Introduction
Since the 1990s, rapid growth in farmers’ markets, community supported agriculture (CSA) farms, and local cooperative supermarket sales has been observed across the USA, with thousands of small, diversified vegetable farms using these direct market opportunities to sell their produce (Uva, Reference Uva2002; USDA, 2009a, 2013; Johnson et al., Reference Johnson, Cowan and Aussenberg2012). Programs supporting local food systems, such as US Department of Agriculture-Agricultural Marketing Services (USDA-AMS) ‘Know your Farmer, Know your Food’ and various state-sponsored ‘Buy Local’ initiatives, increasingly advocate for smaller scale vegetable growers to use direct marketing strategies (USDA, 2009b; WI-DATCP, 2013). These programs implicitly assume that direct marketing enables farmers to produce smaller volumes of vegetables while capturing a larger profit margin (net return per dollar of sales). However, research investigating the ability for direct marketing channels to contribute to profitable farm businesses varies in its conclusions (Hardesty and Leff, Reference Hardesty and Leff2010; LeRoux et al., Reference LeRoux, Schmit, Roth and Streeter2010; Silva et al., Reference Silva, Dong, Mitchell and Hendrickson2014).
To increase the success of diversified farms selling into local food systems, farmers must have the ability to appropriately price their crops, accounting for production costs including their labor as a farm operator. Economic decision-making tools, such as enterprise budgets, can help farmers estimate the costs and returns associated with the production of specific crops and estimate break-even prices (Chase et al., Reference Chase, Smith and Delate2006). However, for these assessments to be accurate, farmers must have specific knowledge of their production and marketing costs (Chase, Reference Chase2008).
Labor expenses are the major cost associated with production and marketing on diversified vegetable farms and may account for a significant proportion of a farm's deviation from the financial estimates derived from standardized enterprise budgets (Hendrickson, Reference Hendrickson2005; Hardesty, Reference Hardesty2007; Ali and Lucier, Reference Ali and Lucier2008; LeRoux et al., Reference LeRoux, Schmit, Roth and Streeter2010). Labor efficiencies can vary widely across farms (Lohr and Park, Reference Lohr and Park2009); a University of Wisconsin-Madison study documented that the total labor hours required for production of diversified vegetables can vary from 187 to 1211 labor hours per ha (Hendrickson, Reference Hendrickson2005). This variability can significantly impact the accuracy of generalized enterprise budgets as compared with the realized values for a specific farm, highlighting the importance of farm-specific labor values in determining farm budgets and related pricing decisions. However, capturing accurate labor inputs can be difficult, with retrospective estimates of labor inputs often resulting in inflated and inaccurate values. More frequent documentation in the form of time diaries has been demonstrated to produce better measurements, however, than retrospective accounting (Juster and Stafford, Reference Juster and Stafford1991).
In order to provide strategies to assess labor costs, we undertook a participatory research project with organic farmers to capture labor hours required throughout the production season using a time diary approach. The primary objectives of the project were to assess the process of data collection as well as document the range of production and postharvest processes, with the goal of better refining data collection recommendations to farmers and establishment of benchmark labor input values. Farmers and employees on working organic vegetable farms (both wage and non-wage employees) collected labor hours associated with field production, harvesting and packing of specific crops grown during three production seasons. Additionally, farmer observations and experiences from the participatory data collection effort were captured in order to develop best practice recommendations for other farmers undertaking labor data collection efforts.
Materials and Methods
To compare the labor hours required to produce, harvest and pack individual crops on diversified organic vegetable farms, we began an intensive labor data collection effort in partnership with 12 organic farmers across the upper Midwestern USA. Farmers were self-selected based upon their willingness to record production, harvesting and packing labor inputs for nine crops throughout the growing season. Seven farms were located in Wisconsin, two in North Dakota and one each in Minnesota, Iowa and Illinois. The 12 farms included in this study varied by size of operation, with six of the farms totaling 2 ha or less of production area, three farms totaling 2–27 ha of production area, and three farms totaling 28–40 ha of production area. Crops produced and marketing channels used were numerous, with farms growing 24–37 different crops and selling into CSA, farmers’ markets, farm stands, direct wholesale markets, restaurants, local institutions and cooperative grocers. Farmers focused their data collection efforts on nine crops commonly grown on diversified vegetable farms in the upper Midwestern USA: (snap beans (Phaseolus vulgaris), broccoli (Brassica oleracea var. botrytis), carrots (Daucus carota), garlic (Allium sativum), lettuce (Lactuca sativa), onion (Allium cepa), potatoes (Solanum tuberosum), spinach (Spinacia oleracea), and tomatoes (Solanum lycopersicum)). These crops differ in their production requirements and growth habits, with some crops (i.e., tomatoes) requiring very intensive and specific field activities such as trellising and mulching.
Data were compiled using a participatory research model, with farmers and researchers partnering on the data collection and analysis. We followed methodology aligning with the ‘time diary method’ for behavioral measurement (Niemi, Reference Niemi1993), which has been found to yield data consistent with respondent behavior. Throughout three growing seasons (2010–2012), farmers and/or farm employees recorded the amount of time each employee (including the primary farmer and worker shares, if applicable) worked to complete field production and harvest/pack labor tasks directly related to the nine vegetable crops. Farmers received one-on-one training by a member of the university research team at the beginning of the production season, outlining the types of tasks for which labor data should be reported and the frequency of reporting. Each farm completed daily labor data collection as best appropriate for their farm crews; in some cases, farmers or farm crew leads collected data daily for all farm employees, while in other cases, each individual employee recorded their individual time spent on a task. Labor data were returned to the research team and standardized to represent the total labor hours contributed by the equivalent of one individual employee per 100 m of crop row of production or harvested area. Field production and harvest/packing labor hours per 100 m of crop row were averaged both each year and across years. The coefficient of variation for each crop was calculated and reported as a percent. All data were analyzed using the Wilcoxon–Mann–Whitney rank sums test, with a subsequent post hoc analysis conducted using Wilcoxon's matched pairs, sign-rank test, with P-values of a = 0.05 considered significant. Following the 3-year data collection exercise, informal exit interviews were conducted with each of the farmers included in the participatory research effort. Farmers were asked to provide their perspectives on successes, failures and utility of the data collection exercise, as related to the impact on their farm management decisions and ability to assess their cost-of-production and profitability.
Results
Labor data collection
Season-long labor inputs required for in-field production of nine vegetable crops over the 3 years of this study varied from 2.3 to 29.8 h 100 m−1 of crop row (Table 1). Averaged over all growing seasons, tomatoes required the greatest number of field production labor hours per row m, averaging 15.1 h 100 m−1 of crop row per production season. Coefficients of variation (cv) differed between crops and between years. For example, in 2010, the cv calculated for hours per 100 m of crop row for field production of onions on the 12 farms was low, at 11.4%; conversely, in 2012, the coefficient of variation for this same crop was 119.8%, demonstrating high variability between farms in the number of hours per 100 m of crop row needed for in-field production. Similarly, labor inputs required for harvest and packing of the nine crops varied widely, from 1.6 to 22.6 h 100 m−1 of crop row of production (Table 2). Averaged over all seasons, tomatoes required the greatest number of harvest and pack labor hours per 100 m of crop row of production as compared with broccoli (17.1 h 100 m−1 of crop row versus 3.3 h 100 m−1 of crop row), although not differing statistically from the labor required for the harvest and pack of the other crops. As with hours of labor required for crop production in the field, high variability was observed in the calculated coefficients of variation for labor hours required for the harvest and pack of 100 m of crop row of crop, both between crops and between years.
Table 1. Comparison of seasonal mean labor hours required for field production (labor hours per 100 m of crop row of production) of nine vegetable crops (snap beans, broccoli, carrots, garlic, lettuce, onion, potatoes, spinach and tomatoes) grown on 12 organic diversified vegetable farms in the upper Midwestern USA, 2010–2012. Labor hours were compiled by recording all time contributed by each farmer, farm employee and worker share related to the field production activities of each specific crop up until the time of harvest.
h, hours; cv, coefficient of variation.
1 Value represents hours of labor per 100 row meter required by a single farm employee.
2 Coefficient of variation.
3 Mean ± Standard Error of the Mean.
4 Numbers in columns followed by a different lowercase letter were significantly different at P < 0.05 according to post hoc analysis using Wilcoxon's matched pairs, sign-rank test.
Table 2. Comparison of seasonal mean labor hours required for harvest and packing (labor hours per 100 m of crop row of production) of nine vegetable crops (snap beans, broccoli, carrots, garlic, lettuce, onion, potatoes, spinach and tomatoes) grown on 12 organic diversified vegetable farms in the upper Midwestern USA, 2010–2012. Labor hours were compiled by recording all time contributed by each farmer, farm employee and worker share related to the harvesting and packing activities of each specific crop up until marketing of the crop.
h, hours; cv, coefficient of variation.
1 Value represents hours of labor per 100 row meter required by a single farm employee.
2 Coefficient of variation.
3 Mean ± Standard Error of the Mean.
4 Numbers in columns followed by a different lowercase letter were significantly different at P < 0.05 according to post hoc analysis using Wilcoxon's matched pairs, sign-rank test.
To compare the impact of both farm size and sales volumes (as a proxy for harvested yield) on labor efficiency, farms were grouped into categories representing three scales: small market farms (2 ha or less), mid-sized farms (2–10 ha), and larger farms (28–40 ha) (Table 3). Hours required for field production and harvest and packing tended to decrease as farm size increased, possibly reflecting the higher levels of mechanizations and additional efficiencies allowed by larger crew sizes. However, with the exception of field production of broccoli where smaller farms appeared to require significantly higher labor inputs, these differences were not significant. Sales volumes for each crop across farm sizes did not significantly differ, nor demonstrate any clear trends with labor inputs.
Table 3. Comparison of mean labor hours contributed by each farmer, farm employee and worker share required for field growing, and harvest/packing per 100 m of crop row of seven crops and their sales volumes (as kg produce per 100 m of crop row) on 12 organic diversified vegetable farms of three size categories (small, medium and large) in the upper Midwestern USA, 2010–2012.
1 ‘Small’ represents farms of 0–2 ha scale.
2 ‘Medium’ represents farms of 227 ha scale.
3 ‘Large’ represents farms of 28–40.5 ha scale.
4 Value represents hours of labor per 100 m of crop row performed by a single farm employee.
5 Mean ± Standard Error of the Mean.
6 Numbers in rows followed by a different lowercase letter were significantly different at P < 0.05 according to post hoc analysis using Wilcoxon's matched pairs, sign-rank test.
7 Value reported as heads sold per 100 m of crop row of production.
Farmer experiences
Farmers reported several strategies to facilitate labor data collection on their farms. One farmer integrated labor data collection into his CSA inventory system, with employees supplying labor data to the dock manager. Work team leaders, each spent 8 min per day completing labor hour forms, followed by a single employee completing the crop-by-crop analysis at the end of the production and harvest seasons, requiring 50 h. A second farmer reported that the crew recorded their work hours as a group after a specific task was completed. Another farm linked labor data recording to payroll entry, requiring employees to report labor hours in a detailed fashion on their timesheets.
Farms of differing scales reported unique challenges associated with labor data collection. Larger farms with more employees found greater difficulty ensuring that all workers consistently maintained the necessary records. However, some larger farms designated this responsibility to a manager or payroll coordinator, which created more oversight in the collection of data. In contrast, while small farms had fewer employees, the farm owner, often solely responsible for the majority of production and marketing work, may have limited resources to oversee the recording of labor hours.
Farmers reported several different outcomes resulting from data collection activities and subsequent analysis. By allowing for the calculation of detailed and accurate cost-of-production figures, farmers could more effectively price individual crops in each of their markets. One of the interviewed farmers used the data to negotiate a better price for his product, citing more effective negotiations with the availability of ‘real, substantive information to show buyers’. Another grower, upon evaluation of the data and its contribution to cost of production, discovered that several of his wholesale crops were either losing money or breaking even. After assessing the most profitable crops for wholesale markets, he shifted his crop mix to expand production of those crops for both wholesale and CSA, while restricting less profitable crops solely to his CSA market. Farmers also reported that data collection allowed for more effective labor management. One farmer noted that due to high labor costs, the most sensible financial decision may be to leave a low-yielding crop unharvested, as the cost to harvest over a widely-dispersed area would exceed the price obtained for that crop at market. Additionally, with data clearly illustrating the impact of labor costs on profit margins, farmers also highlighted the importance of a well-trained, efficient labor crew. One farmer commented on the value of the data with respect to formulated informed management decisions, including the assignment of more efficient crew members to certain crops, where labor inputs were high and margins were low. Furthermore, another farmer commented on the value of the data for machinery purchase decisions, allowing him to assess of the cost of a piece of equipment and subsequent depreciation against the cost of maintain the labor crew.
Discussion
Organic vegetable farms in the upper Midwestern USA integrate a wide range of production strategies into their farm operations. As such, labor inputs required to grow, harvest and pack crops across farms can vary widely. While accurate calculations of break-even prices are essential to assess profitability, the determination of the cost of production for specific crops remains a significant challenge on diversified organic vegetable farms. The real-world data collected from this study, in addition to our ongoing efforts, indeed highlight the significant variation existing in labor inputs required to complete activities associated with the production, harvest and packing of vegetable crops commonly grown on diversified organic farms. With this high level of variability, for farms to accurately assess their cost of production and break-even prices, farm-specific assessment of labor inputs is crucial. Additionally, while general benchmark values for labor inputs for specific crops might serve to guide farmers as to their general labor efficiencies for various crops, benchmark values may not provide the resolution needed to allow farmers to make more informed pricing decisions.
The farmers involved in this study expressed frustration regarding the level of effort and coordination required for the collection of these values throughout the production season. Although recordkeeping is an integral part of the organic certification process, farmers varied greatly in their approaches and attentiveness to recordkeeping. Despite efforts to simplify record-keeping approaches throughout this project with farmer input, labor values were difficult to capture with multiple paid and volunteer employees, frequent transitions between activities and difficulty in recording information in the field. However, our ongoing efforts to compare labor inputs across a wide range of organic vegetable farms demonstrate that identifying methods for farm businesses to accomplish farm-specific data collection is essential for accurate cost-of-production calculations. Additional tools that integrate with other farm recordkeeping activities, such as employee time sheets or organic certification records, may facilitate labor data collection.
While the time diary methods for labor data collection described in this paper offer one approach for farmers to more precisely calculate their costs of production, with the substantial degree of effort and commitment required to complete data collection throughout the production season, it may not be the optimal approach for all farms. Ongoing and future research will continue to investigate strategies to facilitate on-farm data collection allowing for more accurate estimations of cost-of-production. This may include more manageable data collection approaches, such as focus on shorter periods of intensive data collection specific to certain crops and activities, grouping of crops based on similar production and harvest strategies, and refining benchmark values based on level of farm mechanization. Additionally, more efficient recordkeeping and data analysis tools, such as the Veggie Compass tool developed by the project team (http://www.veggiecompass.com), will allow farmers to derive the most value from these data collection efforts.
Acknowledgement
This research was supported by the US Department of Agriculture Risk Management Agency and the Ceres Trust.
