Data collection.

The data for this analysis came from a 2012 web-based survey of fruit and vegetable farmers who were part of the Food Industry MarketMaker database. The database provided us with 4312 mail addresses and 3015 e-mail addresses of growers located in 16 states (Alabama, Arkansas, District of Columbia, Florida, Georgia, Iowa, Illinois, Indiana, Kentucky, Michigan, Mississippi, Nebraska, New York, Ohio, Pennsylvania, and South Carolina). This database facilitated the access to direct-to-consumer (DTC) markets channels by connecting farmers and buyers. Thus, farmers in our sample tended to have small- and medium-sized operations and sell directly to consumers.

We conducted a web-based survey using a mixed-mode design. To increase participation rate, we included a two-dollar bill in the invitation letter and sent e-mail reminders with intervals of two weeks. We received 1559 farmer responses, an effective response rate (36%) for similar online surveys (Dillman et al., 2014). The survey asked for farm characteristics, farmer’s demographics, source of information, and beliefs and perceptions toward organic agriculture and certification. We also asked farmers the percentage of their farm that was conventional, certified organic, transitioning to organic, or using organic practices but not certified.

The subsample focused on for the study included 234 farmers. Of them, 64% farmers were organic certified (149) and 36% were certified but decided to drop out from certification (85). To obtain a clear analysis of the drivers of organic decertification, the subsample excluded conventional farmers and organic farmers who were never certified.

Model specification.

We used a standard probit regression to assess the drivers of organic decertification. The probit estimated how farmer’s demographics, farm characteristics, sources of information, and farmer’s perceptions drove the decision to decertify. For instance, certified farmers selling through DTC market channels may have perceived that alternative labels (e.g., Certified Naturally Grown and Indiana Grown) were less costly or more effective at capturing price premiums. Alternatively, larger certified farmers selling wholesale may have perceived they needed the organic label to capture price premiums and decided to remain certified. The study also provided estimates of a robust probit regression. We used robust standard regressions to deliver valid statistical inferences in case the survey yielded variables that were independent but not identically distributed. The robust probit can also help address potential heteroscedasticity (Wooldridge, 2010).

The standard probit was based on an observed binary variable y and latent continuous variable . Although y^* was not observed, we were able to observe

View Expanded

Given the variables y and y^*, we can estimate:

View Expanded

Equation [1] illustrates the standard probit regression. The dependent variable was the binary decision to decertify among organic certified farmers. We grouped producers in two categories. The first category was the group of farmers who were USDA organic certified or were in the 3-year transition period to certification. The second group was the group of farmers who were certified and dropped the certification program. Thus, the dependent variable decertified = 1 if the farmer decertified and decertified = 0 if the farmer was certified or in transition to certification at the moment of the survey.

View Expanded

where is the standard normal cumulative distribution function, X = (1, dtc, X₂) is a vector of covariates, and is a vector of unknown constants.

Table 1 illustrates the set of covariates in Eq. [1]. A correlation test performed to the set of covariates in Eq. [1] indicated the lack of correlation. The covariates demog corresponded to demographic factors such as educational attainment, gender, race, if the respondent farmed part-time, and farm location. The study grouped farmers in four geographic regions such as South, Delta, Northeast, and Midwest. The South region corresponded to operations located in Florida, Georgia, and South Carolina. The Delta region grouped farms located in Alabama, Arkansas, Mississippi, and Louisiana. Farms in the Northeast region were located in New York and Pennsylvania. Last, Midwest farms were located in Iowa, Illinois, Indiana, Michigan, Nebraska, Ohio, and Kentucky.

Table 1.

Categories and descriptions of the variables used to investigate the drivers of organic decertification among fruit and vegetable farmers of 16 states in the United States.

View Table

The set of variables farm illustrated farm characteristics. We categorized farm size as exempt, small, and medium according to their gross annual sales. We followed Torres et al. (2016) to base sale cutoffs for farm size categories. Exempt farmers reported sales less than $5,000 in gross annual sales, small farms reported sales between $5,000 and $50,000, and medium and large farms were those with gross annual sales of more than $50,000. Farm characteristics also included the number of crops, if the farmer grew vegetable crops, if the farmer used DTC market channels, the average distance to markets, the number of acres, if the farmer rented land, if the business structure of the farm was a sole proprietorship, the number of years farming, and the average hours per week the farmer worked on farm business.

The set of variables information included the sources of information that farmers found useful. The list of sources of information include price reporting services, directories of product buyers, university extension, farmers and growers’ associations, organic certifiers, and wholesalers or retailers. We followed Constance and Choi (2010), Sierra et al. (2008), Walz (2004), and York et al. (2007) to analyze the farmers’ attitudes and beliefs toward the organic industry. For example, the set of variables perceptions included farmers’ attitudes and beliefs toward organic markets reliability, the riskiness of organic agriculture, the perception of organic as a diversification strategy, the uncertainty to obtain organic price premiums, and if the farmer supported the philosophy of organic farming. The index variable bcerti measured farmers’ perceptions toward the process of organic certification. The use of an index helped minimize the number of explanatory variables and the number of missing observations (Kim and Mueller, 1978). This index included farmer’s perception that certification was a barrier due to loss of freedom, paperwork, cost of certification, interaction with the certifier, and lack of information. The index grouped the variables based on a 3-point Likert-like scale from not a barrier (1) to a severe barrier (3).

Summary statistics.

Table 2 provides the mean differences for all the variables used in the model, by type of producer. More than 63% of farmers in the sample had college or postgraduate education, with a similar proportion among certified farmers and those who decertified. Only 38% of farmers were female, with similar percentages for both certified and decertified growers. These results are consistent with Walz (2004), who found that women tend to be a minority among organic farmers.

Table 2.

Comparison of respondents of the Purdue 2012 survey of MarketMaker growers categorized by producer type (full sample, N = 234; decertified farmers, N = 85; certified farmers, N = 149).

View Table

There were significantly more decertified farmers (56%) located in the Midwest region in states including Iowa, Illinois, Indiana, Michigan, Nebraska, Ohio, and Kentucky, when compared with certified farmers (36%) (P < 0.01). The concentration of decertified farmers in the Midwest may be a response of farmers to the increased popularity of DTC outlets and initiatives to promote local markets. For instance, we found that a third of Midwest decertified farmers sell directly to consumers, compared with only a fourth of certified farmers. On the other hand, certified farmers are concentrated in the Northeast region in states such as Pennsylvania and New York (P < 0.01). Guptill (2009) reported that organic certified agriculture is attractive to New York farmers and Dimitri and Greene (2002) illustrated that organic production in New York rated high compared with the rest of the country. High-volume and high-value markets in large population centers may be offering economic incentives to organic certified fruits and vegetables operations in the Northeast.

Less than a third of farmers in our sample operated part-time. Part-time farming was significantly more common among decertified farmers than their counterparts (P < 0.1). For instance, more than a third of decertified farmers responded that farming was their part-time job, whereas this was only true for less than a fourth of certified farmers. Our data showed that there were significantly more small farms that have decertified than their counterparts (P < 0.01). This finding is consistent with Constance and Choi (2010), who reported that organic operations tended to be smaller in size. Alternatively, there were significantly more medium (and large) farms that remained certified than their counterparts (P < 0.01). Table 2 shows that more than 50% of farms in our sample were medium (>$50,000 gross annual sales) and most of them remained certified (60%), whereas only 35% of medium farms decided to decertify. Consistently, Sierra et al. (2008) found that decertification rates were smaller for larger operations.

The average distance to market for farmers in our sample was 39 miles. The data show that more decertified farmers sold their produce to markets closer to their operations than certified growers (P < 0.1). It may be that the lack of distribution channels has motivated growers sell directly to consumers (Sierra et al., 2008), where the certified label may not be a requirement to access price premiums. More than half of decertified growers were sole proprietors (P < 0.1). The sources of information that most farmers considered useful were university extension services and farmers’ and growers’ associations. Price reporting services were reported as significantly more useful by certified farmers (42%) than their counterparts (P < 0.01). Similarly, significantly more certified farmers (68%) perceived that information from certifiers is useful (P < 0.01). About 74% of farmers in our sample believed that organic markets were reliable; and this was significantly higher for certified farmers (79%) than their counterparts (P < 0.05). The perception that certification process as a barrier to certify was significantly higher among decertified farmers than their certified counterparts (P < 0.01).

We asked decertified farmers to rank the most important reasons to decertify. Following Sierra et al. (2008), we provided three broad barrier categories to farmers: market barriers, certification process barriers, and production barriers. Certification barriers was ranked as the most important barrier by 72% of decertified farmers, followed by market barriers (23%) and production barriers (5%).

The farming practices of decertified farmers.

This study contributes to the literature by exploring the farming practices adopted by decertified farmers. The results provide empirical evidence that most decertified farmers remained growing organically and selling through DTC outlets. For instance, 72% of decertified farmers maintained organic practices, whereas only 16% of decertified farmers reverted to conventional agriculture and 12% of decertified producers reported growing under both conventional and organic practices. In other words, less than 30% of growers reverted their operation to conventional agriculture. In contrast, most of the decertified farmers in Sierra et al. (2008) reverted to conventional practices.

It is possible that farmers dropped out of certification because of the barriers of certification but decided to maintain organic practices as a strategy to obtain price premiums for locally- and organically grown produce. It is also possible that decertified farmers were “committed organic” farmers, who are defined as those invested in the philosophy of the organic movement (Darnhofer et al., 2005). The results also provide insight on the marketing practices of decertified operations. About 39% of decertified farmers sold only through DTC markets, 7% sold only wholesale, and 54% sold through both DTC and wholesale markets. In other words, most decertified farmers used direct sales for their produce.

Regression results.

The main contribution of this article is the empirical evidence of the main drivers of organic decertification of fruit and vegetable farmers in the United States. Table 3 displays the coefficients and marginal effects from the robust probit model. Although estimates and marginal effects were similar between the standard and robust probit regressions, this study used the robust results to provide conservative estimates and address potential heteroscedasticity. The results showed that demographics, farm characteristics, sources of information, and perceptions were key factors influencing the decision of certified farmers to drop the program.

Table 3.

Robust probit estimation results of the drivers to organic decertification (certified operations were uses as base with a coefficient of zero).

View Table

There is evidence from the regression analysis that demographics variables are driving farmers’ decision to opt out from the certification program. For instance, female farmers were 16% less likely to decertify than men (P < 0.01). An explanation is that women tend to have a greater affinity to organic agriculture (Parrott et al., 2006) because of their focus on the nutrition and health benefits for their families (Padel, 2001). These findings were consistent with Sierra et al. (2008) who found that fewer women tend to leave certification programs. Farmer’s ethnicity and the decision to decertify seemed to be related. Farmers who identified themselves as black or African American, American Indian, Asian or Pacific Islander, or multiracial were 38% more likely to decertify when compared with white farmers (P < 0.05).

Results suggest that farm location was a significant driver for the farmers’ decision to decertify organic. Farmers in the Midwest region were 14% more likely to decertify with respect to their Northeast counterparts (P < 0.01). Torres et al. (2016) reported that farmers located in Northeast states such as Pennsylvania and New York were more likely to adopt certification, which may be due to the existence of larger population centers offering access to high-value and high-volume markets. Other studies have also reported that organic agriculture tends to concentrate in areas where the demand and processing and distribution chains exist (Dimitri and Oberholtzer, 2009; Klonsky, 2000). The absence of numerous urban markets in the Midwest may be motivating farmers to sell locally, where the USDA organic label not as necessary to obtain price premiums (Torres et al., 2016). On the other hand, farmers in the Delta region were 29% less likely to decertify. In our sample, farmers in Delta states such as Alabama, Arkansas, Mississippi, and Louisiana sold their produce to more distant markets. One explanation is that as farmers’ distance to market increased, they relied on the USDA organic label to access price premiums. This is especially true as the results suggested that farmers located farther from their markets were less likely to decertify (P < 0.1).

Results from the robust probit suggested that farm management and farm characteristics were important factors driving the decision to decertify among certified farmers. Farmers selling vegetables were 30% more likely to decertify when compared with fruit growers (P < 0.05). Dimitri and Greene (2002) reported that smaller organic operations tend to grow mainly vegetables. Similarly, our sample suggested that operations growing vegetables had fewer acres and lower sales. It may be that farmers perceived organic certification as costly or risky among smaller operations growing vegetables. This is especially true as the markets for organic vegetables tended to become easily saturated and had a higher price volatility (Klonsky and Smith, 2002).

Surprisingly, farmers selling directly to consumers were 58% less likely to decertify (P < 0.01). An explanation may be that most of farms in our MarketMaker database were smaller or medium, who sold directly to consumers. These operations may be able to capture higher price premiums in local markets for their certified produce, especially if they sold through large value markets in urban centers.

Sources of information and farmer’s perceptions toward organic agriculture influenced their decision to decertify. For instance, farmers were 27% (P < 0.01) and 14% (P < 0.01) less likely to decertify if they reported that information from price reporting services and organic certifiers were useful, respectively. An explanation is that price reporting services are generally available for large mainstream markets (e.g., wholesale); thus, farmers accessing these price reports may need to rely on the USDA organic label to access organic price premiums in wholesale markets. On the other hand, farmers who received useful information from other farmers were 12% more likely to decertify (P < 0.1). It is likely that building connections with other farmers may be helping producers access local markets, where there are enough economic incentives for noncertified products.

Farmers who perceived that organic markets are reliable and transitioning to organic is risky were 29% and 25% less likely to decertify, respectively (P < 0.01). Our data showed that farmers perceiving organic agriculture as risky tended to own larger operations, specialized in fewer crops, obtained information from certifiers, and sold through wholesale channels. Although these farmers may have perceived organic as risky, they seemed to have implemented strategies to manage risk and uncertainty (Hanson et al., 2004). Last, it seems that the certification process was a major barrier to remain certified. Farmers who perceived that certification was a barrier were 26% more likely to decertify (P < 0.01).