Cost management strategy for energy use in agriculture : EU sample

Efficiency analysis is useful for all kinds of production. Outputs and inputs can be optimized by this analysis. In this way costs are also minimized. Many inputs are used in agriculture production and only purpose is to produce more. In agricultural production there are desirable and undesirable outputs. The amount of emissions comes from fossil fuels is an undesirable output. We must maximize the desirable outputs and minimize the undesirable outputs. Also, we need to minimize the inputs to minimize the costs. In this study, Data Envelopment Analysis and partial efficiency analysis was applied to determine the cost management. Two outputs and four inputs were detemined for the measurement of the cost efficiency. According to analysis, it is determined that, the EU countries must protect the capital stock optimization and decrease the energy usage about 29.81%, land usage about 64.80% and employment in agriculture about 35.62%.


Introduction
Agricultural production has always been there since the beginning of humanity.Today, industry and services sectors may seem like surpassed the agricultural sector, but it's a fact that the importance of sustainable agriculture is increasing.Many inputs are used in agriculture production and only purpose is to produce more.There are many components of sustainable agriculture.The nvironment and energy use are just some of them.However, the economy should also be included as a separate component.Is it?Like any production, agricultural production also has a cost.We need agricultural production, because we need food to live.In order to produce more, how much we need to spend more?
The greatest danger to the environment arises from the use of fossil fuels.Fossil fuels are used in many areas of agricultural production.Also a large portion of the pollution caused by agriculture, stems from fossil fuels.Here we encounter three questions.How much we spend?How much we produce?How much we pollute?After these questions, the fourth and more important question may be asked.Can we produce the same, by polluting less and spending less?Efficiency analysis can answer this question.
This study is about partial energy efficiency in agriculture.So we can determine the sufficient amount of usage.On the other hand, we will compare the production value of agricultural, pollution comes from the consumption of the fossil fuels in agriculture, employment in agriculture, capital stock in agriculture and energy use in agriculture as cost efficiency.Thus, we think we will get the answers to our questions.

Data and Methods
In this study, 28 countries of the EU were determined as decision making units (DMU).Cost efficiency scores were determined by output oriented Data Envelopment Analysis (DEA) model under the assumption of constant returns to scale.All data were obtained from the Food and Agriculture Organization (FAO) for the year 2012.Missing data were estimated based on historical data.

Data Set
In the production process, there are desirable and undesirable outputs.This also applies to agricultural production.In this study, we determined two outputs and four inputs: Outputs:  Gross production value for agricultural production (million $)  Amount of emissions for agriculture caused by fossil fuells (CO2 equivalent) (gigagrams): Includes, gas-diesel oil, motor gasoline, natural gas, liquefied petroleum gas, fuel oil and coal) Inputs:

Methods
Partial efficiency was used in order to estimate the energy efficiency of EU countries.Partial efficiency is a simple ratio of a single output and a single input ( Ozden, 2016).The purpose of this model is to maximize the output without changing the amount of inputs.But we have desirable and undesirable outputs.When we include both of them, model tries to maximize both of them.So if we want to analyze both desirable and undesirable outputs we need to maximize the desirable and minimize the undesirable output.There are several models for this situation.But these models are not suitable for basic softwares.Therefore a new simple index was designed to analyze the desirable and undesirable outputs together by Ozden (2016).When we change the place of the numerator and denominator the problem will be solved.So we include the undesirable output for analysis as undesirable output -1 .In this case when the model tries to maximize the undesirable output, undesirable output will be minimized.In this study the undesirable output is the amount of emissions for agriculture caused by fossil fuells (CO2) (Ozden, 2016).And we include the undesirable output for analysis as 1/ CO2.

Results and Discussion
In this study, we determined two outputs and four inputs to estimate the efficiency scores.Descriptive statistics for all variables can be seen in Table 1.Primarily, we must calculate the partial energy efficiencies for desirable and undesirable outputs.Gross production value OF agriculture is a desirable output and the amount of emissions from agriculture (CO2 eq) is an undesirable output.Partial energy efficiency for desirable output is a simple ratio of energy use in agriculture and gross production value of agriculture.Partial energy efficiency for undesirable output is also a simple ratio of energy use in agriculture and the amount of emissions from agriculture .Thus, we can determine the energy efficient countries in EU.As seen in Table 2, the most efficient country for desirable output is Germany and UK following it.The most inefficient countries are, respectively, Cyprus, Latvia and Malta.Energy efficient countries for undesirable output are Luxembourg, and Malta respectively.And the inefficient ones are Netherlands, Poland and France (Table 3).These scores are not relative efficiency scores, they are countries own efficiencies.The next step is, to calculate the percentage of energy use in agriculture of total energy use.
Here, a different partial energy index has been introduced for the countries.In terms of energy use in agriculture, it will not wrong to say that the countries with a low percentage, are more efficient.As seen in Table 4, Germany and Malta are the most efficient countries and Poland and Netherlands are the most inefficient countries.On the basis of three calculated partial efficiency scores, the most efficient countries in terms of energy use are, Germany and Malta and the most inefficient ones are Poland and the Netherlands.After the partial energy efficiency estimation the next step is the estimation of cost efficiency with desirable and undesirable outputs.
When we consider the results of cost efficiency estimation, the reference countries are Germany, Malta and Belgium.It means, these countries are determined as full efficient.And Poland is again determined as the most inefficient country.The mean of cost efficiency scores are "0.653"(Table 5).Finally, the last step is the determination of improvements for inputs and outputs.Improvements are given for each input, output and country.Improvements for outputs shows us the percentage that how much we can increase our production, without changing the amount of inputs.As previously mentioned, we take the output, amount of emissions from agriculture (CO2 eq), as 1/ CO2.So it means that, how much we can decrease the amount of emissions from agriculture without changing the amount of inputs.As a result, in EU, we can increase the gross production value for agriculture 73.42 % and we can decrease the amount of emissions from agriculture 84.30%.The striking result is about Bulgaria.Bulgaria can decrease its emission release 229.40%without changing the amount of inputs (Table 6, Table 7).
Ozden and Dios-Palomares, 2015).Output oriented DEA model was used in order to estimate cost efficiency scores.After the introduction of the DEA model by Farrel (1957), Charnes, Cooper and Rhodes (1978) were designed CCR for efficiency estimation.CCR model is based on assumptions of Constant Returns to Scale (CRS).Detailed description about this model can be seen in many literatures (Ozden and Armagan, 2005; Martic et al., 2015; Ozden, 2010; Cooper et al., 2011; Ozden and Dios-Palomares, 2015;