observed in the bid/acceptance markets may not be the same for producers who sell their rough rice where there is no competition between buyers for the two types of milling processes. From the perspective of the rice producer selling in the bid/acceptance markets, however, such distinctions between type of buyer are irrelevant since he/she cannot control who bids on the rice; in fact, the only point of concern is that several pmential buyers bid in order to provide a competitive market atmosphere. A technical discussion of the analytical method used is presented in Appendix A. 1*‘ Table l. Federal grades and grade requirements for classes of rough rice o“. ~ Maximum limits of-- Seeds and heat- Percent damaged kernelsa chalky kernels Heat Percent damaged red rice kernels and In and objec- damaged medium Total tionable kernels In or (singly seeds (singly long short Percent or com— (singly or or com— grain grain other Color Grade bined) combined) bined)b rice rice typesc requirements U.S. No. l h 3 0.5 1.0 2.0 l.0 Shall be gray or creamy U.S. No. 2 7 5 l.5 2.0 b.0 2.0 May be slightly QFBY U.S. No. 3 l0 8 2.5 h.0 6.0 3.0 May be light QFEY U.S. No. h 27 22 h.0 6.0 8.0 5.0 May be gray or slightly rosy U.S. No. 5 37 32 6.0 l0.0 l0.0 l0.0 May be dark gray or rosy U.S. No. 6 75 75 l5.0d l5.0 l5.0 l0.0 May be dark gray or rosy U.S. Sample gradee Number in 500-gram sample. b Includes peck damaged kernels. C These limits do not apply to the class mixed rough rice. d Rice in grade U.S. No. 6 shall contain not more than 6.0 percent of damaged kernels. e U.S. sample grade shall be rough rice which: (a) does not meet the requirements for any of the grades from U.S. No. 1 to U.S. No. 6, inclusive; (b) contains more than 14.0 percent of moisture; (c) is musty, sour, or heating; (d) has any commercially objectionable foreign odor; or (e) is otherwise of distinctly low quality. Source: USDA 1977. discounts were multiplied by average Texas rice yields to measure the per acre impact of each quality variable. Data from ARI were grouped by regions and counties according to storage location (Figure 2). Each quality attribute was arrayed and summarized by county. Since ARI rough rice price is based on a pool rice and predetermined formula, t ese data were not included in t e price-qualit analysis; rather, their use was limited to demonstrating the prevalence o the respective quality factors across the Texas Rice Belt beyond the west side counties represented by the bid/acceptance markets. RICE QUALITY The more im ortant quality factors, such as whole kernel ield, brokens, eck, weed seed, eat damage, and test weight, were observeclin each id/acceptance market during the study period. The data set was incomplete, however, for red rice, smut (damage caused by fungus), chalk, reen rice, and a miscellaneous "other" category. If values for peck, weed see , heat damage, red rice, smut, and chalk were not recorded on the grade sheets for an individual lot, then values for these quality characteristics were assumed to be zero.3 Where test weight was not recorded on the grade sheet for an individual lot, the average test weight for that market and year was assumed. Green rice and the miscellaneous "other" category were not included in the analysis since data were not recorded on the grade sheets in all markets. If no settlement rice data were available for a given lot of rice, the observation was deleted betidre anal sis. Data were weighted by the quantity (i.e., pounds of rough rice) in each ot. Quality Variables The quality factors which can be controlled with production management decisions are peck, red rice, weed seeds, smut, and green or immature rice kernels. Each of these quality factors gives unpleasant appearance on the grocery shelf. Thus, rice millers try to remove them in t e milling process. This removal costs the rice miller additional processing expense and results in a loss in finished product volume. Rough rice prices are discounted to cover these additional costs. Peck refers to damage caused by stink bugs, among other reasons (Luh 1980; Stansel 1983). Stink bug damage is a discolored mark or fissure on the kernel which sometimes results in breakage, thereby reducing head yield. Stink bug damage can prevent normal development of a grain, resulting in field yield loss (Parker 1983). Field yield loss is not measured in the bid/acceptance market data. Current recommendations for control of the rice stink bug include the insecticides methyl parathion and carbaryl (Drees 1983) (Table 5.27).“ 3 The bid/acceptance market managers indicated this is the appropriate approach. Tables B.1-B.38 are found in Appendix B. Houston 4 2 Alvin )5 ' wanton Q) / Campy l‘ ® ' was“ Danbury Z A/ (9 MCKSON Q MATAGORDA Ganado Bay City WCTfiRIA ’ Q Figure i. Location of the bid/acceptance markets within the Texas rice production area. R Red rice, or rice with a red colored pericar , comes from an off-type rice that is considered a weed by rice producers an processors. lt usually enters a field through impure seed and then becomes a continual problem through volunteer reinfestation (USDA 1973). Complete removal of the red layer increases breakage of cultivated white rice, lowering milling ield (Smith 1981). Current control practices include crop rotations, removal of badly infested fields from rice production, and a reliable seed source (Eastin 1983). The variable for weed seeds is the number of seeds per 500-gram sample. The discounts associated with weed seeds are dependent on s ecies of weed (Parker 1983). Federal grades distinguish between noxious an other weeds (USDA 1977). In this study, however, it was only possible to consistently observe the total number of weed seeds. In addition to the market discounts for weed seeds, weed populations are negatively related to rice field ields (Smith 1968). Current control measures include rotation, removal of adly infected fields from rice production, and herbicide programs (Eastin 1983). Kernel smut is a disease occurring as black spores on the endosperm of rough rice and results in a grayish discoloration o milled rice (Leonard and Martin 1970). High nitrogen fertilization increases the incidence of smut. Most current long grain varieties are susceptible (Atkins and Marchetti 1979). Green rice refers to the percent of kernels that are immature. Usually the lighter immature kernels are removed at harvest or during drying, resultin in rice field yield loss (Parker 1983). The impact of this quality factor on market price is minor, since it is not reported on grade sheets for most markets. Quality factors affected by post-harvest management decisions should also concern producers. Rice is arvested when the moisture content of the grain is high (e.g., 13 to 20 percent) and then is dried to 13 ercent moisture efore storing. If the rice is dried too quickly, some whole ernel damage occurs during the millin process, i.e., hi h heat induced stress cracks contribute to additional reakage. The w ole kernel is the most valuable product to the miller. Thus, discounts occur for the additional breakage. Heat dama e is an internally generated chemical process during which the rice kerne changes color (e.g., from its normal white appearance to a light or dark tint) as a result of being stored at a hi h moisture content for an excessive amount of time. Proper post-harvest andling of rough rice lowers the amount of damage of these two types. A third category of quality factors involves those factors that are more affected by the environment than by pre- or post-harvest management decisions. Chalk refers to undeveloped or immature areas present in rice kernels which result in a chalky appearance. Long grain varieties currently grown in Texas are inherently translucent (Bollich 1983). Thus, chalk is not as much a problem as it has been in the past. Degree of chalkiness, however, is influenced by the environment; and therefore, any variety can show chalk (Hodges et al. 1979; Bollich 1983). The "other" category includes miscellaneous non-insect oriented kernel damage, i.e., water marks in the bran, sprouting, etc. (Parker 1983). Many of these damages are caused by lodging, a problem related to variety, high nitrogen fertilization, and weather conditions. Varietal selection and fertilization management help reduce these types of damage. However, Northwest C entra/ South R Figure 2. Texas Rice Belt production regions. adverse weather (hurricanes) is still the major causal factor. Comparison of Means Across Bid/Acceptance Markets During the 3 market years for which rice bid/acceptance market data were analyzed, mean settlement prices were highest in 1983-84 and lowest in 1982-83 (Tables B.1-B.13). The average mill price at time of settlement-was not the highest in 1983-84 but in 1981-82, indicating either margins and/or discounts may have decreased during the study eriod. In comparing the five markets across the 3 market years, the Alvin mar et had a higher mean final settlement price than the other four markets during each year of the stud period. Yet this market had a much lower percentage of grades 1 and 2 t an the other markets (Tables B.14-B.18). Overall, quality averaged about one grade lower at Alvin compared with quality at Bay Cit , Ganado, and El Campo markets. Nearness to the Houston milling center, differences in proportion of rice parboiled, and/or differences in grading procedures may have influenced price in this market. The proportion of grades 1 and 2 was much higher in the El Campo, Ganado, and Bay City markets (79.7 to 94.8 percent) than in the Alvin and Danbur markets (30.4 to 55.3 percent). Rice from the latter two areas was affected/more by the hurricane Alicia in 1983 than was rice moving through the three former markets. However, even in 1981-82 and 1982-83 the number of grades 1 and 2 was still appreciably lower in the Alvin and Danbury markets. Means (avera es) of quality data across the bid/acceptance markets and years indicate low evels of poor quality factors, exce t peck.5 Producer control measures for red rice, weed seeds, and post- arvest handlin -related damages appear to be relatively effective in maintaining comparativjy low levels of these undesirable attributes. Thus, stink bug damage appears to be the quality factor least effectively controlled by producers. Comparison of Quality Variables Across the Texas Rice Belt The bid/acceptance markets for which data were analyzed in this study are all located west of Houston. The quality levels observed in these areas may not be representative of the entire Texas Rice Belt. Therefore, quality data from throughout the Texas Rice Belt for 1982-83 and 1983-84 were collected from individual lots of rice marketed by ARI in an effort to investigate this concern. The ARI data were grouped into four regions (Figure 2) and statistics for the respective quality factors were calculated (Tables B.19—B.25). These data indicate damage from stink bugs was prevalent in all areas during both years, further confirmin the importance of stink bug damage represented in data for the rice markete through the bid/acceptance markets. ARI rice data also point to a slightly higher level of red rice east of 5 Although the reported means indicate small levels of poor quality factors, several lots of rice exhibited quality problems. Proportionally, these lots were a small part of the total. Discounts could be severe at the maximum level of quality factors occurring in each market (Tables B.1-B.13). Houston than in the other areas. The other quality factors do not appear to be different by regions, again, indicating ‘producers are effectively controlling most of the qua ity limiting actors (except or peck in all areas and red rice on the east side). Calculated statistics for ARI data indicate 36.6 percent of the 1982-83 and 1983-84 crops contained over 1.5 percent peck damage (Table 2). The bid/acceptance market data analyzed for al 3 market years indicate a sli htly higher proportion (49.4 percent) with over 1.5 percent eck damage (Ta le 2). Percent peck damage reported in the ARI data ranged rom 0 to 20.8 percent, and from 0 to 9.9 percent in the bid/acceptance market data. Means for the two sets of data are similar. The frequency tables for red rice and smut indicate very little rice in either data source above the 1.5 percent level (Table 2). Chalk shows a lar e concentration, between 1.5 and 2.0 ercent, in the bid/acceptance market ata. The ARI data does not, however, in icate chalk to be a problem. Rice with the number of weed seeds above U.S. Number 2 grade (greater than 5 per 500-gram sample) was relatively low in both data sets (Table 3). Modeling Quality Factors The quality factor levels of a given lot of rice affect the value. Buyers offer premiums for high quality rice while generally discounting poor quality rice. Before determinin the effects of quality factors on rough rice price it should be determined i market location affects the price offered for rough rice. Statistical tests indicate that market location made a difference in price received for rough rice in all years except 1981-82,'implyin a price advantage (or disadvantage) in one or more markets relative to t e ot er markets in the study (Table 5.34).“ The assumption that buyers discounted a given quality factor the same across all markets within a given year was also tested (Table B.34). These results indicate the following: 1) The effects of mill price, an indicator of rice supply/demand conditions, were not the same across markets in each year; 2) Buyers discounted red rice the same across all markets in a given year; 3) The discounts for heat damage were different across markets in each yean 4) Chalk was discounted the same in each market during a given year; 5) The premium/discount for broken kernels was the same across markets in 1981-82, but different in 1982-83 and 1983-84; 6) Buyers discounted for peck the same across markets in 1981-82, but differently across markets in 1982-83 and 1983-84; 6 This also implies that data cannot be combined across markets in 1982-83 and 1983-84 for the analysis. Table 2. Percent of selected quality factors in lots of rice in American Rice, Incorporated and bid/acceptance markets American Rice, lncorporateda Bid/acceptance marketsb Percent _ range Peck Red Smut Chalk Peck Red Smut Chalk 0 2.8 78.h 3h.h 0.0 6.5 73.0 h8.2 60.h 0.0l-0.50 lh.l l5.5 h7.l 93.8 6.7 l6.h 3l.0 0.0 0.5l-l.00 27.l 2.9 l0.2 5.0 l9.2 7.2 l3.3 0.2 l.0l-l.50 l9.h l.O 3.8 l.l l8.2 0.8 3.7 0.0 l.5l-2.00 l5.0 0.6 2.0 0.l 20.9 0.3 1.3 32.0 2.0l-3.00 l3.9 0.5 l.8 0.0 18.7 l.h l.O 6.7 >3.00 7.7 l.l 0.7 0.0 9.8 0.9 l.5 0.7 10 Includes 1982-83 and 1983-84 combined. Bid/acceptance markets located at Alvin, Danbury, Bay City, El Campo, and Ganado. Includes 1981-82, 1982-83, and 1983-84 combined. \ Table 3. Percent of weed seed in lots of rice in the American Rice, Incorporated and bid/acceptance markets Percent range American Rice, lncorporateda Bid/acceptanceb Number O 80.8 77.3 0.01-2.50 0.5 O.h 2.51-5.00 9.h 9.0 5.01-7.50 1.0 0.2 7.51-10.00 0.9 h.7 10.01-15.00 1.0 3.0 >15.00 6.h 5.h a Inciudes 1982-83 and 1983-84 combined. b Bid/acceptance markets located at Aivin, Danbury, Bay City, E1 Campo, Inciudes 1981-82, 1982-83, and 1983-84 combined. and Ganado. 11 7) Discounts for weed seed were different across markets during 1981-82 and 1982-83, but the same across markets in 1983-84; 8) Test weight discounts were different across markets during 1981-82 and 1982-83; however, buyers reacted similarly across markets for test weight during 1983-84; 9) Buyers discounted for smut differently across markets only in 1983-84; 10) The premiums for head yield were different across markets in 1981-82 and 1983-84, but the same across markets in 1982-83. The analytical method discussed in Ap endix A, restricted by the assumption tests discussed above, was use to estimate the premium/discount associated with a one-unit change in a quality variable. The results from these estimates are discussed below. Value of Quality Characteristics A hedonic rough rice price function is a regression of the formz7 PF é bo + b1M|LL + b2HEAD + b3BR0KENS + bhSEED + b5RED + b5PECK + b7SMUT + bgCHALK + b9HEAT + b]0TEST where PF is the observed final settlement rough rice price for a given lot of rice in the bid/acce tance markets, b1...,b10 are the premiums or discounts associated with each quality factor, MILL is the milled rice price in Houston during the week the rough rice was sold, and HEAD, BROKENS, SEED (weeds), RED, PECK, SMUT, CHALK, HEAT (damage), and TEST (wei ht) measure the level of the respective quality factors analyzed for each infiividual lot sold discussed earlier. The b0 value is different if the market location made a difference in price. Test results shown in Table B.34 indicate the b0 coefficient is different by market location in 1982-83 and 1983-84. The b1...,b10 values can be different across markets within a given year if rice buyers discount differently by market. Of the independent variables considered, only head yield and brokens are highly correlated (Table 8.35).‘; The estimated coefficients (i.e., remiums or discounts (dollars per hundredweight) for each quality varia le) for the price-quality relationships associatedwith each market year are presented in Tables 4-6. Even though the statistical tests discussed above indicated the coefficients were not the same across markets in some situations, the hedonic price-quality relationships across all markets are also included in Tables 4-6. Because the pooled 7 See Appendix A for a technical discussion of the analytical method used in estimating the hedonic rough rice price function. The rice industry commonly perceives a relationship between test weight and brokens. However, the bid/acceptance market data for each year showed no strong correlation between the two factors. 12 Table b. Estimated coefficients for the hedonic price equations for rough rice, l98l—82 (absolute t-ratios in parentheses)a Quality All variable Alvin Ganado Bay City marketsb S/cwt Intercept -9.h588 -9.h588 -9.h588 -9.2887 (11-56) ='< (h -56) >'< (h -56) ='= (h -51+) >'= Mill price 0.hh78 0.hh78 0.hh78 0.h5l5 (29-57) ='= (29 ~57) ='= (29 -57) ='= (31 -68) >’< Head yield 0.1381 0.0723 0.ll02 0.1229 (h.07)* (2.0l)* (2.82)* (3.92)* Brokens 0.0359 0.0359 0.0359 0.0519 (0-95) (0-95) (0-95) (1 1+7) Seed —0.007l —0.0l97 —0.0l97 —0.0l22 (3.0h)* (3.6h)* (h.h6)* (6.29)* Red rice -0.1716 -0.1716 -0.l7l6 -0.1831 (6-35)* (6-35)* (6-35)* (6-7h)* Peck -0.2897 -0.2897 -0.2897 -0.2858 (7 -86) ='= (7-86) ='= (7-86) ='= (8 -03) >'< Smut 0.0099 0.0099 0.0099 0.0027 (0.25)) (0.2l+) (0.211) (0.07) Chalk —0.lhh8 ~0.lhh8 -0.lhh8 0.0506 (0.82) (0.82) (0.82) l.l7 Heat damage —0.0037 0.0200 -0.1692 —0.0036 (-2.82)* (l.8h) (1.02) (2.6h)* Test weight 0.053h 0.1255 0.09b3 0.05h8 (1-19) (2.73)='= (1-96)=‘= (1-311) * Indicates rejection of the null hypothesis at the 5 percent level of significance. Rejection of the null hypothesis implies that the quality characteristic affected the rough rice settlement price by the estimated coefficient amount for each unit change in the quality factor. The covariance analysis across markets had R2 =0.8148 and F-ratio = 109.53; the critical F value is 1.66 (5 percent level of significance). Ordinary least squares analysis for all markets combined had R2 = 0.8006 and F-ratio = 183.09; the critical F value is 1.91 (5 percent level of significance). There were 467 observations in the data set. D All data pooled and coefficients estimated by ordinary least squares. 13 Table 5. Estimated coefficients for the hedonic price equations for rough rice, 1982-83 (absolute t—ratios in parentheses)a Quality A11 variable Alvin Danbury El Campo Ganado Bay City marketsb $/cwt Intercept 10.10b2 —2.9066 -8.5372 -3.1690 2.39hh -h.519b (l.79) (0.h8) (3.59)* (1.28) (0.26) (2.8h)* Mill price 0.h173 0.710h 0.6560 0.0h19 0.3h51 0.3127 (2 - 29) ='< (2-95) >'= (9-81) >'< (0-65) (6.811) ='= (9-21) ='= Head yield 0.0920 0.0920 0.0920 0.0920 0.0920 0.12h9 (5-32)* (5-32)* (5-32)* (5-32)* (5-3Z)* (7»53)* Brokens 0.1339 0.0272 0.039h 0.0539 0.0352 0.0759 (2.9l)>'< (0.91) (1.811) (2.6l+)='< (1.61) (l1.07)='< Seed -0.0259 —0.0015 —0.0083 c -0.0120 -0.0083 (l.h2) (0.26) (6.07)* (h.l5)* (6.55)* Red rice -0.2267 —0.2267 -0.2267 -0.2267 -0.2267 -0.1010 (h.ll)* (h.ll)* (h.ll)* (h.ll)* (h.ll)* (2.l8)* Peck -0.3676 —0.0895 —0.2l79 -0.0367 —0.1057 -0.0815 (2. 11+) >'= (1.28) (11.67) >'< (1.12) (11.67) ='< (11.67) ='= Smut -0.0002 —0.0002 —0.0002 —0I00O2 —0.0002 -0.0023 (0.01) (0.01) (0.01) (0.01) (0.01) (0.10) Chalk 0.1627 0.1627 0.1627 0.1627 0.1627 0.0187 (3-10) ='< (3-10) ='= (3-10)>'< (3-10) ='< (3-10) ='= (0-71) Heat damage 0.0052 —0.0357 0.0068 0.0321 c 0.0012 (0.39) (2 .70) ='= (0.511) (0.112) (0.16) Test weight -0.3299 -0.1537 0.006h 0.1230 —0.1113 0.0006 (2.711) ='< (1.76) (0.18) (3.11.)='= (0.56) (0.02) Indicates rejection of the null hypothesis at the 5 percent level of significance. Rejection of the null hypothesis implies that the quality characteristic affected the rough rice settlement price by the estimated coefficient amount for each unit change in the quality factor. The covariance analysis across markets had R2 = 0.5102 and F—ratio = 18.67; the critical F value is 1.50 (5 percent level of significance). Ordinary least squares analysis for all markets combined had R2 = 0.4002 and F-ratio = 44.76; the critical F value is 1.91 (5 percent level of significance). There were 682 observations in the data set. All data pooled and coefficients estimated by ordinary least squares. Data not reported. Table 6. Estimated coefficients for the hedonic price equations for rough rice, l983-8h (absolute t—ratios in parentheses)a Quality All variable Alvin Danbury El Campo Ganado Bay City marketsb $/cwt Intercept , 19.8329 5.1323 -2.0116 -u.o7n7 -h.l0h7 9.9906 (h-l2)* (0-56) (0-#7) (0-78) (0-85) (2.78)* Mill price -1.0276 -0.6588 0.2611 -0.0736 -0.2839 -ofl537h (5-75)>'< (1-53) (1.66) (0-37) (1-113) (5-7l)>'< Head yield 0.1393 0.2230 0.2203 0.1951 0.2620 0.2001 (3-30)* (4-73)* (5-82)* (5-23)* (6-72)* (5-55)* Brokens 0.1795 0.1795 0.1795 0.1795 0.1795 0.2007 (h.h0)* (h.h0)* (h.h0)* (h.hO)* U+.h0)* U+.59)* Seed -0.0077 -0.0077 -0.0077 "0.0077 -0.0077 -0.0070 (h.8l)* (h.8l)* (h.8l)* (h.8l)* U+.8l)* (3.97)* Red rice -0.1701 "0.1701 -0.1701 "0.1701 -0.1701 *0.2973 (2.l6)='< (2.l6)='< (2.l6)='= (2.l6)='< (2.l6)=’= (3.l+l)='= Peck -0.6572 "0.0313 -0.2650 -0.3805 ~0.1672 "0.3521 (7-13) i‘ (0-28) (3=11)"< (3-35) "= (21111) ‘i (3-20) >2’ Smut -3.3530 -1.3706. -0.1352 0.1620 “0.2Z86 -0.0757 (lO.l+6) =’< (3. l0) ='= (0.62) (0.711) (l .92) (£1.86) ='< Chalk 0.0501 0.0601 0.06h1 0.0691 0.0601 0.2113 (0 .60) (0 . 60) (0 . 60) (O . 60) (0 .60) (Li . l l) >'< Heat damage -0.0033 c c —0.0366 c -0.0113 »(1.29) (3.oo)* (h.i2)* Test weight 0.0773 0-0773 0.0773 0.077} 0.0773 "0.0%78 (1-51) (1-51) (1-51) (1-51) (1-51) (Q~95) * Indicates rejection of the null hypothesis at the 5 percent level of significance. Rejection of the null hypothesis implies that the quality characteristic affected the rough rice settlement price by the estimated coefficient amount for each unit change in the quality factor. ’ The covariance analysis across markets had R2 = 0.4451 and F—ratio = 21 81; the critical F value is 1.55 (5 percent level of significance). Ordinary least squares analysis for all markets combined had R2 = 0.2770 and F—ratio = 33.11; the critical F value is 1.91 (5 percent level of significance). There were 875 observations in the data set. All data pooled and coefficients estimated by ordinary least squares. Data not reported. M; U’! all-markets regression (ordinary least squares) imposes restrictions which may not be true (i.e., the coefficients are the same across markets) the results are sometimes quite different from the covariance analysis. Mill prices, a general indicator of the changes in supply/demand during the marketing year, impacted all markets in 1981-82, all mar ets except Ganado in 1982-83, and only the Alvin market in 1983-84. The general level of rice prices fell abruptly) during 1981-82, continued a modest decline in 1982-83, and eld relatively sta le in 1983-84. The relative stability in rice prices during 1983-84 could explain why mill prices were not im ortant during this season. A coefficient of 0.4478 for mill price in 1981 -82 in icates a $1/cwt change in the mill price during the year affected rou h rice prices $0.4478/cwt in the same direction. The negative coefficients or milled rice prices in 1983-84 for many markets were not expected. Whole kernel yield (head rice yield), the most important quality characteristic in terms of revenue derived from millin , was statistically significant in all markets and years. Whole kernel yiegd coefficients varied from 0.0723 in the Ganado market in 1981-82 to 0.2624 in the Bay City market in 1983-84, indicating that the value of a one unit increase in head yield (i.e., 1 lb whole kernels) varied from $0.0723 t0 $0.2624. Price of U.S. No. 2 milled long rain rice, free on board (FOB) mill, Houston, ranged from $17.75 to %22.00/cwt during the study period. The impact of broken rice uantity in a milled sample, the difference between total mill yield and whole kernel yield, was insignificant on rough rice prices at the bid/acceptance markets during 1981-82. However, the amount of rokens affected prices in the Alvin and Ganado markets during 1982-83 and in all markets during 1983-84. Discounts for weed seed occurred in all markets during 1981-82 and 1983-84. During 1982-83, however, this quality factor was significant only in the El Campo and Bay City markets. The discounts measured by dollars per hundredwei ht in rough rice price per one weed seed per 500-gram sample, varied from 0.0071 to $0.0197. Discounts durin 1983-84 were the same across markets, but nearer the lower portion of t e 1981-82 range. Red rice affected the rough rice price in all markets durin all years of the study period. Discounts measured by dollars per hundredweig t for a 1 percentage point increase in red rice in a sample, were the same across markets in a iven ear, but varied from $0.1701 in 1983-84 to $0.2267 in 1982-83. ln t e big/acceptance markets studied, avera e red rice levels were low. Some individual lots ran as high as 25 percent re rice, however, bringing sizeable price discounts, i.e., for a 25-percent level of red rice, the estimated discount in the rough rice price was $5.67/cwt during 1982-83. Peck damage, rimarily caused by stink bugs, affected rough rice prices in all markets during all years except at Danbury and Ganado during 1982-83 and at Danbury durin 1983-84. Discounts for peck were more variable than those for red rice, but t e average levels of these discounts were similar. Thus, a kernel with peck damage results in about the same discount as a kernel of red rice if all other factors are the same, i.e., in either case the kernel is undesirable. In the data analyzed, reported peck damage ranged from 0 to 9.9 16 percent. Buyers discounted rough rice prices appreciably in the upper levels of this range. Smut was not a significant quality factor exce t in the Alvin and Danbury markets during 1983-84. Hurricane Alicia moved t rough these areas in August 1983, creating quality problems with unharvested rice. Rice harvested west of these two market areas suffered less quality damage. Chalk, heat damage, and test weight impacted on rough rice price only on a limited basis. Indirect Effects of Peck ln addition to the direct discount associated with the visible kernel damage caused by stink bugs, there is also an indirect discount which is a result of the loss in whole kernel yield, increase in quantity of brokens, and decline in test weight caused by stink bug dama e. The same procedure used to estimate the hedonic price functions was use to estimate the indirect effects of peck damage on whole kernel yield, duality of brokens, and level of test weight. Market location was si nificant in a l situations, indicating the intercepts are not equal across mar ets (Table B.36). The coefficients for the impact of peck on head yield and brokens are not significantly different across markets during I981-82, indicating equal slopes across markets for those two relationships in 1981-82 (Table B.37). Head yield was negatively related to peck in all years and at all markets except Alvin durin 1982-83 and El Campo and Ganado durin 1983-84 (Tables 7-9). For each ad itional percentage point increase in peck, ecreases in head yield range from 0.5653 percentage points at the Bay Cit market to 2.4335 percentage points at the Alvin market during 1983-84. T is range in head yield declines translates into an indirect discount in rough rice prices ranging from $01483 to $0.3390/cwt for a one-unit increase in peck.9 The level of peck affected the level of brokens at all markets during 1981-82; at the El Campo, Ganado, and Bay City markets during 1982-83; and at the Alvin and Danbury markets during 1983-84 (Tables 7-9). This impact varied from 0.2977 to 1.9212 per unit of chan e in peck. Since the effect of peck on brokens is positive and the effect of rokens on rough rice rice is positive as shown in Tables 4-6, the indirect effect of peck (through rokens) on rough rice price will be positive. Brokens and head yield are inversely related, however. The net effect of an increase in peck damage is a decrease in rough rice price, since whole kernels are of greater value than broken kernels. Peck damage affected test weight at Alvin and Ganado markets during 1981-82; at the Danbury, El Campo, and Ganado markets during 1982-83; and at the Alvin, Danbury, El Campo, and Ganado markets during 1983-84 (Tables 7-9). Peck damage lowers the weight of the rain in a bushel of rice, since peck damaged individual kernels are lighter t an normal kernels for the same volume. Test weight was reduced from 0 to 0.6142 lb per each percenta e point increase in peck. Since a standard bushel of rough rice weighs 45 §b, a 9 The effect of peck on head yield (05653 and 2.4335) in Table 8 multiplied by the premium for each unit of head yield ($02624 and 50.1393) in Table 5 gives the indirect discount in rough rice price ($01483 and $03390). 17 i}? Table 7. in parentheses)a Impact of peck on selected quality variables at specified Texas rice bid/acceptance markets during l98l-82 (absolute t-ratios All Item Alvin Ganado Bay City marketsb Head yield: Intercept 60.h222 56.3193 57.9692 56.60hh (33@39)* (l3l-23)* (77-#9)* (l23-55)* Peck —i.l860 —l.l86O -l.l860 -0.5057 (SQMW (5.lll)='< (5-l+l)>'= (2-57)* Brokens: intercept l0.6é86 l3.0379 ll.9h23 l2.8603 (l7=2@)* (35-h9)* (l3-6h)* (3%-9l)* Peck 0.65%? O.65h7 O.65h7 O.25h3 6A9) is‘ (3-li9) ='< (3 - b9) >'= (l -55) Test weight: intercept %5.]308 h6.2lO3 h5.8756 h6.0282 (239-33)* (333-50)* (177-66)* (522-#6)* ?eck "0.3é@2 —O.5h56 —0.l653 -0.3857 {Essi-éli} s (6.810 ='< (l .63) (9.80) >'= * Indicates rejection of the null hypothesis at the 5 percent level of significance. Rejection of the null hypothesis implies that peck affected the quality characteristic by the amount of the estimated coefficient for each unit change in peck. all markets combined had R2 = The covariance analysis-across markets had R2 peck-head yield; R2 = 0.04 and F—ratio = 7.19 for peck—brokens; and R2 = 0.21 and F—ratiQ = 24.76 For peck-test weight. 0.01 and F-ratio = 0.01 and F~ratio = 2.3% for peck-brokens; and R2 0.10 and F—ratio = 16.44 for Ordinary least squares analysis for 6.61 for peck-head yield; R2 0.17 and F-ratio = 96.05 for peck—test weight._The critical F value is 3.92 (5 percent level of significance). There were A67 observations in the data set. D All data pooled and coefficients estimated by ordinary least squares. Table 8. parentheses)a Impact of peck on selected quality variables at specified Texas rice bid/acceptance markets during l982—83 (absolute t—ratios in All Item Alvin Danbury El Campo Ganado Bay City marketsb Head yield: Intercept 6l.96lh 62.60h9 59.69ll 57.0590 56.9703 58.1136 (#2-57)* (69-39)* (127-09)* (l7#-Z3)* (153-32)* (257-20)* Peck -l.3369 —l.8bh0 -l.836l —l.3330 —0.7h08 -0.8776 (1.78) (5-03)* (6-37)* (5-96)* (h-87)* (7-6l)* Brokens: Intercept lO.5203 9.l5l5 l0.03ll ll.5255 l2.29ll ll.2583 (8.lh)* (l0.60)* (2h.0h)* (39.62)* (38.h5)* (59.65)* Peck 0.lh69 0.7981 l.302h 0.9289 0.2977 0.5225 (0-22) (1-96) (5~09)* (h-67)* (2-20)* (5-h3)* Test weight: Intercept h5.3h79 h6.5073 .h6.l729 h5.6822 h5.h70l b5.6592 (l33.87)* (205.5l)* (h22.38)* (600.39)* (5h2.9l)* (870.66)* Peck -0.3227 -0.2789 -0.6lh2 —0.l365 0.0007 —0.l3h2 (l.85) (2.62)* (9.l6)* (2.62)* (0.02) (5.02)* * Indicates rejection of the null hypothesis at the 5 percent level of significance. Rejection of the null hypothesis implies that peck affected the quality characteristic by the amount of the estimated coefficient for each unit change in peck. The covariance analysis across markets had R2 = 0.26 and F-ratio = 27.28 for peck- head yield; R2 = 0.13 and F-ratio - 11.74 for peck-brokens; and R2 = 0.22 and F- ratio = 21.89 for peck-test weight. Ordinary least squares analysis for all markets combined had R2 = 0.08 and F-ratio = and F-ratio = 29.45 for peck—brokens; and R2 test weight. The critical F value is 3.92 (5 percent level of significance). There were 708 observations in the data set. b All data pooled and coefficients estimated by ordinary least squares. 57.98 for peck—head yield; R2 = 0.04 0.03 and F-ratio = 25.17 for peck- 19 Table 9. parentheses)a Impact of peck on selected quality variables at specified Texas rice bid/acceptance markets during 1983-8h (absolute t—ratios in _ All Item Alvin Danbury El Campo Ganado Bay City marketsb Head yield: Intercept 59.7797 60.7835 59.6286 57.5l9h 57.8h3h 59.2981 (73-2l)* (92-73)* (112-09)* (93-90)* (97-7l)* (213-6#)* Peck -2.h335 -l.l59h -0.6501 -0.2122 -0.5653 -l.3590 (7-29)* (3-2#)* (1-92) (0-#7) (2-ll)* (9-23)* Brokens: Intercept l0.682l 9.8308 l0.0537 ll.66l3 ll.3608 l0.3007 (l6.89)* (l8.lh)* (22.8h)* (23.0l)* (23.l9)* (h3.b3)* Peck l.92l2 0.8857 0.2100 —0.2565 0.23h2 l.0l9l (6.96)* (2.99)* (0.75) (0.68) (l.06) (8.lh)* Test weight: Intercept hh.9782 h6.2986 h6.3l60 h6.3757 h5.h690 h6.02l6 (331-98)* (398-7l)* (#91-l6)* (hh8-l5)* (#33-3l)* (815-35)* Peck -0.32lh —O.30hh -0.h922 -0.h979 0.0000 —0.hl9h (5.h3)* (h.80)* (8.20)* (6.h6)* (0.00) (lh.06)* Indicates rejection of the null hypothesis at the 5 percent level of significance. Rejection of the null hypothesis implies that peck affected the quality characteristic by the amount of the estimated coefficient for each unit change in peck. The covariance analysis across markets had R2 = head yield; R2 = 0.22 and F-ratio = 26.96 for peck-brokens; and R2 = 0.44 and F- ratio = 77.69 for peck-test weight. Ordinary least squares analysis for all markets combined had R2 = 0.09 and F-ratio = 86.07 for peck-head yield; R2 = 0.07 and F-ratio = 66.28 for peck-brokens; and R2 = 0.18 and F-ratio = 198.37 for peck- test weight. The critical F value is 3.92 (5 percent level of significance). There 0.18 and F-ratio = 21.67 for peck- were 889 observations in the data set. b All data pooled and coefficients estimated by ordinary least squares. one-unit increase in peck damage could lower test weight to 44.3858 lb/bu. Such a decrease in test weight translates to indirect discounts across markets and years in rou h rice price ranging from $0 to $00688 per percentage point of peck damage. ‘l Analysis of the relationshi between head yield-peck, brokens-peck, and test weight-peck with the ARI ata indicates similar coefficients and statistical significance levels. These results, presented in Table B.38, were used to verify the bid/acceptance market relationship. Peck Related Field Losses In addition to the implicit price discounting associated with peck damage, rice producers often suffer physical yield losses as a result of stink bug infestations. Stink bugs in the nymphal and adult sta es feed on rice as the panicle develops (Bowlin 1963). During the early mfik sta e of grain development, stink bu amage can prevent normal grain gevelopment, resulting in an empty ume or shriveled grain (i.e., yield loss). During the dough stage of grain Eevelopment, stink bu dama e weakens the grain structurally, resulting in breaka e during mil in an /or development of a black spot on the grain, both of whic contribute to ower quality. Bowling (1963) found that yields were decreased by 7.2 percent when stink bugs were maintained at four per square foot during grain forming, as compared to a no stink bug situation. Using Bowling’s test data and regressing percent peck damage against yield expressed as a percent of control yield (no stink bug) indicate a loss of 17.1 percent in field yield for each percentage point of peck damage." The equation estimated was: Yield (percent of control) = 103.3 - 17.0981 peck level (11.01) R2 = 0.98 where the figure in parenthesis is t-ratio. Swanson and Newsom (1962) reported Louisiana yield losses associated with stink bug damage in 1960 and 1961 cage tests. Their results were similar to those found by Bowling (1963). No recent studies were found relating to field loss as a result of stink bugs. Applying 1 percent peck damage to the estimated equation and assuming the Texas yie d with no peck damage at 5,500 lb, indicates a 759-lb field loss as a result of peck (Table 10). During crop years 1982-83 and 1983-84, ARI 10 The effect of peck on test weight (0 to 0.6142) in Tables 8 and 9 multiplied by the premium for each unit of test weight ($00773 to $00064) in Tables 5 and 6 produces the indirect discount in rough rice price ($0 to $00688). A word of caution: Bowling’s results recorded peck damage up to 0.59 percent with four bugs per square foot, about one-half of the mean damage found in the ARI and bid/acceptance market data analyzed in this study. Extrapolations beyond the test levels could be in error. 21 Table l0. Estimated field yields with selected levels of peck damagea Percent peck damage 0 0.5 l.0 l.5 2.0-Q lb/A 5,500 h,26h 3,879 3,b9h 3,llO 5,000 h,738 h,3l0 3,883 3,h55 5,500 5,2ll h,7hl h,27l 3,80l 6,000 5,685 5,172 h,659 h,lh6 a Estimated from Bowl1ng’s (1963) data: Yield (percent of control) = 103.3 - 17.0981 peck level (11.01) R2 = 0.98 t-value = ( ). reported a weight average peck dama e in Texas rice of 1.41 percent (Table B.33). With an average yie d of 45.48 undredweight/acre (cwt/A) across both crop years, Bowling’s data implies an average fieldg/ield loss due t0 stink bug damage of almost 12 cwt/A. With rough rice price between $7 and $11/cwt, field losses as a result of stink bug damage represent a sizeable cost to the producer. This discussion is extrapolating from 1963 data. Additional research is needed to determine the impact of stink bug numbers on rice quality and field loss with current varieties. Red Rice and Weed Related Field Losses Weeds reduce rice yields by competing for growth requirements. The competitive effects vary by rice variety, weed type, the environment, and relative time of emergence of weeds and rice (Diarra et al. forthcoming). As weed density increases, crop yields decrease. Smith (1968) found that rice field yields were decreased by 4-10 percent for one weed per 8 square feet and 19-40 percent for one weed per square foot. Variability depended upon weed type. Diarra et al. (forthcoming) indicate five red rice plants per square meter reduce rice yields 21-23 percent. No published research was found relating rice field loss to number of weed seed per 500-gram sam le of rough rice or percent red rice in a sam le of rough rice. Part of the problem of attemptin to relate these factors is t at red rice or weed seeds may shatter in the field efore harvest, may be partially cleaned from the rough rice during combining, or may not be harvested during hand harvesting of research plots. Nevertheless, red rice and/or weed seeds in a sample of rough rice indicates their presence in the field and a previous rice field yield loss. While no data are available to develop field loss estimates associated with either red rice or noxious weeds, it is important to recognize that the economic consequences of these quality factors are somewhat different, although related, to the losses associated with eck. It should be noted that the carryover effect of red rice and other wee s through germination, vegetative growth, and ro agation in subsequent years extends their potential impact (i.e., detrimenta ef ects) beyond a 1-year phenomenon as assumed with peck and stink bugs. The number of seed reproduced by each red rice or noxious weed plant suggests these phenomena are most likely geometric in nature through time, with some degree of miti ation occurring in association with normal cultural management practices suc as tillage, herbicide treatments, rotations, etc. A proposed method of accounting for the net degree of economic loss associated with an observed level of red rice and for weed seeds in a rough rice sample is: YL; (gib/il) - E(P;) N NL = [YL] (X) ' Pi] + Z = (1+r) l'l with y; = h; (Z) where 23 NL = net present value of economic losses (dollars per acre); YL1(x) = yield losses in current year associated with a sample level of x red rice or weed seed (pounds per acre); P, = current rough rice price (dollars per pound); h,-(Z) = function relating mitigation effects of tillage, herbicide treatment, rotations, etc. on level of red rice or weed seed infestation in year i; g;(h,(Z)) = function relating geometric explosion of red rice'or weed infestations in year i; YL,-(g;(h,-(Z))) = function relating yield losses associated with red rice or weed infestations in year i; E(P;) = expected rough rice price in year i (dollars per pound» r == discount rate (i.e., opportunity cost of capital) (percent); and N = length of planning horizon. Substantial research is needed to clarify the mathematical properties of the respective yield loss and red rice/weed perpetuative function. One conclusion that can be reached at this point, however, is that rice producers must reco nize the subsequent benefits of control treatments for red rice and/or wee s beyond the immediate year. Marginal Implicit Prices The estimated hedonic functions in Tables 4-6 and the indirect impacts of peck shown in Tables 7-9 describe the pricing structure for rough rice in id/acceptance markets in Texas. These data can be used to derive estimates of the premium or discount (dollars per hundredwei ht of rough rice) associated with a one-unit change in the quality variaqale." The discounts (per hundredwei ht and per acre) for peck, weed seed, red rice, chalk, heat damage, an smut are given in Tables 11-13. The discounts for peck damage (both direct and indirect) ranged from 12 The direct discount per unit for peck at the Alvin market at 0.2897 is taken directly from Table 4. All direct peck, weed seed, red rice. chalk, heat damage, and smut coefficients per hundredweight in Table 11 are taken directly from Table 4. The indirect discount for peck (whole kernel, brokens, and test weight) is calculated as follows: for 1981-82, the effect of peck on head yield (11860) in Table 7 multiplied by the premium for each unit of head yield ((11381 for Alvin) in Table 4 produces the indirect discount in rough rice price (0.1638) in Table 11. Other markets, years, and indirect effects are calculated similarly. 24 Table ll. at specified Texas rice bid/acceptance markets, l98l-82 Discounts per one unit change for selected quality attributes Quality All attribute Alvin Ganado Bay City markets Premium (+) or discount (') Per hundredweight ($) Peck direct -0.2897* -0.2897* "0.2897* -0.2858* whole kernel —0.l638* -0.0857* -0.l307* -0.0622* brokens 0.0235 0.0235 0.0235 0.0132 test weight —0.0l86 -0.0685* —0.0l56 —0.02ll total -0.hh86 -0.h20h -0.hl25 -0.3559 Weed seed -0.007l* -0.0l97* -0.0l97* '0.0l22* Red rice -0-l7l6* -0.l7l6* -0.l7l6* -0.l83l* Chalk -0.lhh8 ~0.lhh8 -0.lhh8 0.0506 Heat damage -0.0037* 0.0200 -0.1692 -0.0036 Smut 0.0099 0.0099 0.0099 0.0027 Premium (+) or discount (t) Per acre ($)a Peck direct l3.62>'< —l3.62='< -i3.62='< —13.h3>'= whole kernel -7.70% —b.O3* -6.lh* -2.93* brokens l.lO l.lO l.lO 0.62 test weight -0.87 -3.22* -0.73 -0.99 total 21.09 -19.76 -19.39 -16.73 Weed seed -0.33* -0.93* -0.93* -0.57* Red rice -8.07* -8.07* -8.07* —8.6l* Chalk -6.81 -6.81 -6.81 2.38 Heat damage -O.l7* O.9h -7.95 -O.l7* Smut O.h7 O.h7 O.h7 O.l3 a weighted by state yield in 1981 (47 cwt) (USDA 1984). * Coefficients are significant at 5 percent level. 25 Table 12. Discounts per one unit change for selected quality attributes at specified Texas rice bid/acceptance markets, 1982-83 Quality A11 attribute Alvin Danbury E1 Campo Ganado Bay City markets Premium (+) or discount (—) per hundredweight ($) Peck direct -0.3676 -0.0895 -0.2179 -0.0367 -0.1057* -0.0815* whole kernei -0.1230 -0.1696 -0.1689 -0.1226* -0.0682* -0.1096* brokens 0.0197 0.0217 0.0513 0.0501* 0.0105 0.0397* test weight 0.1065 0.0h29 -0.0039 -0.0168* -0.0001 -0.0001 total -0.36hh -0.19h5 -0.339h -0.1260 -0.1635 -0.1515 Weed seed -0.0259 -0.0015 -0.0083 b -0.0120* -0.0083* Red rice -0.2267 -0.2267 -0.2267 -0.2267* -0.2267* -0.1010* Chalk 0.1627 0.1627 0.1627 0.1627* 0.1627* 0.0187 Heat damage 0.0052 -0.0357 0.0068 0.0321 b 0.0012 Smut -0.0002 -0.0002 -0.0002 -0.0002 -0.0002 -0.0023 Premium (+) or discount (') per acre ($)a Peck » direct 17.2h* -h.20 -10.22* -1.72 -h.96* -3.82* whole kerne1 -5.77 -7.95* -7.92* '5.75* -3.20* -5.1h* brokens 0.92 1.02 2.h1 2.35* 0.99 1.86 test weight 5.99 2.01 -0.18 -0.79* -0.00 -0.00 total 17.09 -9.12 -15.55 -5.91 -7.67 -7.10 Weed seed -1.19 -0.07 —O.38* b —O.55* —0.39* Red rice -10.h1* -10.h1* -10.h1* -10.h1* 10.h1* -h.7h Chaik 7.h7* 7.h7* 7.h7* 7.h7* 7.h7* 0.88 Heat damage 2.2h -1.6h* 0.31 1.h7 b 0.06 Smut -0.01 -0.01 -0.01 -0.01 -0.01 -0.11 a Weighted by state yieid in 1982 (46.9 cwt) (USDA 1984). D Data not reported. * Coefficients are significant at 5 percent 1eve1. 26 Table l3. Discounts per one unit change for selected quality attributes at specified Texas rice bid/acceptance markets, l983-8h Quality All attribute Alvin Danbury El Campo Ganado Bay City markets Premium (+) or discount (-) per hundredweight (S) Peck direct -0.6572 -0.0313 -0.266b* -0.38h5* -0.l672* —0.352l* whole kernel -0.3390 -0.2585* -0.lh32 -0.0hlh -0.lh83* -0.277h* brokens 0.3449 0.l590* 0.0377 -0.0h60 0.0h20 0.2086* test weight -0.02b8 -0.0235 -0.0380 -0.0385 0.0000 0.0200 total -O.676l -0.l5h3 -0.4099 -0.5l0h -0.2735 -0.4009 weed seed -0.0077 -0.0077* -0.0077* -0.0077* -0.0077* —0.0070* Red rice -0.l70l -0.l70l* -0.l70l* -0.l70l* -0.l70l* -0.2973* Chalk 0.0641 0.0641 0.0641 0.0641 0.06hl 0.2ll3* Heat damage -0.0033 b b -0.0366* b -0.0ll3* Smut -3.3430 -l.3706* -0.l352 0.l620 -0.2286 -0.h757* Premium (+) or discount (-) per acre ($)a Peck direct -28.52% -l.36 -ll.56* -l6.69* -7.26* -l5.28* whole kernel -lh.7l* -ll.22* -6.2l -l.80 -6.44% -l2.0h* brokens l4.97* 6.90% l.6h -2.00 1.82 9.05* test weight -l.O8 -l.02 -l.65 -l.67 0.00 0.87 total -29.34 -6.70 -l7.78 -22.16 -ll.88 -l7.h0 Weed seed -0.33% -0.33* -0.33* -0.33* -0.33% -0.30* Red rice -7.38* -7.38* -7.38* -7.38% -7.38* -l2.90* Chalk 2.78 2.78 2.78 2.78 2.78 9.l7* Heat damage -0.lb b b -l.59* —0.h9* Smut lh5.09* -59.48% -5.87 7.03 -9.92 —20.65* a weighted by state yield in 1983 (43.4 cwt) (USDA 1984). D Data not reported. * Coefficients are significant at 5 percent level. 27 $0.4125 t0 $0.4486/cwt or $19.39 t0 $21.09/A in 1981-82 across markets.“ This range across markets was slightly lower during 1982-83. Discounts for peck across markets during 1983-84 were larger and more variable than for t e 2 previous years, rangin from $0.1543 to $0.6761/cwt or $6.70 to $29.34/A. The discounts for peck in icate a 1 percentage point reduction in peck damage could have raised the price received per hundredweight for rough riceby $0.1260 to $0.6761 across all markets and years ($5.91 to $29.34/A). i Applying the discount for peck to the average level of peck in each market per year indicates an average discount ranging from $0.12 to $1.40/cwt of rough rice or $5.63 to $60.76/A (Tables 14-16). Average peck damage across each market and year ranged from 0.9 to 2.6 percent. Individual lots in the bid/acceptance markets were reported, however, with up to 9.9 percent peck damage (i.e., $58 to $290/A). The data obtained from ARl indicate peck damage as high as 20.8 percent in 1982-83 (i.e., $123 to $355/A) (Table B.19). These levels of peck damage indicate sizeable discounts in the rough rice market. Discounts in the rough rice markets coupled with stink bug induced field losses point to sizeable losses in revenue where peck damage is a problem. The discount for one weed seed per 500-gram sample across markets and years averaged from $0.00 to $0.0259/cwt ($0.00 to $1.19/A) (Tables 11-13). Combinin the discounts per unit of weed seeds with the average level of weed see s reported by market and year shows discounts ranging from $0.00 to $0.13/cwt ($0.00 to $6.11/A) (Tables 14-16). The average number of weed seeds across markets and years ranged from 1.9 to 12.8/500-gram sam le, with most of the markets avera ing below the number of seeds permitted or U.S. No. 2 rice (i.e., seven). ln ividual lots ranged, however, from 0 to 550 weed seeds er 500-gram sample. The lots with hi h weed seed numbers brought sizeab e discounts in the markets and probab%y large reductions in rough rice field yields. The discount for red rice was relatively stable across the bid/acceptance markets for all years, ranging from $0.1701 to $0.2267/cwt ($7.38 to $10.41/A) (Tables 11-13). A plying the discount er unit of red rice to the average level of red rice in eac market and ear indicates discounts rangin from $0.00 to $0.17/cwt ($0.00 to $7.97/A) (Ta les 14-16). The average samp es for the bid/acceptance markets in this study all fell within the red rice uality requirements for U.S. No. 2 or better. The general levels of red rice in some of the areas served by the bid/acceptance markets in this study were so low that the data were not recorded. The highest average levels of red rice were in the Alvin area, though these levels were lower than U.S. No. 2 rice. The ARI data (as indicated earlier) shows more red rice present in the eastern portion of the Texas Rice Belt. The averages during 1982-83 and 1983-84, even for these areas, however, were better than the requirements for U.S. No. 2 rice. As previously discussed, the presence of red rice in the sample indicates a lowering of rice field yields as a result of competition from red rice. Discounts for smut ranged from $0.00 to $3.34/cwt ($0.00 to $145.09/A) 13 State average yield for each year was multiplied by the quality discount per hundredweight to derive discounts per acre. Texas rice yields average 4,700 lb, 4,790 lb, and 4,340 lb during 1981, 1982, and 1983, respectively (USDA 1984). 28 Table lb. butes at specified Texas rice bid/acceptance markets, l98l-82 Economic impacts at the means of selected quality attri- Quality All attribute Alvin Ganado Bay City markets Premium (+) or discount (') per hundredweight ($)a Peck direct -0.76* -0.35* -0.67* -0.53* whole kernel —O.h3* —O.lO* —0.30* —0.ll* brokens 0.06 0.03 0.05 0.02 test weight -0.05 -0.07* -0.l3 -0.0h total -l.l8 -0.h9 -1.05 -0.66 weed seed -o.o9>'= -o.ou=~< -o.13>'< -o.07>'< Red rice -0.l2* 0.00* -0.0l* -0.0h Chalk -0.31 0.00 ~0.29 0.05 Heat damage 0.00* 0.00 -0.0l -0.0l* Smut 0.0l 0.0l 0.00 0.00 Premium (+) or discount (') Per acre ($)b Peck direct *35.72* -l6.h5* -3l.h9* -2h.9l* whole kernel -20.2l* -h.70* -lh.l0* -5.l7* brokens 2.82 l.hl 2.35 O.9h test weight -2.35 -3.29* —6.ll —l.88 total -55.h6 -23.03 -h9.35 -31.02 Weed seed -h.Z3* -l.88* -6.ll* -3.29* Red rice -5.6h* 0.00* -O.70* -l.88* Chalk -15.57 0.00 -13.63 2.35 Heat damage 0.00* 0.00 0.h7 -O.h7* Smut O.h7 0.h7 0.00 0.00 a Discount per unit of quality variable (Table 11) of that quality variable in the specified market (Tables 8.1, B.8, b Weighted by state yield in 1981 (47 cwt) (USDA 1984). * Coefficients are significant at the 5 percent level. multiplied by average level and 8.11). 29 Table l5. Economic impacts at the means of selected quality attributes at specified Texas rice bid/acceptance markets, l982-83 ‘H Quality All attribute Alvin Danbury El Campo Ganado Bay City markets Premium (+) or discount (-) per hundredweight ($)a Peck direct -0.b8* -0.17 -0.29* -0.03 -0.20* -0.l2* whole kernel -0.Z3* -0.33* -0.Z2* -0.ll* -0.l3* -0.l7* brokens 0.0h 0.0h 0.07 0.0h* 0.02 0.06* test weight 0.20 -0.08 -0.01 -0.02* 0.00 0.00 total -0.67 -0.5h -0.h5 '0.lZ -0.31 -0.23 Weed seed -0.06 0.00 -0.05% c —0.02* -0.02* Red rice -0.l7* -0.0h* 0.00* 0.00* 0.00* -0.0l* Chalk 0.36* 0.35* 0.00* 0.00* 0.0Z* 0.0] Heat damage 0.00 0.00* 0.00 0.01 0.00 0.00 Smut 0.00 0.00 0.00 0.00 0.00 0.00 Premium (+) or discount (-) Per acre ($)b Peck direct -3l.89* -7.97 -l3.60* -l.hl -9.38* *5.63* whole kernel -l0.79* -l5.h8* -l0.3Z* -5.l6* -6.l0* -7.97* brokens l.88 l.88 3.28 l.88* O.9h 2.8l* test weight 9.38 -3.75 -0.h7 -0.9h* 0.00 0.00 total -3l.hZ -25.32 -Zl.l0 -5.63 ~lh.5k -10.79 Weed seed -2.81 0.00 —2.3h* c —O.9h* -O.9h* Red rice -7.97* -l.88* 0.00* 0.00* 0.00* -0.h7* Chalk lb.88* l6.hZ* 0.00* 0.00* 0.9b* 0.h7 Heat damage 0.00 0.00* 0.00* O.h7 0.00 0.00 Smut 0.00 0.00 0.00* 0.00 0.00 0.00 Discount per unit of quality variable (Table 12) multiplied by the average level of that quality variable in the specified market (Tables B.2, 8.4, B.6, B.9, and B.12). Weighted by state yield in 1982 (46.9 cwt) (USDA 1984). Data not reported. * Coefficients are significant at the 5 percent level. 30 Table l6. specified Texas rice bid/acceptance markets, i983-8h Economic impacts at the means of selected quality attributes at Quality All attribute Alvin Danbury El Campo Ganado Bay City markets Premium (+) or discount (") Per hundredweight (S)a Peck direct —l.36* -0.04 —0.36* —0.hh* -0.32* -0.55* whole kernel -0.70* -0.3l% -0.l9 -0.05 —0.l7* -0 44* brokens 0.7l* 0.l9* 0.05 0.05* 0.08 0.33* test weight -0.05 -0.03 -0.05 -0.04 0.00 0.03 total -l.h0 -0.l9 -0.55 -0.48 —0.hl -0.63 weed seed -0.0h* -0.04% -0.06* -0.02* -0.0h* -0.04% Red rice -0.03* -0.05* 0.00* 0.0l* 0.00* —0.02* Chalk 0.l5 0.l0 0.00 0.00 0.00 0 lh* Heat damage 0.00 c c 0.00* c -0 02* Smut —0.8h* -0.28% -0.0l 0.03 -0.05 -0 09* Premium (+) or discount (') Per acre ($)b Peck direct -59.02* -l.7h —l5.62* -l9.l0* —l3.89* -23.87* whole kernel —30.38* -l3.h5* -8.25 -2.17 —7.38* -l9.09* brokens 30.8l* 8.25* 2.l7 2.l7* 3.47 lh.32* test weight -2.17 —l.30 -2.l7 -l.7h 0.00 l.30 total -60.76 -8.25 -23.87 20.83 -l7.79 -27.34 Weed seed —l.7h* -l.7h* -2.60* —0.87* -l.8l* —l.7h* Red rice -l.30* -2.l7* 0.00* -0.h3* 0.00% -0.87% Chalk 6.5l h.3h 0.00 0.00 0.00 6.08* Heat damage 0.00 c c 0.00* c — 0.87* Smut —36.h6* —l2.l5* -0.h3 l.30 -2.17 —3.9l* a Discount per unit of quality variable (Table 13) multiplied by the average level of that quality variable in the specified market (Tables 8.3, 8.5, 8.7, 8.10, and B43). b Weighted by state yield in 1983 (43.4 cwt) (USDA 1984). C Data not reported. * Coefficients are significant at the 5 percent level. 31 (Tables 11-13), but were significant only in the Alvin and Danbury markets during 1983-84. Hurricane Alicia moved through these areas in August 1983, lowering quality of unharvested rice. Discounts per acre at the sample means were $12.15 (Danbury) and $36.46 (Alvin) (Tables 14-16). No quality problem with smut was detected in the other markets during the time period analyzed. Discounts for chalk and heat damage are presented in Tables 11-13. These two quality factors had little effect on rough rice rices in the ' bid/acceptance markets studied, however, as further ref ected in Tables 14-16. ECONOMIC QUALITY CONTROL MEASURES The range in magnitude of the per acre discounts presented in the previous section suggests the incidence and related cost of quality damage are not consistent across all markets analyzed during all years of the study. In several instances, the imputed discounts are at significant levels on a dollars per hundredweight and per acre basis, inferring concurrent control measures may be occurring at less than optimal levels, timing, and conditions. These results are highly supportive of a control program targeted at individual producers and locations rather than a broad all-encompassing effort. Regardless of the absolute magnitude of discounts across each market year, peck dama e is relatively more costly per hundredweight and per acre than red rice and weed seeds, at least with respect to price discount. Table B.27 summarizes recommended measures of insecticide control (Drees 1983). With respect to the control of rice stink bugs, the recommended practices range in cost from $3.70 to $7.78 per application. Texas Rice Belt producers are using from one to four applications of methyl parathion or from one to three applications of Sevin in their rice stink bug control program (Engbrock 1984); total control costs range from $3.70 to $14.80/A for methyl parathion to $7.78 to $23.34/A for Sevin.“ This is what producers are doing; but, what level of control maximizes net returns to the producer? The information required to analyze that question is twofold: 1) What are the benefits of control on a per quality unit basis?, and 2) What are the costs of control on a per quality unit basis? Tables 14-16 identify the average potential total dollar returns associated with complete control of the elements responsible for poor quality attributes in rough rice. Such absolute control is generally recognized as not economically feasible, i.e., diminishing production of additional control inputs commences at some point, resulting in marginal costs of additional control eventually exceeding-the associated marginal returns. Tables 11-13 indicate the mar inal discounts associated with a one-unit change in the respective quality attri utes, i.e., marginal returns associated with controlling the responsible elements such that the attribute is reduced ( eck, weed seeds, red rice, chalk, smut, and heat damage) or increased (whole ernel yield, total milling yield, and test weight) by one unit. 14 Field records during 1979-84 for Wharton County producers indicate 74 percent of the acreage received two or more applications of insecticides (Gerlow 1985). Methyl parathion was the major control measure used. 32 Economic decisions regarding control measures affectin the quality attributes discussed here should be evaluated on the basis otgtheir cost per marginal unit of control relative to the associated marginal returns such as those identified in Tables 11-13. There is no information currently available regarding the cost of control on a per quality unit basis. Additional research in this area by Texas Agricultural Experiment Station and Texas Agricultural Extension Service scientists is underway in the Texas Rice Belt. LIMITATIONS OF THE STUDY Several important limitations of this study should be kept in mind. Discounts estimated in this analysis are only for the bid/acce tance markets indicated. Limited data were found measuring the impact o quality variables on field yield. Additionally, these data were taken during the early 1960s. Yield losses associated with the factors contributing to poor quality attributes in rou h rice samples represent additional significant costs to producers. These osses should be considered when an attempt is made to identify economic control measures (Bowling 1963; Eastin 1983; Stansel 1983). When the number of bidders is low and bidders have close contacts with each other, the markets may not be operating competitively. The number of bidders is generally lower on poorer quality rice; thus, discounts associated with lower quality rice may be partly a result of the thin market existing in the rice industry. Nevertheless, this does not bias the results reported here, since these discounts are real for the five bid/acceptance markets analyzed. SUMMARY AND CONCLUSIONS This paper reports results of analyses of 1981-82, 1982-83, and 1983-84 data from five rough rice bid/acceptance markets in the Texas Rice Belt. The objective was to determine the premium/discounts associated with various rou h rice quality factors. These five markets are located on the western side of t e Texas Rice Belt. To determine the extent of rice quality problems across thej region, quality data were obtained on rice marketed by ARI for 1982-83 an 1983-84. Quality factors, whole kernel yield, brokens, peck, red rice, weed seed, smut, chalk, heat damage, and test weight were analyzed to determine their impact on rough rice price. The premium per unit of whole kernel yield varied from $0.0723 at Ganado during 1981-82 to $0.2624 at Bay City durin 1983-84. The premium er unit of brokens averaged $0.1795 in each market uring 1983-84. Tota discounts per unit of peck varied from $0.4125 to $0.4486 during 1981-82. The range across markets was slightly lower during 1982-83. Peck discounts during 1983-84 were larger and more variable, however, than durin the 2 previous years. Discounts in the rough rice markets coupled with stink bug-induced field losses point to sizeable losses in revenue where peck damage is a problem. Discounts per unit of red rice were relatively stable across the bid/acceptance markets for all years, ranging from $0.1701 to $0.2267. The 33 occurrence of red rice was low, however, in the bid/acceptance markets. ARI data show more red rice being present in samples of red rice grown in the eastern portion of the Texas Rice Belt. Presence of red rice in a sample also indicates a lowering of rice field yields as a result of competition from red rice. The discount per weed seed in a 500- ram sample varied from $00071 to $00197. The average number of weed seegs per sample across markets and years ranged from 1.9 to 12.8/500-gram sample, with most of the markets averaging below the number of seeds permitted for U.S. No. 2 rice. Lots with high weed seed numbers brought sizeable discounts in the markets and also large reductions in rough rice field yields. Discounts for smut were only si nificant in the Alvin and Danbury markets during 1983-84. Hurricane A icia moved through these areas in August 1983. Discounts for chalk and heat damage had little effect on rough rice prices in the bid/acceptance markets studied. Depending on the costs associated with controlling the respective (‘uality characteristics, rice producers may be experiencing significant economic osses as a result of the price discounts associated with peck, red rice, weed seed, chalk, heat damage, and smut, among other quality attributes. Additional research is required to identify the aggregate impact of yield losses associated with several factors contributing to poor rough rice uality and identify the appropriate economic levels of control which affect t e specific quality attributes of rou h rice. This will require research by entomologists and economists on e ficient use of various stink bug control tactics and impact of stink bu level on field yields, peck damage, and milling characteristics of damage rice. 34 REFERENCES Atkins, John G. and Marco A. Marchetti. Rice Diseases, United States De artment of Agriculture, Science and Education Administration, Farmers Bul etin No. 2120, rev. April 1979. Bollich, Charles. United States Department of Agriculture, Beaumont, Texas, Personal Communication, 1983. Bowling, C.C. "Cage Tests to Evaluate Stink bug Damage to Rice." Journal of Economics Entomology. 56-2(1963):197-200. Brorsen, B. W. "A Study of the Efficiency and Dynamics of Rice Prices." Unpublished Ph. D. Dissertation, Texas A&M University, May 1983. Brorsen, B. W., W. R. Grant, and M. E. Rister. "Economic Values of Rice Quality Factors," Texas Agricultural Experiment Station, PR-4202, June 1984. Deaton, Angus and John Muellbauer. Economics and Consumer Behavior. New York: Cambridge University Press, 1980. Diarra, Amadou, Roy J. Smith, Jr., and Ronald E. Talbert. "Inference of Red Rice with Rice," Weed Science (forthcoming). Drees, Bastiaan M., Rice lnsect Management, Texas Agricultural Extension Service, 1983. Eastin, Ford. Texas Agricultural Experiment Station, Beaumont, Texas, Personal Communication, 1983. Engbrock, James E., Texas Agricultural Extension Service, Bay City, Texas, Personal Communication, 1984. Ethridge, Don E. and Bob Davis. "Hedonic Price Estimation for Commodities: An Application to Cotton." Western Journal of Agricultural Economics, 7(1982):293-300. Freund, Rudolf J., and Ramon C. Littell, SAS for Linear Mode/s: A Guide to the ANOVA and GLM Procedures. Cary, North Carolina: SAS Institute, lnc., 1981. Gerlow, Arthur, Texas Agricultural Extension Service, Bryan, Texas, Personal Communication, 1985. Griliches, Zvi. "Introduction: Hedonic Prices Revisited." Price Indexes and Quality Change. Zvi Griliches, ed., Cambridge, Massachusetts: Harvard University Press, 1971. Hodges, R.J., C.N. Bollich, M.A. Marchetti, and B.D. Webb, "Rice Varieties." Texas Agricultural Extension Service, B-1239, 1979. Ladd, George W. and Marvin B. Martin. "Prices and Demands for Input Characteristics." American Journal of Agricultural Economics. 58(1976):21-30. 35 Lancaster, K. J. Consumer Demand: A New Approach. New York, 1971. Leonard, Warren H. and John H. Martin. Cereal crops. New York: The Macmillan Company, 1970. Lucas, Robert E. B. "Hedonic Price Functions." Economic lnquiries. 13(1975):157-178. Luh, Bor S. Rice: Production and Utilization, Westport, Connecticut: AVI Publishing Company, lnc., 1980. Martinez, Adolfo, Harlan Traylor, and Lonnie L. Fielder. "Analysis of Quality and Non-Quality Factors on Prices of Medium and Long Grain Rough Rice in Louisiana." Louisiana State University, D.A.E. Research Report No. 507, Sept1976. Mullins, Troy, Warren R. Grant, and Ronald D. Krenz. "Rice Production Practices and Costs in Major U.S. Rice Areas, 1979." Arkansas Agricultural Experiment Station Bulletin No. 851, Mar. 1981. Parker, Melvin. Rice Belt Warehouses, lnc., El Campo, Texas, Personal Communications, I983. Rosen, Sherwin. "Hedonic Prices and Implicit Market Product Differentiation in Pure Competition." Journal of Political Economy. 82(1974):34-55. Smith, Roy J., Jr. "Control of Red Rice (Oryza sativa) in Water Seed Rice 0. sativa," Weed Science, 29(1981):663-666. . "Weed Competition in Rice," Weed Science, 1 6-2(1968):252-255. Stansel, James W. Texas Agricultural Experiment Station, Beaumont, Texas, Personal Communication, 1983. Swanson, M.C. and L.D. Newsom. "Effect of Infestation by the Stink bug Oeba/us pugnax, on Yield and Quality in Rice," Journal of Economic Entomology, 55(1962):877-79. United States Department of Agriculture. Crop Production, 7983 Annual Summary. SRS, Washington, D.C. Jan. 1984. . Rice in the United States: Varieties and Production ARS, Washington, D.C., Ag. Handbook No. 289, Revised June,1973. . Rice Market News. AMS, Little Rock, selected issues. . United States Standards for Rough Rice, Brown Rice for Processing, and Mi//ed Rice. FCIS, Washington, D.C. Aug. 1977. 36 APPENDIX A HEDONIC PRICING Hedonic price functions are regressions of the form (Lucas 1975): (i) P; = P(V;],...,V;_j;u;) where P,- is the observed price of commodity i, Vi] measures the amount of some "intrinsic uality" j per unit of commodit i, and u; is a disturbance term. This type of model can be derived from a non- inear programming model (Lancaster 1971; Lucas 1975; Ladd and Martin 1976). This type of interpretation implies a linear model (1). Most researchers, however, have used a semi-logarithmic relationship between prices and characteristics (Griliches 1971). The analysis discussed here is performed using a linear specification of (1). Ty ically, estimated hedonic price functions identify neither demand nor supply unctions (Rosen 1974). Both observed prices and implicit prices of embodied attributes may be affected by aggregate demand/suppl conditions. The implied values of an embodied quality attribute may not be t e same across marketing years and may also vary with the specific market (location) being analyzed. The previous discussion is appropriate for the analysis of cross-section data. The data sets used in this study, however, are pooled time-series/cross-section data. The hedonic estimation technique must, therefore, be adusted for differences in market forces over time. Ethridge and Davis (1982) an Martinez et al. (1976) accounted for temporal price changes by including some combination of linear and quadratic time trends and dummy variables for month or year in the model. Deaton and Muellbauer (1980) suggest using some type of an index variable and propose the following semi-logarithmic model: (2) ln(P;t) = ln(lt) + f(V;],...,V;J'; ug) where l. is the price of some reference commodity that can serve as a measure of the general price level. Since no weekly farm price is available for rice in the study area, the Texas weekly mill price is used in this analysis as the index variable (USDA). Farm and mill prices move very closely together (Brorsen 1983). No a pr/ori information is available regarding whether the farm/mill margin is an absolute markup, a constant percentage, or some combination of both. For this study, equation (2) implies the margin between farm and mill prices is a constant percentage. 37 ln this study a linear specification is used and the mill price is included as one of the regressors.“ The uality factors, thus, can be inter reted as discounts or premiums from t e base rice. The question stil remains as to how data should be analyzed under a bid/acceptance system as exists in rou h rice markets. Martinez et al. (1976) discarded the observations where the bi was not accepted; i.e., they assumed such observations were not reflective of an effective market. The bid price, however, can be viewed as representing existing demands. A considerable amount of information is eliminated if the observations are discarded where the bid was reected. The bid price represents the highest price any partici ating bi der is willin to pay for a given lot of rice on a given day within t e constraints of the id/acceptance market. In Brorsen et al. 1984 demand was estimated by including the highest bid price, regardless of whether the bid was accepted, as the dependent variable. However, in this analysis, the final settlement price for each lot of rice was obtained and used as the dependent variable. Only settlement prices are used, since some high bid prices are not serious bids. The discounts associated with quality are still expected to vary from year to year depending on aggregate supply and demand. The data consisted of a cross section of observations for a given sale. However, the number of cross-sectional observations was not equal across markets or time periods. Separate coefficients could have been estimated for each sale and market, but the larger number of coefficients estimated would make the interpretation of the results difficult. Additionally, the limited number of individual lots sold during some of the sales would restrict the quality variables analyzed. As an alternative, the cross-sectional data for each sale were pooled for the crop years, resulting in the estimated hedonic price function for each crop year. Hypotheses that the intercept and slope coefficients were the same across markets were tested using the pooled crop year data.“ Analysis of covariance was used to test these hypotheses (Freund and Littell 1981). Thus, the resulting model is: (3) Pimtk = alk +n§§ anknn + bmkP?ill tjé, cjmkvjimtk + Uimtk with i = 1,...,|mtk; m = l,...,Nk; t = l,...,52; and k = 1, 2, and 3, where Pimtk is the settlement price for lot numbenrii in market m during week t of year k, Dn is a dummy variable for market, Ptk is the milled rice price in 15 The results in this study are similar regardless of whether a linear or semi-logarithmic specification is used. The linear model was selected partly because of its theoretical interpretation and ease of explanation to members of the rice industry. Tests using the Box-Cox transformation also showed the linear specification to be more appropriate. The functional form of the model for each market and year were tested using the Box-Cox transformation. Results indicate a linear model was appropriate in all cases (Table 8.26). 38 week t of year k, V is quality factor j for lot number i in market m during week t of year k, the a’s, b’s and c’s are parameters to be estimated, lmtk is the number of lots sold in market m during week t of year k, and Nk is the number of markets for which data was analyzed from year k. Three markets are included for 1981-82 (N1 = 3) and five markets are included for 1982-83 and 1983-84 (N2 = 5 and N3 = 5). The model provides a framework for testing whether the slopes of the quality variables and the intercept term are the same across markets. 39 APPENDIX B Table B.1. Weighted means, standard deviations, and ranges of selected variables at Alvin, 1981-82 Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 170 Head yield 2 57.33h5 6.212h 19.00 66.00 Mi11 yield % 69.70h6 1.6669 62.00 72.00 Grade no. 2.9hhh 1.2865 2.00 7.00 Seed no. 12.8h59 28.98h8 0.00 2h5.00 Peck % 2.6138 1.h023 0.80 8.60 Red rice % 0.71h6 2.3135 0.00 25.00 Smut % 0.69h5 1.0315 0.00 5.00 Chaik 2 2.1513 0.3632 2.00 h.00 Stack % 0.8325 b.6525 0.00 33.00 Test weight 1b/bu hh.8128 1.1001 h1.00 h7.00 Settlement price S/cwt 10.8078 1.8327 5.25 13.63 Mill price at sale S/cwt 211.0559 2.10M 20.00 26.00 40 Table B.2. Weighted means, standard deviations, and ranges of selected variables at Alvin, l982-83 Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 58 Head yield 2 59.9730 2.1988 55.00 62.00 N111 yield % 70.2668 1.2913 7.00 72.00 Grade no. 2.59h8 0.7952 2.00 7.00 Seed n0. 2.2357 9.2017 0.00 30.00 Peck % 1.8613 0.5392 0.80 3.70 Red rice % 0.7713 1.2229 0.00 7.70 Smut % 0.5968 0.9513 0.00 3.70 Chalk % 2.2121 0.9599 2.00 5.00 Stack % 0.2898 0.9881 0.00 5.00 Test weight lb/bu hh.7h7l 0.6h88 hl.50 h6.00 Settlement price S/cwt 9.3789 0.3973 6.80 l0.l3 Mill price at sale S/cwt l8.5690 0.5537 l7.75 20.00 41 Table B.3. Weighted means, standard deviations, and ranges of selected variables at Alvin, l983-8h Standard Variable Unit Mean Deviation Minimum 7 Maximum Observation no. l59 Head yield % 5h.6627 6.2365 25.00 63.00 Mill yield % 69.3692 l.h7h2 60.00 72.00 Grade no. 2.956h 0.97l3 l.00 7.00 Seed no. h.779h lh.0l8h 0.00 l85.00 Peck % 2.07l6 0.9186 0.60 8.00 Red rice % 0.l96h 0.8l07 0.00 ll.00 Smut % 0.25l0 0.2376 0.00 l.l0 Chalk % 2.h050 0.57lh 2.00 h.00 Stack % 0.l836 0.856l 0.00 5.50 Test weight lb/bu hh.3857 0.823h h0.00 h5.50 Settlement price S/cwt ll.2785 2.3257 6.55 l7.l6 Mill price at sale S/cwt l9.705l 0.h367 l9.00 20.25 42 Table B.h. Weighted means, standard deviations, and ranges of selected variables at Danbury, 1982-83 Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 66 Head yield 8 58.9hlh 3.h503 h6.00 6h.00 Mill yield % 69.7316 1.2713 66.00 71.00 Grade no. 2.7713 1.1860 2.00 7.00 Seed n0. 3.3188 9.9695 0.00 77.00 Peck % 1.9281 0.8155 0.80 3.80 Other damage % 0.0051 0.0315 0.00 0.20 Color grade 2.795h 7.1352 0.00 h6.00 Red rice % 0.1836 0.32h0 0.00 2.20 Smut 8 0.9718 1.8032 0.00 12.00 Green % 0.1760 0.8123 0.00 5.00 Chalk % 2.1772 0.6689 0.00 h.00 Stack % 0.07lh 0.h030 0.00 h.h0 Test weight lb/bu h5.9h86 .0.7383 b3.00 h7.00 Moisture % 12.266h 0.5b10 11.10 1h.h0 Settlement price $/cwt 9.3332 0.6827 5.85 10.50 11111 price at sale S/cwt 18.8788 0.3289 18.00 19.00 43 Table B.5. Weighted means, standard deviations, and ranges of selected variables at Danbury, l983-8h Standard Variable Unit Mean Deviation Minimum . Maximum Observation no. 72 Head yield % 59.39l9 3.758h 30.00 6h.00 Mill yield % 70.2858 l.2063 59.00 72.00 Grade no. 3.1736 1.6355 2.00 7.00 Seed no. b.7689 l6.22l9 0.00 250.00 Peck % l.2002 l.38hh 0.00 9.00 Color grade 0.l952 0.3963 0.00 l.00 Red rice % 0.2997 0.66lh 0.00 3.70 Smut % 0.20h9 0.3250 0.00 l.l0 Chalk % l.56h6 0.972l 0.00 b.O0 Moisture % l2.l763 0.5200 l0.h0 lh.20 Test weight lb/bu h5.9332 l.03l2 hl.00 h8.00 Odor % 0.l258 0.33l7 0.00 l.00 Bugs % 0.l758 0.3806 0.00 1.00 Grass % 0.029l 0.l682 0.00 1.00 Mud % O.l305 0.3368 0.00 l.00 Shelled % b.8235 l5.2636 0.00 60.00 Sprout % 0.07l2 0.257l 0.00 l.00 Green % 0.l375 0.3hhh 0.00 l.00 Settlement price S/cwt l0.h290 0.98l3 6.83 ll.56 Mill price at sale $/cwt 20.0069 0.28h5 l9.00 20.25 Table B.6. Weighted means, standard deviations, and ranges of selected variables at El Campo, l982—83 Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 22h Head yield % 57.2611 9.3038 35.00 66.00 M111 yield % 69.0092 1.2317 61.00 72.00 Grade no. l.8237 l.356l l.00 7.00 Peck 2 1.3165 0.9185 0.10 h.50 Other damage % 0.8902 1.9059 0.00 9.50 Red rice 8 0.0189 0.0997 0.00 0.80 Total damage % 2.20l0 2.l0h0 0.10 9.80 Smut 8 0.3782 0.8169 0.00 5.00 Moisture % ll.2l9l 0.8589 '8.90 l6.60 Test weight lb/bu 95.3586 1.2578 50.00 98.00 Seed HO. 5.6891 27.6967 0.00 350.00 Stack % 0.0099 0.2992 0.00 9.00 Settlement price S/cwt 9.2830 0.7253 6.05 l0.87 Mill price at sale S/cwt l8.5507 0.565l l7.75 20.00 45 Table 8.7. Weighted means, standard deviations, and ranges of selected variables at El Campo, 1983-8h Standard Variable Unit Mean Deviation Minimum 9 Maximum Observation no. 311 Head yield % 58.7606 3.6687 36.00 66.00 M111 yield % 69.0965 1.16h9 61.00 71.00 Grade no. 1.9859 1.h1h6 1.00 7.00 Peck 2 1.3361 0.8253 0.00 5.50 Other damage % 1.1771 l.5h67 0.00 9.00 Total damage % 2.6llh 2.223h 0.20 12.00 Smut % 0.0870 0.3097 0.00 3.30 Test weight lb/bu h5.6593 0.9860 91.00 h8.00 Moisture % 11.5665 0.6560 10.00 12.60 Seed no. 8.0111 33.3190 0.00 550.00 Settlement price S/cwt l0.7l96 0.7369 7.00 ll.h5 Mill price at sale S/cwt l9.82h5 0.hl26 19.00 20.25 46 ~\ Table B.8. Weighted means, standard deviations, and ranges of selected variables at Ganado, l98l—82 Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 208 Head yield % 5h.8739 k.9h22 h0.00 66.00 M111 yield % 68.7099 1.60h5 62.00 72.00 Grade HO. 1.6h18 0.997h 1.00 7.00 Seed n0. 1.8613 8.83h3 0.00 205.00 Peck % 1.2191 0.8593 0.00 h.00 Red rice % 0-0063 0.0568 0.00 0.80 Smut % 0.5882 0.9706 0.00 7.20 Stack % 0.1690 0.h733 0.00 h.60 Test weight lb/bu h5.5h52 1.2238 h0.50 57.50 Moisture % 11.2585 0.6h99 9.10 13.60 Other damage % l.0002 l.3lb7 0.00 6.50 Settlement price S/cwt l0.207l l.h226 h.95 l3.8O Mill price at sale $/cwt 2l.92l9 2.3267 l9.00 27.00 \. 47 Table 8.9. Weighted means, standard deviations, and ranges of selected ‘Ea variables at Ganado, l982-83 Standard 4 Variable Unit Mean Deviation Minimum _i Maximum Observation no. l90 Head yield % 55.86l2 h.3728 36.00 65.00 Mill yield % 68.22l3 l.6h90 52.00 7l.00 Grade no. l.532h 0.98h3 l.00 7.00 Seed no. 0.0000 0.0000 0.00 0.00 Peck % 0.8965 l.l5hh 0.00 7.00 Red rice % 0.0lll 0.0723 0.00 0.60 Smut % 0.ll36 0.2655 0.00 l.50 Stack 2 0.036h 0.h665 0.00 6.00 Test weight lb/bu h5.5598 0.9801 h0.00 h7.50 Moisture % ll.5932 0.36l0 l0.50 l3.00 Other damage % 0.3277 0.93h7 0.00 6.20 Settlement price S/cwt 8.97h3 0.5770 3.50 l0.50 Mill price at sale $/cwt l8.h895 0.557h l7.75 l9.00 48 \~ Table B.lO. Weighted means, standard deviations, and ranges of selected variables at Ganado, l983-8h Standard Variable Unit Mean Deviation Minimum Maximum Observation no. l53 Head yield 2 57.2987 h.07h0 h5.00 66.00 Mill yield % 68.6385 l.h52l 63.00 7l.00 Grade no. l.82hh l.02h7 l.00 6.00 Peck % l.l3hh 0.7hl8 0.00 h.50 Other damage % 0.99h8 l.5l75 0.00 7.50 Red rice % 0.0705 0.3080 0.00 h.20 Smut % 0.l763 0.35h0 0.00 l.50 Test weight lb/bu h5.8l20 0.8588 h2.00 h7.00 Seed no. 2.5863 8.7358 0.00 75.00 Stack % 0.0937 0.6365 0.00 6.50 Settlement price $/cwt l0.7l8h 0.8057 7.50 ll.60 Mill price at sale S/cwt l9.8650 0.38h9 l9.00 20.25 \\ 49 Table B.11. Weighted means, standard deviations, and ranges of selected variables at Bay City, 1981-82 Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 92 Head yield % 55.2580 5.7629 25.00 62.00 N111 yield 2 68.6880 1.1022 58.00 70.00 Grade no. 2.15h6 0.6713 1.00 6.00 Seed n0. 6.7961 19.9508 0.00 275.00 Peck % 2.2960 1.1206 0.80 6.50 Other damage Z 0.0000 0.0000 0.00 0.00 Color grade 0.0000 0.0000 0.00 0.00 Red rice % 0.0866 0.3033 0.00 b.00 Smut 2 0.3799 0.h907 0.00 9.00 Green % 2.6378 1.0627 0.00 8.00 Chalk % 2.0199 0.1628 0.00 3.00 Stack % 0.0600 0.9989 0.00 10.00 Test weight lb/bu h5.h950 0.9935 fi0.00 97.00 Settlement price S/cwt 9.9216 1.750h h.70 13.h5 Mill price at sale S/cwt 21.8262 2.5272 19.00 26.00 50 \» Table B.12. Weighted means, standard deviations, and ranges of selected variables at Bay City, 1982-83 Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 17h Head yield % 55.5991 2.7001 3h.00 63.00 M111 yield % 68.9199 0.9205 69.00 71.00 Grade no. 1.9853 0.6122 1.00 6.00 Seed no. 2.0333 10.2881 0.00 135.00 Peck % 1.9329 1.9820 0.00 9.90 Color grade 0.0000 0.0000 0.00 0.00 Red rice % 0.0000 0.0026 0.00 0.10 Smut % 0.0791 0.2931 0.00 2.00 Green % 0.1757 0.7871 0.00 7.00 Chalk % 0.1097 0.hh5h 0.00 2.00 Stack % 0.0019 0.0538 0.00 2.00 Test weight lb/bu b5.h693 0.169h h3.00 51.00 Other damage % 0.0000 0.0000 0.00 0.00 Settlement price S/cwt 9.1122 0.5739 7.10 11.00 Mill price at sale S/cwt 18.5833 O.60h3 17.75 20.00 51 Table B.l3. Weighted means, standard deviations, and ranges of selected variables at Bay City, l983-8h Standard Variable Unit Mean Deviation Minimum § Maximum Observation no. l88 Head yield % 56.77h3 3.335h h0.00 65.00 Mill yield % 68.578l l.l6h0 63.00 7l.00 Grade no. 2.33h3 0.8h5h l.00 6.00 Peck % l.89l3 l.lh75 0.00 7.00 Smut Z 0.2237 0.6227 0.00 6.00 Seed no. 5.h2l7 23.h8l3 0.00 550.00 Settlement price S/cwt l0.h6ll 0.8050 6.32 ll.h8 Mill price at sale $/cwt l9.5788 0.3825 l9.00 20.25 52 Table B.lh. Federal grades at Alvin for the market years l98l-82, l982-83, and l983-8h Variable Unit l98l-82 1982-83 1983-85 Observation no. l70 58 l59 Grade l % 0.00 0.00 2.5h Grade Z % 5l.85 53.50 27.88 Grade 3 % 25.62 37.79 58.81 Grade h % 9.60 6.25 l6.98 Grade 5 2 5.98 l.l6 l.7l Grade 6 % 6.02 0.78 0.08 Sample Grade % l.93 0.52 2.00 53 54 Table B.l5. Federal grades at Danbury for the market years l982—83 and l983—8h Variable Unit l982-83 l983-8h Observation no. 66 72 Grade l % 0.00 0.00 Grade Z % 55.27 52.60 Grade 3 % 30.02 l8.86 Grade h Z 3.03 lO.38 Grade 5 Z 5.98 h.lh Grade 6 % 3.h7 h.75 Sample Grade X 2.23 9.27 Table B.16. Federal grades at E1 Campo for the market years 1982-83 and 1983—8h Variable Unit 1982-83 1983-81-1 Observation no. 22h 311 Grade 1 % 60.99 h5.21 Grade 2 % 18.73 37.23 Grade 3 % 8.28 7.66 Grade h % 6.h1 3.02 Grade 5 % 1.76 0.h9 Grade 6 % 2.16 3.18 Sample Grade 2 1.67 3.20 55 56 Table B.17. Federal grades at Ganado for the market years 1981-82, 1982-83, and 1983-8h Variable Unit 1981-82 1982-83 1983-8h Observation no. 208 190 153 Grade 1 % 60.92 66.62 h5.63 Grade 2 % 22.61 22.90 39.19 Grade 3 X 11.h7 5.39 9.1h Grade h % 2.16 2.57 2.87 Grade 5 % 2.15 1.39 1.35 Grade 6 % 0.58 0.h8 1.82 Sample Grade % 0.10 0.65 0.00 Table B.18. Federal grades at Bay City for the market years 1981-82, 1982-83, and 1983-8% Variable Unit 1981-82 1982-83 1983-8h Observation no. 92 172 188 Grade 1 2 1.75 11.50 0.65 Grade 2 2 90.13 83.32 80.39 Grade 3 % 3.75 3.13 10.65 Grade h % 0.00 0.28 3.79 Grade 5 % 3.98 0.82 2.36 Grade 6 Z O.hO 0.9h 2.18 Sample Grade % 0.00 0.00 0.00 57 Table B.l9. Weighted means, standard deviations, and ranges of selected variables at American Rice, Incorporated east Texas warehouses, l982—83 Standard _ Variable Unit Mean Deviation Minimum Q Maximum Observation no. 5l7 Head yield % 59.2231 b.9191 26.00 66.00 Mill yield Z 69.8852 1.38h2 56.00 72.00 Grade no. 1.9255 1.h271 1.00 7.00 Loan value $ 8.6l28 2.l958 0.00 l0.l2 Peck % 1.3036 1.1229 0.00 20.80 Total damage % 2.3301 2.0578 0.00 28.80 Smut 8 0.3317 0.9929 0.00 h.h0 Red rice % 0.9095 1.1813 0.00 59.00 Other damage % 0.7279 0.8815 0.10 21.50 Chalk 8 0.2192 0.0616 0.10 0.90 Stack light no. 3.2629 27.9063 0.00 626.00 Stack dark no. 0.0507 0.8795 0.00 26.00 Mix long grain Z 0.0000 0.0000 0.00 0.00 Mix medium grain % 0.0881 0.35l0 0.00 h.90 Test weight lb/bu h5.ll55 0.9900 h0.l0 h9.l0 Seed no. 7.2128 h7.2685 0.00 1509.00 Settlement price S/cwt 8.7h22 0.6557 3.78 9.62 58 \\ Table B.20. Weighted means, standard deviations, and ranges of selected variables at American Rice, Incorporated northwest Texas warehouées; 1982-83 Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 827 Head yield % 57.7165 5.6575 29.00 66.00 Mill yield 2 69.2650 l.5672 62.00 72.00 Grade no. 1.5500 1.1208 1.00 7.00 Loan vaiue S 8.8580 1.5183 0.00 10.07 Peck % l.5782 0.9l57 0.00 8.50 Total damage Z 2.5357 1.8936 0.00 12.80 Smut % 0.55l5 0.7225 0.00 6.70 Red rice 2 0.0555 0.l959 0.00 5.80 Other damage % 0.5635 0.5282 0.10 5.90 Chalk % 0.2522 0.0757 0.l0 0.70 Stack light n0. 0.7627 l2.2389 0.00 550.00 Stack dark n0. 0.0026 0.l02l 0.00 5.00 Mix long grain Z 0.0000 0.0000 0.00 0.00 Mix medium grain % 0.0135 0.0781 0.00 0.90 Test weight 1b/bu 55.3862 l.l535 39.70 59.30 Seed 2 5.5955 59.l900 0.00 l788.00 Settlement price S/cwt 8.7655 0.5805 5.90 9.62 59 Table B.2l. Weighted means, standard deviations, and ranges of selected variables at American Rice, Incorporated south Texas warehouses, l982—83 Standard Variable Unit Mean Deviation Minimum 1 Maximum Observation no. l028 Head yield % 57.8099 3.8982 29.00 66.00 M111 yield 2 68.8219 1.5580 61.00 72.00 Grade nO. 1.5800 1.2hh6 1.00 7.00 Loan value $ 8.7h85 1.7927 0.00 10.07 Peck % 1.1828 0.9523 0.10 9.30 Total damage 2 2.2098 2.0690 0.30 17.50 Smut % 0.3379 0.5689 0.00 6.10 Red rice % 0.0697 0.5138 0.00 20.70 Other damage % 0.hlhl 0.h690 0.l0 7.30 Chalk % 0.22h7 0.0718 0.10 0.50 Stack light HO. 0.5212 11.8226 0.00 10h3.00 Stack dark HO. 0.0000 0.0000 0.00 0.00 Mix long grain % 0.0000 0.0000 0.00 0.00 Mix medium grain % 0.0l89 0.2228 0.00 22.hO Test weight lb/bu h5.56bh l.07h8 39.80 b8.h0 Seed n0. 0.6519 27.9132 0.00 961.00 Settlement price S/cwt 8.7h36 0.h558 5.23 9.62 60 ‘\ Table 8.22. Weighted means, standard deviations, and ranges of selected variables at American Rice, Incorporated east Texas warehouses, 1983-85 Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 593 Head yield 8 57.1328 5.1691 25.00 66.00 Mill yield % 69.5399 l.2585 59.00 7l.00 Grade no. 2.0861 1.3603 1.00 7.00 Loan value S 8.6932 1.9958 0.00 10.12 Peck 2 l.9205 l.5356 0.00 l5.80 Total damage % 2.7533 2.2252 0.00 17.20 Smut % 0.3593 0.6085 0.00 5.50 Red rice 2 0.5137 2.0797 0.00 76.20 Other damage 2 1.0275 1.6776 0.10 56.20 Chalk % 0.2759 0.1800 0.l0 l.70 Stack light no. 25.5000 253.5276 0.00 9999.00 Stack dark no. 0.3526 l5.6523 0.00 736.00 Mix long grain 2s 0.0059 0.21187 0.00 111.20 Mix medium grain % 0.1229 0.5370 0.00 10.50 Test weight lb/bu 55.8965 l.ll2l 35.50 58.50 Seed no. 5.0663 25.0w. 0.00 372.00 Settlement price S/cwt 10.0733 1.0561 5.88 12.00 61 Table B.23. Weighted means, standard deviations, and ranges of selected variables at American Rice, Incorporated central Texas warehouses, 1983-8h Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 13 Head yield % 59.1h82 h.1693 h9.00 6h.00 Mill yield % 70.0195 1.0163 68.00 71.00 Grade no. 1.9777 1.2501 1.00 6.00 Loan value S 8.8207 2.h0h6 0.00 9.95 Peck 2 1.9576 0.9088 0.60 3.60 Total damage % 3.5018 1.h557 1.20 5.80 Smut % 0.163h 0.2109 0.00 0.80 Red rice % 0.15h6 0.182h 0.00 0.50 Other damage % 0.771h 0.5061 0.20 1.90 Chalk Z 0.2226 0.0798 0.10 0.h0 Stack light no. 0.0000 0.0000 0.00 0.00 Stack dark no. 0.0000 0.0000 0.00 0.00 Mix long grain % 0.0000 0.0000 0.00 0.00 Mix medium grain % 0.0623 0.107b 0.00 0.30 Test weight lb/bu hh.3872 1.0206 h2.h0 h5.70 Seed no. 6.1097 16.7317 0.00 68.00 Settlement price S/cwt 10.2h02 0.8391 8.5% 11.80 62 i. -\ Table B.2b. Weighted means, standard deviations, and ranges of selected variables at American Rice, Incorporated northwest Texas warehouses, l983-8h Standard Variable Unit Mean Deviation Minimum Maximum Observation no. 7l8 Head yield % 56.1501 5.2789 18.0 65.00 M111 yield 2 69.3780 1.0926 52.0 72.00 Grade HO. 1.7131 1.2100 1.0 7.00 Loan value $ 8.8197 1.7201 0.0 10.09 Peck % 1.5270 1.1090 0.2 11.20 Total damage % 2.h639 2.0309 O.h 25.80 Smut % 0.382% 0.6278 0.0 6.60 Red rice % 0.0333 0.3893 0.0 19.20 Other damage % 0.58l8 0.7022 0.l 9.00 Chalk 24 0.3880 0.30116 0.1 1.60 Stack light % 12.9936 8h.1703 0.0 895.00 Stack dark % 0.0550 1.0171 0.0 23.00 Mix long grain Z 0.0000 0.0000 0.0 0.00 Mix medium grain % 0.0068 0.0607 0.0 l.h0 Test weight lb/bu h5.2385 0.9959 39.l h7.l0 Seed n0. 9.5076 95.5096 0.0 1796.00 Settlement price S/cwt lO.h587 0.9660 6.2 l2.02 63 Table B.25. Weighted means, standard deviations, and ranges of selected variables at American Rice, Incorporated southern Texas warehouses, l983-8h Standard Variable Unit Mean Deviation Minimum :1 Maximum Observation no. 687 Head yield % 57.3550 3.6951 30.00 65.00 M111 yield % 69.6021 1.2256 62.00 71.00 Grade no. 1.5689 1.2222 1.00 7.00 Loan value S 8.9hh5 l.7337 0.00 l0.0h Peck 24 1.11.99 0.9818 0.20 7.60 Total damage 2s 2.2555 2.6200 0.30 19.80 Smut % 0.0912 0.2051 0.00 2.50 Red rice % 0.0283 0.2h55 0.00 5.h0 Other damage % 0.h72h 0.6620 0.l0 6.h0 Chalk % 0.3539 0.2722 0.10 1.60 Stack light no. h.575h fi2.965h 0.00 850.00 Stack dark no. 0.0056 0.1059 0.00 2.00 Mix long grain % 0.0000 0.0000 0.00 0.00 Mix medium grain % 0.0086 0.0603 0.00 l.O0 Test weight lb/bu 55.6207 1.0756 37.50 h7.90 Seed no. h.lh05 22.7781 0.00 387.00 Settlement price $/cwt l0.8l09 0.9hlh 7.l0 ll.9l Table B.26. Maximum likelihood estimates of the Box-Cox Model at the Alvin, Danbury, El Campo, Ganado, and Bay City markets for the years 1981-82, 1982-83, and 1983-8h Combined Market 1981-82 1982-83 1983-8h years A1V1n 1.00 1.02 1.01 1.10 Danbury -- 1.01 1.02 1.01 E1 Campo -- 0.81 1.00 0.80 Ganado 0.85 1.00 0.11 0.93 Bay City 0.8h 1.00 1.01 1.00 Note: A value of 1.00 means a linear model is selected while a value of 0 means a log—linear model is preferred. Table B.27. Summary of 1985 rice stink bug control costsa Formu- Active lation ingredients Cost of Spray Flying Total (unit) (lb) insecticide volume cost) cost lnsecticideb rate/A S/unit S/A gal/A S/A S/A Methyl parathion h EC 3/M pt 0.25 l.3h/pt l.O0 l—5 2.70 3.70 Sevin XLR-hlb/gal l qt l.0O l8.70/gal h.68 2-3 3.l0 7.78 The reported costs are based on a survey of several west and east sides of the Texas Rice Belt. Thus, quoted costs should be fairly representative of the range of costs for the entire rice growing area. companies located on both the 1 Methyl parathion is the least expensive and Sevin XLR is the most expensive treatment recommended for rice stink bug control; thus, these insecticides represent the largest range in stink bug control costs. Source: Drees 1983. 66 Table B.28. Percent peck damage by county, American Rice, Incorporated data, 1982-83 and 1983-89 Range of peck damage (%) County 0 .01-.50 .51'1.00 1.01-1.50 1.51-2.00 2.00-3.00 >3.00 percent Brazoria 0.0 11.9 19.1 21.8 20.5 18.9 7.8 Harris 3.3 22.7 21.7 19.8 19.9 7.3 5.3 Matagorda 0.0 10.7 31.0 19.6 13.0 15.8 9.9 Ca1houn 0.0 16.7 29.1 10.2 21.1 8.3 19.6 Jackson 0.0 29.9 37.9 15.2 10.5 7.3 5.2 Victoria 0.0 77.6 0.0 0.0 0.0 7.9 19.5 Wharton 0.0 30.9 33.7 18.7 8.3 6.5 2.9 Colorado 0.1 23.7 37.9 18.8 12.0 5.9 1.6 ,Fort Bend 0.0 3.0 15.5 21.8 18.1 31.5 10.1 Wa11er 0.0 1.1 16.2 27.7 17.9 25.1 12.0 Madison 5.2 3.2 36.7 15.2 27.0 7.3 5.9 Austin 0.0 0.0 17.2 0.0 58.5 0.0 29.3 Hardin 87.3 0.0 0.0 9.5 0.0 0.0 8.2 Orange 100.0 0.0 0.0 0.0 0.0 0.0 0.0 Jefferson 9.7 9.2 16.8 19.9 18.7 20.9 15.8 Liberty 8.9 9.6 29.6 21.5 16.7 12.3 6.9 Chambers 11.6 6.1 11.6 9.2 16.7 29.8 20.0 Texas 2.8 19.1 27.1 19.9 15.0 13.9 7.7 67 Table B.29. 1982-83 and 1983-85 Percent red rice by county, American Rice, Incorporated data, Range of red rice (%) County .01-.50 .51-1.00 1.01-1.50 1.51-2.00 2.00-3.00 >3.00 percent Brazoria 93.3 6.7 0.0 0.0 0.0 0.0 0.0 Harris 61.1 26.7 8.3 0.3 2.8 0.0 0.9 Matagorda 93.7 5.5 0.8 0.1 0.0 0.0 1.0 Calhoun 98.3 1.7 0.0 0.0 0.0 0.0 0.0 Jackson 87.9 9.2 2.0 0.5 0.3 0.2 0.0 Victoria 100.0 0.0 0.0 0.0 0.0 0.0 0.0 Wharton 86.5 11.3 1.3 0.7 0.1 0.0 0.2 Colorado 92.3 6.8 0.5 0.2 0.2 0.1 0.0 FOFt Bend 86.9 8.5 3.2 0.3 0.7 0.2 0.2 Waller 91.5 8.1 0.5 0.0 0.0 0.0 0.0 Madison 76.5 19.6 3.9 0.0 0.0 0.0 0.0 Austin 82.8 17.2 0.0 0.0 0.0 0.0 0.0 Hardin 51.8 58.8 9.5 0.0 0.0 0.0 0.0 Orange 33.7 35.6 15.8 5.0 3.0 0.0 6.9 Jefferson 38.8 51.0 8.3 3.9 2.2 1.8 5.0 Liberty 55.8 32.1 5.6 1.8 1.0 1.7 3.0 Chambers 55.2 30.2 11.5 3.2 2 6 3.5 5.0 Texas 78.5 15.5 2 9 1.0 0 6 0.5 1.1 68 f1 Table B.30. Percent smut damage by county, American Rice, Incorporated data, 1982-83 and 1983-85 Range of smut damage (%) County 0 .01".50 .51-1.00 1.01-1.50 1.51-2.00 2.00-3.00 >3.00 percent Brazoria 62.0 38.0 0.0 0.0 0.0 0.0 0.0 Harris 32.2 52.9 5.8 12.1 5.2 2.8 0.0 Matagorda 51.1 52.0 5.0 0.8 0.5 0.7 0.0 Calhoun 33.5 50.6 12.7 0.0 3.3 0.0 0.0 Jackson 57.3 36.5 6.6 5.9 1.8 2.7 0.2 Victoria 53.5 56.5 0.0 0.0 0.0 0.0 0.0 Wharton 55.0 51.9 7.8 3.0 0.8 1.3 0.2 Colorado 33.5 50.5 10.1 2.5 2.1 0.5 0.9 Fort Bend 30.3 39.6 17.9 5.1 3.5 1.6 2.1 Wa11er 23.3 39.5 15.0 8.5 7.7 5.0 2.1 Madison 35.6 50.5 9.8 10.5 2.8 2.0 0.0 Austin 17.2 29.3 53.5 0.0 0.0 0.0 0.0 Hardin 80.6 19.5 0.0 0.0 0.0 0.0 0.0 Orange 32.2 55.2 9.8 3.8 0.0 0.0 0.0 Jefferson 27.0 52.6 16.9 6.5 1.0 3.5 2.7 Liberty 27.6 55.5 10.5 3.1 .0 2.1 0 3 Chambers 55.0 53.5 5.5 5.3 1.7 0.0 0 0 Texas 35.5 57.1 10 2 3 8 2.0 1 8 0 7 69 Table B.31. Percent chalk by county, American Rice, Incorporated data, 1982-1983 and 1983-8% Range of chalk (Z) County 0 .01-.50 .51-1.00 1.01-1.50 1.51-2.00 2.00-3.00 >3.00 percent Brazoria 0.0 100.0 0.0 0.0 0.0 0.0 0.0 Harris 100.0 0.0 0.0 0.0 0.0 0.0 0.0 'Matagorda 0.0 98.8 1.2 0.0 0.0 0.0 0.0 Calhoun 0.0 100.0 0.0 0.0 0.0 0.0 0.0 Jackson 0.0 92.8 5.2 1.3 0.7 0.0 0.0 Victoria 0.0 75.3 25.7 0.0 0.0 0.0 0.0 Wharton 0.0 91.7 6.9 11.9 0.0 0.0 0.0 Colorado 0.0 89.7 11.8 3.3 0.2 0.0 0.0 Fort Bend 0.0 95.1 2.9 2.0 0.0 0.0 0.0 Waller 0.0 99.3 5.8 0.9 0.0 0.0 0.0 Madison 0.0 96.8 3.2 0.0 0.0 0.0 0.0 Austin 0.0 82.8 17.2 0.0 0.0 0.0 0.0 Hardin 0.0 100.0 0.0 0.0 0.0 0.0 0.0 Orange 0.0 100.0 0.0 0.0 0.0 0.0 0.0 Jefferson 0.0 98.9 1.6 0.0 0.0 0.0 0.0 Liberty 0.0 96.5 3.2 0.9 0.0 0.0 0 0 Chambers 0.0 96.5 2.h 1. 0.1 0.0 O O Texas 0.0 93.8 5.0 1.1 0 1 0.0 0 0 70 ‘\ Table B.32. Number of seed by county, American Rice, Incorporated data, 1982-83 and 1983-85 Range of seed (no.) County 0 .01-2.50 2.51-5.00 5.01-7.50 7.51-10.00 10.01-15.00 >15.00 percent Brazoria 90.3 0.0 9.7 0.0 0.0 0.0 0.0 Harris 56.7 0.0 57.0 0.0 0.0 0.0 6.3 Matagorda 77.0 0.0 12.7 0.5 1.5 1.9 6.5 Ca1houn 96.2 0.0 0.8 1.1 0.0 1.5 0.5 Jackson 86.8 1.0 5.9 0.1 0.7 0.5 5.0 Victoria 100.0 0.0 0.0 0.0 0.0 0.0 0.0 Wharton 78.5 0.5 9.3 1.5 1.3 1.5 7.8 Colorado 82.8 0.5 7.6 0.3 0.8 1.0 7.1 Fort Bend 81.0 0.0 11.8 1.0 0.5 0.6 5.2 Waller 85.1 2.2 10.0 0.2 0.6 0.3 2.6 Madison 89.5 0.0 2.8 5.5 0.0 0.0 3.5 Austin 100.0 0.0 0.0 0.0 0.0 0.0 0.0 Hardin 80.5 0.0 19.5 0.0 0.0 0.0 0.0 Orange 83.9 0.0 7.1 0.0 0.0 0.0 9.0 Jefferson 85.2 0.0 8.2 0.2 1.8 0.5 5.1 Liberty 75.6 0.5 13.6 1 5 0.3 1.5 7.2 Chambers 80.2 0.5 10.6 11 2 0.2 0.2 7.2 Texas 80.8 0.5 9.5 1.0 0.9 1.0 6.5 71 Table B.33. Mean and standard deviation for rice quality factors, American Rice, Incorporated data, 1982-83 and 1983—8h Peck Red Rice Smut Chalk Seed Standard Standard Standard Standard 8 Standard County Mean Deviation Mean Deviation Mean Deviation Mean Deviation Mean Deviation percent number Brazoria 1.63 1.22 0.02 0.09 0.06 0.08 0.20 0.09 0.37 1.16 Harris 1.29 1.19 0.20 0.56 0.h5 0.65 0.20 0.05 7.53 2h.78 Matagorda 1.57 1.12 0.07 0.65 0.19 0.31 0.22 0.10 3.23 11.33 Calhoun 1.76 1.73 0.01 0.05 0.30 0.38 0.19 0.08 0.38 2.22 Jackson 1.15 0.90 0.05 0.22 0.30 0.58 0.28 0.21 3.51 33.37 Victoria 0.98 1.3h 0.00 0.00 0.05 0.06 0.b7 0.20 0.00 0.00 Wharton 1.01 0.78 0.05 0.27 0.2h O.h7 0.28 0.19 5.67 2h.93 Colorado 1.05 0.7h 0.02 0.15 0.30 0.55 0.3h 0.26 h.8h 21.65 Fort Bend 1.95 1.00 0.08 0.53 0.51 0.87 0.27 0.17 11.51 110.83 Waller 1.91 1.02 0.02 0.07 0.61 0.76 0.29 0.16 1.75 9.57 Madison 1.33 0.77 0.06 0.1h 0.h1 0.5h 0.26 0.17 1.h8 5.85 Austin 2.29 1.21 0.02 0.0h 0.53 0.31 0.22 0.18 0.00 0.00 Hardin 0.h0 1 1h 0.23 0.27 0.02 0.0h 0.2h 0.05 0.63 1.32 Orange 0.00 0.00 0.71 2.09 0.2h 0.28 0.22 0.07 5.53 22.89 Jefferson 1.91 1.h1 0.h8 1-06 0.51 0.76 0.21 0.10 b.1h 22.30 Liberty 1.h2 1.23 0.h1 1.82 0.32 0.h9 0.25 0.13 7.37 h6.31 Chambers 2.0h 1.63 0.63 1.5h 0.22 0.38 0.25 0.16 h.69 18.32 Texas 1.h1 1.06 0.07 0.h3 0.3h 0.58 0.28 0.19 5.06 39.11 at“ a... .._.t...........w..u 72 Table B.3h. Tests of hypothesis of consistency in quality effects across Texas bid/acceptance markets (F-ratio)a Market yearb Quality factor 1981-82 1982-83 l983—8h Market 1.97 ll.O6* lZ.hQ* Mill price 2.57mi 13.1w 7.o7>'< Head yield 8.9 * l.88 h.68* Brokens l.7h 2.20** 2.2l* Weed seed 5.93* 2.76* 0.98 Red rice 0.89 2.06 1.87 Peck 2.12 3.hQ* l2.97* Smut l.l6 O.6h 2h.h8* Chalk O.h6 l.Z6 l.lh Heat damage 3.0h* 2.39% 6.30* Test weight h.l0* 2.hl* O.7l * Indicates rejection of the null hypothesis (i.e., quality effects are equal across markets) at the 5 percent level of significance. **Indicates rejection of the null hypothesis at the 10 percent level of significance. a Bid/acceptance markets were located at Alvin, Danbury, Bay City, E1 Campo, and Ganado. b Rejection of the null hypothesis indicates that premiums/discounts for that quality factor are different across markets within a given year. 73 Table B.35. Correlation between quality factors Market Quality factor year and Mill Head Red Heat quality factor price yield Brokens Seed rice Peck Smut Chalk damage 1981-82: Head yield 0.10 Brokens -0.03 -0.96 Seed -0.12 -0.06 0.01 Red rice 0.16 -0.00 0.02 0.13 Peck -0.05 -0.10 0.03 0.50 0.17 Smut 0.10 -0.15 0.13 -0.17 -0.05 -0.12 Chalk 0.2h 0.07 -0.0h 0.3h 0.28 0.59 -0.06 Heat damage -0.08 -0.01 0.01 0.01 -0 01 0.0h -0.03 0.07 Test weight 0.05 0.31 -0.17 -0.38 -0 10 -0.5h -0.12 -0.31 -0.11 Q U 1982-83: Head yield -0.01 Brokens -0.01 -0.9h Seed -0.09 -0.05 0.03 Red rice 0.06 0.05 -0.02 0.02 Peck 0.07 -0.28 0.21 0.09 0.10 Smut 0.07 0.1h -0.13 -0.06 0.06 0.05 Chalk 0.1h 0.21 -O.1h -0.03 0.b5 0.20 0.2h Heat damage 0.03 0.01 0.01 0.01 0.02 0.0h 0.13 0.lh Test weight -0.11 0.25 -0.16 -0.05 -0.10 -0.23 -0.0h -0.01 -0.09 1983-8h: Head yield 0.06 Brokens -0.08 -0.96 Seed 0.1h -0.03 -0.01 Red rice 0.00 -0.10 0.09 -0.03 Peck 0.09 -0.33 0.29 0.12 0.05 Smut -0.05 -0.07 0.07 -0.08 0.08 0.11 Chalk -0.05 -0.36 0.118 -0.07 0.21 0.18 0.12 Heat damage 0.02 0.0h -0.02 -0.02 0.00 0.0h 0.01 0.12 Test weight -0.02 0.h7 -0.h5 -0.03 -0.1L -0.36 -0.09 -0.h7 -0.09 74 Table B.36. Tests of significance of classification variables (Texas bid/acceptance markets) on selected quality variables (F—ratio) Market yeara Classification Selected variable variables l98l-82 l982-83 l983-8h Market Head yield 2l.OZ* 39.h9* l8.65* Brokens 9.5h* lh.39* 32.72* Test weight 7.32* 2h.h3* 38.72% * Indicates rejection of the null hypothesis at the 5 percent level of significance. a Rejection of the null hypothesis indicates that the market intercept term is not equal across markets for that year. 75 76 Table B.37. Tests of hypothesis of constancy in peck coefficients on selected variables across Texas bid/acceptance markets (F—ratio)a Market yearb Quality factor affected by peck 1981-82 1982-83 1983+8$ Head yie1d 0.b6 3.97* 6.33* Brokens 0.88 3.98* 8.h0* Test weight #.51* 17.13* 13.97* * Indicated rejection of the nu11 hypothesis at the 5 percent 1eve1 of significance. a Bid/acceptance markets were iocated at Aivin, Danbury, Bay City, E1 Campo, and Ganado. b Rejection of the hu11 hypothesis indicates that the effect of peck on head yie1d, brokens, or test weight was different across markets within a given year. Table B.38. Impact of peck on selected quality variables, American Rice, Incorporated (t-ratios in parentheses) Item I982-83 I983-8h I982-83/I983-8h Head yield: Intercept 59.6259 57.7725 58.9856 (#71-79)* (359-79)* (595-50)* Peck -0.9588 -O.6Ih6 -O.8hO9 ("I3-39)* ('7-79)* (-I5-68)* F-ratio I79.2h* 60.63* 2h5.77* R2 0.052 0.026 0.062 Brokens: Intercept 10.3730 I2.I0h2 iO.9b55 (9b.05)* (88.62)* (I25.92)* Peck O.h6bO 0.3393 O.hb5I (7 J12) ='= (l1 -92) >'= (9 111+) >'= F-ratio 55.I2* 2h.T8* 89.I3* R2 0.016 0.011 0.016 Test weight: Intercept h5.9359 h5.h722 h5.7586 (I591-32)* (I293-59)* (20hI-I7)* Peck -0.6225 -0.2069 -0.3320 (-25-83)* (-II-5b)* (-27-30)* F—ratio 667.00* 133.13* 765.536 R2 0.168 0.055 0.118 * Indicates rejection of the nuii hypothesis at the 5 percent Ievei of significance. 77 [Blank Page in Orifiafl Bulletin] ‘ , ‘ _. r ,3 M!‘ , ,4‘,- » é fl‘ ,., \ . \ , r h. u’ ‘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 , 45 Station and does not imply its approval to the exclusion of other products that also may be suitable. i " 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. ECONO~RICE y