Agronomy Journal 93:827-835 (2001)
© 2001 American Society of Agronomy
PRODUCTION PAPER
Analysis of Seedbeds and Maturity Groups for Dryland Soybean on Clayey Soil
Michael P. Popp*,a,
Terry C. Keislingb,
Lawrence R. Oliverc,
Carl R. Dillond and
Patrick M. Manninga
a Dep. of Agric. Econ. and Agribusiness, 220 Agric. Building, Univ. of Arkansas, Fayetteville, AR 72701
b Dep. of Crop, Soil, and Environ. Sci., Univ. of Arkansas, Northeast Res. and Ext. Cent., P.O. Box 48, Keiser, AR 72351
c Dep. of Crop, Soil, and Environ. Sci., 276 Altheimer Drive, Univ. of Arkansas, Fayetteville, AR 72701
d Dep. of Agric. Econ., 403 Agric. Eng. Building, no. 2, Univ. of Kentucky, Lexington, KY 40546-0276
* Corresponding author (mpopp{at}comp.uark.edu)
Received for publication May 8, 2000.
 |
ABSTRACT
|
|---|
Soybean [Glycine max (L.) Merr.] production systems on clayey soils are difficult to manage. With improvements in no-till planting equipment and herbicide technologies in the early 1990s, no-till production has become a viable alternative to the traditional tilled seedbed. Therefore, the relative economic performance of tilled and no-till seedbeds with respect to profitability, sensitivity to input price changes, and risk is assessed for maturity group (MG) IV, V, and VI soybean. Field experiments using split plots (main plots were MG and subplots were seedbeds) and a randomized complete block design with four replications were conducted from 1992 to 1994 at Rohwer, AR and from 1990 to 1997 at Keiser, AR on Sharkey and Sharkey silty clay, respectively. The importance of weather conditions is highlighted in the varied seedbed preparation effect on grain yields, with no clear advantage to either method. On average, yields were higher for MG IV at Rohwer and MG VI at Keiser. The breakeven price and yield analysis suggested that MG selection had a larger economic impact than seedbed preparation, regardless of location. This analysis also showed the extent of production cost differences and associated risk of loss by location. Risk analysis revealed that optimal production strategies changed when input costs were added to yield information and further confirmed that MG selection affects profitability more than seedbed preparation. Production practices that better exploit the yield potential of various MG cultivars (as related to weather conditions) therefore deserve further research attention.
Abbreviations: CDFs, cumulative distribution functions • FSD, first-degree stochastic dominance • GSD, generalized stochastic dominance • MG, maturity group • MSBG, Mississippi State Budget Generator • NEREC, Northeast Research and Extension Center • NRAT, net returns above total specified cost • RACs, risk aversion coefficients • SSD, second-degree stochastic dominance
|
INTRODUCTION
|
|---|
CLAYEY SOILS have historically been challenging to manage, especially those with montmorillonitic clays. Their plastic nature and stickiness are the primary physical properties that make management difficult. In addition, the window of appropriate soil moisture for performing machinery operations is very short. In the lower Mississippi River Delta, there are more than 3.25 million ha of clayey soils that are cropped primarily to soybean (Boul, 1973). Thus, their management in a humid environment is very important.
With the advent of equipment and herbicides that allow effective management of no-till production on these soils, there is the question of whether a no-till seedbed system has advantages over the tilled seedbed system. These potential advantages might include higher average profitability as well as reduced financial and production risk.
The tilled seedbed system on these soils is prepared well ahead of planting time and is referred to as a stale seedbed (Oliver et al., 1993). This dictates that the soil be disked or field-cultivated to destroy existing vegetation and to incorporate dinitroaniline herbicides (Baldwin et al., 1991). The field must then receive rain so that the clods formed during the tillage process will break down enough to yield a continuous soil surface. Planting then takes place into the stale seedbed, which will be dry and crusted on the very top (usually 1 cm deep). This dry surface crust keeps the clay soil from adhering to the planting equipment, and therefore allows seed placement into soil moisture without continually cleaning the equipment.
The no-till planting system eliminates preplant soil tillage so that planting can begin when the soil would otherwise be tilled for preparation of a stale seedbed. Weather permitting, this can lead to an opportunity to move the planting date earlier into the season without having to coordinate with prior field operations. This earlier planting can also be coupled with an early soybean planting system (Heatherly, 1999). Heatherly recommends a system for no-tilling on clay that also provides field drainage from the tractor wheels (Heatherly et al., 1986; Heatherly et al., 1994; Heatherly et al., 1993) at planting.
Factors that affect the implementation of early planting systems are assessed here. The analysis was undertaken to (i) assess the yield response and associated profitability of these systems, (ii) investigate the economic risk associated with these systems through break-even price and yield analysis, and (iii) identify changes in the choice of optimal production method across different risk preferences of decision makers. Specific hypotheses to be tested are that there are no differences in agronomic and economic performance across (i) maturity group (MG) and (ii) seedbed preparation method.
|
MATERIALS AND METHODS
|
|---|
Agronomic
A field trial was established at the University of Arkansas Northeast Research and Extension Center (NEREC) at Keiser, AR in 1990 (35°40'56''N, 90°6'17''W) on Sharkey silty clay (very-fine, smectitic, thermic, Chromic Epiaquerts). The experimental design was a randomized complete block with four replications. The treatment design was a split plot with soybean MG as main plots and seedbed preparation as subplots. Plot size was 3.9 by 30 m. Soybean cultivars of MG IV, V, and VI were planted each year at the recommended time for each MG as weather and equipment scheduling permitted (Table 1). At Keiser, general planting date recommendations (Ashlock, 1999) for MG IV soybean are the first opportunity in April. Full-season MG V and VI soybeans are preferably planted between 25 April and 10 June. As shown in Table 1, this lead to earlier planting dates for MG IV compared with MG V and VI (for most years, weather permitting) and, in the odd case, earlier planting in the no-till category due to lesser tillage requirements before planting.
No-till treatments received a broadcast vegetation-burndown herbicide application before planting. An additional burndown treatment was needed in certain years when weather delayed planting. Pre-emergence herbicide treatments were applied at planting, and postemergence herbicide treatments were applied as needed to control later emerging broadleaf and grass weeds. These treatments are summarized in Table 2.
Stale seedbed plots were tilled and received preplant-incorporated herbicide 2 to 6 wk before planting to reduce vegetation before planting. As a result, less preplant burndown herbicides were used than in the no-till system. Pre-emergence and postemergence herbicides were applied as they were in the no-till plots.
Plot centers from all plots were harvested, and the seed yields were adjusted to 13% moisture. Analysis of variance was done using SAS (1989). The design was treated as a split plot over years.
A second field trial was established at the Southeast Branch Experiment Station at Rohwer, AR in 1992 (33°48'6''N, 91°2'38''W) on Sharkey clay. The experimental design and plot size were the same as the field trial at the NEREC, and the same soybean cultivars were planted (Table 1). At Rohwer, the MG IV planting date recommendation starts after 15 March, and MG V and VI recommendations are the same as those at Keiser (Ashlock, 1999). At Rohwer, equipment scheduling problems occurred in 1994, and weather conditions resulted in late planting in 1992.
Vegetation control and field preparation were quite similar to the field trial at the NEREC; however, no burndown herbicide applications were made to stale seedbed plots. Also, in 1993, a herbicide application was required to desiccate existing vegetation in MG IV and MG V plots to permit timely harvesting. Harvest and statistical procedures were the same as those at the NEREC.
Economic
The economic analysis was based on enterprise budgets generated by the Mississippi State Budget Generator (MSBG) (Spurlock and Laughlin, 1992). A total of 60 enterprise budgets were generated, one for each MG–seedbed preparation combination for each year at each location. The MSBG was used to calculate only direct and fixed expenses while net returns (returns to land, labor, and management) were calculated for each yield observation using a spreadsheet. The enterprise budgets were developed using yields, input requirements, and field operations obtained from the records of the agronomic experiment.
A soybean price of $0.234 kg-1 was used to calculate gross receipts, representing a 10-yr (1988–1997) average of the statewide soybean price based on values reported in the Arkansas Agricultural Statistics for 1997 (Arkansas Agric. Statistics Serv., 1998). The statewide average price (quantity weighted by region) was used to eliminate any market effects due to years with abnormally high or low prices and to remove potential regional price differences. Further, prices were not expected to be affected by MG, seedbed preparation, or yield for the individual producer selling in a competitive market place. The input prices included in the version of MSBG issued by the Arkansas Cooperative Extension Service were used for the field operations (Windham and Brown, 1997). These costs include seed, herbicide, custom-applied fertilizer, hauling, interest on operating capital, hired labor, fuel, repair and maintenance as well as depreciation and interest on equipment. The implication for risk analysis is that net return variability is driven by yield and input cost changes across treatments and not by soybean price.1
A relatively basic machinery complement was used in the budgeting process. Machinery used was based on the actual operations performed in the field study. Equipment size and custom operations are those reflective of a representative farm situation in eastern Arkansas. The main differences in equipment requirements were in preplant operations where mechanical weed control with tillage was substituted with chemical weed control in the no-till plots. This resulted in two to four and zero to two fewer passes across the field for the no-till method at Keiser and Rohwer, respectively. The difference in field passes varied more by year than by MG cultivar. Associated total specified cost savings for no-till compared with stale plots ranged from $0.51 to $37.39 ha-1, with an average difference of $23.79 ha-1 in favor of no-till production at Keiser. At Rohwer, the difference in total specified costs was $36.25 ha-1 on average in favor of the stale seedbed production method. However, there were also observations that favored the no-till method by as much as $47.22 ha-1. Differences in total specified costs across the no-till and stale seedbed production methods were therefore mixed across location and production year (see Table 3).
The General Linear Models (GLM) procedure in SAS (1989) was used to analyze seed yield and net returns. This procedure was used due to missing yield data for several treatments in various replications. The model utilized yield and net returns above total specified costs as dependent variables.
To gain more insights on the economic implications of this study, break-even analyses were conducted. Break-even analysis was conducted for prices and yields above both direct and total specified expenses to capture economic feasibility in the short and long term, respectively.
Additional risk analyses were performed to take risk preferences of producers into account. To this end, stochastic dominance methods can be used to differentiate among various strategies by taking into account both net returns and risk preferences of the decision maker. Through multiple pair-wise comparisons of cumulative distribution functions (CDFs) of the production strategies (such as those shown in Fig. 1 and 2), leading strategies can be identified as those that offer (i) higher returns at a given level of risk or (ii) the lowest risk at a given level of return.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 1. Comparison of cumulative probability density functions of no-till and stale seedbed preparation methods using net returns above total specified cost (NRAT) across all maturity groups (MGs) for each planting method at Rohwer, AR, 1992–1994. Cumulative probability density functions were plotted using the fractile method espoused by Schlaifer (1959).
|
|
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 2. Comparison of cumulative probability density functions of net returns above total specified costs (NRAT) for maturity group (MG) IV, V, and VI for (A) stale and (B) no-till seedbed preparation methods at Keiser, AR, 1990–1997. Cumulative probability density functions were plotted using the fractile method espoused by Schlaifer (1959).
|
|
For example, looking at Fig. 1, the likelihood of a loss, achieving net returns above total specified cost (NRAT) <$0 ha-1, is greater for CDFs that generally lie up and to the left of others. At Rohwer, the stale seedbed preparation strategy is preferred over the no-till strategy because the likelihood of a loss is roughly 30% lower (note arrow in Fig. 1) for the stale option.
To differentiate strategies across entire CDFs rather than just one point on the CDF, three typical stochastic dominance procedures exist: Using first-degree stochastic dominance (FSD), a decision maker always prefers more to less [even if a strategy (X) is highly risky, as long as X's CDF is entirely to the right of a less risky strategy (Y). An example is the preference of the stale seedbed preparation method compared with no till in Fig. 1]. As soon as the CDFs cross (as shown in Fig. 2), however, FSD cannot be used to categorize strategies. This is where second-degree stochastic dominance (SSD) can help to differentiate strategies by assuming risk aversion of the decision maker. The decision maker still prefers more to less, but not if the riskiness exceeds that of a production practice with somewhat lower returns. This method leads to a smaller set of leading strategies but is restrictive in the sense that risk aversion (as opposed to risk-neutral or risk-seeking behavior) is assumed. A third alternative is generalized stochastic dominance (GSD). Here, strategies are evaluated over a range of risk preferences. A risk-averse producer, faced with two scenarios, would examine the expected values and be willing to sacrifice expected value to attain less variability. The opposite is true for a risk-seeking producer. The decision of a risk-neutral producer will be based solely on expected value. First-degree stochastic dominance and SSD can thus be thought of as subsets of GSD. Because CDFs typically intersect, the FSD decision rule will not be able to differentiate among strategies. Using the more complex SSD method narrows choices but assumes risk aversion. Therefore, GSD analysis is performed to provide a comprehensive analysis of strategies in a fashion similar to Segarra et al. (1991).
For this study, the decision to pool the data at each location (i) across seedbed preparation method to select an appropriate MG strategy and (ii) across MG to select preferred seedbed preparation strategies is based on the results of the GLM procedure discussed above. Should the interaction between MG and seedbed preparation be insignificant, pooling would be supported without loss of information. Similarly, significant impacts of both replication and year would justify treating each observation as equally likely, regardless of year and replication. In that case, risk analysis would represent spatial risk (difference in location within a year for replications at the same experiment station) and temporal risk (to a different extent at Keiser and Rohwer as 6 and 3 yr of data are used, respectively).
Generalized stochastic dominance would be employed to identify a singular strategy by seedbed preparation method and MG using yield, net return above direct cost, and NRAT data over various ranges of risk preferences, which are described using risk aversion coefficients (RACs).
Risk aversion coefficients can theoretically range from negative infinity (strongly risk seeking) to positive infinity (extremely risk averse), with risk neutrality defined by a RAC of zero. McCarl's (1988) RISKROOT program allows the calculation of break-even RACs that identify risk preference boundaries over which a single preferred strategy exists. This approach differs from Meyer's (1977) approach where RAC ranges are selected and then preferred strategies over that range are identified. The difficulty lies with the selection of good RAC ranges because RACs differ across different sets of decision makers and situations (Raskin and Cochran, 1986). McCarl's (1988) procedure was chosen because decision makers can see which strategies are in the preferred set over a range of risk preferences. Finally, the range of RACs across which strategies are evaluated is set using the methodology of McCarl and Bessler (1989). Minimum and maximum RACs are used to restrict the range of feasible risk preferences, and thereby the number of preferred strategies.
|
RESULTS AND DISCUSSION
|
|---|
Agronomic
Analysis of variance results and actual yield levels for all locations and years are reported in Tables 4 and 5, respectively. Yield had an interaction between year and seedbed preparation at Keiser and between MG and year at both locations (Table 4). On average, stale seedbeds resulted in slightly higher yields than no till at both locations. Annual differences across seedbed preparation methods were significant. Similar results were found with year x MG interactions. On average, MG IV yielded higher than MG VI and V at Rohwer while at Keiser, MG VI outperformed MG V and IV in that order.
The importance of the weather scenario is illustrated by the way yields change from one growing season to the next. For example, in 1997, weather caused a large delay in planting date; therefore, the potential yield advantage of MG IV through drought avoidance was not realized and actually lead to more drought stress because soybean was maturing in the middle of August. By comparison, MG VI had superior yield results with an early rain in mid-September—too late for MG IV but beneficial for MG VI. Weather impacts also manifested themselves in the planting dates across all years of the study. While MG IV did allow for earlier planting compared with MG V and VI, this was not the case all of the time. Further, given weather conditions, the expected advantage of earlier planting due to no-till seedbed preparation occurred less than one quarter of the time; therefore, drought avoidance due to early planting may not necessarily be touted as an advantage of no-till production in this study.
Economic
Annual weather patterns also impacted NRAT by seedbed preparation and MG (Table 4). At Keiser, MG VI soybean resulted in the highest NRAT on average while MG IV soybean provided the highest average returns at Rohwer (Table 6).
View this table:
[in this window]
[in a new window]
|
Table 6. Net returns above total specified costs (NRAT) as affected by seedbed preparation and cultivar maturity group (MG).
|
|
The economic results were similar to the yield results at Keiser. However, at Rohwer, a highly significant main effect on seedbed preparation for NRAT (Tables 4 and 6) was observed with the stale seedbed preparation method outperforming the no-till method by $50.37 ha-1.
On average, seedbed preparation had no effect on break-even prices above total specified costs at Keiser while the stale seedbed method had considerably lower break-even prices at Rohwer mainly due to lower direct expenses but also due to slightly higher yields (Table 7). At both locations, there was considerable variation in year-to-year break-even prices, with ranges (second to last column) from $0.04 to $0.16 kg-1 depending on location and seedbed preparation. Comparing the seedbed preparation and MG break-even price results, it appears that MG had a larger impact on break-even prices than seedbed preparation. At Keiser, this is determined by yield and direct cost differences, whereas at Rohwer, differences in yield response are more influential than direct cost. Also, differences in year-to-year break-even prices ranged from $0.06 to $0.28 kg-1 (second to last column). This is nearly twice as large as the range observed across seedbed preparation results.
View this table:
[in this window]
[in a new window]
|
Table 7. Break-even prices above total specified expenses as affected by seedbed preparation and cultivar maturity group (MG).
|
|
Further analysis of production costs involved with the different treatments revealed that (i) fixed costs differed only marginally across MG and (ii) the stale seedbed preparation method was associated with higher fixed costs than the no-till method at both locations. This is largely because of differences in the number of passes across the field and in associated equipment needs. At Rohwer, lower direct expenses, especially herbicide costs, for the stale compared with the no-till seedbed preparation method were sufficient to offset fixed cost differences that are largely a function of the number of equipment passes across the field. Because the stale method only required approximately one more pass across the field than no till and herbicide costs were $35.53 ha-1 higher for the no-till method, the stale seedbed preparation method was superior to the no-till method. Overall, direct expense differences played a more important role at Rohwer in choosing seedbed preparation method while fixed costs were more influential at Keiser.
Close monitoring of costs associated with equipment needs and passes across the field as well as herbicide costs therefore play an important and different role at the two locations analyzed. At Keiser, more emphasis needs to be placed on controlling equipment cost, whereas relatively closer scrutiny over herbicide applications appears appropriate at Rohwer.
While a break-even price is affected by both changes in yield and cost over time, an analysis of break-even yields above total specified expenses (Table 8) focuses on changes in total cost only. The no-till seedbed preparation method resulted in the lowest average break-even yield at Keiser while the stale seedbed method resulted in the lowest average break-even yield at Rohwer. Changes in total specified cost over time were less variable for the stale seedbed method at both locations, especially at Rohwer. This was mainly due to changes in herbicide requirements over time for the no-till method at Rohwer. Therefore, primarily analyzing costs, the use of the stale seedbed production method may be less risky than the no-till production method. The results show that yield increases of approximately 400 kg ha-1 are required to remain in business in the long run at Keiser as a direct result of higher production costs.
View this table:
[in this window]
[in a new window]
|
Table 8. Break-even yields above total specified expenses as affected by seedbed preparation and cultivar maturity group (MG).
|
|
Using McCarl's (1988) RISKROOT program, Table 9 shows preferred production strategies over a range of risk preferences for Keiser and Rohwer when comparing seedbed preparation and MG strategies. Given the results of Table 4 and the discussion in the Materials and Methods section, the data was pooled to compare seedbed preparation strategies (i.e., no till with MG IV, V, and VI vs. stale with MG IV, V, and VI) and the MG strategies (i.e., MG IV using no till and stale vs. MG V using no till and stale vs. MG VI using no till and stale). Results are presented for the McCarl and Bessler (1989) RAC ranges and may be used to identify a strategy likely to be adopted across all risk preferences by selecting the strategy that dominates most of the RAC range. For example, in the first row of Table 9, the results of the seedbed preparation analysis indicate that the stale method at Keiser is preferred according to the yield criterion over the RAC range of -0.0102 to 0.399. This is a range that is consistent with a decision maker that demonstrates behavior ranging from risk seeking to strong risk aversion. Because this strategy covers most of the RAC range of -0.399 to 0.399, it is argued that, on the basis of yield, decision makers would pick the stale seedbed production method as superior. By similar reasoning, decision makers would pick the no-till strategy using the net return above direct cost and NRAT criteria at Keiser. At Rohwer, the stale strategy is preferred across all criteria. The bottom half of the table indicates that MG VI is superior over MG IV and MG V at Keiser. At Rohwer, MG VI would prevail for decision makers who are slightly risk averse. However, MG IV prevails for a wider range of risk preferences and would likely be the choice for a broader set of decision makers at Rohwer. On the basis of these numbers, an agronomically oriented people might suggest the use of MG VI and MG IV soybean with a stale seedbed at Keiser and Rohwer, respectively, if they were using the yield criterion alone. Taking economic returns (net returns above direct cost or NRAT) into account, the preferred strategy switches to the no-till option for Keiser. In other words, both yield and input cost fluctuations impact production practice decisions when taking risk preferences into account. (Note that average yield for the no-till practice is only marginally lower than the average stale yield at Keiser in Table 5.)
View this table:
[in this window]
[in a new window]
|
Table 9. RISKROOT program results using yield, net returns above direct cost, and net return above specified cost (NRAT) to identify preferred seedbed preparation and maturity group (MG) strategies over a range of risk preferences, Keiser and Rohwer, AR.
|
|
Even though statistical tests did not suggest performing a six-way comparison (i.e., MG IV, V, and VI each for no till and stale seedbed preparation), additional testing was done to identify whether MG selection or rankings would differ across seedbed preparation methods using NRAT at Keiser. Cumulative distribution functions of returns across MGs are plotted in Fig. 2 across both seedbed preparation methods. In both panels (Fig. 2A and 2B), the MG-VI CDF lies mostly to the right and below the other MG varieties. Further, the MG IV variety is least favorable in both panels because their CDFs lie to the left and above the CDFs of the other MG. This further supports the statistical results of Table 4 and suggests that MG differences are similar regardless of seedbed preparation method.
Finally, in order to examine the sensitivity of the risk analysis results to alternative temporal and spatial risks, two stochastic dominance experiments were conducted for the Keiser data set. (Rohwer data was not analyzed in this fashion because too few observations were available.) The first examined the removal of the 1990 data because the replication data used to calculate the average yield were no longer on record. The use of the average would thus eliminate the spatial risk for that year, and therefore might bias the results. Without the 1990 data, the Keiser risk analysis results for net returns resulted in identical qualitative results and nearly identical quantitative results. (RISKROOT output is available upon request from the lead author for this scenario.) The second experiment examined the elimination of spatial risk altogether by conducting risk analysis using only replication averages for each year (now only seven observations were available for each strategy). Results demonstrated that exclusion of spatial risk led to a switch in preference to the stale seedbed preparation method at Keiser. Furthermore, while the MG VI cultivar choice remains stochastically dominant for nearly all risk preferences, MG V dominates for the strongly risk-averse individual when only considering the temporal risk of yield and NRAT. Producers with little variation in field conditions (i.e., little spatial risk) may thus prefer the stale seedbed preparation method and MG V cultivars if they are strongly risk averse. Given the likelihood that field conditions of producers exhibit some spatial differences, the results were presented for the complete data set. Further, the 1990 data was included to use the same data set throughout the results discussion (Tables 4– 9), with little evidence of bias as discussed above.
|
INTERPRETIVE SUMMARY
|
|---|
A superior production strategy is one that provides the highest profit with the least amount of risk. Selecting on the basis of profit, the study identifies the no-till and MG VI production strategy as best for Keiser while the best production strategy for Rohwer was the complement of stale seedbed preparation and MG IV. While the soil type is similar for both locations, Rohwer, 320 km south of Keiser, typically provides earlier planting windows, and therefore offers more drought avoidance for MG IV cultivars. Producers may also partially adopt the above recommendations by choosing to plant a portion of their acreage to another MG cultivar that allows timely completion of field operations given the resources under their control. This implies that production practices that take full advantage of the yield potential of various MG cultivars (as related to weather conditions) would be the most productive research area for future attention. On the other hand, planting date differences across seedbed preparation methods may not yield much additional insight because weather conditions did not generally allow for differences in planting dates, at least in this study.
The break-even analysis suggested that MG selection had a larger impact on both break-even prices and yields than seedbed preparation, regardless of location. Further examination of costs showed that herbicide costs played a larger role at Rohwer than at Keiser. Fixed cost differences, due to the number of passes across the field and associated equipment needs, are more important at Keiser than at Rohwer.
Analysis of GSD revealed that risk-return analysis may lead to different decisions than looking at yields alone. In addition, the results also showed that MG selection and/or rankings were independent of seedbed preparation method.
Further research to identify which of the above production strategies would most likely be adopted by producers would consider the timing and number of field operations. Various seedbed preparation methods with different equipment use and scheduling may impose unrealistic restrictions on operation size, given weather conditions and recommended planting dates.
|
ACKNOWLEDGMENTS
|
|---|
The authors acknowledge the Arkansas Soybean Promotion Board for financial support and the staffs of the Northeast Research and Extension Center and Southeast Research and Extension Center. Special thanks also goes to Alan Pearce.
|
NOTES
|
|---|
1 While some regional price differences may exist across locations in Arkansas due to differences in transportation costs to final export market destinations, these price differences are assumed to play a negligible role in risk analysis of net returns in this study. This does not diminish the usefulness of the study because decision makers would look at comparisons of local prices and break-even prices to make production decisions (to produce or not to produce and/or what production practice to select). 
|
REFERENCES
|
|---|
- Arkansas Agricultural Statistics Service. 1998. Arkansas agricultural statistics for 1997. Misc. Publ. 400. Univ. of Arkansas Coop. Ext. Serv., Little Rock.
- Ashlock, L. 1999. Arkansas soybean handbook. Misc. Publ. 197. Univ. of Arkansas Coop. Ext. Serv., Little Rock.
- Baldwin, F.L., J.W. Boyd, and C.B. Guy. 1991. Recommended chemicals of weed and brush control. Misc. Publ. 44. Univ. of Arkansas Coop. Ext. Serv., Little Rock.
- Boul, S.W. (ed.) 1973. Soils of the southern United States and Puerto Rico. Southern Coop. Ser. Bull. 174. USDA-ARS, Washington, DC.
- Heatherly, L.G. 1999. Early soybean production system. p. 103–118. In L.G. Heatherly and H. Hodges (ed.) Soybean production in the midsouth. CRC Press, Boca Raton, FL.
- Heatherly, L.G., C.D. Elmore, and S.R. Spurlock. 1994. Effect of irrigation and weed control treatment on yield and net return from soybean (Glycine max). Weed Technol. 8:69–76.
- Heatherly, L.G., J.A. Musick, and J.G. Hamill. 1986. Economic analysis of stale seedbed concept of soybean production on clay soil. Bull. 944. Mississippi Agric. and Forestry Exp. Stn., Mississippi State.
- Heatherly, L.G., R.A. Wesley, C.D. Elmore, and S.R. Spurlock. 1993. Net returns from stale seedbed plantings of soybean (Glycine max) on clay soil. Weed Technol. 7:972–980.
- McCarl, B.A. 1988. Preference among risky prospects under constant risk aversion. South. J. Agric. Econ. 20:25–33.
- McCarl, B.A., and D. Bessler. 1989. Estimating an upper bound on the Pratt risk aversion coefficient when the utility function is unknown. Aust. J. Agric. Econ. 33:56–63.
- Meyer, J. 1977. Choice among distributions. J. Econ. Theory 14:326–336.
- Oliver, L.R., T.E. Klingaman, M. McClelland, and R.C. Bozsa. 1993. Herbicide systems in stale seedbed soybeans (Glycine max) production. Weed Technol. 7:816–823.
- Raskin, R., and M.J. Cochran. 1986. Interpretation and transformations of scale for the Pratt-Arrow absolute risk aversion coefficient: Implications for stochastic dominance. West. J. Agric. Econ. 11:204–210.
- SAS Institute. 1989. SAS/STAT user's guide. Ver. 6. 4th ed. SAS Inst., Cary, NC.
- Schlaifer, R. 1959. Probability and statistics for business decisions: An introduction to managerial economics under uncertainty. McGraw-Hill, New York.
- Segarra, E., J. Keeling, and J. Abernathy. 1991. Tillage and cropping system effects on cotton yield and profitability on the Texas southern high plains. J. Prod. Agric. 4:566–571.
- Spurlock, S.R., and D.H. Laughlin. 1992. Mississippi State Budget Generator user's guide. Ver. 3.0. Agric. Econ. Tech. Publ. 88. Mississippi Agric. and Forestry Exp. Stn., Mississippi State.
- Windham, T.E., and C.D. Brown. 1997. Soybeans, stale seedbed non-irrigated, clay and mixed soils, Arkansas. Enterprise Budget AG-492-11-97. Arkansas Coop. Ext. Serv., Little Rock.
TOP
ABSTRACT
INTRODUCTION
