Advertisement

Short communication: Effect of carbon emission trading on European Union butter prices

Open ArchivePublished:January 10, 2019DOI:https://doi.org/10.3168/jds.2018-15788

      ABSTRACT

      Since its inception, the European Union (EU) carbon emission market has been vastly successful in reducing greenhouse gas emissions. Accordingly, the usage of environmentally friendly fuels (e.g., ethanol, biodiesel) has increased significantly over the last few years. Given that EU biodiesel is mainly produced from rapeseed oil and soybean oil, higher carbon taxes are likely to increase the demand of these important vegetable oils, which further affects the prices of its close substitute such as butter. Nevertheless, the association between the EU emission trading scheme and butter prices remains understudied. In this paper, we aim to fill this vacuum in the existing literature. Applying the autoregressive distributed lag bound testing procedure, we show that emission market seems to have a long-term effect on EU butter prices, implying that changes in the levels of carbon taxes will lead to changes in the price level of butter. These results are of vital importance to policymakers and investors.

      Key words

      Short Communication

      Since its inception, the European Union (EU) emission trading scheme (ETS) has been remarkably efficient in the reduction of greenhouse gas (GHG) emissions. In 2015, for instance, GHG emissions in the EU-28 countries were down by 22% compared with 1990 levels, representing an absolute reduction of 1,265 million tonnes of CO2-equivalents, putting the EU on track to surpass its 2020 target, which is to reduce GHG emissions by 20% by 2020 and by 40% by 2030 compared with 1990 (data from https://ec.europa.eu/eurostat/statistics-explained/pdfscache/1180.pdf).
      The success of the EU allowance (EUA) market has promoted the use of environmentally friendly fuels all over the world. For instance, biodiesel has emerged as the leading biofuel used in the EU transport sector. A recent report published by the European Commission indicates that biodiesel production capacity has increased to about 26.3 billion liters, with an annual production of about 10.5 billion liters (8.3 million tonnes of oil equivalents) or 40% of total capacity. Such significant growths in biodiesel could be due to the concerns about energy security, GHG emissions, and oil market uncertainty (
      • Chiu F.P.
      • Hsu C.S.
      • Ho A.
      • Chen C.C.
      Modeling the price relationships between crude oil, energy crops and biofuels.
      ;
      • Dutta A.
      Cointegration and nonlinear causality amongst ethanol-related prices: Evidence from Brazil.
      ).
      In the EU, biodiesel is mainly produced from rapeseed oil and soybean oil. At present, rapeseed oil accounts for about 80% of the biodiesel feedstock. Given that vegetable oil is the main feedstock for producing biodiesel in the EU, it is likely that higher carbon permit prices would raise the demand for vegetable oil, which further affects the prices of a close substitute such as butter.
      • Crago C.L.
      • Khanna M.
      Carbon abatement in the fuel market with biofuels: Implications for second best policies.
      also argue that as biofuel production competes for land with agricultural production, higher carbon tax could increase land rents, which would then raise feedstock prices. Nevertheless, investigation into the linkage between carbon emission and butter prices is understudied in the existing literature. This study aims to fill this void. As understanding the dynamics of butter prices plays a pivotal role in designing the EU dairy policy, our research could be of vital importance to policymakers and investors.
      This paper uses monthly prices of the EU-ETS and butter covering the period from July 2009 to December 2017. The sample consists of 102 observations. The data were retrieved from Thomson Reuters DataStream (https://eikon.thomsonreuters.com/index.html). For the EU emission market, the data for 3 different periods are available. Phase I covers the period from January 2005 to December 2007, and phase II covers January 2008 to December 2012. At present, the EU-ETS is completing the third phase. The data for phase I are not included because the allowance prices were close to zero toward the end of phase I as a consequence of interphase EUA banking restrictions (
      • Tian Y.
      • Akimov A.
      • Roca E.
      • Wang V.
      Does the carbon market help or hurt the stock price of electricity companies? Further evidence from the European context.
      ).
      Table 1 reports the descriptive statistics for EUA and butter price indexes. These numbers suggest that food prices are more volatile than the emission market as evidenced by the corresponding standard deviations. Moreover, both indexes are negatively skewed and kurtosis was >3 for each series, implying that the data follow leptokurtic distributions. The Jarque-Bera test suggests that the data are normally distributed only for the butter price index.
      Table 1Main descriptive statistics (values in parentheses denote P-values) of the 2 indices under study
      ΔEUA and ΔButter = first-order differences in European Union allowance and butter prices, respectively. .
      IndexMeanSDSkewnessKurtosisJarque-Bera test
      ΔEUA−0.05301.0631−0.41344.239.29 (0.00)
      Significant at the 1% level.
      ΔButter1.44358.0547−0.03613.481.01 (0.60)
      1 ΔEUA and ΔButter = first-order differences in European Union allowance and butter prices, respectively. .
      ** Significant at the 1% level.
      Table 2 shows the unit root test results for the indexes used. The results of both augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) tests indicate that although the data follow a nonstationary process at levels (i.e., raw data), the stationary condition is satisfied when first differences are considered. The ADF and PP tests are used to examine whether the data contain any trend. Nonstationary processes usually have trends, whereas stationary processes do not contain such trends.
      Table 2Results of unit root tests
      EUA = European Union allowance. The values in parentheses denote P-values.
      IndexAugmented Dickey-Fuller testPhillips-Perron test
      LevelFirst differenceLevelFirst difference
      EUA−1.65 (0.45)−12.29 (0.00)***−1.55 (0.45)−12.28 (0.00)**
      Butter−1.44 (0.56)−12.63 (0.00)***−1.31 (0.62)−12.45 (0.00)**
      Significant at the **1% or ***0.1% level.
      1 EUA = European Union allowance. The values in parentheses denote P-values.
      Methodologically, we used the autoregressive distributed lag (ARDL) model to examine the long-term association between carbon emission trading and EU butter prices. The ARDL bound test has several attractive features. First, all the testing equations are allowed to have different lags. Second, it can be applied regardless of whether the underlying variables are stationary, that is, I(0); integrated of order 1, that is I(1); or fractionally integrated (
      • Bouri E.
      • Jain A.
      • Biswal P.C.
      • Roubaud D.
      Cointegration and nonlinear causality amongst gold, oil, and the Indian stock market: Evidence from implied volatility indices.
      ). Finally, the test is free from the spurious regression problem. Note that this test has a prerequisite that the variables under study should not be integrated of order 2 or higher.
      Now, we construct the following unrestricted ARDL regressions without any time trend component:
      ΔEUAt=ω1+i=1nα1,iΔEUAti+i=1nβ1,iΔButterti+a1EUAt1+b1Buttert1+ɛ1t,
      [1]


      ΔButtert=ω2+i=1nα2,iΔEUAti+i=1nβ2,iΔButterti+a2EUAt1+b2Buttert1+ɛ2t,
      [2]


      where Δ indicates the first difference operator: ΔEUA indicates the first-order difference for EUA prices, ΔButter is the first-order difference for EU butter prices, α and β refer to the lag coefficients for ΔEUA and ΔButter, and ε1 and ε2 are error terms. To examine whether cointegrating relationship exists among the series under study, it is sufficient to test the null hypothesis H0: a = b = 0, where a and b refer to the first-order lag coefficients for the EUA and butter prices. The general F-statistics are further calculated and compared with 2 different sets of critical values provided by
      • Pesaran M.H.
      • Shin Y.
      • Smith R.J.
      Bounds testing approaches to the analysis of level relationships.
      . One of these sets is used as the upper bound for purely I(1) series, and the other is used as the lower bound for purely I(0) series. Cointegration or a long-run relationship is said to be present only if the computed F-statistic surpasses the upper bound critical value.
      Table 3 shows the results of the ARDL bound tests. These findings reveal that cointegration is present when butter prices act as the dependent variable. We do not find any significant linkage when emission market is the dependent variable. Such results indicate that causality runs from emission to butter price index, implying that changes in the price level of the EUA market cause significant changes in the butter price level.
      Table 3Results of autoregressive distributed lag bound tests
      The critical F-statistic at the 5% level for model with all I (1) series is 5.73. See Table CI(iii) with k = 1 (where k = number of explanatory variables) on page 300 of Pesaran et al. (2001). ΔEUA and ΔButter = first-order differences in European Union allowance and butter prices, respectively.
      Dependent variableTest statisticsDecision
      ΔEUA1.70No cointegration
      ΔButter5.85
      Significant at the 5% level.
      Cointegration exists
      1 The critical F-statistic at the 5% level for model with all I (1) series is 5.73. See Table CI(iii) with k = 1 (where k = number of explanatory variables) on page 300 of
      • Pesaran M.H.
      • Shin Y.
      • Smith R.J.
      Bounds testing approaches to the analysis of level relationships.
      . ΔEUA and ΔButter = first-order differences in European Union allowance and butter prices, respectively.
      * Significant at the 5% level.
      In addition to the ARDL bound tests, we also employ the Toda-Yamamoto (TY) version of Granger causality test to investigate whether any short-term linkage exists between these 2 sectors. See
      • Toda H.Y.
      • Yamamoto T.
      Statistical inference in vector autoregressions with possibly integrated processes.
      for more details about the TY test. The results of the TY test, displayed in Table 4, reveal that the EUA market and the butter industry do not move together in the short run.
      Table 4Results of the Toda-Yamamoto test
      The lag used in the Toda-Yamamoto test is 1. The Akaike information criterion and Bayesian information criterion are used to determine the required number of lags. EUA = European Union allowance.
      CausalityTest statisticsP-value
      EUA to butter1.320.25
      Butter to EUA0.880.35
      1 The lag used in the Toda-Yamamoto test is 1. The Akaike information criterion and Bayesian information criterion are used to determine the required number of lags. EUA = European Union allowance.
      Overall, these results suggest that the EUA market seems to have a long-term effect on the butter price index. Such findings are not surprising given that higher carbon taxes promote the production of biodiesel and hence the demand for vegetable oil, which is the main feedstock for the EU biodiesel industry, would increase considerably. Because butter and vegetable oil appear to be substitute commodities, the growing demand for vegetable oil leads to a significant rise in the butter price index. Thus, the carbon emission trading has an indirect, but prominent effect on EU butter prices.
      The outcomes of this empirical research could be important for policymakers in formulating appropriate hedging strategies to avoid the contagious risk emanating from the EU emission trading scheme. In addition, the information on the EUA prices could be useful in predicting the price variations of the EU butter industry as we find a significant long-run linkage between these 2 markets. Moreover, in the near future, the production of biodiesel could undergo incredible growth, which could, in turn, cause uncertainties in the butter industry. It is important therefore to develop and improve the futures market for butter to precisely measure the risk associated with this important food commodity. Such monitoring system could play a crucial role in planning EU food policy.

      REFERENCES

        • Bouri E.
        • Jain A.
        • Biswal P.C.
        • Roubaud D.
        Cointegration and nonlinear causality amongst gold, oil, and the Indian stock market: Evidence from implied volatility indices.
        Resour. Policy. 2017; 52: 201-206
        • Chiu F.P.
        • Hsu C.S.
        • Ho A.
        • Chen C.C.
        Modeling the price relationships between crude oil, energy crops and biofuels.
        Energy. 2016; 109: 845-857
        • Crago C.L.
        • Khanna M.
        Carbon abatement in the fuel market with biofuels: Implications for second best policies.
        J. Environ. Econ. Manage. 2014; 67: 89-103
        • Dutta A.
        Cointegration and nonlinear causality amongst ethanol-related prices: Evidence from Brazil.
        Glob. Change Biol. Bioenergy. 2018; 10: 335-342
        • Pesaran M.H.
        • Shin Y.
        • Smith R.J.
        Bounds testing approaches to the analysis of level relationships.
        J. Appl. Econ. 2001; 16: 289-326
        • Tian Y.
        • Akimov A.
        • Roca E.
        • Wang V.
        Does the carbon market help or hurt the stock price of electricity companies? Further evidence from the European context.
        J. Clean. Prod. 2016; 112: 1-8
        • Toda H.Y.
        • Yamamoto T.
        Statistical inference in vector autoregressions with possibly integrated processes.
        J. Econom. 1995; 66: 225-250