Technology Adoption and Off-Farm Household Income: The Case of Herbicide-Tolerant Soybeans
Fernandez-Cornejo, Jorge, Hendricks, Chad, Mishra, Ashok, Journal of Agricultural and Applied Economics
We model the interaction of off-farm work and adoption of agricultural technologies and the impact of adopting these technologies on farm household income from on farm and off-farm sources after controlling for such interaction, and estimate the model for the case of adoption of herbicide-tolerant (HT) soybeans using a nationwide survey of soybean farms for 2000. We find that adoption of HT soybeans is positively and significantly related to off-farm household income for U.S. soybean farmers, after controlling for other factors. In addition, while on-farm household income is not significantly related to adoption, total household income increases significantly with adoption.
Key Words: agricultural household model, biotechnology, herbicide tolerant soybeans, off-farm income, technology adoption
JEL Classifications: O33, Q12.
Herbicide-tolerant (HT) crops contain traits that allow them to survive certain herbicides that previously would have destroyed the crop along with the targeted weeds.1 This allows farmers to use more effective postemergent herbicides, expanding weed-management options (Gianessi and Carpenter). Adoption of HT crops has risen dramatically, particularly for HT soybeans, since commercial availability in 1996. HT soybeans use rose quickly to about 17% of U.S. soybean acreage in 1997 and reached 81 % in 2003 (Fernandez-Cornejo and McBride; U.S. Department of Agriculture, National Agricultural Statistics Service).
A major element in assessing the farm-level impacts of HT crops is their microeconomic impact. Faced with reduced returns to crop production caused by low commodity prices, farmers were said to have viewed biotechnology as a potential means for reducing costs and/or increasing yields, thereby improving financial performance (Fernandez-Cornejo et al. 2002). In particular, rapid adoption of HT soybeans by U.S. farmers was seen as evidence that the perceived benefits of this technology outweighed the expected costs.
However, recent research showed no statistically significant differences between the net returns (both at the enterprise and whole-farm level) from using HT and conventional soybeans (Fernandez-Cornejo and McBride). This suggests that other considerations may be driving adoption. In particular, some researchers believe that adoption of HT soybeans is driven by the relative simplicity and flexibility of the weed-control program. HT programs allow growers to apply one herbicide product over the soybean crop at any stage of growth instead of using several herbicides to control a wide range of weeds "without sustaining crop injury" (Gianessi and Carpenter). In addition, using HT soybeans is said to make harvest "easier" (Duffy).
While difficult to measure, simplicity and flexibility translate into reduced management time employed to supervise production, freeing time for other uses (Fernandez-Cornejo and McBride). An obvious important alternative use of operators' time (and their spouses', if married) is off-farm employment. However, despite the likelihood of a strong interaction between adoption of management-saving agricultural technologies and off-farm employment by both the operator and his/her spouse, the role of off-farm activities has been largely neglected in studies of technology adoption in agriculture. Moreover, as Smith observed, standard measures of farm profitability, such as net returns (to management), give an incomplete picture of economic returns because they exclude the value of management time.
Using the appropriate measure of economic performance for farm households is essential, given the importance of off-farm income in U.S. agriculture. Made possible by alternative employment opportunities and facilitated by labor-saving technological progress, such as mechanization, off-farm work by farm operators and their spouses has risen steadily over the past decades, becoming the most important component of farm household income. As Mishra et al. show, total net income earned by farm households from farming grew from about $15 billion in 1969 to nearly $50 billion in 1999. However, off-farm earned income, which began at a roughly comparable figure in 1969 ($15 billion), soared to about $120 billion in 1999. Moreover, as Mishra et al. note, as women's wages have risen, married women have become more likely to work in the paid labor market and household tasks are now shared between spouses.
This article has two main objectives: (1) to develop an econometric model to examine the interaction of off-farm work and adoption of agricultural technologies and the impact of technology adoption on farm household income (from on-farm and off-farm sources) after controlling for such interaction and (2) to estimate the model for the case of adoption of HT soybeans using data from a nationwide survey of soybean farms for 2000. The main research question is whether or not HT adoption has a significant effect on each of the components of household income.
Mean Household Income for Adopters and Nonadopters of HT Soybeans
The household income measure used to examine the impact of adopting HT soybeans on household income has two components: household income from farming and off-farm household income. Actual mean household income, calculated directly from a nationwide U.S. Department of Agriculture (USDA) survey of soybean farmers in 2000, differs for adopters and nonadopters of HT soybeans. As shown in the table below, total household income is much higher for adopters than for nonadopters. Moreover, most of the difference is due to the off-farm income component.
These results are consistent with the notion that the management time saved by the adoption of HT soybeans is used to increase household off-farm activities, as discussed in the introduction. Thus, it appears that, to measure the economic benefits from HT soybean adoption, one also needs to look off the farm.
However, while illustrative, a comparison of means can only lead to a definite conclusion in an ideal experimental setting, where factors other than adoption are controlled by making them as similar as possible.2 Unlike controlled experiments, conditions other than the treatment are not equal in farm surveys. Thus, these differences in mean household income cannot necessarily be attributed to adoption of HT soybeans because survey results are influenced by many other factors, including operator characteristics and management practices. Moreover, farmers are not assigned randomly to the two groups (adopters and nonadopters of the HT technology), but make the adoption choices themselves. Therefore, adopters and nonadopters may be systematically different and these differences may manifest themselves in farm performance and could be confounded with differences due purely to adoption. This situation, called self-selection, would bias the statistical results, unless it is corrected.
For these reasons, an econometric impact model is specified, which statistically controls for factors considered relevant, and for which there are data, by holding them constant, so that the effect of adoption can be estimated. The model developed takes into consideration that farmers' adoption and off-farm employment participation decisions may be simultaneous. The model also corrects for self-selection to prevent biasing the results (Greene).
The Conceptual Framework
When an interior solution for M occurs, Equations (7) and (8) can be solved together, independently of the rest of the Kuhn-Tucker conditions, to obtain the demand functions for on-farm labor, i.e., the optimal production and consumption decisions can be separated because the off-farm wage determines the value of the operator's and spouse's time (W = μ/ λ) (Huffman 1991; Huffman and Lange).4
Based on the above theoretical discussion and given the cross-sectional nature of the data, one can use the implicit function theorem to deduce expressions for off-farm labor supply for farm operator and spouse and technology adoption (which affects off-farm labor supply of farm of operators and spouses) that are functions of wages, prices, human capital, nonlabor income, and other exogenous factors. These factors are replaced in reduced-form representations of labor supply and adoption by observable farm, operator, and household characteristics, including human capital. Ambient variables (family size, access to urban areas), which might affect the productive capacity of the farm operator and the spouse, are also included. The following section outlines the empirical model and estimation method used to conduct the analysis.
The Empirical Model
A two-stage econometric model is specified. The first stage, the decision model, examines the off-farm work participation decisions and the technology adoption decision. The second stage is used to estimate the impact of adoption on household income.
The Decision Model
The joint estimation of three or more probit equations was computationally unfeasible until recently because of the difficulty of evaluating high-order multivariate normal integrals. Over the past decade, however, the estimation has been made possible with Monte Carlo simulation techniques (Geweke et al.; Greene).
The vector Z, includes: (i) farm factors, such as farm size and complexity of the operations, (ii) human capital (operator age/experience and education), (iii) household characteristics (such as the number of children), (iv) off-farm employment opportunities, which will depend on the farms' accessibility to urban areas and the change in the rate of unemployment in nearby urban areas, (v) farm typology, (vi) government payments.5,6 The factors or attributes influencing adoption of HT soybeans, included in the vector Z^sub a^ include farm factors, human capital, farm typology, a proxy for risk (risk-averse farmers are less likely to adopt agricultural innovations, Fernandez-Cornejo et al., 1994), as well as crop and seed prices.
The Income Impact Model
The second stage is the income impact model, which provides estimates of the impact of adoption on household income after controlling for other factors. The empirical representation of this model based on Equation (17), the reduced-form expression of household income, is NI* = NI(W^sub x^, P^sub q^, P^sub g^, A, H, ψ, Γ, R, T).
To correct for self-selection bias, we follow Maddala (p. 260) and Greene (p. 642, 643) and obtain consistent estimates of the parameters θ and α by regarding self-selection and simultaneity (discussed earlier) as sources of endogeneity. Because the dummy variable I cannot be treated as exogenous, instrumental variable techniques are used to purge the dependence of I. The predicted probability of adoption, obtained from the decision model, is used as an instrument for I in Equation (21).
Note that, unlike the traditional selectivity model, in which the effects are calculated (separately) using the subsamples of adopters and nonadopters, the impact model uses all the observations and is known as a treatment-effects model, used by Barnow, Cain, and GoIdberger, and others. The treatment-effects model consists of the regression Y = θV' + αI + ε, where the observed indicator variable I (I = 1 if I* > 0 and I = 0 if I* ≤ 0), indicates the presence or absence of some treatment (adoption of HT crops in this case) and the unobserved or latent variable I* is given by I* = δZ'^sub α^ + v (Greene).
Total household income (NI*), as represented in Equation (17), has two components: household income from farming (FARMHHI) and off-farm household income (TOTOFI). Household income from farming includes farm business household income, operator-paid farm income, household members paid farm income, etc. (see detailed definitions in Table 1). Off-farm household income includes off-farm business income, income from operating other farm business, off-farm wages and salaries, etc. (Table 1).
The components of vector V include farm location and typology, operator age, education and experience, number of children, price of soybeans, a measure of specialization on soybean production, a measure of the extent of livestock operations, farm size, and proxies for local labor market conditions.
Data and Estimation
The model is estimated using data obtained from the nationwide Agricultural Resource Management Survey (ARMS) developed by the Economic Research Service (ERS) and the National Agricultural Statistics Service (NASS) of USDA and conducted in 2000 (USDA, ERS). The ARMS survey is designed to link data on the resources used in agricultural production to data on use of technologies (including the use of genetically engineered crops), other management techniques, chemical use, yields, and farm financial/economic conditions for selected field crops. The survey includes three phases (screening, obtaining production practices and cost data, and obtaining financial information). The ARMS is a multiframe, probability-based survey in which sample farms are randomly selected from groups of farms stratified by attributes such as economic size, type of production, and land use.
The data set includes 17 soybean producing states: Arkansas, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Mississippi, Michigan, Minnesota, Missouri, Nebraska, North Carolina, Ohio, South Dakota, Tennessee, and Wisconsin. After selecting those farms that planted soybeans in 2000 and eliminating those observations with missing data, there were 2,258 observations available for analysis. Table 2 shows the definitions as well as the sample averages of the main variables used in the model.
Because of the complexity of the survey design, a weighted least squares (WLS) technique is used to estimate the parameters using full-sample weights developed by the NASS of the USDA.
Two methods are used for the calculation of variances (and standards errors). The standard procedure is simple but does not account for the effect of stratification. However, as Carrington et al. note, "... ignoring stratification in the estimation of variances is not likely to cause substantial bias." The alternative method of variance estimation accounts for the survey design and involves a delete-a-group jackknife method. While this alternative method "can be used meaningfully in a remarkably broad range of settings" in complex survey designs, such as ARMS (Kott 1998), it has also a drawback. It is overly conservative (estimated variances and standard errors are higher than the true variances and standard errors; Kott, 1998), which may underestimate the significance of a variable under some circumstances.
The alternative method follows the logic of the standard jackknife method except that a group of observations is deleted in each replication. It consists of partitioning the sample data into r groups of observations (r = 15 in this survey) and resampling, thus forming 15 replicates and deleting one group of observations in each replicate (Kott; Kott and Stukel; Rust). A set of sampling weights is calculated by NASS for each replicate. The model is run first with the full-sample weights to obtain the parameter estimates. The model is then run 15 additional times (using each of the 15 replicate weights) and the vector of parameters obtained in each case is compared with the full-sample parameter vector to calculate the standard errors.7
Decision Model Results
The maximum likelihood estimates of the decision model, i.e., the three-equation multivariate probit model (Equations [20a]-[20c]) are shown in Table 3. This decision model is mainly used to estimate the predicted probabilities of adoption, accounting for the interaction with off-farm employment decisions. Table 3 also shows the standard errors and f-statistics calculated using the standard variance calculation procedure and the delete-agroup jackknife variance calculation procedure.
Beginning with the operator's off-farm work participation decision and considering the significant variables under the standard variance calculation procedure, the operator's decision to work off-farm is positively related to age but negatively related to age squared, indicating that off-farm work participation increases with age up to a certain point and then declines. Operator's off-farm work is also positively related to his/her education, to the operator's spouse making day-to-day decisions in the farm, and to two farm-typology variables (operating residential and limited-resource farms). On the other hand, the operator's decision to work off-farm is negatively related to farm size and complexity (as measured by the number of commodities produced), to the number of children in the household, and to increases in unemployment in areas within commuting distance from the farm. The operator's off-farm work decision is also negatively related to the share of the farm's land owned by the operator, but this relationship is not statistically significant (p-value = 0.14). The operator's off-farm work decision is not significantly related to the farm being located in a particular region of the country.
The off-farm work participation decision of the operator's spouse is positively related to age and negatively related to age squared, indicating that spouse's off-farm work participation also increases with age up to a certain age and then declines. The spouse's off-farm work decision is also positively related to operating residential farms (typology variable). The spouse's off-farm work decision is negatively related to the spouse making day-to-day decisions in the farm and it is also negatively related to farm size, but, unlike the operator's case, it is not significantly related to farm complexity, number of children in the household, and changes in unemployment within commuting distance from the farm. Also, the spouse's off-farm work decision is negatively related to the land ownership share but, unlike the operator's case, this relationship is statistically significant. On the other hand, like the operator's, the spouse's off-farm work decision is not significantly related to location in a particular region of the country.
Adoption of HT soybeans is significantly positively related to age (but negatively related to age squared), to location in the heartland, and to the price of soybeans. Adoption is negatively related to farm size, to the number of children in the household, to operating retirement farms (typology variable), and to the percent of land owned by the operator.
Under the most strict, although perhaps too conservative, delete-a-group jackknife variance calculation procedure, the operator's decision to work off-farm is negatively related to farm size and complexity (as measured by the number of commodities produced) and positively related to the operators' education and to farm typology variables (retirement, residential, and limited-resource farms). The operator's spouse's off-farm work participation decision is negatively related to farm size, but, unlike the operator's case, it is not significantly related to farm complexity. The spouse's off-farm work decision is also positively related to operating residential farms. On the other hand, the spouse's off-farm work is positively related to age and negatively related to age squared, indicating that the spouse's off-farm work increases with age but only up to a certain point.
Impact Model Results
The results of this model are shown in Table 4. The relationship of adoption of HT soybeans with off-farm household income is positive and statistically significant under both the standard and jackknife variance calculation procedures (Table 4). The elasticity of offfarm household income with respect to the probability of adoption of herbicide-resistant soybeans (calculated at the mean) is +0.843. That is a 10% increase in the probability of adoption of HT soybeans is associated with an increase in off-farm household income of 8.4%.8 On the other hand, adoption of HT soybeans did not have a significant effect on household income from farming under either the standard or jackknife variance calculation procedures (Table 4).
Adoption of HT soybeans is positively and significantly associated with total household income (from off-farm and on-farm sources). The calculated elasticity of total household income with respect to the probability of adoption of herbicide-resistant soybeans (calculated at the mean) is +0.643. This means that a 10% increase in the probability of adoption of herbicide-resistant soybeans is associated with an increase in total household income of 6.4%.
Regarding the influence of farm size on off-farm household income, we find that, after controlling for other factors, both the linear and quadratic coefficients are significant for farm size (Table 4), implying that income from off-farm sources increases with size, but only up to a certain point, at which off-farm income reaches a maximum, and then declines as size increases further. This maximum was reached at a size of about 2,670 acres, which is five times the average farm size in the sample. On the other hand, the relationship between size and income from on-farm sources was not significant after controlling for other factors (Table 4). Because of the predominance of offfarm income, total household income increases with size but up to a maximum of 1,846 acres, which is lower than the size at which the maximum for off-farm income was obtained.
This article develops a model that incorporates the technology-adoption decision into the agricultural household. Two contributions to traditional research on technology-adoption analysis are introduced in this article: the unit of analysis is the household rather than the farm business and the metric for economic performance is household income rather than farm profits.
This article finds statistically significant relationships between off-farm income, adoption of HT soybeans, and structural characteristics, such as farm size. U.S. farm operators and their spouses are more likely to work off-farm and together are more likely to obtain a higher household income from off-farm sources when they operate small soybean farms (lacking economies of scale). Adoption of HT soybeans is significantly and positively associated with off-farm household income for U.S. soybean farmers, after controlling for other factors. In addition, while on-farm household income is not significantly affected by adoption, there is a significant relationship between adoption and total household income. Thus, our findings also suggest a rationale for the rapid HT adoption by U.S. farmers. Farmers may adopt herbicide-tolerant soybeans because the simplicity and flexibility of the weed-control program permits them to save management time, allowing farmers to obtain a higher income from off-farm activities.
Interpreted in a broader sense, our findings illustrate the importance of accounting for household and firm interactions in modeling farmer adoption decisions. In particular, our findings suggest that the tradeoff between the time spent working on the farm and off the farm translates into a substitution of economies of scope (derived from engaging in multiple income-generating activities, on and off the farm) for economies of scale. Thus, our findings appear to provide empirical confirmation to Kitty Smith's observation that, like the economists' perceived link between capital intensity and scale dependency of technologies, "... perhaps management intensity should also be viewed as a potential source of scale bias."
[Received January 2005; Accepted March 2005.]
1 The most common HT crops are resistant to glyphosate, an herbicide effective on many species of grasses, broadleaf weeds, and sedges. Glyphosate tolerance has been incorporated into soybeans, corn, canola, and cotton. Other genetically modified HT crops include corn resistant to glufosinate ammonium.
2 For example, means can be compared for yields of two groups of soybean plots that are equal in soil type, rainfall, sunlight, and all other respects, except that one group receives a treatment (e.g., uses HT varieties) and the other group does not. As an alternative to controlled experiments, the subjects that receive treatment and those that don't can be selected randomly.
3 The marginal value of time of the farm operator (or spouse) when all his/her time is allocated to farm work and leisure and none is allocated to off-farm work (P^sub q^(∂Q/∂F^sub i^\M^sub i=o^), represents the shadow value of farm labor and is called the reservation wage for offfarm work for the operator (i = o) or spouse (i = s). In this context, the operator (or spouse) will work offfarm when his/her reservation wage is less than the anticipated off-farm wage rate and will not work offfarm otherwise. Note also that assuming that both the operator and spouse face wages that are dependent on their marketable human capital characteristics H, local labor market conditions and job characteristics Ω, but not on the amount of off-farm work (Huffman, 1991; Huffman and Lange, 1989; Tokle and Huffman, 1991), the off-farm market labor demand functions are W^sub i^ = W^sub i^(H, Ω) (i = o, s).
4 Moreover, when an interior solution occurs, from Equations (10), (11a), and (11b), we obtain V^sub L^/U^sub G^ = W/P^sub g^; that is, the marginal rate of substitution of between consumption goods and leisure is equal to the ratio of the wage rate and the price of the consumption goods.
5 Following Goodwin and Holt (2002), some prices are not included in our empirical models because prices are approximately constant across households when data consist of cross-sectional observations taken at a point in time.
6 Farm typology classification is based on the occupation of the farm operator and includes mutually exclusive typology categories, such as limited-resource, retirement, residential lifestyle, or a nonfamily farm. Limited-resource farms are constrained by low levels of assets and household income. Retirement farms are those with operators who report that they are retired (excluding limited-resource farms). Residential lifestyle farms are those with operators who report a major occupation other than farming (excluding limited-resource farms) (Hoppe et al.).
8 Results are typically expressed as elasticity-the percentage change in a particular effect (e.g., income) relative to a small percentage change in adoption of the technology from current levels. The results can be viewed in terms of the aggregate effect (across a region or sector) from aggregate increases in adoption. However, in terms of a typical farm-that has either adopted or not, the elasticity is usually interpreted as the (marginal) farm-level effect associated with an increase in the probability of adoption, away from a given, e.g., current level of adoption.
Carrington, W.G., J.L. Eltinge, and K. McCue. An Economist's Primer on Survey Samples. Center for Economic Studies, U.S. Bureau of the Census CES 00-15, October 2000. http://148.129. 75.160/paper.php?paper= 101613
Duffy, M. "Who Benefits from Biotechnology?" Paper presented at the American seed Trade Association meeting, Chicago, IL, December 5-7, 2001.
Fernandez-Cornejo, J., E.D. Beach, and Wen-Yuan Huang. "The Adoption of IPM Techniques by Vegetable Growers in Florida, Michigan, and Texas." Journal of Agricultural and Applied Economics 1(1994): 158-72.
Fernandez-Cornejo, J., C. Klotz-Ingram, and S. Jans. "Farm-Level Effects of Adopting Herbicide-Tolerant Soybeans in the U.S.A." Journal of Agricultural and Applied Economics 34(1)(2002):149-63.
Fernandez-Cornejo, J., and W.D. McBride. Adoption of Bioengineered Crops. Agricultural Economic Report No. 810. U.S. Department of Agriculture, ERS, Washington, DC, 2002.
Geweke, J., M. Keane, and D. Runkle. "Alternative Computational Approaches to Inference in the Multinomial Probit Model." Review of Economics and Statistics 76(4)(1994):609-32.
Gianessi, L.P., and J.E. Carpenter. Agricultural Biotechnology: Insect Control Benefits. National Center for Food and Agricultural Policy, Washington, DC. 1999. http://www.bio.org/food&ag/ ncfap.htm.
Goodwin, B.K., and M.T. Holt. "Parametric and Semiparametric Modeling of the Off-Farm Labor Supply of Agrarian Households in Transition Bulgaria. American Journal of Agricultural Economics 84(2002): 184-209.
Goodwin, B.K., and A.K. Mishra, "Farming Efficiency and the Determinants of Multiple Job Holding by Farm Operators." American Journal of Agricultural Economics 86(2004): 722-29.
Greene, W.H. Econometric Analysis, 3rd ed. Upper Saddle River, NJ: Prentice-Hall, 1997.
Hoppe, R.A., J. Perry, and D. Banker. "ERS Farm Typology: Classifying a Diverse Ag. Sector." Agricultural Outlook, AGO-266, Department of Agriculture, ERS, Washington, DC. 1999.
Huffman, W.E. "Farm and Off-Farm Work Decisions: The Role of Human Capital." Revue of Economics and Statistics 62(1)(1980): 14-23.
Huffman, W.E. "Agricultural Household Models: Survey and Critique." Multiple Job Holding Among Farm Families. M.C. Hallberg, J.L Findeis, and D.A. Lass, eds. Ames: Iowa State University Press, 1991.
Huffman, W.E., and M.D. Lange. "Off-Farm Work Decisions of Husbands and Wives: Joint Decision Making." Revue of Economics and Statistics 71(1989):471-80.
Huffman, WE., and H. El-Osta. Off-Farm Work Participation, Off-Farm Labor Supply and OffFarm Labor Demand of U.S. Farm Operators. Staff paper 290. Ames: Iowa State University, 1997.
Kimhi, A. "Quasi Maximum Likelihood Estimation of Multivariate Probit Models: Farm Couples' Labor Participation." American Journal of Agricultural Economics 76(1994):828-35.
Kimhi, A. "Family Composition and Off-Farm Participation Decisions in Israeli Farm Households." American Journal of Agricultural Economics 86(2004):502-12.
Kott, P.S., and D.M. Stukel. "Can the Jackknife Be Used With a Two-Phase Sample?" Survey Methodology 23(2)(1997):81-89.
Kott, RS. Using the Delete-A-Group Jackknife Variance Estimator in NASS Surveys. RD Research Report RD-98-01, Washington, DC. USDA, NASS, 1998.
Lass, D.A., J.L. Findeis, and M.C. Hallberg. "OffFarm Labor Employment Decisions by Massachusetts Farm Households." Northeastern Journal of Agricultural and Resource Economics 18(1989):149-59.
Lass, D.A., and C.M. Gempesaw II. "The Supply of Off-Farm Labor: A Random Coefficients Approach." American Journal of Agricultural Economics 74(1992) :400-11.
Maddala, G.S. Limited-Dependent and Qualitative Variables in Econometrics. Cambridge, UK: Cambridge University Press, 1983.
Mishra, A.K., H.S. El-Osta, MJ. Morehart, J.D. Johnson, and J.W. Hopkins. Income, Wealth, and the Economic Well-Being of Farm Households. Agricultural Economic Report 812. U.S. Department of Agriculture, ERS, Washington, DC, 2002.
Rust, K. "Variance Estimation for Complex Estimators in Sample Surveys." Journal of Official Statistics (1)(1985):381-97.
Singh, I., L. Squire, and J. Strauss. (Eds.), Agricultural Household Models: Extensions, Applications, and Policy. Baltimore, MD: Johns Hopkins University Press, 1986.
Smith, K.R. "Does Off-Farm Work Hinder 'Smart' Farming?" Agricultural Outlook AGO-294, (2002):28-30.
Stefanides, Z., and L.W. Tauer. "The Empirical Impact of Bovine Somatropin on a Group of New York Dairy Farms." American Journal of Agricultural Economics 81(1999):95-102.
Tokle, J.G., and W.E. Huffman. "Local Economic Conditions and Wage Labor Decisions of Farm and Rural Nonfarm Couples." American Journal of Agricultural Economics 73(1991):52-70.
U.S. Department of Agriculture, Economic Research Service. Agricultural Resource Management Survey (ARMS). Briefing Room. March 11, 2003. http://www.ers.usda.gov/Briefing/ARMS/
U.S. Department of Agriculture, National Agricultural Statistics Service. Prospective Plantings. Washington, DC, March 31, 2003, pp. 20-21. http://usda.mannlib.cornell.edu/reports/nassr/ field/pcp-bbp/psp10303.txt
Jorge Fernandez-Cornejo and Ashok Mishra are economists with the Economic Research Service (ERS), U.S. Department of Agriculture. Chad Hendricks was an intern at ERS at the time of the study.
The views expressed are those of the authors and do not necessarily represent the views or policies of ERS or the U.S. Department of Agriculture.…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Technology Adoption and Off-Farm Household Income: The Case of Herbicide-Tolerant Soybeans. Contributors: Fernandez-Cornejo, Jorge - Author, Hendricks, Chad - Author, Mishra, Ashok - Author. Journal title: Journal of Agricultural and Applied Economics. Volume: 37. Issue: 3 Publication date: December 2005. Page number: 549+. © Southern Agricultural Economics Association Apr 2008. Provided by ProQuest LLC. All Rights Reserved.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.