Measuring Farmer Response to the Rate of Agricultural Innovation: Experimental Evidence from Michigan

Final report for GNC17-244

Project Type: Graduate Student
Funds awarded in 2017: $11,979.00
Projected End Date: 12/31/2020
Grant Recipient: Michigan State University
Region: North Central
State: Michigan
Graduate Student:
Faculty Advisor:
Dr. Nicole Mason
Michigan State University
Expand All

Project Information

Summary:

The development and release of innovative new agricultural technologies has provided farmers with numerous tools to increase yields while reducing input use and maintaining the integrity of the natural resource base.

Despite a vast literature on farmer demand for sustainable technologies, little work has been done to understand how the rate of innovation influences farmer decision-making about whether and when to adopt new technologies. This is an important problem due to large increases in the number of new technologies made available on the marketplace that may be driving farmers to delay or to forego the adoption of new sustainable technologies. This research will develop theoretical links between the rate of innovation of sustainable agricultural technologies and optimal farmer adoption decisions. Through this research, farmers will learn how the rate of innovation may alter profit-maximizing adoption behavior and we anticipate they will incorporate this information into their choices over new technologies. Through the analysis of results, policymakers and researchers will better understand farmer decision-making processes and the strength of experimental methodologies to capture agricultural innovation mechanisms. Utilizing a novel lab-in-the-field experiment and survey, we will test these hypotheses with Michigan farmers of wheat, corn, and soybeans. Farmers will make investment choices during a dynamic economic experiment allowing for the observation of farmer adoption behavior under different innovation scenarios. A survey will help to contextualize the experimental findings, and gain insight into how farmers perceive and plan new technology adoption decisions. Farmers, industry partners, and MSU Extension faculty will play key roles in evaluating the experimental design to ensure it is reflective of how farmers make adoption decisions in the field.

This research will provide insight into how farmers might increase expected profits and best implement sustainable agricultural technologies by adjusting adoption decisions based upon the observed rate of agricultural innovation.

Project Objectives:

The goal of this research is to investigate the effect of the rate of technological innovation on farmer adoption of sustainable, productivity-enhancing agricultural technologies. There are three primary learning outcomes associated with this project. First, farmers will learn how the rate of innovation influences optimal timing of agricultural technology adoption decisions to maximize profit. Data from an economic experiment will be used to observe farmer behavior under different technological innovation scenarios. Second, agricultural educators and policymakers will learn what factors farmers consider when making a sustainable technology adoption decision. Extending beyond expected profitability, this research will directly ask farmers about their decision-making processes. The primary action outcome of this research is that farmers will take into account information about the rate of technological innovation when choosing what technologies to adopt and the timing of new investments.

Research

Materials and methods:

The experiment is implemented using oTree experimental software and run using a personal internet connected device. All sessions in this experiment include two stages. Stage 1 involves a standard lottery choice experiment to elicit a measure of risk preference. In Stage 2, subjects participate in a 26 round technology choice environment with the between-subject treatment conditions (high innovation versus low innovation) imposed. At the end of the experiment, subjects complete a wheat production, technology knowledge/information, and demographic survey and are paid their earnings from their decisions in the form of a gift card to an online retailer. 

We use two treatments over the rate of technological innovation. In the high innovation treatment, there is an 80% chance that a new technology is invented at the beginning of a round. In the low innovation environment, the probability of a new technology is set at 20%. Subjects are provided with a menu of technology options that can be used in the production environment. Once generated, a new technology is immediately added to a subject’s choice set and remains available for use throughout the duration of the experiment. Table 1 presents the innovation patterns generated using these parameters and faced by subjects in each treatment.

Excluding the baseline technology immediately available during the first round, subjects in the high innovation treatment observe 19 new technologies during the course of the 26 round experiment compared to 10 in the low innovation treatment. Subjects are informed about the rate of innovation they will face at the beginning of the experiment, and the probability of observing a new technology in any period is independent of previous innovations.

Table 1: Technology generation patterns

Round

High Innovation (p=0.8)

Low Innovation       (p = 0.2)

1

Baseline

Baseline

2

Technology A

Technology A

3

Technology B

Technology B

4

Technology C

 

5

 

Technology C

6

Technology D

Technology D

7

Technology E

 

8

Technology F

 

9

Technology G

Technology E

10

Technology H

 

11

 

 

12

 

Technology F

13

 

 

14

Technology I

 

15

Technology J

 

16

 

Technology G

17

Technology K

 

18

Technology L

 

19

 

Technology H

20

Technology M

 

21

Technology N

Technology I

22

Technology O

Technology J

23

Technology P

 

24

Technology Q

 

25

Technology R

 

26

Technology S

 

 

Returns to a technology in each round are determined by a draw from a discrete uniform distribution. Subjects are presented with the mean and range of possible returns associated with each technology on the decision screen. We allow the mean and variance of returns to new innovations to evolve over time where, on average, we can expect that new technologies will have higher mean returns than older technologies, but this is not guaranteed. The range of returns (variance) follows a similar evolution process where we expect new technologies to have a smaller range of returns, on average, compared to existing technologies, reflecting more predictable outcomes when using new innovations. But, again, it is not guaranteed that every new technology will have a small range of returns. Full details on the mean and range of returns to each new technology can be found in Table 2.  Note that we hold the stochastic process governing the evolution of returns to new innovations constant across the two treatments. A subject using Technology C in the high innovation treatment will face the same uniform distribution for returns as a subject using Technology C in the low innovation treatment. This ensures that we are only changing the rate of the innovation process between the two treatment groups.

Technology adoption in this experiment requires the payment of a fixed cost, making the adoption decision only partially reversible. Subjects seeking to adopt a new technology that they have not used in a previous round must pay an up-front cost of 50 points. Once paid, an individual can continue using the technology in future rounds at no cost. This cost is the same if the technology arrived in the current round or was added to the choice set previously. Individuals choosing to adopt a technology previously used in another period can switch back for a cost of 25 points. Subjects are provided with a starting endowment of 100 points to cover the cost of two changes to technologies not used previously. This is to prevent individuals who would like to adopt a new technology as soon as it becomes available from having to delay the decision until they have accumulated enough points to pay the fixed cost. The baseline technology used in both rounds is provided at no cost to participants.

Table 2: Technology point returns

Technology

Mean returns

Minimum

Maximum

Range of returns

Baseline

80

0

160

160

A

91

9

174

165

B

103

28

178

150

C

114

40

189

149

D

139

70

209

139

E

154

92

217

125

F

159

101

217

116

G

153

92

214

122

H

156

94

218

124

I

165

98

233

135

J

163

100

226

126

K

172

115

230

115

L

172

114

230

116

M

171

117

225

108

N

186

128

245

117

O

176

118

234

116

P

175

114

236

122

Q

184

129

240

111

R

180

131

230

99

S

186

140

233

93

At the end of each round, the players view a summary screen with the following information: (i) the technology they used during the round; (ii) the returns generated by the technology; (iii) whether or not they switched production technologies, and the level of the switching cost incurred; and (iv) total cumulative earnings after the current round. Point earnings are totaled across all 26 rounds (and the lottery choice experiment), and converted to dollars at a rate of $1.00 per 125 points. At the end of the experiment, subjects are presented with a performance summary describing their total earnings, then they complete the short survey collecting demographic information.

In 2019-2020 we completed three significant tasks related to the materials and methods based upon presentations of the pilot and feedback from the Michigan Wheat Program. First, the online software which we used to run the technology adoption was changed, and required significant updates to the programming to match with the new architecture of the software. This was completed during the Spring/Summer of 2019. 

Second, upon receiving the contact information for members of the Michigan Wheat Program we were able to identify 160 producer email addresses that were complete and suitable targets for the survey. We also invited a random sample of ~600 Michigan producers through farmer contact listings to participate in the survey. This listing included producers of any agricultural crops or products. Survey invitations were sent out to the Michigan Wheat Program list of producers in Fall/Winter 2019 and to the additional listing in Winter/Spring 2020.  

Third, we added three modules to the survey focusing on several different aspects of production and technology information/awareness to capture a more complete view of the farmers who participate in the research and their behavior with new technologies. In the first production module, we focus specifically on 2019 land use and the number of acres/yields planted to wheat, corn, soybeans, and beans during the production year. In the second technology adoption module, we focus on how producers learn about and share information about new production technologies. This includes where they learn about information, who they share information with, and how often they actively seek out new information. In the final module we focus on specific production technology adoption timing. 

Research results and discussion:

Surveys were sent to the first wave of Michigan Wheat Program contacts (160 farmers) during Winter 2019-2020. We also worked with MSU faculty to identify another email list of 600 Michigan farmers who can also serve to increase the sample size. We also adjusted the payment schedule to increase incentives to participate. Farmers now earn a guaranteed $25 for participating and up to an additional $50 based on their performance in the experiment (up to a total of $75). After sending out approximately 760 emailed surveys to Michigan producers, we obtained 19 completed responses for a response rate of 2.5%. While this response rate was lower than we had hoped, it remains in a normal range for email surveys with farmers. Additionally, our response rate may have been affected by both adverse weather events and the COVID-19 pandemic which occurred during the study period. 

Table 3 presents summary results from the survey portion of the experiment. Our respondents were, on average, 50 years old, overwhelmingly male, and had some college education. Approximately 58% of our sample had farming as their primary occupation, 47% had off-farm employment, and our respondents had an average of 24 years of farming experience. The average farmer participating in the survey owned 366 acres and rented in an additional 527 acres. Conditional on growing a specific crop, in 2019 producers planted an average of 241, 374 188, and 461 acres of wheat, corn, soybeans, and dry beans respectively. 

Table 3 Summary statistics

  Full Sample (N=19)
  mean sd min max
Demographic Characteristics        
Age 50.95 14.73 27 84
Gender (1=Male, 0=Female) 0.947 0.229 0 1
Education levela 3.263 1.046 1 4
Farming as principal occupation (1=Yes, 0=No) 0.579 0.507 0 1
Has off-farm employment (1=Yes, 0=No) 0.474 0.513 0 1
Years engaged in farming 24.47 17.54 3 60
Relative risk aversion coefficient  0.676 0.653 -0.950 1.370
         
Farm Characteristics        
Land owned in 2019 (acres) 366.1 560.1 0 2,000
Land rented in 2019 (acres) 527.3 717.0 0 2,400
Area planted to wheat in 2019 if producing wheat (acres) 241.7 288.8 2.5 900
Area planted to corn in 2019 if producing corn (acres) 373.8 352 20 1,200
Area planted to soybeans in 2019 if producing soybeans (acres) 187.6 154.7 8 600
Area planted to dry beans in 2019 if producing dry beans (acres) 460.6 348.9 34 1,000
Purchased crop insurance in 2019 (1=Yes, 0=No) 0.667 0.485 0 1
Average years using new wheat variety 4.083 1.832 2 8
         
Precision agriculture technologies used in 2019 (1=Yes, 0=No)        
GPS guidance with manual control 0.263 0.452 0 1
GPS guidance with autosteer 0.526 0.513 0 1
GPS-enabled sprayer control 0.263 0.452 0 1
GPS for logistics 0.0526 0.229 0 1
Variable rate application services 0.526 0.513 0 1
Satellite/aerial imagery for internal use 0.474 0.513 0 1
GIS field mapping 0.421 0.507 0 1
Drone/Unmanned Aerial Vehicle (UAV) 0.0526 0.229 0 1
Soil electrical conductivity mapping 0 0 0 0
Other vehicle-mounted soil sensor 0 0 0 0
Chlorophyl/greenness sensors 0 0 0 0
Other precision agriculture technology 0.0526 0.229 0 1
         
Where do you learn about new agricultural technologies? (1=Yes, 0=No)        
Trade journals 0.474 0.513 0 1
Manufacturers advertisement 0.316 0.478 0 1
Agricultural consultants 0.368 0.496 0 1
Agricultural input suppliers 0.684 0.478 0 1
Other farmers 0.789 0.419 0 1
Trade shows 0.316 0.478 0 1
Extension agents 0.158 0.375 0 1
Farmer organizations 0.421 0.507 0 1
Websites/web forums 0.421 0.507 0 1
Social Media 0.368 0.496 0 1
         
Which of the following features of a new agricultural technology would you rank as top three in order of importance? (1=Yes, 0=No)        
Increased yields 0.947 0.229 0 1
Decreased costs  0.842 0.375 0 1
Contributes to on-farm sustainability 0.263 0.452 0 1
Compatibility with other on-farm technologies 0.158 0.375 0 1
Longevity 0.105 0.315 0 1
Sold at local input supplier 0 0 0 0
Used by other farmers in my community 0.105 0.315 0 1
Brand of the technology 0 0 0 0
Improved pest control 0.105 0.315 0 1
Access to consulting services 0 0 0 0
Decreased yield variability 0.211 0.419 0 1
Availability of financing  0.105 0.315 0 1
Recommended by university extension 0 0 0 0
         
Technology seeking behaviors        
How often do you seek out information about new agricultural technologies?b 3.667 1.680 1 6
Have you ever delayed a technology purchase? (1=Yes, 0=No) 0.389 0.502 0 1
How likely are you to adopt a new technology as soon as it becomes available?c 2.412 1.064 1 4
         
With whom do you discuss new agricultural technologies? (1=Yes, 0=No)        
Input supplier 0.474 0.513 0 1
Other farmers 0.789 0.419 0 1
Agricultural consultants 0.368 0.496 0 1
Agricultural extension agents 0.263 0.452 0 1
Company sales representatives 0.474 0.513 0 1
I don’t discuss new agricultural technologies with anyone 0.0526 0.229 0 1
         
How do you share information about new agricultural technologies? (1=Yes, 0=No)        
In person 0.842 0.375 0 1
Text message 0.211 0.419 0 1
Phone calls 0.316 0.478 0 1
Social media 0.105 0.315 0 1
Web forums 0 0 0 0
Email 0.158 0.375 0 1
Farmer/organization meetings 0.368 0.496 0 1
I don’t share information about new agricultural technologies 0.105 0.315 0 1

Notes: a1=High school/GED, 2=Some college, 3=Associates, 4=Bachelors, 5=Graduate/Professional degree. b1=Only when new technology needs replaced, 2=Rarely, 3=Annually, 4=Monthly, 5=Weekly, 6=Daily. c1=Extremely unlikely, 2=Unlikely, 3=Neither likely nor unlikely, 4=Likely, 5=Extremely likely. Source: Authors calculations. 

Our main focus in this study is understanding behaviors surrounding the adoption and use of new agricultural technologies among Michigan farmers. We first focused on understand what precision agriculture technologies were being used across our sample. In our sample, respondents reported using an average of approximately 3 precision agriculture technologies, with individual farmers using up to 6 technologies at the same time. The most commonly used technologies in our sample were GPS guidance with autosteer (52.6%), variable rate application services (52.6%), satellite/aerial imagery (47.4%), and GIS field mapping (42.1%). This data suggests that farmers in our sample were receptive to new agricultural technologies and the majority used precision agriculture technology in their operation. 

We next focused on where farmers learned about new agricultural technologies. Two information sources stood out among our sample. First, 78.9% of farmers indicated that they relied on other farmers as a source of information about new agricultural technologies. This highlights the importance of early adopters in a farming community to raise awareness of a potential technology. Second, 68.4% of our sample receive information on new technologies from agricultural input suppliers. Given that suppliers often stock new technologies (e.g. seed varieties) and can provide information on complementary technologies, this is an important information source for producers. Only 15.8% of farmers indicated that they learn about new technologies from extension agents. This may indicate that commodity extension programs are more focused on practices and markets rather than promoting the uptake of new agricultural technologies.

We also asked producers to rank the most important features of new agricultural technologies that they may consider adopting. The majority of our sample indicated that increased yields (94.7%) and decreased costs (84.2%) were among their top three most important features. Additionally contributing to on-farm sustainability (26.3%) and decreased yield variability (21.1%) were also important features. These findings emphasize the yield and cost concerns are at the forefront of the technology adoption decision. 

We find that farmers tend to discuss new agricultural technologies with other farmers (78.9%), input suppliers (47.4%), and company sales representatives (47.4%) with the vast majority sharing this information in person (84.2%) or at farmer/organization meetings (36.8%). These findings emphasize the importance of face-to-face interaction for the dissemination of information about new agricultural technologies. 

We next turn to analyzing the results from our economic experiment with the Michigan farmers. Each subject makes technology adoption decisions in every round of the experiment. We consider the four types of adoption scenarios that can occur within a round: No Change, Adopt Now, Adopt Delay, and Adopt Used. No Change occurs when a subject chooses to continue using a current technology in a given round – there is no technology switch and no fee incurred. Adopt Now is when a subject adopts a new technology in the round when it is first made available. Adopt Delay is when a subject adopts a new technology in any round after the round in which it is first made available. Finally, Adopt Used is when a subject re-adopts a technology that was previously used (and abandoned) in earlier rounds. Together, these four measures are mutually exclusive and cover all of the possible adoption decisions of a participant in a given round. 

Table 4 presents summary statistics of our key outcome variables for the full sample, the low innovation treatment, and the high innovation treatment. Our key outcome variables are experimental earnings (number of points) as well as the number of different technology adoption decisions made by farmers during the game. In the the low innovation treatment, we find that producers adopt an average of 7.18 technologies during the experiment compared to 4.88 technologies among producers in the high innovation treatment. However, while we predicted that the number of technology adoptions would be higher in the low innovation treatment, these two values are statistically indistinguishable (p=0.373) from one another. Similarly across all of our key experimental outcome variables, we find no statistically significant differences between the high innovation and low innovation treatments. One explanation for this finding may be that due to our low sample size, we lack the statistical power to detect an effect.

Table 4: Differences in means of experimental outcomes

  Full Sample (N=19) Low Innovation (N=11) High Innovation (N=8) Difference p-value
  Mean Standard Dev. Mean Standard Dev. Mean Standard Dev.
Experimental earnings (points) 3,523 617.4 3,576 479.2 3,450 801.0 126.1 0.673
Number of technology adoptions 6.211 5.401 7.182 5.811 4.875 4.824 2.307 0.373
Number of technologies adopted immediately  5.211 3.102 5.273 1.737 5.125 4.518 0.148 0.922
Number of technologies adopted with delay  0.947 1.224 1.091 1.375 0.750 1.035 0.341 0.564
Number of previously used technologies adopted  1 3.249 1.727 4.197 0 0 1.727 0.264

Notes: P-values from two-sided t-test. Significance denoted by *** p<0.01, ** p<0.05, * p<0.1, respectively.

We also use multinomial logistic regression regression analysis to investigate the probability that a farmer makes a specific adoption decision (No Change, Adopt Used, Adopt Delay, Adopt Now) within a given round of the experiment. These regression models include several key variables of interest: an indicator for treatment status (1=High innovation treatment, 0=Low innovation treatment), farmer risk aversion levels (Relative risk aversion coefficient), whether or not a farmer observed a higher than average technology return in the previous round (Good tech return), whether a new innovation became available in the previous round (Innovation arrival), and whether a farmer switched their technology in the previous round (Technology switch). We restrict the analysis to only the technologies farmers had in common across the two treatments.

Table 5 presents the marginal effects of each of these variables on the probability of a farmer’s decision. We find that being in the high innovation treatment significantly increases the probability that a farmer chooses not to change the technology that they’re using within a given round by 10.5 percentage points and decreases the probability that a farmer adopts a technology they’ve used before by 6.4 percentage points. In other words, when new technologies are arriving at a faster rate we find that farmers are more likely to stick with the technology they’re already using. This is somewhat consistent with our expectations, as the faster arrival of new technologies may make a “wait and see” approach the optimal strategy in the presence of fixed and sunk costs to adoption. 

Table 5: Multinomial Logit Marginal Effects for the Effect of Treatment on the Probability of Technology Adoption Behavior

    Common technologies
    No change Adopt used  Adopt  delay Adopt now
High innovation treatment 0.105*** -0.064*** -0.017 -0.025
    (0.005) (0.000) (0.296) (0.444)
Relative risk aversion coefficient -0.044 0.067*** -0.016 -0.007
    (0.163) (0.009) (0.169) (0.759)
Good tech return (t-1) 0.086** -0.026 -0.020 -0.041
    (0.012) (0.115) (0.254) (0.185)
Innovation arrival (t-1) 0.027*** 0.007** -0.005** -0.029***
    (0.000) (0.012) (0.046) (0.000)
Technology switch (t-1) -0.093** 0.076*** -0.004 0.020
    (0.011) (0.000) (0.824) (0.525)
Observations 474 474 474 474

Note: Dependent variable is the adoption behavior of an individual in a given round. In each round the individual has 4 choices: use their current technology (No Change), adopt a previously used technology (Adopt used), adopt a new technology from a previous round (Adopt delay), or adopt an innovation that just arrived (Adopt now). This model uses data on only common technologies between the two treatments (Baseline Technology through Technology J). Significance denoted by *** p<0.01, ** p<0.05, * p<0.1, respectively.

Focusing on other variables of interest, we find that more risk averse farmers are more likely to adopt a technology that they’ve used before. A farmer who earned an above average return in a previous round with a given technology is more likely to stick with that technology in the current round. If a new technology arrived in the previous round, we find that farmers are significantly more likely to stick with their current choice or adopt something they’ve used before and significantly less likely to adopt something new immediately or with delay. Finally, if a farmer switched technologies in the previous round they are less likely to stick with their current choice and more likely to adopt something that they’ve used before. 

Participation Summary
27 Farmers participating in research

Educational & Outreach Activities

8 Webinars / talks / presentations
2 Pilot behavioral experiments with Michigan wheat growers and with university undergraduates.

Participation Summary

8 Farmers
50 Ag professionals participated
Education/outreach description:

In May 2018, we piloted the experimental design and farmer survey with several members of the Michigan Wheat Board/Program as a test for subsequent experiments with growers and students. This activity prompted discussions with the Michigan Wheat Program about the experimental and survey design and highlighted variations in risk preferences and adoption behavior among program members. 

Other outreach activities have included presenting preliminary results from this work at the 2018 Agricultural and Applied Economics Association (AAEA) meetings. In 2019 presentations of this work included the University of Florida (UF) Biocomplexity Research Seminar, UF Experimental Economics Lab Seminar, Southern Agricultural Economics Association (SAEA) meetings, and the AAEA meetings. In 2020 this research was presented at the USDA Economic Research Service. 

Due to COVID-19, our planned outreach at the Winter/Summer Michigan Wheat Program Meetings was cancelled. Continued planned activities from this work include publishing a journal article and updating membership of the Michigan Wheat Program about our findings.

Project Outcomes

1 New working collaboration
Project outcomes:

This project focuses on understanding individual and farmer responses to innovation through economic experiments. Our primary contribution is through increased knowledge of how individuals might respond to increased innovation of new sustainable agricultural technologies.

Overall, these results provide some support for our hypothesis that faster rates of agricultural innovation may affect farmer adoption behavior. In the experimental game, we find evidence to suggest that farmers exposed to a higher rate of innovation are more likely to continue using their current technology rather than switch to something else. Combined with our survey results, this suggests that farmers may be more likely to adopt new technologies when the rate of innovation is slower and when farmers have a chance to learn about and discuss the new technologies with fellow farmers in their community. Additionally, higher rates of innovation may result in fewer technologies being adopted in the short-run, as individuals bypass some incremental technological improvements in the favor of future gains. This could be problematic for sustainable technological innovations that are not associated with dramatic cost or resource savings at the individual level. Reducing the costs of switching or releasing technologies with more significant improvements over those currently in use may incentivize more rapid technology adoption. Programs to lower adoption costs might include information dissemination via extension or even public-private partnerships to reduce search costs and improve access in retail outlets.

Knowledge Gained:

Over the course of the project, my awareness of sustainable agriculture as a complex set of complementary management decisions has increased significantly. While individual agricultural technologies can and do promote sustainable agricultural change on their own (e.g. new varieties of seed), a large part of sustainable change depends on how those technologies are used as part of a broader agricultural production process and land management plan. While my experimental design continues to focus on the innovation process of a single technology, I have tried to incorporate questions on complementary technology use into the farmer survey instrument. 

I’ve also learned about the challenges facing extension and other agricultural specialists with having up to date contact information for producers. This can affect information delivery about sustainable agriculture. In our context, having an online game/survey instrument is more flexible for producers, but many farmers may not have provided a current email address or contact information to their producer group. 

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.