- Email: [email protected]
Policy impact analysis of penalty and reward scenarios to promote flowering cover crops using a business simulation game
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Available online at www.sciencedirect.com
ScienceDirect http://www.elsevier.com/locate/biombioe
Policy impact analysis of penalty and reward scenarios to promote flowering cover crops using a business simulation game Gesa Sophie Holst*, Oliver Musshoff, Till Doerschner Georg-August-University of Goettingen, Faculty of Agricultural Sciences, Department of Agricultural Economics and Rural Development, Farm Management Group, Platz der Goettinger Sieben 5, D-37073 Goettingen, Germany
article info
abstract
Article history:
Due to an increasing number of biogas plants and the positive qualities of maize as a biogas
Received 3 June 2013
substrate, the cultivation of silage maize has risen in Germany. However, there are still various
Received in revised form
reasons for the limitation of the cultivation area of silage maize. Hence, policymakers are
6 June 2014
currently discussing various alternative biogas substrates and ways to promote their cultiva-
Accepted 19 August 2014
tion. One possible alternative is the use of special flowering cover crops with additional
Available online 18 September 2014
ecological benefits. Using a business simulation game conducted with farmers, the present study investigates whether the implementation of a reward and penalty policy will improve
JEL classification:
the uptake of flowering cover crops in the production programs of farmers. The results indicate
C90
that the implementation of these policy measures was followed by a significant increase in the
Q24
cultivation area of flowering cover crops. The penalty policy leads to a stronger increase in the
Q28
size of the cultivation area of flowering cover crops than the reward policy, even though the
Keywords: Flowering cover crops
policies have the same income effect for farmers. Furthermore, the results reveal that the cultivation of flowering cover crops is influenced by various socio-demographic variables. © 2014 Elsevier Ltd. All rights reserved.
Alternative biogas substrate Business simulation game Policy impact analysis Reward policy Penalty policy
1.
Introduction
Fossil energy sources are finite resources that contribute significantly to anthropogenic global warming through the release of CO2 emissions. Therefore, the German government has decided to promote the expansion of renewable energy sources with the ‘Act on granting priority to renewable energy sources’ [1]. This act determines the remuneration for electricity generated from renewable energy sources and aims to
realize a 35% and 85% share of renewable energy in total electricity generation by 2020 and by 2050, respectively. Due to the profitability and possible income stabilization, many farmers have invested in renewable energies [2]. In Germany, the number of biogas plants increased between 2002 and 2012 from 1600 to 7515 with a total installed capacity of 3.352 GWh [3]. Moreover, the total electricity production from biomass is the second most important source of renewable energy with a share of 27.7%, after wind power with a share of 35.3% [4]. Thus, the cultivation of energy crops,
* Corresponding author. Tel.: þ49 0 551 39 4836. E-mail addresses: [email protected], [email protected] (G.S. Holst). http://dx.doi.org/10.1016/j.biombioe.2014.08.009 0961-9534/© 2014 Elsevier Ltd. All rights reserved.
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
such as maize and whole crop silage, has increased sharply. In 2013, energy crops for fermentation in biogas plants were grown on a total surface area of 11,570 km2 [5]. The cultivation of energy maize accounts for 8000 km2 of the aforementioned surface area [6]. Due to its high yield of dry matter and its energy content, maize is the preferred crop for biogas production [7]. The expansion of energy production from biomass, however, is not necessarily considered to be positive as it can lead to serious environmental problems, such as the pollution of ground water by nutrients or the loss of organic matter in farmland [7]. For the future development of electricity production from biomass, it is essential to reconcile the production of biogas substrate and nature conservation. An initial step for policymakers in Germany is to restrict the use of silage maize and grain, including corn-crop mix, grain maize, and ground ear maize in biogas plants to 60 mass fraction, which is anchored in the ‘Act on granting priority to renewable energy sources’ and became effective on 01.01.2012 [1]. A variety of alternative biogas substrates, such as cup plant, Sudan grass and sorghum, have been discussed [8e10]. In addition, the fermentation use of silage produced from flowering cover crops in biogas plants is being explored. Initial results show that flowering cover crops are well suited for fermentation in biogas plants [11,12]. Further advantages include low input, the creation of habitats for wildlife, as well as the increasing acceptance shown by the positive public response to fields that are used for, or surrounded by, flowering cover crops. Bees and other insects use the flowering crops as a food source and wildlife offer these crops cover along with a secure place to deliver. The harvest takes place in October when most flowers are withered and the animals and insects have left the field [11,12]. For the aforementioned reasons, a political goal could be the integration of flowering cover crop cultivation into the production programs of farmers. However, the introduction of a new policy is accompanied by high costs [13]. Prior to the introduction of a policy, a policy impact analysis is essential in evaluating whether a policy measure is effective. The development of models which simulate the consequences of a policy implementation may be an opportunity [14,15]. Frequently used models for policy impact analysis assume a perfect rationally behaving profit maximizer [16]. However, it is often discussed that assuming profit maximizing behavior is not appropriate [17e19]. Explanatory approaches regarding the concept of utility maximization suppose that people maximize their utility and, therefore, various objectives, such as making profits, risk aversion, traditions, recreational activities, or social recognition are pursued [20]. Another explanatory approach is the bounded rationality [21]. Thus, models assuming rational decision makers could reflect the distorted consequences of policy implementation. These limitations can be addressed by experiments which are not based on decisions given through exogenously predetermined theories. Instead, we observe the real decisions of real people. In both laboratory experiments and in business simulation games, it is possible to set incentives for motivating participants to make well-conceived decisions [22]. Business simulation games make it possible to design a realistic decisionmaking situation, resulting in a significant advantage
197
compared to classical laboratory experiments [23]. Thus, they seem to be especially suitable for policy impact analysis. This study explicitly examines the farmers' reaction to the implementation of policies in order to promote the share of flowering cover crops in the agricultural landscape. For this purpose, the multi-period, single-person business simulation game is developed. Additionally, the participant farmers will be confronted with different policy measures. The following questions will be addressed in the business simulation game: 1. Does the implementation of reward and penalty policies have an impact on the proportion of flowering cover crops in the production programs of the farmers? 2. Is either a reward or a penalty policy with the same income effect more effective? 3. Does the policy change lead to the cultivation of flowering cover crops as a biogas substrate? The novelty of this paper lies in the policy impact analysis that is geared towards implementing flowering cover crops in the production programs of farmers. In recent years, studies have dealt with flowering crops and their environmental benefits, with primary research focusing on the nature conservation concept and the impacts on biodiversity [24,25]. Vollrath et al. [11] and Vollrath and Werner [12] investigated the benefits of flowering cover crops as a biogas substrate. To our knowledge, there are no publications that address the individual effect of policies to increase the quantity of flowering cover crops in the agricultural landscape. Furthermore, a new aspect to this research is that a business simulation game conducted with real decision makers e in our case farmers e is used for the policy impact analysis. The article is structured as follows: First, the behavioral theoretical hypotheses are derived (Section 2). Sections 3 and 4 explain the experimental design and sample characteristics. The results are presented in Section 5 and the article concludes with a summary and a discussion of future opportunities (Section 6).
2.
Hypothesis generation
Reward and penalty strategies encourage human compliance through the use of rules and laws and, in this way, establish a social order [26]. Penalty payments pursue a strategy of deterrence in order to prevent rules being broken, whereas rewards present an incentive to direct human behavior in a desired direction [27]. Accordingly, it is assumed that a reward and a penalty policy strategy can direct the behavior of farmers, so that the cultivation of flowering cover crops is extended. Thus, the following hypothesis can be derived: H1. Regardless of whether the policymakers introduce a reward for growing flowering cover crops or a penalty with the same income effect for not growing flowering cover crops, the share of flowering cover crops in the production program of farmers will increase. In the economic literature, it has been hinted that the effects of penalty policies differ from reward policies [26]. Kahneman and Tversky [28], as well as Kahneman et al. [29] found
198
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
evidence for loss aversion meaning that people weigh a monetary loss more heavily than an equally high profit. Concerning a reward and a penalty policy with the same income effect means that the loss of the penalty payment is weighed higher than the reward payment. Even the opportunity cost effect supports the assumption that reward and penalty policies differ in their effect on human behavior. Out of pocket costs, such as penalty payments, are given a higher weight than opportunity costs, which correspond to a foregone reward payment [29]. Another reason for the different impact of penalty and reward policies is the normative commitment of laws. If laws make sense for people and they accept them, they do not break the laws even if it is possible without facing consequences [26]. Therefore, the following hypothesis can be derived: H2. The penalty policy changes the cultivation behavior of farmers regarding the scope of flowering cover crops better than a reward policy with the same income effect. Flowering cover crops can be cultivated as nature conservation measures [30,31], or even as a biogas substrate for energy production with additional ecological benefits [12]. If new policy measures promote the use of flowering cover crops as a biogas substrate, we assume that the cultivation area of flowering cover crops used as a biogas substrate increases because of the higher revenues that are generated compared to the cultivation of these crops for nature conservation. Hence, flowering cover crops are used to complement the production of biogas substrate. Thus, the following hypothesis is derived: H3. The implementation of reward and penalty policies would promote the cultivation of flowering cover crops for biogas substrate. Farmers that cultivate flowering cover crops behave ecoconsciously, because they create habitats for wildlife and insects. Regarding the socio-demographic and socio-economic variables and their influence on the environmental behavior of farmers, only limited literature exists. Buttel et al. [32] investigated the impact of age, education, income, size of agricultural enterprises, and the organic or conventional land use on the environmental behavior. As significant factors, only the negative effect of age and the positive effect of the organic operation focus can be detected. Carr and Tait [33] indicate that the productivity and effectiveness idea can affect the management behavior of farmers with detriment to nature conservation (see also Ref. [34]). The connectedness of farmers to nature influences the environmental concern [35]. This leads to the following hypothesis: H4. Socio-demographic and socio-economic variables impact the cultivation of flowering cover crops for nature conservation and for the production of biogas.
3.
Design of the experiment
The experiment is divided into three sections. In the first section, the incentive compatible, multi-period, one-person business simulation game is conducted. Subsequently, a Holt-
and-Laury lottery (HLL) is carried out [36] to determine the risk attitude of the participants (Section 2). Finally, sociodemographic and socio-economic data from the participants are collected (Section 3).
3.1.
General structure of the business simulation game
The design of the business simulation game aims to enable the participants to experience business situation that is as realistic as possible with realistic decision situations [38]. However, business simulation games can only model real events in its essential characteristics. Therefore, the market price levels and yields are reflected as uncertainties. The experiment is conducted with farmers who know that they are participating in an experiment and that their behavior will be recorded and analyzed. Hence, a non-standard subject pool is used. In addition, the experiment has been conducted at an exhibition and not in a laboratory situation. Consequently, the business simulation game can be classified according to Charness et al. [39] as an extra-laboratory experiment. The participants are told to assume that they are a manager of a virtual farm with 100 ha of farmland. They have to run the farm for 12 production periods. Each round of the business simulation game stands for a production period and requires the following basic decisions of the participants: 1. Production program decision: Configuration of the production program for the cultivation of the farmland with the production activities winter wheat, silage maize, sorghum and flowering cover crops. 2. Contract decision: Acceptance of a substrate delivery contract for an adjacent biogas plant for 0 t, 1500 t, 3000 t or 4500 t of fresh matter. For fulfillment, it is possible to use silage maize, sorghum and flowering cover crops. In the business simulation game, (1) deterministic and (2) stochastic parameters are given. Ad (1): The deterministic parameters are communicated at the beginning of the business simulation game, do not change randomly and apply to all participants. When the business simulation game starts, each participant has a starting capital of V 100,000. In each production period, withdrawals in the amount of V 30,000 are made to cover the costs of living. To guarantee a crop rotation with at least two cultures, minimum cultivation requirements exist for winter wheat and silage maize. No minimum cultivation requirements exist for growing sorghum and flowering cover crops. All production activities can be grown on a maximum of 70 ha of farmland (Table 1). It is noteworthy that the entire available land (100 ha) must be cultivated by using the aforementioned four crops; there is no option to set aside some land for other uses. The processing options of each crop are shown in Table 1. Winter wheat is only used for selling on the spot market, whereas silage maize is sold and used for fulfilling the biogas substrate delivery contract. The cultivation of sorghum is solely for the production of biogas substrate. Flowering cover crops, however, have two alternative forms of use; they can be grown as a substrate for biogas plants with additional ecological benefits, or they may remain unused for ecological reasons and are then compensated for by a payment from the
199
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Table 1 e Overview of the four crops for farmland cultivation. Wheat
Silage maize
Sorghum
Processing option (market price)
Spot market (volatile price)
Biogas (35 V t1); Spot market (volatile price)
Biogas (35 V t1)
Biogas (35 V t1)
Flowering cover crops Nature conservation (640 V ha1)
Cost of cultivation
970 V ha1
832 V ha1
800 V ha1
880 V ha1
340 V ha1
Minimum extent Maximum extent
5 ha 70 ha
5 ha 70 ha
0 ha 70 ha
0 ha 70 ha
Weather impact on yield above-average (20%) average (60%) below-average (20%)
9.6 t ha1 8.0 t ha1 6.4 t ha1
60.5 t ha1 55.0 t ha1 49.5 t ha1
53.5 t ha1 48.0 t ha1 42.5 t ha1
46.0 t ha1 40.0 t ha1 34.0 t ha1
No yield measured, biodiversity
Table 2 e Market price developments of winter wheat (left) and silage maize (right) from period 0 to period 1. Current wheat price (period 0) 1
200.00 V t
Uncertain winter wheat price in the following production period (period 1) 220.00 V t1 probability of 50% 180.00 V t1 probability of 50%
state. The contracted substrate delivery is remunerated at 35 V t1, regardless of whether the biomass is provided with maize silage, sorghum or flowering cover crops. Ad (2): The stochastic parameters change randomly from one production period to another and, therefore, vary among the participants. The market prices for winter wheat and silage follow an arithmetic Brownian motion [37] starting with an identical initial value for all participants (Table 2). The market prices decrease or increase, starting with the current price in any production period with a probability of 50% by 20 V t1 for winter wheat and by 1.50 V t1 for silage maize. Furthermore, weather conditions will affect yield and, accordingly, gross margins. There is a distinction between above-average, average, and below-average weather conditions (Table 1). The periods of above-average and belowaverage weather occur each with a probability of 20%. Average weather is expected with a probability of 60%. Aboveaverage weather has the consequence that the yields of all production activities per hectare achieve its maximum, whereas below-average weather leads to a yield drop to the minimum. Both the probabilities of the weather conditions and the yields per hectare are communicated at the beginning of the business simulation game. Despite the uncertain yields, the chosen delivery contract has to be fulfilled by 100%. If this is not the case, the missing substrate amount has to be bought from the market for twice the current market price of silage maize. Each participant receives a transfer payment of 300 V ha1 of farmland after each completed period of production. At the same time, the participants are informed that amendments may occur in the course of the business simulation game. Since there are no storage possibilities for crops, all goods are sold at the end of each production period at the current market prices. The current prices and the weather conditions of the previous period are communicated at the beginning of each new production period. Moreover, the participants receive further information about the profit of the realized
Current silage maize price (period 0) 25.00 V t
1
Uncertain silage maize price in the following production period (period 1) 26.50 V t1 probability of 50% 23.50 V t1 probability of 50%
production program and the contract decisions of the previous production periods.1
3.2.
Changes in the policy framework
At the beginning of the business simulation game, participants are randomly assigned to one of three policy scenarios. During the first six production periods of the business simulation game, the policy framework is identical for all three policy scenarios. However, in the subsequent production periods, 7 through 12, the policy framework conditions change whereby the policy scenarios are defined as follows: Scenario 1 (reference scenario): The policy framework remains unchanged over the entire duration of the business simulation game. Scenario 2 (reward scenario): The participants are informed that the transfer payment decreases by 10% to 270 V ha1. Simultaneously, policymakers introduce an additional premium of 300 V for each ha used for growing 1
Experiments aim to model real events in its essential characteristics. To handle the conflict of real characteristics on the one hand and the virtual world of business simulation games on the other hand, we made some simplifications. One simplification is the same price per ton of biogas substrate whereby 35 V t1 is a realistic value taken from the literature [40,12]. Furthermore, we have chosen an arithmetic Brownian motion [37] with zero drift to indicate that the prices are uncertain. Further simplifications are made with respect to the weather conditions. Weather conditions consist of various factors influencing the yields of crops, primarily through temperature, draughts [41], precipitation and radiation [42]. The interaction of these factors and the stage of plant growth have an important influence on yield effects [41,42]. Therefore, we summarize factors which result in higher yields as above-average conditions and weather conditions resulting in low yields as below-average conditions. Additionally, we abstained from the opportunity of storage for the purpose of simplicity and for the reason that we do not receive any relevant information for answering our research questions.
200
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
flowering cover crops. The state pays a maximum amount of V 3000 per farm and, thus, subsidizes a maximum cultivation area of 10 ha of flowering cover crops. Scenario 3 (penalty scenario): The policy punishes all participants who use less than 10% of their farmland for the cultivation of flowering cover crops. Each hectare below the fulfillment requirement will incur a penalty of V 300. Scenarios 2 and 3 do not differ in their income effect for farmers. In order to compare the participants' behavior in the scenarios, we construct triplets of randomly selected participants, i.e., there is always three individual participants playing the business simulation game with the same price and weather developments, but with different underlying policy scenarios.
3.3.
Holt and Laury lottery
The Holt-and-Laury lottery is an experimental method used to identify the risk attitude of people. This methodology presented by Holt and Laury [36] has not been changed and is already established in agricultural economics [43,44]. First, participants are introduced to the game with the explanation of the lottery whereby participation is completely voluntary and the outcomes are always positive. The participants are asked to choose between the two lotteries: A (i.e., the safer alternative) and B (i.e., the riskier alternative). In lottery A, the participants can either win V 200 or V 160, whereas the riskier lottery B offers prize money of either V 385 or V 10. The probabilities of achieving the two possible prize moneys in the lotteries are systematically varied at 10% intervals beginning at either V 200 in lottery A (V 385 in lottery B) with a possibility of 10%, or V 160 (V 10 in lottery B) with a probability of 90%. Consequently, there exist ten different decision situations whereby the expected value increases in both lotteries. In decision situation five, the expected value of lottery B becomes higher than the expected value of lottery A. The HLLvalue indicates the number of safe choices when the decision from lottery A changes to the more risky lottery B and represents the risk attitude. Farmers who chose lottery A four times are risk neutral. When farmers chose the safer lottery A one to three times, it indicates risk seeking whereby an HLLvalue between five and nine shows risk aversion.
3.4.
Incentives for well-conceived decisions
In order to attract participants, each participant receives a representation allowance of V 10. The planned playing period of 30 min corresponds to an average hourly wage of V 20. In the German agricultural sector, the average hourly wage is V 9.92 [45]. Hence, the allowance is more than twice as high as the average hourly wage and should motivate the farmers to participate in the business simulation game. For ensuring incentive compatibility, additional cash prizes can be won, V 2005 in total. This amount of money corresponds to an expected value of V 16.71 for each of the planned 120 participants. Since the playing period is already paid by the representation allowance, the expected gain covers the opportunity costs of additional 50 min to encourage well-conceived decisions. We did not define what is meant by ‘well-conceived’ decisions in the instructions for the
participants for the reason that we did not want to influence the participants' decisions, however, the incentives depend on the financial success of the virtual farms led by the participants. As in reality, aiming for multiple objectives, e.g., nature conservation or risk reduction decreases the expected profits. Hence, if participants aim for other objectives than profits in the business simulation game, the expected incentive decreases and our experiment becomes more realistic. To set incentives in economic experiments is common [46e48]. Former studies have shown that incentives influence the behavior of participants and help to improve decision making [46,48]. They work harder and think twice if the incentive depends on their own performance [49,50], thus, problem solving capabilities can be improved. Nevertheless, incentives can also cause biases [47,51]. Camerer and Hogarth [49], however, summarize that incentives more often result in better experimental findings than in causing biases. Different payment systems for incentives exist. On the one hand, each participant receives a small incentive while, on the other hand, a fraction of all participants receives a high incentive. However, participants overestimate the probability of being selected if only one participant is paid [49]. Therefore, we decided that only a fraction of participants are able to receive high incentives. Out of the 120 participants that are anticipated to take part in the business simulation game, four of these participants will be drawn by lottery to win a cash prize. The first three cash prizes depend on the participants' financial success in the business simulation game and amount to a maximum of V 540. The three randomly drawn winners receive the share of the maximum monetary gain that corresponds to their financial success.2 The fourth cash prize depends on the decisions in the Holtand-Laury lottery. For the determination of the cash prize, one participant is randomly drawn by a lottery and will have the chance to win between V 10 and V 385 e depending on his risk attitude e in a Holt-and-Laury lottery carried out solely for that participant.
4.
Description of the sample
The experiment was conducted at the agricultural exhibition “EuroTier” that took place between the 13th and 16th of November, 2012 in Hanover (Germany). Nine hundred and forty-six visitors of the exhibition were randomly chosen and directly contacted and were invited to participate in the business simulation game. In total, 123 farmers (13% of those contacted) successfully completed the experiment, with 41 farmers playing in each policy scenario. To complete the experiment, the participants needed 43 min on average. The socio-demographic and socio-economic characteristics of the participants in the three policy scenarios are summarized in Table 3. 2
We do not rank the participants with regard to their business success and do not pay money to the first ranked farmers as is often done in stock exchange simulation games [52]. Paying the best could lead to risk seeking behavior. As the winners are randomly drawn through a lottery by us, we assume that the risk preference is not influenced.
201
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Table 3 e Socio-demographic and socio-economic characteristics of the participants (41 participants in each policy scenario). Characteristics
Age (in years) Percent of female participants Years of education HLL-valuea Farm manager Farms farmed on a part-time basis Farmland (in ha) a
Policy Scenario 1
Policy Scenario 2
Policy Scenario 3
Mean
SD
Mean
SD
Mean
SD
31.8 12.2% 13.6 5.9 39.0% 19.5% 225.5
12.7 e 3.3 2.0 e e 392.3
28.8 14.6% 13.7 5.2 34.1% 7.3% 196.9
10.3 e 3.5 2.2 e e 220.4
27.7 9.8% 13.8 5.1 29.3% 14.6% 312.5
9.4 e 3.2 1.6 e e 617.4
1e3 ¼ risk seeking, 4 ¼ risk neutral, 5e9 ¼ risk averse.
On average, the participants were 29 years old, including the youngest participant at 16 years and the oldest at 62 years. A possible reason for the low average age may be the computer-based aspect of the experiment, which might discourage older farmers from participating [53]. The agricultural enterprises of the participants are on average larger than the average German farm with 56 ha of farmland [54]. Moreover, 55% of the farms in Germany are farmed on a part-time basis [54] whereas only 14% of the participants in our experiment run their agricultural businesses on a parttime basis. Hence, the sample characteristics do not match with the common characteristics of German farmers. However, the aim of our experiment is not the representativeness of the sample, but rather to test economic theories, i.e., the effects of the implementation of reward and penalty policies with the same income effect. There is a widespread consensus that the benefits of internal validity are more important to people than the lack of external validity if the experiments aim is to test theories, as is the case in our study [55]. Using the H-test according to Kruskal and Wallis, it can be shown that the socio-demographic and socio-economic characteristics of the participants in the three policy scenarios do not differ significantly in age (p-value ¼ 0.140), years of education (p-value ¼ 0.961), the HLL-value (p-value ¼ 0.228) and farmland (p-value ¼ 0.759). In addition, a chi-square-test indicates that there are no significant differences between the policy scenarios in the percentage of female participants (p-value ¼ 0.800), farm managers (p-value ¼ 0.855) and parttime farmers (p-value 0.273).
5. Behavioral control impacts of different policy measures on farmers In the first six production periods, the basic conditions are the same for all participants. It can therefore be analyzed how the cultivation area of flowering cover crops changes when the reward or penalty policy appears. The average size of the cultivation area of flowering cover crops is presented in Table 4 for the production periods 1e6 and 7e12 for each of the three policy scenarios. The H-test, according to Kruskal and Wallis, shows that the cultivation areas of flowering cover crops do not differ significantly between the three policy scenarios in periods 1e6 (p-value ¼ 0.812).
For Scenario 1 (reference scenario), the paired t-test shows that the size of the cultivation area in the first six periods does not differ significantly from the size of the cultivation area in periods 7e12 (p-value ¼ 0.732). In contrast, the cultivation area changed significantly with the introduction of a reward (Scenario 2) or penalty policy (Scenario 3) (p-value ¼ 0.080 and pvalue ¼ 0.001, respectively). Fig. 1 provides an overview regarding the success of the implementation of the reward and the penalty policy. Due to the same income effect of the reward and the penalty policy in 21.54% of all cases, the policy compliance is economically advantageous, while, in 78.46% of all cases, the policy compliance is not economically advantageous when aiming at profit maximization. If policy compliance is economically advantageous, 75.47% of the participants grow 10 ha or more of flowering cover crops in the reward policy and 83.02% of the participants in the penalty policy. With regard to ‘policy noncompliance is economically advantageous’ in the reward policy, 54.92% of the participants grow 10 ha or more of flowering cover crops. Comparing the penalty policy to these results, 77.72% of the participants grow 10 ha or more of flowering cover crops when ‘non-compliance is economically advantageous’ to maximize profits. This makes clear that the penalty scenario has a stronger effect on compliance. The difference between maximum possible and maximum achievable profit with policy compliance depicts the profit differential which is used for result estimation. The average profit differential is V 563.47. When policy compliance is economically advantageous for maximizing profits, the profit differential is V 0, per definition. The maximum profit differential is V 2044.33.
Table 4 e Size of the cultivation area of flowering cover crops in the business simulation game in ha (41 participants in each policy scenario). Policy Scenario
1 2 3 Mean
Area of flowering cover crops
Area of flowering cover crops
Periods 1e6
Periods 7e12
Mean
SD
Mean
SD
10.21 7.92 8.22 8.78
13.55 11.86 10.74 12.13
9.88 10.01 12.73 10.87
12.68 9.87 10.31 11.09
Change in average
0.33 2.09 4.51 2.09
202
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Fig. 1 e Success of the implementation of the reward and the penalty policy in production periods 7e12 (41 participants in each policy scenario). (a) For the percentages serve policy compliance economically advantageous as a baseline. (b) For the percentages serve policy non-compliance economically advantageous as a baseline.
Due to having 12 observations in the production periods of the business simulation game for each of the 123 participants, our dataset depicts panel data. The BreuschePagan-test indicates that a random effects regression has to be favored with regard to an OLS regression (p-value < 0.001). However, a random effects regression is not consistent according to the Hausman-test (p-value ¼ 0.047) and a fixed effect model is not suitable because time-constant variables relating to the sociodemographic and socio-economic characteristics of the participants, need to be included in the estimation. Hence, we estimated another model which is a hybrid regression with random slopes. Time-varying covariates are included with a time-constant effect which is the individual average of the variable and a context variable which is the value of each period adjusted by the individual average. Unobserved heterogeneity can be controlled by the context variable and the integration of random slopes whereupon the estimation becomes more precise indicated by the AIC. From the business simulation game, the following four initial situations and are used for the result estimation: The cropping areas of flowering cover crops for all three policy scenarios for the first six periods represent the reference situation (Situation 1). The policy Scenarios 1e3, which occur in the seventh production period, reflect the policy implementation effects (Situations 1e3). In Table 5, they are referred to as dummy Scenarios 1, 2 and 3 and deal with the baseline of the first six production periods in the estimated models. Three different models have been estimated. Model 1 reflects the effects of the independent variables on the total cropping area of flowering cover crops. Furthermore, Models 2 and 3 highlight the effects separated for the cultivation of flowering cover crops for biogas production or nature conservation purposes. Hypothesis H1 which assumes that reward and penalty policies promote the cultivation of flowering cover crops, is confirmed: In both, the reward and the penalty scenario, the share of flowering cover crops increases significantly (Model 1). Participants confronted with the reward policy grow 2.285 ha more on average of flowering cover crops in comparison to the reference periods 1e6. Whereas participants confronted with the penalty policy increase the share of flowering cover crops by 4.584 ha on average. Compared to the
reference situation of periods 1 thought 6, no significant changes occur in the reference scenario and periods 7 through 12. Consequently, the incentive strategy of the reward policy and the deterrence strategy of the penalty policy caused a change in the participants' behavior regarding the cultivation of flowering cover crops, this result can also be seen in Fig. 2. From the sixth to the seventh production period, the cultivation area of flowering cover crops increases for the reward and the penalty policy. However, it can also be seen that the profit differential affects the behavior of cultivating flowering cover crops. For each V 1000 profit differential, the participants grow 5.291 ha on average less of the flowering cover crops. Hypothesis H2 assuming that penalty policies have a stronger impact on the cultivation area of flowering cover crops than reward policies is confirmed: The implementation of a penalty policy leads to a stronger increase in the size of the cultivation area of flowering cover crops than the implementation of a reward policy (Model 1), although the policies do not differ in their income effect. A linear restriction reveals that the effects of the reward and the penalty policy differ from each other at a significance level of 10%. Thus, the loss aversion and the opportunity cost effect, as well as the normative commitment of laws have to be taken into account when estimating policy consequences. These effects influence the awareness of reward and penalty scenarios. Hypothesis H3 which supposes that reward and penalty policies lead farmers to grow flowering cover crops as a biogas substrate, is confirmed: A penalty policy results in an increase in the cultivation of flowering cover crops to produce biogas substrate by 2.535 ha (Model 2), this increase can also be seen in Fig. 2. In contrast, the introduction of a reward policy has no significant effect on the cultivation of flowering cover crops for biogas plants. Despite having the same income effect of policy Scenarios 2 and 3, only the penalty policy (Scenario 3) achieved a significant increase in the cultivation of flowering cover crops in order to produce biogas substrate. Socio-demographic and socio-economic variables affect the cultivation of flowering cover crops is assumed in hypothesis H4 which is confirmed: The model results in Table 5 show that socio-demographic and socio-economic variables have a significant influence on the cultivation area of flowering cover crops. Female participants grow significantly more
203
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Table 5 e Hybrid regression with random slopes to explain the cultivation area of flowering cover crops (N ¼ 1476).a
Constant Dummy Scenario 1 Dummy Scenario 2 Dummy Scenario 3 Average individual profit differentialb Profit differentialc Risk attituded Substrate delivery amount 1500 t1 Substrate delivery amount 3000 t1 Substrate delivery amount 4500 t1 Age Gendere Years of education Earning powerf Farmingg Renewable energyh Profit flowering cover cropsi Nature flowering cover cropsj Agri-environmentk AIC a b c d e f g h i j k
Model 1 Dependent variable: land used for flowering cover crops
Model 2 Dependent variable: land used for flowering cover crops, biogas plant
Model 3 Dependent variable: land used for flowering cover crops, nature conservation
Coefficient
t-Statistic
Coefficient
t-Statistic
Coefficient
t-Statistic
30.023 0.164 2.285 4.584 5.291 0.178 0.098 0.309 0.304 1.232 0.082 7.281 0.656 6.273 4.742 2.546 2.872 1.855 1.000
5.003*** 0.235 2.706*** 5.947*** 1.974* 0.031 0.269 0.257 0.257 0.989 1.272 3.373*** 2.655*** 3.103*** 1.760* 1.757* 1.928* 1.171 1.234
20.258 0.098 0.880 2.535 2.527 0.570 0.105 2.793 3.004 3.469 0.088 3.779 0.271 2.474 2.889 1.711 0.747 2.032 1.015
5.547*** 0.184 1.516 4.330*** 1.644 1.268 0.504 3.126*** 3.431*** 3.784*** 2.380** 3.064*** 1.928* 2.152** 1.884* 2.073** 0.880 2.249** 2.199**
8.999 0.004 1.176 2.049 2.718 0.595 0.006 2.670 2.925 5.206 0.006 3.380 0.380 3.773 1.921 0.807 2.147 0.189 0.003
2.194** 0.008 2.109** 3.645*** 1.506 1.394 0.026 3.063*** 3.406*** 5.770*** 0.129 2.327** 2.288** 2.777*** 1.061 0.828 2.144** 0.178 0.006
10,617.69
9784.86
9635.57
*p-value < 0.10; **p-value < 0.05; ***p-value < 0.01. Individual average of the difference between maximum possible and maximum achievable profit with policy compliance in V 1000. Observed profit differential minus individual average profit differential. 1e3 ¼ risk seeking, 4 ¼ risk neutral, 5e9 ¼ risk averse. 1 ¼ female, 0 ¼ male. 1 ¼ mainstay farm, 0 ¼ part-time farm. 1 ¼ conventional, 0 ¼ organic. Have you invested in renewable energies aside from biogas plants in reality? 1 ¼ yes, 0 ¼ no. Do you think it is possible to earn money by growing flowering cover crops in reality? 1 ¼ yes, 0 ¼ no. Do you think the cultivation of flowering cover crops for nature conservation makes sense in reality? 1 ¼ yes, 0 ¼ no. Do you agree or disagree with agri-environmental measures in reality? 1 ¼ total disagree to 5 ¼ total agree.
flowering cover crops for biogas production and for nature conservation than men (Models 1, 2 and 3). Consequently, female farmers may be more open to alternative methods of cultivation, as well as demonstrating concerns for nature
conservation. Participating farmers, for whom the farm is the main source of income, grow significantly fewer flowering cover crops for biogas plants and for nature conservation. Participants who responded to the questionnaire that it is
Fig. 2 e Average cultivation area of flowering cover crops and average cultivation area of flowering cover crops for biogas substrate in the production periods 1e12 (41 participants in each policy scenario).
204
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
possible to make profits with the cultivation of flowering cover crops in reality grow more of the flowering cover crops for nature conservation. However, there is no significant effect upon the size of the cultivation area of flowering cover crops for biogas plants. The variables ‘gender’, ‘years of education’, ‘earning power’ and ‘farming’ have a highly significant influence on the size of the cultivation area of flowering cover crops (Model 1). Interestingly, these effects differ if the two kinds of flowering cover crops are separately estimated (Models 2 and 3). The covariates ‘gender’, ‘years of education’ and ‘earning power’ have a significant impact on the size of the cultivation area of flowering cover crops on both variations but in varying degrees. Considering Model 2, the variable ‘age’ is significant. In contrast, there is no significant influence of this variable on the size of the cultivation area of flowering cover crops for nature conservation (Model 3). The negative impact of ‘age’ on the size of the cultivation area of flowering cover crops for energy production implies that older farmers may be more skeptical towards alternative biogas substrates. The effect of the substrate delivery contracts differs with regard to the production alternative of flowering cover crops. Whereas the decision for delivery contracts of 1500 t, 3000 t and 4500 t has a positive influence on the cultivation of flowering cover crops as a biogas substrate. It has a negative impact on the size of the cultivation area of flowering cover crops for nature purposes. It becomes clear that there are different factors influencing the reaction and implementation of policies.
6.
Conclusion and outlook
The cultivation of silage maize has risen due to the increasing number of biogas plants and the good qualities of silage maize as a biogas substrate. However, the increased cultivation of maize is not always positively considered from the viewpoint of society and can lead to ecological problems. For reaching the goal of the German government that is the sustainable development of renewable energy, the cultivation of alternative biogas substrates has to be promoted. Flowering cover crops are a promising biogas substrate with additional ecological benefits. In the present study, a business simulation game is used to investigate whether the implementation of a reward and a penalty policy will improve the integration of flowering cover crops into the production programs of farmers. Participating farmers of the business simulation game manage a virtual farm and have to make cultivation and contract decisions through the duration of several production periods. The reward and the penalty policy which occur during the business simulation game do not differ in terms of their income effect. The results reveal that both the reward and the penalty policy lead to an increase in the cultivation of flowering cover crops. The implementation of a penalty policy has a more positive effect on the cultivation of flowering cover crops than the implementation of a reward policy, even though both policies do not differ in their income effect. Hence, it can be concluded that human behavior is influenced more by loss aversion and the opportunity cost effect, as well as the normative commitment of laws. Moreover, the results show
that the implementation of the reward policy has no effect on the cultivation of flowering cover crops to be used as a biogas substrate. In contrast, the introduction of the penalty policy promotes the use of flowering cover crops for biogas substrate. On the supposition of equal costs of policy implementation and the goal to increase the use of flowering cover crops as a biogas substrate, the implementation of a penalty policy seems to be more promising than a reward policy. Another finding is that socio-demographic and socio-economic variables influence the behavior towards different policy measures. The business simulation game can be seen as an important step to investigate the influence of sociodemographic and socio-economic effects on cultivation decisions. However, the results apply for young, well-educated farmers and are not representative of all German farmers. For policy impact analysis, the business simulation game is the first step to predicting the behavior of farmers regarding the implementation of a reward and a penalty policy. The results are very promising and provide initial suggestions and starting points for policymakers. Further research is needed concerning the group of participants. For instance, the experiment can be repeated with participants from different cultural and religious backgrounds or even by non-farmers to compare the decision behavior. Additionally, the height of the incentives can be analyzed with regard to the decisions. Furthermore, the question arises of whether our results are transferable to other policy measures and policy implementations.
Acknowledgements Thanks are extended to the anonymus referees and the editor for helpful advice and comments on a earlier version of this paper. The authors gratefully acknowledge financial support from the German Research Foundation (DFG).
references
[1] Act on granting priority to renewable energy sources (Renewable Energy Sources Act e EEG). [Internet] [cited 2013 Feb 12]. Available from: http://www.bmu.de/fileadmin/bmuimport/files/english/pdf/application/pdf/eeg_2012_en_bf.pdf. [2] Mbzibain A, Hocking TJ, Tate G, Ali S. Renewable enterprises on UK farms: assessing levels of uptake, motivations and constraints to widespread adoption. Biomass Bioenergy 2013;49(1):28e37. [3] German Biogas Association (Fachverband Biogas e.V.). Branchenzahlen 2013 und Prognose der Branchenentwicklung 2014. Entwicklung der Anzahl Biogasanlagen und der gesamten installierten elektrischen Leistung in megawatt [MW]. [Internet]. Freising: German Biogas Association; [updated 2013 Nov; cited 2014 Apr 30]. Available from: http://www.biogas.org/edcom/webfvb.nsf/ id/DE_Branchenzahlen/$file/13-11-11_Biogas% 20Branchenzahlen_2013-2014.pdf. [4] DESTATIS (Statistisches Bundesamt). Production. Gross electricity production in Germany from 2011 to 2013. [Internet]. Wiesbaden: DESTATIS; [cited 2014 Apr 30]. Available from: https://www.destatis.de/EN/FactsFigures/
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
[5]
[6]
[7]
[8]
[9] [10]
[11]
[12] [13]
[14]
[15]
[16] [17] [18] [19] [20] [21] [22]
[23] [24]
EconomicSectors/Energy/Production/Tables/ GrossElectricityProduction.html. FNR (Fachagentur Nachwachsende Rohstoffe e.V.). Cultivation of renewable resources in Germany 2012. [Internet]. Gu¨lzow-Pru¨zen: FNR; [cited 2014 Apr 30]. Available from: http://mediathek.fnr.de/media/downloadable/files/ samples/r/z/rz_fnr4_0252_grafik_nawaro_anbau_101013_ eng.jpg. FNR (Fachagentur Nachwachsende Rohstoffe e.V.). Cultivation area of maize in Germany. Year of cultivation 2013. [Internet]. Gu¨lzow-Pru¨zen: FNR; [cited 2014 Apr 15]. Available from: http://mediathek.fnr.de/media/ downloadable/files/samples/m/a/maisgrafik_2013_300_ cmyk.jpg. Herrmann A. Biogas production from maize: current state, challenges and prospects. 2. Agronomic and environmental aspects. Bioenergy Res 2013;6(1):372e87. German Biogas Association (Fachverband Biogas e.V.). Wasserschutz beim Anbau von Energiepflanzen fu¨r Biogasanlagen. Hinweispapier mit Handlungsempfehlungen. Freising: German Biogas Association; 2013 Jul [cited 2014 Apr 30]. Available from: http://sojus.dbfz.de/sites/default/files/ 13-07-11_Papier%2520Gew%C3%A4sserschutz_end.pdf. Weiland P. Biogas production: current state and perspectives. Appl Microbiol Biotechnol 2010;85(4):849e60. Hartmann A. Sorghum als Biogassubstrat e Ertragspotential bei optimaler und reduzierter Wasserverfu¨gbarkeit, vol. 24. Nu¨rthingen: Mitteilungen der Gesellschaft fu¨r Pflanzenbauwissenschaften; 2012. p. 238e9. Vollrath B, Kuhn W, Werner A. “Wild” statt “Mono” e neue € chsheim: Wege fu¨r die Biogaserzeugung [Internet]. Veitho Bayerische Landesanstalt fu¨r Weinbau und Gartenbau; 2010 [cited 2012 Aug 28]. Available from:, http://www.lebensraumbrache.de/_downloads/poster_und_veroeffentlichungen/ WildstattMono_LandinForm.pdf. € ren. DLZ Vollrath B, Werner A. Wildpflanzen rentabel verga Agrar 2012;12(12):42e6. Gong Y, Janssen M. From policy implementation to business process management: principles for creating flexibility and agility. Gov Inf Q 2012;29(1):61e71. Lensink S, Londo M. Assessment of biofuels supporting policies using the BioTrans model. Biomass Bioenergy 2010;34(2):218e26. Helming K, Diehl K, Kuhlman J, Jansson T, Verburg PH, Bakker M, et al. Ex ante impact assessment of policies affecting land use, part B: application of the analytical framework. Ecol Soc [Internet] 2011;16(1):29 [cited 2013 May 5]. Available from: http://www.ecologyandsociety.org/vol16/ iss1/art29/. Veetil VP. Conceptions of rationality in law and economics. Eur J Law Econ 2011;31(2):199e228. Camerer CF. Behavioral economics. Princeton: Princeton University Press; 2001. Camerer CF, Fehr E. When does “Economic Man” dominate social behavior. Science 2006;5757(311):47e52. Fehr E, Gaechter S. Fairness and retaliation: the economic of reciprocity. J Econ Perspect 2000;14(3):159e81. Benz M. Entrepreneurship as a non-profit-seeking activity. Int Entrep Manage J 2009;5(1):23e44. Selten R. Bounded rationality. J Inst Theor Econ 1990;146(4):649e58. Hertwig R, Ortmann A. Experimental practices in economics: a methodological challenge for psychologists? Behav Brain Sci 2001;24(3):383e451. Levitt SD, List JA. Field experiments in economics: the past, the present, and the future. Eur Econ Rev 2009;53(1):1e18. Haenke S, Scheid B, Schaefer M, Tscharntke T, Thies C. Increasing syrphid fly diversity and density in sown
[25]
[26] [27]
[28] [29]
[30]
[31]
[32]
[33]
[34]
[35]
[36] [37] [38] [39]
[40]
[41]
[42]
[43]
[44]
[45]
205
flowering cover crops within simple vs. complex landscapes. J Appl Ecol 2009;46(5):1106e14. Haaland C, Gyllin M. Butterflies and bumblebees in greenways and sown wildflowering cover crops in southern Sweden. J Insect Conserv 2010;14(2):125e32. Tyler TR. Why people obey the law. New Jersey: Princeton University Press; 2006. Tyler TR, Blader SL. Can businesses effectively regulate employee conduct? the antecedents of rule following in work settings. Acad Manage J 2005;48(6):1143e58. Kahneman D, Tversky A. Prospect theory: an analysis of decision under risk. Econometrica 1979;47(2):263e92. Kahneman D, Knetsch JL, Thaler RH. Anomalies. The endowment effect, loss aversion, and status quo bias. J Econ Perspect 1991;5(1):193e206. Wratten SD, Gillespie M, Decourtye A, Mader E, Desneux N. Pollinator habitat enhancement: benefits to other ecosystem services. Agr Ecosyst Environ 2012;159:112e22. Thomas F, Denzel K, Hartmann E, Luick R, Schmoock K. Kurzfassung der Agrarumwelt- und Naturschutzprogramme, Darstellung und Analyse der Entwicklung der Maßnahmen der Agrarumwelt- und Naturschutzprogramme in der Bundesrepublik Deutschland. Bonn: Bundesamt fu¨r Naturschutz; 2009. Buttel FH, Gillespie GW, Larson III OW, Harris CK. The social bases of agrarian environmentalism: a comparative analysis of New York and Michigan farm operators. Rural Sociol 1981;46(3):391e410. Carr S, Tait J. Differences in the attitudes of farmers and conservationists and their implications. J Environ Manage 1991;32(3):281e94. Luetz M, Bastian O. Implementation of landscape planning and nature conservation in the agricultural landscape e a case study from Saxony. Agric Ecosyst Environ 2002;92(2e3):159e70. Goslin E, Williams KJH. Connectedness to nature, place attachment and conservation behavior: testing connectedness theory among farmers. J Environ Psychol 2010;30(3):298e304. Holt CA, Laury SK. Risk aversion and incentive effects. Am Econ Rev 2002;92(5):1644e55. Dixit AK, Pindyck RS. Investment under uncertainty. Princeton: Princeton University Press; 1994. Harrison GW, List JA. Field experiments. J Econ Lit 2004;42(4):1009e55. Charness G, Gneezy U, Kuhn MA. Experimental methods: extralaboratory experiments e extending the reach of experimental economics. J Econ Behav Organ 2013 Jul;91:93e100. FNR (Fachagentur Nachwachsende Rohstoffe e.V.). Prices of biogas substrate [V/t FM]. [Internet]. Gu¨lzow-Pru¨zen: FNR; [cited 2014 Apr 15]. Available from: http://mediathek.fnr.de/ grafiken/daten-und-fakten/preise-und-kosten/preisebiogassubstrate-interaktiv.html. € hn W. Impact of weather and yield Chmielewski FM, Ko components of winter rye over 30 years. Agric For Meteorol 2000;102(4):253e61. De Wit AJW, Boogaard HL, Van Diepen CA. Spatial resolution of precipitation and radiation: the effect on regional crop yield forecasts. Agric For Meteorol 2005;135(1e4):156e68. Brick K, Visser M, Burns J. Risk aversion: experimental evidence from South African fishing communities. Am J Agric Econ 2012;94(1):133e52. Maart-Noelck SC, Musshoff O. Investing today or tomorrow? An experimental approach to farmers' decision behavior. J Agric Econ 2013;64(2):295e318. DESTATIS (Statistisches Bundesamt). Verdienste und Arbeitskosten, Verdienste in der Landwirtschaft, 2010.
206
[46]
[47]
[48] [49]
[50]
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Wiesbaden: Statistisches Bundesamt; 2011. Fachserie 16, Reihe 1. Cooper DJ, Kagel JH, Lo W, Gu QL. Gaming against managers in incentive systems: experimental results with Chinese students and Chinese managers. Am Econ Rev 1999;89(4):781e804. Friedman D. Monty Hall's three doors: construction and deconstruction of a choice anomaly. Am Econ Rev 1998;88(4):933e46. Harrison GW. Expected utility theory and the experimentalists. Empir Econ 1994;19(2):223e53. Camerer CF, Hogarth RM. The effects of financial incentives in experiments: a review and capital-labor-production framework. J Risk Uncertain 1999;19(1e3):7e42. Duersch P, Oechssler J, Schipper BC. Incentives for subjects in internet experiments. Econ Lett 2009;105(1):120e2.
[51] Ashton RH. Pressure and performance in accounting decision settings: paradoxical effects of incentives, feedback and justification. J Acc Res 1990;(28 Suppl.):S148e80. [52] Bothos E, Apostolou D, Mentzas G. Collective intelligence with web-based information aggregation markets: the role of market facilitation in idea management. Expert Syst Appl 2012;39(1):1333e45. [53] Cutler SJ, Hendricks J, Guyer A. Age differences in home computer availability and use. J Gerontol B Psychol Sci Soc Sci 2003;58(5):271e80. [54] DESTATIS (Statistisches Bundesamt). Landwirtschaft auf einen Blick. Wiesbaden: Statistisches Bundesamt; 2011. [55] Schram A. Artificiality: the tension between internal and external validity in economic experiments. J Econ Methodol 2005;12(2):225e37.
Available online at www.sciencedirect.com
ScienceDirect http://www.elsevier.com/locate/biombioe
Policy impact analysis of penalty and reward scenarios to promote flowering cover crops using a business simulation game Gesa Sophie Holst*, Oliver Musshoff, Till Doerschner Georg-August-University of Goettingen, Faculty of Agricultural Sciences, Department of Agricultural Economics and Rural Development, Farm Management Group, Platz der Goettinger Sieben 5, D-37073 Goettingen, Germany
article info
abstract
Article history:
Due to an increasing number of biogas plants and the positive qualities of maize as a biogas
Received 3 June 2013
substrate, the cultivation of silage maize has risen in Germany. However, there are still various
Received in revised form
reasons for the limitation of the cultivation area of silage maize. Hence, policymakers are
6 June 2014
currently discussing various alternative biogas substrates and ways to promote their cultiva-
Accepted 19 August 2014
tion. One possible alternative is the use of special flowering cover crops with additional
Available online 18 September 2014
ecological benefits. Using a business simulation game conducted with farmers, the present study investigates whether the implementation of a reward and penalty policy will improve
JEL classification:
the uptake of flowering cover crops in the production programs of farmers. The results indicate
C90
that the implementation of these policy measures was followed by a significant increase in the
Q24
cultivation area of flowering cover crops. The penalty policy leads to a stronger increase in the
Q28
size of the cultivation area of flowering cover crops than the reward policy, even though the
Keywords: Flowering cover crops
policies have the same income effect for farmers. Furthermore, the results reveal that the cultivation of flowering cover crops is influenced by various socio-demographic variables. © 2014 Elsevier Ltd. All rights reserved.
Alternative biogas substrate Business simulation game Policy impact analysis Reward policy Penalty policy
1.
Introduction
Fossil energy sources are finite resources that contribute significantly to anthropogenic global warming through the release of CO2 emissions. Therefore, the German government has decided to promote the expansion of renewable energy sources with the ‘Act on granting priority to renewable energy sources’ [1]. This act determines the remuneration for electricity generated from renewable energy sources and aims to
realize a 35% and 85% share of renewable energy in total electricity generation by 2020 and by 2050, respectively. Due to the profitability and possible income stabilization, many farmers have invested in renewable energies [2]. In Germany, the number of biogas plants increased between 2002 and 2012 from 1600 to 7515 with a total installed capacity of 3.352 GWh [3]. Moreover, the total electricity production from biomass is the second most important source of renewable energy with a share of 27.7%, after wind power with a share of 35.3% [4]. Thus, the cultivation of energy crops,
* Corresponding author. Tel.: þ49 0 551 39 4836. E-mail addresses: [email protected], [email protected] (G.S. Holst). http://dx.doi.org/10.1016/j.biombioe.2014.08.009 0961-9534/© 2014 Elsevier Ltd. All rights reserved.
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
such as maize and whole crop silage, has increased sharply. In 2013, energy crops for fermentation in biogas plants were grown on a total surface area of 11,570 km2 [5]. The cultivation of energy maize accounts for 8000 km2 of the aforementioned surface area [6]. Due to its high yield of dry matter and its energy content, maize is the preferred crop for biogas production [7]. The expansion of energy production from biomass, however, is not necessarily considered to be positive as it can lead to serious environmental problems, such as the pollution of ground water by nutrients or the loss of organic matter in farmland [7]. For the future development of electricity production from biomass, it is essential to reconcile the production of biogas substrate and nature conservation. An initial step for policymakers in Germany is to restrict the use of silage maize and grain, including corn-crop mix, grain maize, and ground ear maize in biogas plants to 60 mass fraction, which is anchored in the ‘Act on granting priority to renewable energy sources’ and became effective on 01.01.2012 [1]. A variety of alternative biogas substrates, such as cup plant, Sudan grass and sorghum, have been discussed [8e10]. In addition, the fermentation use of silage produced from flowering cover crops in biogas plants is being explored. Initial results show that flowering cover crops are well suited for fermentation in biogas plants [11,12]. Further advantages include low input, the creation of habitats for wildlife, as well as the increasing acceptance shown by the positive public response to fields that are used for, or surrounded by, flowering cover crops. Bees and other insects use the flowering crops as a food source and wildlife offer these crops cover along with a secure place to deliver. The harvest takes place in October when most flowers are withered and the animals and insects have left the field [11,12]. For the aforementioned reasons, a political goal could be the integration of flowering cover crop cultivation into the production programs of farmers. However, the introduction of a new policy is accompanied by high costs [13]. Prior to the introduction of a policy, a policy impact analysis is essential in evaluating whether a policy measure is effective. The development of models which simulate the consequences of a policy implementation may be an opportunity [14,15]. Frequently used models for policy impact analysis assume a perfect rationally behaving profit maximizer [16]. However, it is often discussed that assuming profit maximizing behavior is not appropriate [17e19]. Explanatory approaches regarding the concept of utility maximization suppose that people maximize their utility and, therefore, various objectives, such as making profits, risk aversion, traditions, recreational activities, or social recognition are pursued [20]. Another explanatory approach is the bounded rationality [21]. Thus, models assuming rational decision makers could reflect the distorted consequences of policy implementation. These limitations can be addressed by experiments which are not based on decisions given through exogenously predetermined theories. Instead, we observe the real decisions of real people. In both laboratory experiments and in business simulation games, it is possible to set incentives for motivating participants to make well-conceived decisions [22]. Business simulation games make it possible to design a realistic decisionmaking situation, resulting in a significant advantage
197
compared to classical laboratory experiments [23]. Thus, they seem to be especially suitable for policy impact analysis. This study explicitly examines the farmers' reaction to the implementation of policies in order to promote the share of flowering cover crops in the agricultural landscape. For this purpose, the multi-period, single-person business simulation game is developed. Additionally, the participant farmers will be confronted with different policy measures. The following questions will be addressed in the business simulation game: 1. Does the implementation of reward and penalty policies have an impact on the proportion of flowering cover crops in the production programs of the farmers? 2. Is either a reward or a penalty policy with the same income effect more effective? 3. Does the policy change lead to the cultivation of flowering cover crops as a biogas substrate? The novelty of this paper lies in the policy impact analysis that is geared towards implementing flowering cover crops in the production programs of farmers. In recent years, studies have dealt with flowering crops and their environmental benefits, with primary research focusing on the nature conservation concept and the impacts on biodiversity [24,25]. Vollrath et al. [11] and Vollrath and Werner [12] investigated the benefits of flowering cover crops as a biogas substrate. To our knowledge, there are no publications that address the individual effect of policies to increase the quantity of flowering cover crops in the agricultural landscape. Furthermore, a new aspect to this research is that a business simulation game conducted with real decision makers e in our case farmers e is used for the policy impact analysis. The article is structured as follows: First, the behavioral theoretical hypotheses are derived (Section 2). Sections 3 and 4 explain the experimental design and sample characteristics. The results are presented in Section 5 and the article concludes with a summary and a discussion of future opportunities (Section 6).
2.
Hypothesis generation
Reward and penalty strategies encourage human compliance through the use of rules and laws and, in this way, establish a social order [26]. Penalty payments pursue a strategy of deterrence in order to prevent rules being broken, whereas rewards present an incentive to direct human behavior in a desired direction [27]. Accordingly, it is assumed that a reward and a penalty policy strategy can direct the behavior of farmers, so that the cultivation of flowering cover crops is extended. Thus, the following hypothesis can be derived: H1. Regardless of whether the policymakers introduce a reward for growing flowering cover crops or a penalty with the same income effect for not growing flowering cover crops, the share of flowering cover crops in the production program of farmers will increase. In the economic literature, it has been hinted that the effects of penalty policies differ from reward policies [26]. Kahneman and Tversky [28], as well as Kahneman et al. [29] found
198
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
evidence for loss aversion meaning that people weigh a monetary loss more heavily than an equally high profit. Concerning a reward and a penalty policy with the same income effect means that the loss of the penalty payment is weighed higher than the reward payment. Even the opportunity cost effect supports the assumption that reward and penalty policies differ in their effect on human behavior. Out of pocket costs, such as penalty payments, are given a higher weight than opportunity costs, which correspond to a foregone reward payment [29]. Another reason for the different impact of penalty and reward policies is the normative commitment of laws. If laws make sense for people and they accept them, they do not break the laws even if it is possible without facing consequences [26]. Therefore, the following hypothesis can be derived: H2. The penalty policy changes the cultivation behavior of farmers regarding the scope of flowering cover crops better than a reward policy with the same income effect. Flowering cover crops can be cultivated as nature conservation measures [30,31], or even as a biogas substrate for energy production with additional ecological benefits [12]. If new policy measures promote the use of flowering cover crops as a biogas substrate, we assume that the cultivation area of flowering cover crops used as a biogas substrate increases because of the higher revenues that are generated compared to the cultivation of these crops for nature conservation. Hence, flowering cover crops are used to complement the production of biogas substrate. Thus, the following hypothesis is derived: H3. The implementation of reward and penalty policies would promote the cultivation of flowering cover crops for biogas substrate. Farmers that cultivate flowering cover crops behave ecoconsciously, because they create habitats for wildlife and insects. Regarding the socio-demographic and socio-economic variables and their influence on the environmental behavior of farmers, only limited literature exists. Buttel et al. [32] investigated the impact of age, education, income, size of agricultural enterprises, and the organic or conventional land use on the environmental behavior. As significant factors, only the negative effect of age and the positive effect of the organic operation focus can be detected. Carr and Tait [33] indicate that the productivity and effectiveness idea can affect the management behavior of farmers with detriment to nature conservation (see also Ref. [34]). The connectedness of farmers to nature influences the environmental concern [35]. This leads to the following hypothesis: H4. Socio-demographic and socio-economic variables impact the cultivation of flowering cover crops for nature conservation and for the production of biogas.
3.
Design of the experiment
The experiment is divided into three sections. In the first section, the incentive compatible, multi-period, one-person business simulation game is conducted. Subsequently, a Holt-
and-Laury lottery (HLL) is carried out [36] to determine the risk attitude of the participants (Section 2). Finally, sociodemographic and socio-economic data from the participants are collected (Section 3).
3.1.
General structure of the business simulation game
The design of the business simulation game aims to enable the participants to experience business situation that is as realistic as possible with realistic decision situations [38]. However, business simulation games can only model real events in its essential characteristics. Therefore, the market price levels and yields are reflected as uncertainties. The experiment is conducted with farmers who know that they are participating in an experiment and that their behavior will be recorded and analyzed. Hence, a non-standard subject pool is used. In addition, the experiment has been conducted at an exhibition and not in a laboratory situation. Consequently, the business simulation game can be classified according to Charness et al. [39] as an extra-laboratory experiment. The participants are told to assume that they are a manager of a virtual farm with 100 ha of farmland. They have to run the farm for 12 production periods. Each round of the business simulation game stands for a production period and requires the following basic decisions of the participants: 1. Production program decision: Configuration of the production program for the cultivation of the farmland with the production activities winter wheat, silage maize, sorghum and flowering cover crops. 2. Contract decision: Acceptance of a substrate delivery contract for an adjacent biogas plant for 0 t, 1500 t, 3000 t or 4500 t of fresh matter. For fulfillment, it is possible to use silage maize, sorghum and flowering cover crops. In the business simulation game, (1) deterministic and (2) stochastic parameters are given. Ad (1): The deterministic parameters are communicated at the beginning of the business simulation game, do not change randomly and apply to all participants. When the business simulation game starts, each participant has a starting capital of V 100,000. In each production period, withdrawals in the amount of V 30,000 are made to cover the costs of living. To guarantee a crop rotation with at least two cultures, minimum cultivation requirements exist for winter wheat and silage maize. No minimum cultivation requirements exist for growing sorghum and flowering cover crops. All production activities can be grown on a maximum of 70 ha of farmland (Table 1). It is noteworthy that the entire available land (100 ha) must be cultivated by using the aforementioned four crops; there is no option to set aside some land for other uses. The processing options of each crop are shown in Table 1. Winter wheat is only used for selling on the spot market, whereas silage maize is sold and used for fulfilling the biogas substrate delivery contract. The cultivation of sorghum is solely for the production of biogas substrate. Flowering cover crops, however, have two alternative forms of use; they can be grown as a substrate for biogas plants with additional ecological benefits, or they may remain unused for ecological reasons and are then compensated for by a payment from the
199
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Table 1 e Overview of the four crops for farmland cultivation. Wheat
Silage maize
Sorghum
Processing option (market price)
Spot market (volatile price)
Biogas (35 V t1); Spot market (volatile price)
Biogas (35 V t1)
Biogas (35 V t1)
Flowering cover crops Nature conservation (640 V ha1)
Cost of cultivation
970 V ha1
832 V ha1
800 V ha1
880 V ha1
340 V ha1
Minimum extent Maximum extent
5 ha 70 ha
5 ha 70 ha
0 ha 70 ha
0 ha 70 ha
Weather impact on yield above-average (20%) average (60%) below-average (20%)
9.6 t ha1 8.0 t ha1 6.4 t ha1
60.5 t ha1 55.0 t ha1 49.5 t ha1
53.5 t ha1 48.0 t ha1 42.5 t ha1
46.0 t ha1 40.0 t ha1 34.0 t ha1
No yield measured, biodiversity
Table 2 e Market price developments of winter wheat (left) and silage maize (right) from period 0 to period 1. Current wheat price (period 0) 1
200.00 V t
Uncertain winter wheat price in the following production period (period 1) 220.00 V t1 probability of 50% 180.00 V t1 probability of 50%
state. The contracted substrate delivery is remunerated at 35 V t1, regardless of whether the biomass is provided with maize silage, sorghum or flowering cover crops. Ad (2): The stochastic parameters change randomly from one production period to another and, therefore, vary among the participants. The market prices for winter wheat and silage follow an arithmetic Brownian motion [37] starting with an identical initial value for all participants (Table 2). The market prices decrease or increase, starting with the current price in any production period with a probability of 50% by 20 V t1 for winter wheat and by 1.50 V t1 for silage maize. Furthermore, weather conditions will affect yield and, accordingly, gross margins. There is a distinction between above-average, average, and below-average weather conditions (Table 1). The periods of above-average and belowaverage weather occur each with a probability of 20%. Average weather is expected with a probability of 60%. Aboveaverage weather has the consequence that the yields of all production activities per hectare achieve its maximum, whereas below-average weather leads to a yield drop to the minimum. Both the probabilities of the weather conditions and the yields per hectare are communicated at the beginning of the business simulation game. Despite the uncertain yields, the chosen delivery contract has to be fulfilled by 100%. If this is not the case, the missing substrate amount has to be bought from the market for twice the current market price of silage maize. Each participant receives a transfer payment of 300 V ha1 of farmland after each completed period of production. At the same time, the participants are informed that amendments may occur in the course of the business simulation game. Since there are no storage possibilities for crops, all goods are sold at the end of each production period at the current market prices. The current prices and the weather conditions of the previous period are communicated at the beginning of each new production period. Moreover, the participants receive further information about the profit of the realized
Current silage maize price (period 0) 25.00 V t
1
Uncertain silage maize price in the following production period (period 1) 26.50 V t1 probability of 50% 23.50 V t1 probability of 50%
production program and the contract decisions of the previous production periods.1
3.2.
Changes in the policy framework
At the beginning of the business simulation game, participants are randomly assigned to one of three policy scenarios. During the first six production periods of the business simulation game, the policy framework is identical for all three policy scenarios. However, in the subsequent production periods, 7 through 12, the policy framework conditions change whereby the policy scenarios are defined as follows: Scenario 1 (reference scenario): The policy framework remains unchanged over the entire duration of the business simulation game. Scenario 2 (reward scenario): The participants are informed that the transfer payment decreases by 10% to 270 V ha1. Simultaneously, policymakers introduce an additional premium of 300 V for each ha used for growing 1
Experiments aim to model real events in its essential characteristics. To handle the conflict of real characteristics on the one hand and the virtual world of business simulation games on the other hand, we made some simplifications. One simplification is the same price per ton of biogas substrate whereby 35 V t1 is a realistic value taken from the literature [40,12]. Furthermore, we have chosen an arithmetic Brownian motion [37] with zero drift to indicate that the prices are uncertain. Further simplifications are made with respect to the weather conditions. Weather conditions consist of various factors influencing the yields of crops, primarily through temperature, draughts [41], precipitation and radiation [42]. The interaction of these factors and the stage of plant growth have an important influence on yield effects [41,42]. Therefore, we summarize factors which result in higher yields as above-average conditions and weather conditions resulting in low yields as below-average conditions. Additionally, we abstained from the opportunity of storage for the purpose of simplicity and for the reason that we do not receive any relevant information for answering our research questions.
200
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
flowering cover crops. The state pays a maximum amount of V 3000 per farm and, thus, subsidizes a maximum cultivation area of 10 ha of flowering cover crops. Scenario 3 (penalty scenario): The policy punishes all participants who use less than 10% of their farmland for the cultivation of flowering cover crops. Each hectare below the fulfillment requirement will incur a penalty of V 300. Scenarios 2 and 3 do not differ in their income effect for farmers. In order to compare the participants' behavior in the scenarios, we construct triplets of randomly selected participants, i.e., there is always three individual participants playing the business simulation game with the same price and weather developments, but with different underlying policy scenarios.
3.3.
Holt and Laury lottery
The Holt-and-Laury lottery is an experimental method used to identify the risk attitude of people. This methodology presented by Holt and Laury [36] has not been changed and is already established in agricultural economics [43,44]. First, participants are introduced to the game with the explanation of the lottery whereby participation is completely voluntary and the outcomes are always positive. The participants are asked to choose between the two lotteries: A (i.e., the safer alternative) and B (i.e., the riskier alternative). In lottery A, the participants can either win V 200 or V 160, whereas the riskier lottery B offers prize money of either V 385 or V 10. The probabilities of achieving the two possible prize moneys in the lotteries are systematically varied at 10% intervals beginning at either V 200 in lottery A (V 385 in lottery B) with a possibility of 10%, or V 160 (V 10 in lottery B) with a probability of 90%. Consequently, there exist ten different decision situations whereby the expected value increases in both lotteries. In decision situation five, the expected value of lottery B becomes higher than the expected value of lottery A. The HLLvalue indicates the number of safe choices when the decision from lottery A changes to the more risky lottery B and represents the risk attitude. Farmers who chose lottery A four times are risk neutral. When farmers chose the safer lottery A one to three times, it indicates risk seeking whereby an HLLvalue between five and nine shows risk aversion.
3.4.
Incentives for well-conceived decisions
In order to attract participants, each participant receives a representation allowance of V 10. The planned playing period of 30 min corresponds to an average hourly wage of V 20. In the German agricultural sector, the average hourly wage is V 9.92 [45]. Hence, the allowance is more than twice as high as the average hourly wage and should motivate the farmers to participate in the business simulation game. For ensuring incentive compatibility, additional cash prizes can be won, V 2005 in total. This amount of money corresponds to an expected value of V 16.71 for each of the planned 120 participants. Since the playing period is already paid by the representation allowance, the expected gain covers the opportunity costs of additional 50 min to encourage well-conceived decisions. We did not define what is meant by ‘well-conceived’ decisions in the instructions for the
participants for the reason that we did not want to influence the participants' decisions, however, the incentives depend on the financial success of the virtual farms led by the participants. As in reality, aiming for multiple objectives, e.g., nature conservation or risk reduction decreases the expected profits. Hence, if participants aim for other objectives than profits in the business simulation game, the expected incentive decreases and our experiment becomes more realistic. To set incentives in economic experiments is common [46e48]. Former studies have shown that incentives influence the behavior of participants and help to improve decision making [46,48]. They work harder and think twice if the incentive depends on their own performance [49,50], thus, problem solving capabilities can be improved. Nevertheless, incentives can also cause biases [47,51]. Camerer and Hogarth [49], however, summarize that incentives more often result in better experimental findings than in causing biases. Different payment systems for incentives exist. On the one hand, each participant receives a small incentive while, on the other hand, a fraction of all participants receives a high incentive. However, participants overestimate the probability of being selected if only one participant is paid [49]. Therefore, we decided that only a fraction of participants are able to receive high incentives. Out of the 120 participants that are anticipated to take part in the business simulation game, four of these participants will be drawn by lottery to win a cash prize. The first three cash prizes depend on the participants' financial success in the business simulation game and amount to a maximum of V 540. The three randomly drawn winners receive the share of the maximum monetary gain that corresponds to their financial success.2 The fourth cash prize depends on the decisions in the Holtand-Laury lottery. For the determination of the cash prize, one participant is randomly drawn by a lottery and will have the chance to win between V 10 and V 385 e depending on his risk attitude e in a Holt-and-Laury lottery carried out solely for that participant.
4.
Description of the sample
The experiment was conducted at the agricultural exhibition “EuroTier” that took place between the 13th and 16th of November, 2012 in Hanover (Germany). Nine hundred and forty-six visitors of the exhibition were randomly chosen and directly contacted and were invited to participate in the business simulation game. In total, 123 farmers (13% of those contacted) successfully completed the experiment, with 41 farmers playing in each policy scenario. To complete the experiment, the participants needed 43 min on average. The socio-demographic and socio-economic characteristics of the participants in the three policy scenarios are summarized in Table 3. 2
We do not rank the participants with regard to their business success and do not pay money to the first ranked farmers as is often done in stock exchange simulation games [52]. Paying the best could lead to risk seeking behavior. As the winners are randomly drawn through a lottery by us, we assume that the risk preference is not influenced.
201
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Table 3 e Socio-demographic and socio-economic characteristics of the participants (41 participants in each policy scenario). Characteristics
Age (in years) Percent of female participants Years of education HLL-valuea Farm manager Farms farmed on a part-time basis Farmland (in ha) a
Policy Scenario 1
Policy Scenario 2
Policy Scenario 3
Mean
SD
Mean
SD
Mean
SD
31.8 12.2% 13.6 5.9 39.0% 19.5% 225.5
12.7 e 3.3 2.0 e e 392.3
28.8 14.6% 13.7 5.2 34.1% 7.3% 196.9
10.3 e 3.5 2.2 e e 220.4
27.7 9.8% 13.8 5.1 29.3% 14.6% 312.5
9.4 e 3.2 1.6 e e 617.4
1e3 ¼ risk seeking, 4 ¼ risk neutral, 5e9 ¼ risk averse.
On average, the participants were 29 years old, including the youngest participant at 16 years and the oldest at 62 years. A possible reason for the low average age may be the computer-based aspect of the experiment, which might discourage older farmers from participating [53]. The agricultural enterprises of the participants are on average larger than the average German farm with 56 ha of farmland [54]. Moreover, 55% of the farms in Germany are farmed on a part-time basis [54] whereas only 14% of the participants in our experiment run their agricultural businesses on a parttime basis. Hence, the sample characteristics do not match with the common characteristics of German farmers. However, the aim of our experiment is not the representativeness of the sample, but rather to test economic theories, i.e., the effects of the implementation of reward and penalty policies with the same income effect. There is a widespread consensus that the benefits of internal validity are more important to people than the lack of external validity if the experiments aim is to test theories, as is the case in our study [55]. Using the H-test according to Kruskal and Wallis, it can be shown that the socio-demographic and socio-economic characteristics of the participants in the three policy scenarios do not differ significantly in age (p-value ¼ 0.140), years of education (p-value ¼ 0.961), the HLL-value (p-value ¼ 0.228) and farmland (p-value ¼ 0.759). In addition, a chi-square-test indicates that there are no significant differences between the policy scenarios in the percentage of female participants (p-value ¼ 0.800), farm managers (p-value ¼ 0.855) and parttime farmers (p-value 0.273).
5. Behavioral control impacts of different policy measures on farmers In the first six production periods, the basic conditions are the same for all participants. It can therefore be analyzed how the cultivation area of flowering cover crops changes when the reward or penalty policy appears. The average size of the cultivation area of flowering cover crops is presented in Table 4 for the production periods 1e6 and 7e12 for each of the three policy scenarios. The H-test, according to Kruskal and Wallis, shows that the cultivation areas of flowering cover crops do not differ significantly between the three policy scenarios in periods 1e6 (p-value ¼ 0.812).
For Scenario 1 (reference scenario), the paired t-test shows that the size of the cultivation area in the first six periods does not differ significantly from the size of the cultivation area in periods 7e12 (p-value ¼ 0.732). In contrast, the cultivation area changed significantly with the introduction of a reward (Scenario 2) or penalty policy (Scenario 3) (p-value ¼ 0.080 and pvalue ¼ 0.001, respectively). Fig. 1 provides an overview regarding the success of the implementation of the reward and the penalty policy. Due to the same income effect of the reward and the penalty policy in 21.54% of all cases, the policy compliance is economically advantageous, while, in 78.46% of all cases, the policy compliance is not economically advantageous when aiming at profit maximization. If policy compliance is economically advantageous, 75.47% of the participants grow 10 ha or more of flowering cover crops in the reward policy and 83.02% of the participants in the penalty policy. With regard to ‘policy noncompliance is economically advantageous’ in the reward policy, 54.92% of the participants grow 10 ha or more of flowering cover crops. Comparing the penalty policy to these results, 77.72% of the participants grow 10 ha or more of flowering cover crops when ‘non-compliance is economically advantageous’ to maximize profits. This makes clear that the penalty scenario has a stronger effect on compliance. The difference between maximum possible and maximum achievable profit with policy compliance depicts the profit differential which is used for result estimation. The average profit differential is V 563.47. When policy compliance is economically advantageous for maximizing profits, the profit differential is V 0, per definition. The maximum profit differential is V 2044.33.
Table 4 e Size of the cultivation area of flowering cover crops in the business simulation game in ha (41 participants in each policy scenario). Policy Scenario
1 2 3 Mean
Area of flowering cover crops
Area of flowering cover crops
Periods 1e6
Periods 7e12
Mean
SD
Mean
SD
10.21 7.92 8.22 8.78
13.55 11.86 10.74 12.13
9.88 10.01 12.73 10.87
12.68 9.87 10.31 11.09
Change in average
0.33 2.09 4.51 2.09
202
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Fig. 1 e Success of the implementation of the reward and the penalty policy in production periods 7e12 (41 participants in each policy scenario). (a) For the percentages serve policy compliance economically advantageous as a baseline. (b) For the percentages serve policy non-compliance economically advantageous as a baseline.
Due to having 12 observations in the production periods of the business simulation game for each of the 123 participants, our dataset depicts panel data. The BreuschePagan-test indicates that a random effects regression has to be favored with regard to an OLS regression (p-value < 0.001). However, a random effects regression is not consistent according to the Hausman-test (p-value ¼ 0.047) and a fixed effect model is not suitable because time-constant variables relating to the sociodemographic and socio-economic characteristics of the participants, need to be included in the estimation. Hence, we estimated another model which is a hybrid regression with random slopes. Time-varying covariates are included with a time-constant effect which is the individual average of the variable and a context variable which is the value of each period adjusted by the individual average. Unobserved heterogeneity can be controlled by the context variable and the integration of random slopes whereupon the estimation becomes more precise indicated by the AIC. From the business simulation game, the following four initial situations and are used for the result estimation: The cropping areas of flowering cover crops for all three policy scenarios for the first six periods represent the reference situation (Situation 1). The policy Scenarios 1e3, which occur in the seventh production period, reflect the policy implementation effects (Situations 1e3). In Table 5, they are referred to as dummy Scenarios 1, 2 and 3 and deal with the baseline of the first six production periods in the estimated models. Three different models have been estimated. Model 1 reflects the effects of the independent variables on the total cropping area of flowering cover crops. Furthermore, Models 2 and 3 highlight the effects separated for the cultivation of flowering cover crops for biogas production or nature conservation purposes. Hypothesis H1 which assumes that reward and penalty policies promote the cultivation of flowering cover crops, is confirmed: In both, the reward and the penalty scenario, the share of flowering cover crops increases significantly (Model 1). Participants confronted with the reward policy grow 2.285 ha more on average of flowering cover crops in comparison to the reference periods 1e6. Whereas participants confronted with the penalty policy increase the share of flowering cover crops by 4.584 ha on average. Compared to the
reference situation of periods 1 thought 6, no significant changes occur in the reference scenario and periods 7 through 12. Consequently, the incentive strategy of the reward policy and the deterrence strategy of the penalty policy caused a change in the participants' behavior regarding the cultivation of flowering cover crops, this result can also be seen in Fig. 2. From the sixth to the seventh production period, the cultivation area of flowering cover crops increases for the reward and the penalty policy. However, it can also be seen that the profit differential affects the behavior of cultivating flowering cover crops. For each V 1000 profit differential, the participants grow 5.291 ha on average less of the flowering cover crops. Hypothesis H2 assuming that penalty policies have a stronger impact on the cultivation area of flowering cover crops than reward policies is confirmed: The implementation of a penalty policy leads to a stronger increase in the size of the cultivation area of flowering cover crops than the implementation of a reward policy (Model 1), although the policies do not differ in their income effect. A linear restriction reveals that the effects of the reward and the penalty policy differ from each other at a significance level of 10%. Thus, the loss aversion and the opportunity cost effect, as well as the normative commitment of laws have to be taken into account when estimating policy consequences. These effects influence the awareness of reward and penalty scenarios. Hypothesis H3 which supposes that reward and penalty policies lead farmers to grow flowering cover crops as a biogas substrate, is confirmed: A penalty policy results in an increase in the cultivation of flowering cover crops to produce biogas substrate by 2.535 ha (Model 2), this increase can also be seen in Fig. 2. In contrast, the introduction of a reward policy has no significant effect on the cultivation of flowering cover crops for biogas plants. Despite having the same income effect of policy Scenarios 2 and 3, only the penalty policy (Scenario 3) achieved a significant increase in the cultivation of flowering cover crops in order to produce biogas substrate. Socio-demographic and socio-economic variables affect the cultivation of flowering cover crops is assumed in hypothesis H4 which is confirmed: The model results in Table 5 show that socio-demographic and socio-economic variables have a significant influence on the cultivation area of flowering cover crops. Female participants grow significantly more
203
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Table 5 e Hybrid regression with random slopes to explain the cultivation area of flowering cover crops (N ¼ 1476).a
Constant Dummy Scenario 1 Dummy Scenario 2 Dummy Scenario 3 Average individual profit differentialb Profit differentialc Risk attituded Substrate delivery amount 1500 t1 Substrate delivery amount 3000 t1 Substrate delivery amount 4500 t1 Age Gendere Years of education Earning powerf Farmingg Renewable energyh Profit flowering cover cropsi Nature flowering cover cropsj Agri-environmentk AIC a b c d e f g h i j k
Model 1 Dependent variable: land used for flowering cover crops
Model 2 Dependent variable: land used for flowering cover crops, biogas plant
Model 3 Dependent variable: land used for flowering cover crops, nature conservation
Coefficient
t-Statistic
Coefficient
t-Statistic
Coefficient
t-Statistic
30.023 0.164 2.285 4.584 5.291 0.178 0.098 0.309 0.304 1.232 0.082 7.281 0.656 6.273 4.742 2.546 2.872 1.855 1.000
5.003*** 0.235 2.706*** 5.947*** 1.974* 0.031 0.269 0.257 0.257 0.989 1.272 3.373*** 2.655*** 3.103*** 1.760* 1.757* 1.928* 1.171 1.234
20.258 0.098 0.880 2.535 2.527 0.570 0.105 2.793 3.004 3.469 0.088 3.779 0.271 2.474 2.889 1.711 0.747 2.032 1.015
5.547*** 0.184 1.516 4.330*** 1.644 1.268 0.504 3.126*** 3.431*** 3.784*** 2.380** 3.064*** 1.928* 2.152** 1.884* 2.073** 0.880 2.249** 2.199**
8.999 0.004 1.176 2.049 2.718 0.595 0.006 2.670 2.925 5.206 0.006 3.380 0.380 3.773 1.921 0.807 2.147 0.189 0.003
2.194** 0.008 2.109** 3.645*** 1.506 1.394 0.026 3.063*** 3.406*** 5.770*** 0.129 2.327** 2.288** 2.777*** 1.061 0.828 2.144** 0.178 0.006
10,617.69
9784.86
9635.57
*p-value < 0.10; **p-value < 0.05; ***p-value < 0.01. Individual average of the difference between maximum possible and maximum achievable profit with policy compliance in V 1000. Observed profit differential minus individual average profit differential. 1e3 ¼ risk seeking, 4 ¼ risk neutral, 5e9 ¼ risk averse. 1 ¼ female, 0 ¼ male. 1 ¼ mainstay farm, 0 ¼ part-time farm. 1 ¼ conventional, 0 ¼ organic. Have you invested in renewable energies aside from biogas plants in reality? 1 ¼ yes, 0 ¼ no. Do you think it is possible to earn money by growing flowering cover crops in reality? 1 ¼ yes, 0 ¼ no. Do you think the cultivation of flowering cover crops for nature conservation makes sense in reality? 1 ¼ yes, 0 ¼ no. Do you agree or disagree with agri-environmental measures in reality? 1 ¼ total disagree to 5 ¼ total agree.
flowering cover crops for biogas production and for nature conservation than men (Models 1, 2 and 3). Consequently, female farmers may be more open to alternative methods of cultivation, as well as demonstrating concerns for nature
conservation. Participating farmers, for whom the farm is the main source of income, grow significantly fewer flowering cover crops for biogas plants and for nature conservation. Participants who responded to the questionnaire that it is
Fig. 2 e Average cultivation area of flowering cover crops and average cultivation area of flowering cover crops for biogas substrate in the production periods 1e12 (41 participants in each policy scenario).
204
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
possible to make profits with the cultivation of flowering cover crops in reality grow more of the flowering cover crops for nature conservation. However, there is no significant effect upon the size of the cultivation area of flowering cover crops for biogas plants. The variables ‘gender’, ‘years of education’, ‘earning power’ and ‘farming’ have a highly significant influence on the size of the cultivation area of flowering cover crops (Model 1). Interestingly, these effects differ if the two kinds of flowering cover crops are separately estimated (Models 2 and 3). The covariates ‘gender’, ‘years of education’ and ‘earning power’ have a significant impact on the size of the cultivation area of flowering cover crops on both variations but in varying degrees. Considering Model 2, the variable ‘age’ is significant. In contrast, there is no significant influence of this variable on the size of the cultivation area of flowering cover crops for nature conservation (Model 3). The negative impact of ‘age’ on the size of the cultivation area of flowering cover crops for energy production implies that older farmers may be more skeptical towards alternative biogas substrates. The effect of the substrate delivery contracts differs with regard to the production alternative of flowering cover crops. Whereas the decision for delivery contracts of 1500 t, 3000 t and 4500 t has a positive influence on the cultivation of flowering cover crops as a biogas substrate. It has a negative impact on the size of the cultivation area of flowering cover crops for nature purposes. It becomes clear that there are different factors influencing the reaction and implementation of policies.
6.
Conclusion and outlook
The cultivation of silage maize has risen due to the increasing number of biogas plants and the good qualities of silage maize as a biogas substrate. However, the increased cultivation of maize is not always positively considered from the viewpoint of society and can lead to ecological problems. For reaching the goal of the German government that is the sustainable development of renewable energy, the cultivation of alternative biogas substrates has to be promoted. Flowering cover crops are a promising biogas substrate with additional ecological benefits. In the present study, a business simulation game is used to investigate whether the implementation of a reward and a penalty policy will improve the integration of flowering cover crops into the production programs of farmers. Participating farmers of the business simulation game manage a virtual farm and have to make cultivation and contract decisions through the duration of several production periods. The reward and the penalty policy which occur during the business simulation game do not differ in terms of their income effect. The results reveal that both the reward and the penalty policy lead to an increase in the cultivation of flowering cover crops. The implementation of a penalty policy has a more positive effect on the cultivation of flowering cover crops than the implementation of a reward policy, even though both policies do not differ in their income effect. Hence, it can be concluded that human behavior is influenced more by loss aversion and the opportunity cost effect, as well as the normative commitment of laws. Moreover, the results show
that the implementation of the reward policy has no effect on the cultivation of flowering cover crops to be used as a biogas substrate. In contrast, the introduction of the penalty policy promotes the use of flowering cover crops for biogas substrate. On the supposition of equal costs of policy implementation and the goal to increase the use of flowering cover crops as a biogas substrate, the implementation of a penalty policy seems to be more promising than a reward policy. Another finding is that socio-demographic and socio-economic variables influence the behavior towards different policy measures. The business simulation game can be seen as an important step to investigate the influence of sociodemographic and socio-economic effects on cultivation decisions. However, the results apply for young, well-educated farmers and are not representative of all German farmers. For policy impact analysis, the business simulation game is the first step to predicting the behavior of farmers regarding the implementation of a reward and a penalty policy. The results are very promising and provide initial suggestions and starting points for policymakers. Further research is needed concerning the group of participants. For instance, the experiment can be repeated with participants from different cultural and religious backgrounds or even by non-farmers to compare the decision behavior. Additionally, the height of the incentives can be analyzed with regard to the decisions. Furthermore, the question arises of whether our results are transferable to other policy measures and policy implementations.
Acknowledgements Thanks are extended to the anonymus referees and the editor for helpful advice and comments on a earlier version of this paper. The authors gratefully acknowledge financial support from the German Research Foundation (DFG).
references
[1] Act on granting priority to renewable energy sources (Renewable Energy Sources Act e EEG). [Internet] [cited 2013 Feb 12]. Available from: http://www.bmu.de/fileadmin/bmuimport/files/english/pdf/application/pdf/eeg_2012_en_bf.pdf. [2] Mbzibain A, Hocking TJ, Tate G, Ali S. Renewable enterprises on UK farms: assessing levels of uptake, motivations and constraints to widespread adoption. Biomass Bioenergy 2013;49(1):28e37. [3] German Biogas Association (Fachverband Biogas e.V.). Branchenzahlen 2013 und Prognose der Branchenentwicklung 2014. Entwicklung der Anzahl Biogasanlagen und der gesamten installierten elektrischen Leistung in megawatt [MW]. [Internet]. Freising: German Biogas Association; [updated 2013 Nov; cited 2014 Apr 30]. Available from: http://www.biogas.org/edcom/webfvb.nsf/ id/DE_Branchenzahlen/$file/13-11-11_Biogas% 20Branchenzahlen_2013-2014.pdf. [4] DESTATIS (Statistisches Bundesamt). Production. Gross electricity production in Germany from 2011 to 2013. [Internet]. Wiesbaden: DESTATIS; [cited 2014 Apr 30]. Available from: https://www.destatis.de/EN/FactsFigures/
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
[5]
[6]
[7]
[8]
[9] [10]
[11]
[12] [13]
[14]
[15]
[16] [17] [18] [19] [20] [21] [22]
[23] [24]
EconomicSectors/Energy/Production/Tables/ GrossElectricityProduction.html. FNR (Fachagentur Nachwachsende Rohstoffe e.V.). Cultivation of renewable resources in Germany 2012. [Internet]. Gu¨lzow-Pru¨zen: FNR; [cited 2014 Apr 30]. Available from: http://mediathek.fnr.de/media/downloadable/files/ samples/r/z/rz_fnr4_0252_grafik_nawaro_anbau_101013_ eng.jpg. FNR (Fachagentur Nachwachsende Rohstoffe e.V.). Cultivation area of maize in Germany. Year of cultivation 2013. [Internet]. Gu¨lzow-Pru¨zen: FNR; [cited 2014 Apr 15]. Available from: http://mediathek.fnr.de/media/ downloadable/files/samples/m/a/maisgrafik_2013_300_ cmyk.jpg. Herrmann A. Biogas production from maize: current state, challenges and prospects. 2. Agronomic and environmental aspects. Bioenergy Res 2013;6(1):372e87. German Biogas Association (Fachverband Biogas e.V.). Wasserschutz beim Anbau von Energiepflanzen fu¨r Biogasanlagen. Hinweispapier mit Handlungsempfehlungen. Freising: German Biogas Association; 2013 Jul [cited 2014 Apr 30]. Available from: http://sojus.dbfz.de/sites/default/files/ 13-07-11_Papier%2520Gew%C3%A4sserschutz_end.pdf. Weiland P. Biogas production: current state and perspectives. Appl Microbiol Biotechnol 2010;85(4):849e60. Hartmann A. Sorghum als Biogassubstrat e Ertragspotential bei optimaler und reduzierter Wasserverfu¨gbarkeit, vol. 24. Nu¨rthingen: Mitteilungen der Gesellschaft fu¨r Pflanzenbauwissenschaften; 2012. p. 238e9. Vollrath B, Kuhn W, Werner A. “Wild” statt “Mono” e neue € chsheim: Wege fu¨r die Biogaserzeugung [Internet]. Veitho Bayerische Landesanstalt fu¨r Weinbau und Gartenbau; 2010 [cited 2012 Aug 28]. Available from:, http://www.lebensraumbrache.de/_downloads/poster_und_veroeffentlichungen/ WildstattMono_LandinForm.pdf. € ren. DLZ Vollrath B, Werner A. Wildpflanzen rentabel verga Agrar 2012;12(12):42e6. Gong Y, Janssen M. From policy implementation to business process management: principles for creating flexibility and agility. Gov Inf Q 2012;29(1):61e71. Lensink S, Londo M. Assessment of biofuels supporting policies using the BioTrans model. Biomass Bioenergy 2010;34(2):218e26. Helming K, Diehl K, Kuhlman J, Jansson T, Verburg PH, Bakker M, et al. Ex ante impact assessment of policies affecting land use, part B: application of the analytical framework. Ecol Soc [Internet] 2011;16(1):29 [cited 2013 May 5]. Available from: http://www.ecologyandsociety.org/vol16/ iss1/art29/. Veetil VP. Conceptions of rationality in law and economics. Eur J Law Econ 2011;31(2):199e228. Camerer CF. Behavioral economics. Princeton: Princeton University Press; 2001. Camerer CF, Fehr E. When does “Economic Man” dominate social behavior. Science 2006;5757(311):47e52. Fehr E, Gaechter S. Fairness and retaliation: the economic of reciprocity. J Econ Perspect 2000;14(3):159e81. Benz M. Entrepreneurship as a non-profit-seeking activity. Int Entrep Manage J 2009;5(1):23e44. Selten R. Bounded rationality. J Inst Theor Econ 1990;146(4):649e58. Hertwig R, Ortmann A. Experimental practices in economics: a methodological challenge for psychologists? Behav Brain Sci 2001;24(3):383e451. Levitt SD, List JA. Field experiments in economics: the past, the present, and the future. Eur Econ Rev 2009;53(1):1e18. Haenke S, Scheid B, Schaefer M, Tscharntke T, Thies C. Increasing syrphid fly diversity and density in sown
[25]
[26] [27]
[28] [29]
[30]
[31]
[32]
[33]
[34]
[35]
[36] [37] [38] [39]
[40]
[41]
[42]
[43]
[44]
[45]
205
flowering cover crops within simple vs. complex landscapes. J Appl Ecol 2009;46(5):1106e14. Haaland C, Gyllin M. Butterflies and bumblebees in greenways and sown wildflowering cover crops in southern Sweden. J Insect Conserv 2010;14(2):125e32. Tyler TR. Why people obey the law. New Jersey: Princeton University Press; 2006. Tyler TR, Blader SL. Can businesses effectively regulate employee conduct? the antecedents of rule following in work settings. Acad Manage J 2005;48(6):1143e58. Kahneman D, Tversky A. Prospect theory: an analysis of decision under risk. Econometrica 1979;47(2):263e92. Kahneman D, Knetsch JL, Thaler RH. Anomalies. The endowment effect, loss aversion, and status quo bias. J Econ Perspect 1991;5(1):193e206. Wratten SD, Gillespie M, Decourtye A, Mader E, Desneux N. Pollinator habitat enhancement: benefits to other ecosystem services. Agr Ecosyst Environ 2012;159:112e22. Thomas F, Denzel K, Hartmann E, Luick R, Schmoock K. Kurzfassung der Agrarumwelt- und Naturschutzprogramme, Darstellung und Analyse der Entwicklung der Maßnahmen der Agrarumwelt- und Naturschutzprogramme in der Bundesrepublik Deutschland. Bonn: Bundesamt fu¨r Naturschutz; 2009. Buttel FH, Gillespie GW, Larson III OW, Harris CK. The social bases of agrarian environmentalism: a comparative analysis of New York and Michigan farm operators. Rural Sociol 1981;46(3):391e410. Carr S, Tait J. Differences in the attitudes of farmers and conservationists and their implications. J Environ Manage 1991;32(3):281e94. Luetz M, Bastian O. Implementation of landscape planning and nature conservation in the agricultural landscape e a case study from Saxony. Agric Ecosyst Environ 2002;92(2e3):159e70. Goslin E, Williams KJH. Connectedness to nature, place attachment and conservation behavior: testing connectedness theory among farmers. J Environ Psychol 2010;30(3):298e304. Holt CA, Laury SK. Risk aversion and incentive effects. Am Econ Rev 2002;92(5):1644e55. Dixit AK, Pindyck RS. Investment under uncertainty. Princeton: Princeton University Press; 1994. Harrison GW, List JA. Field experiments. J Econ Lit 2004;42(4):1009e55. Charness G, Gneezy U, Kuhn MA. Experimental methods: extralaboratory experiments e extending the reach of experimental economics. J Econ Behav Organ 2013 Jul;91:93e100. FNR (Fachagentur Nachwachsende Rohstoffe e.V.). Prices of biogas substrate [V/t FM]. [Internet]. Gu¨lzow-Pru¨zen: FNR; [cited 2014 Apr 15]. Available from: http://mediathek.fnr.de/ grafiken/daten-und-fakten/preise-und-kosten/preisebiogassubstrate-interaktiv.html. € hn W. Impact of weather and yield Chmielewski FM, Ko components of winter rye over 30 years. Agric For Meteorol 2000;102(4):253e61. De Wit AJW, Boogaard HL, Van Diepen CA. Spatial resolution of precipitation and radiation: the effect on regional crop yield forecasts. Agric For Meteorol 2005;135(1e4):156e68. Brick K, Visser M, Burns J. Risk aversion: experimental evidence from South African fishing communities. Am J Agric Econ 2012;94(1):133e52. Maart-Noelck SC, Musshoff O. Investing today or tomorrow? An experimental approach to farmers' decision behavior. J Agric Econ 2013;64(2):295e318. DESTATIS (Statistisches Bundesamt). Verdienste und Arbeitskosten, Verdienste in der Landwirtschaft, 2010.
206
[46]
[47]
[48] [49]
[50]
b i o m a s s a n d b i o e n e r g y 7 0 ( 2 0 1 4 ) 1 9 6 e2 0 6
Wiesbaden: Statistisches Bundesamt; 2011. Fachserie 16, Reihe 1. Cooper DJ, Kagel JH, Lo W, Gu QL. Gaming against managers in incentive systems: experimental results with Chinese students and Chinese managers. Am Econ Rev 1999;89(4):781e804. Friedman D. Monty Hall's three doors: construction and deconstruction of a choice anomaly. Am Econ Rev 1998;88(4):933e46. Harrison GW. Expected utility theory and the experimentalists. Empir Econ 1994;19(2):223e53. Camerer CF, Hogarth RM. The effects of financial incentives in experiments: a review and capital-labor-production framework. J Risk Uncertain 1999;19(1e3):7e42. Duersch P, Oechssler J, Schipper BC. Incentives for subjects in internet experiments. Econ Lett 2009;105(1):120e2.
[51] Ashton RH. Pressure and performance in accounting decision settings: paradoxical effects of incentives, feedback and justification. J Acc Res 1990;(28 Suppl.):S148e80. [52] Bothos E, Apostolou D, Mentzas G. Collective intelligence with web-based information aggregation markets: the role of market facilitation in idea management. Expert Syst Appl 2012;39(1):1333e45. [53] Cutler SJ, Hendricks J, Guyer A. Age differences in home computer availability and use. J Gerontol B Psychol Sci Soc Sci 2003;58(5):271e80. [54] DESTATIS (Statistisches Bundesamt). Landwirtschaft auf einen Blick. Wiesbaden: Statistisches Bundesamt; 2011. [55] Schram A. Artificiality: the tension between internal and external validity in economic experiments. J Econ Methodol 2005;12(2):225e37.