The Effect of Commodity Prices and Exchange Rate on the Stock Return of Agriculture and Animal Feed Companies in Indonesia

Oktrevina Oktrevina; Yohannes Kurniawan; Norizan Anwar

doi:10.36923/ijsser.v4i4.149

Open Access Original Research

Oktrevina Oktrevina ⁽¹⁾ , Yohannes Kurniawan ⁽²⁾ , Norizan Anwar ⁽³⁾

(1) School of Accounting, Bina Nusantara University, Jakarta, Indonesia , Indonesia

(2) School of Information Systems, Bina Nusantara University, Jakarta, Indonesia , Indonesia

(3) Faculty of Information Management, Universiti Teknologi Mara, UiTM, Puncak Perdana Campus, 40150, Shah Alam, Selangor, Malaysia , Indonesia

https://doi.org/10.36923/ijsser.v4i4.149

Issue
Vol. 4 No. 4 (2022)

Submitted
February 3, 2022

Published
December 28, 2022

Keywords:

GARCH, Stock Return, Commodity Price, Agriculture, Animal Feed, COVID-19

PDF XML

Abstract

Purpose of the study: This study aims to examine the key drivers for the effect of selected commodity ratios and the exchange rate that haveinfluenced the return on investment of agriculture and animal feed companies in Indonesia.

Methodology: This study uses the GARCH methodology. Using GARCH and IGARCH for monthly and daily data from 2014 until 2021, we also have a different timeline between before COVID and after COVID-19.

Main Findings: According to the results of the study, there are generally significant effects of commodity prices and exchange rates on stock return, as for the COVID timeline, there are several companies that have been affected.

Research limitation: There is a problem with finding the parameter estimates for the error distribution both in GED and Student’s t when finding the best GARCH (1,1) model. The best model finding may not present the best probability due to distributions that affect the tail distribution.

Novelty/Originality of the study: This study suggests to look for more details on the effect of commodity ratio in daily data and monthly data, as thetransaction is mostly using USD rate with IDR rate. There are great possibilities that their movement will impact the return on the investment and people who like to invest in a certain company can use this study as a reference.

Full Article

Generated from XML file

References

A. Bartolome del Canto. (2021). investpy - Financial Data Extraction from Investing.com with Python. GitHub Repository.

Adeleye, B. N. (2018). Time Series Analysis (Lecture 2): Choosing Optimal Lags in EViews. Retrieved from Econometrics Resource for Beginners and Data Analysis: http://cruncheconometrix.blogspot.com/2018/02/time-series-analysis-lecture-2-choosing.html#

Albulescu, C. (2020). COVID-19 and the United States financial markets’ volatility. Finance Research Letters, 38:101699.

Almeida, R. J., Basturk, N., Kaymak, U., & Sousa, J. M. (2014). Estimation of flexible fuzzy garch models for conditional density estimation. Information Sciences, 267 (p):252-264.

Antoniou, A., Koutmos, G. & Pericli, A. (2005). Index futures and positive feedback trading. Evidence from major stock exchanges. Journal of Empirical Finance, 12(2), 219-238.

Baffes, J. (2007). Oil spills on other commodities. Resource Policy, 12 (3), 126-134.

Banerjee, A., & Sarkar, S. (2006). Modeling Daily Volatility of the Indian Stock Market using Intra-Day Data. Retrieved from Working paper No: 588, IIM Calcutta: http://www.iimcal.ac.in/res/upd%5CWPS%20588

Black, F. (1976). Studies of Stock Price Volatility Changes. In: Proceedings of the 1976 Meeting of the Business and Economic Statistics Section. American Statistical Association, Washington DC, 177-181.

Briere, M. & Szafarz, A. (2017). Factor investing. The rocky road from long-only to long-short. In E. Jurczenko (Ed.), Factor Investing. Elsevier, forthcoming.

Chen, J., & Mansa, J. (2021). Investopedia. Retrieved from https://www.investopedia.com/terms/i/investor.asp

Cheteni, P. (2016). Stock market volatility using GARCH models: Evidence from South Africa and China stock markets. MPRA Paper No. 77355.

Chiou, J.-S., & Lee, Y.-H. (2009). Jump dynamics and volatility: Oil and the stock markets. Energy, 34 (6) 788-796.

Chkili, W., Aloui, C. & Nguyen, D. K. (2012). Asymmetric effects and long memory in dynamic volatility relationships between stock returns and exchange rates. . Journal of International Financial Markets, Institutions, and Money, 22, 738–757.

Choi, K. & Hammoudeh, S. (2010). olatility behavior of oil, industrial commodity and stock markets in a regime-switching environment. Energy Policy, 39 (6) 201-228.

Chu, C., S., J., Santoni, G. J. & Liu, T. (1996). Stock market volatility and regime shifts in returns. Information Sciences, 94(1-4):179–190.

Cont, R. (2001). Empirical properties of asset returns: stylized facts and stylized issues. Quantitative Finance, 1(2) :223–236.

Danielsson, J. (2011). Financial Risk Forecasting: The Theory and Practice of Forecasting Market Risk with Implementation in R and Matlab 1st Edition. John Wiley & Sons Ltd.

Ding, ZX, Granger, C., & Engle, R. (1993). A long-memory property of stock market returns and a new model. Journal of Empirical Finance, 1(1):83–106.

Dritsaki, C. (2017). An Empirical Evaluation in GARCH Volatility Modeling: Evidence from the Stockholm Stock Exchange. Journal of Mathematical Finance, 7(2), 366-390.

Eagle, R. & Bollerslev, T. (1986). Modelling the persistence of conditional variances. Retrieved from Econometric Reviews [Inter- net]: https://doi.org/10.1080/07474938608800095

Eviews , V. (2020, Nov 11). ARCH and GARCH Estimation. Retrieved from Estimating ARCH Models In EViews: http://www.eviews.com/help/helpintro.html#page/content/arch-Estimating_ARCH_Models_in_EViews.html

Fan, Y., Zhang, Y.J., Tsai, H.T. & Wei, Y.M. (2008). Estimating ‘value at risk’ of crude oil price and its spillover effect using the ged-garch approach. Energy Economics, 30(6), 3156–3171.

Filis, G., Degiannakis, S., & Floros, C. (2013). Dynamic correlation between stock market and oil prices: The case of oil-importing and oil-exporting countries. International Review of Financial Analysis.

GMPT. (2021). Gabungan Perusahaan Makanan Ternak. Retrieved from http://gpmt.or.id/our-blog

Gorton, G., & Rouwenhorst, G.K. (2006). Facts and fantasies about commodity futures. Financial Analysts Journal, 47-48.

Goyal, A., & Santa-Clara, P. (2003). Idiosyncratic risk matters! Journal of Finance, 58(3), 975-1007.

Greunen, JV, Heymans, A., Heerden, Cv., & Vuuren, Gv.[X] has many roles in the field of ... (2014). THE PROMINENCE OF STATIONARITY IN TIME. Journal for Studies in Economics and Econometrics, 38 (1), 1-16.

Gui, H., & Savickas, R. (2008). Average Idiosyncratic Volatility in G7 Countries. Reviews of Financial Studies, 21, 1259-1296.

Gujarati, D. N., & Porter, D. C. (2009). Basic Econometrics. New York: McGraw-Hill-Irwin.

Hartman, J., Wiklander, O., & Forsberg, L. (2012). Evaluating forecasts from the GARCH(1,1)-model for Swedish Equities. Bachelor Thesis Department of Statistics Uppsala University.

Ildirar, M., & ??can, E. (2015). The Interaction between Stock Prices and Commodity Prices: East Europe and Central Asia Countries. Conference: International Conference on Eurasian Economies, 1(2);13-14.

Jiang, G., & Tian, Y. (2010). Forecasting volatility using long memory and comovements: An application to option valuation under SFAF 123R. Journal of Financial and Quantitative Analysis, 45(2), 502-533.

Joesoef, J. (2008). Pasar uang & pasar valuta asing. Jakarta: Salemba Empat.

Kang, W., & Vespignani, J. L. (2017). Global Commodity Prices and Global Stock Volatility Shocks: Effects Across Countries. SSRN Electronic Journal, 35 (1).

Karmakar, M. (2006). Stock Market Volatility in the Long Run, 1961-2005. Economic and Political Weekly, 41(18), 1796-1802.

Kemenperin, R. (2019). ANALISA STRUKTUR INDUSTRI PAKAN TERNAK DALAM RANGKA PENGEMBANGAN PERWILAYAHAN INDUSTRI. Studi Kasus pada WPPI Jawa Timur, Provinsi Jawa Timur.

Kenton, W. (2020, October 25). What Is the GARCH Process? How It's Used in Different Forms. Retrieved from GARCH Process: https://www.investopedia.com/terms/g/generalalizedautogregressiveconditionalheteroskedasticity.asp

Lingbing , F., & Yanlin, S. (2017). A simulation study on the distributions of disturbances in the GARCH model. Cogent Economics & Finance, 2(1);5.

Ludvigson, S., & Ng, S. (2007). The empirical risk-return relation: A factor analysis approach. Journal of Financial Economics, 83(1), 171-222.

Mabrouk, S., & Saadi, S. (2012). Parametric value-at-risk analysis: Evidence from stock indices. The Quarterly Review of Economics and Finance, , 52, 305–321. Mittnik, S., Doganoglu, T., & Chenyao, D.

Naimy, V., Haddad, O., Ferna ?ndez-Avile ?s, G., & Khoury, E. R. (2021). The predictive capacity of GARCH-type models in measuring the volatility of crypto and world currencies. Applied Economics and Quantitative Methods, 2(7);17.

Nedeljkovic, M., & Maksimovi?, A. (2019). Analysis and Forecast of Foreign Trade Indicators of Corn in Bosnia and Herzegovina. 19(4), 265-274.

Nigatu, G., Badeau, F., Seeley, R., & Hansen, J. (2020). Factors Contributing to Changes in Agricultural Commodity Prices and Trade for the United States and the World. Washington: USDA.

Omari, O. C., Maina, M. S., & Ngina, I. (2020). Forecasting Value-at-Risk of Financial Markets under the Global Pandemic of COVID-19 Using Conditional Extreme Value Theory. Journal of Mathematical Finance, 10(4):569-597.

Putra, A. R., & Robiyanto, R. (2019). The effect of commodity price changes and USD/IDR exchange rate on Indonesian mining companies’ stock return. Jurnal Keuangan dan Perbankan, 28(2),110.

Sabariah, N., Norhafiza, N., & Rusmawati, I. (2014). The Impact of Palm Oil Price on the Malaysian Stock Market Performance . Journal of Economics and Behavioral Studies, 8.

Scotti, C. (2007). Markov switching GARCH models of currency turmoil in southeast Asia. Board of Governors of the Federal Reserve System (US).

Sinbanda, K., & Mlambo, C. (2014). The Impact of Oil Prices on the Exchange Rate in South Africa. Journal of Economics.

The CDMI Consulting Group. (2018-2022). Retrieved from https://www.cdmione.com/buku/animal-feed-industry-in-indonesia/

Tsay, R. S. (2005). Analysis of Financial Time Series. Wiley, New Jersey.

Wan, Y., & Si, Y. W. (2017). A formal approach to chart patterns classification in financial time series. Information Sciences, 411:151–175.

Wenger, F. (2016, May 5). Wenger Feed. Retrieved from Wheat for Poultry and Swine Feeds.

Wooldridge, J. M. (2009). Introductory Econometrics: A Modern Approach. Boston: South Western, Cengage Learning.

Yaffee, R. A., & McGee, M. (2000). An Introduction to Time Series Analysis and Forecasting: With Applications of SAS® and SPSS®. New York: Academic Press, INC.

Zhang, C. (2010). A Reexamination of the Causes of Time-Varying Stock Return Volatilities. Journal of Financial and Quantitative Analysis, 45(3), 663-684.

Zhang, D., Min, H., & Ji, Q. (2020). Financial markets under the global pandemic of COVID-19. Finance Research Letters, Vol 36 (C).

Zhang, Z., Bali, T. G., Cakici, N., & Yan, X. (2005). Does Idiosyncratic Risk Really Matter? Journal of Finance, 60(2), 905-929.

Authors

Oktrevina Oktrevina

School of Accounting, Bina Nusantara University, Jakarta, Indonesia

https://orcid.org/0000-0002-9061-6487

oktrevina@binus.ac.id (Primary Contact)

Yohannes Kurniawan

School of Information Systems, Bina Nusantara University, Jakarta, Indonesia

Norizan Anwar

Faculty of Information Management, Universiti Teknologi Mara, UiTM, Puncak Perdana Campus, 40150, Shah Alam, Selangor, Malaysia

https://orcid.org/0000-0003-1104-1724

Oktrevina, O., Kurniawan, Y., & Anwar, N. (2022). The Effect of Commodity Prices and Exchange Rate on the Stock Return of Agriculture and Animal Feed Companies in Indonesia. Innovation Journal of Social Sciences and Economic Review, 4(4), 1–09. https://doi.org/10.36923/ijsser.v4i4.149

Download Citation

This work is licensed under a Creative Commons Attribution 4.0 International License.

IJSSER is committed to an open-access model for publishing, ensuring that all of its content is freely available under the terms of the Creative Commons Attribution License (CC BY) 4.0. This policy empowers researchers, academicians, and the general public to freely download, copy, print, and access the links to the full texts of published materials at no cost. Furthermore, authors are encouraged to read, distribute, and submit the published works to their institution's library and any other appropriate repositories.

Smart Citations via scite_

Downloads

Download data is not yet available.

1. INTRODUCTION

In this digital area, investment in the stock market is prevalent. An investor is an individual who commits capital that expects a financial return by putting some money into any organization (Chen & Mansa, 2021). All kinds of investors,whether investor, passive, or active, expect a stock return after investing. A study finds that integrating industries and factors leads to a smoother and more extraordinary achievement (Briere & Szafarz, 2017). The primary benefit of investing is the reduction of portfolio risk, and the main advantage of factor investing is the greater expected return. The stock returnis the profit or loss acquired from the investment or stock trading in a period.

Indonesia is a country that is blooming with immeasurable natural prosperity, from natural beauty for tourism to natural wealth that can produce from its energy source, one of which comes from the agricultural sector. Indonesia has always been rich in agricultural products, for example, palm oil, soybeans, wheat flour, corn, and others. With Indonesian agriculture getting huge, this has a positive impact as well, not only for domestic affairs but also for foreign affairs. However, see below, the stock index in Indonesia's agriculture sector shows the stock index's fluctuation from 2010 to2020 (see Figure 1). This condition occurs because Indonesia has a tropical and agricultural background, and the agriculture stock index is not stable.

Figure 1.Graphic of Stock Index of Agriculture in Indonesia (*Source: IDX).

Besides the agriculture sector in Indonesia, another sector that might consider having more enormous dividends is the animal feed sector. Although many other industries experienced severe pressure and some fell to their lowest point, the Indonesian animal feed industry in the last five years (2013-2017),

Its production grew by an average of 8.35%, and its business value grew 11.27%. In 2017 domestic animal feed production was 18.93 million tons with a business value of approximately $9 million. Charoen Pokphand Indonesia dominated this business with a market share of more than 32%, followed by Japfa Comfeed Indonesia with up to 15% (The CDMI Consulting Group, 2018-2022).

Commodity prices can dominate the production decisions of ranchers or farmers on the plant and harvest area of the plant or the livestock record and, hence, affect the supply of agricultural commodities (Nigatu, Badeau, Seeley, & Hansen, 2020).. Another part that generally raises a crucial reason for changes in agricultural commodity prices is trading agricultural goods in the stock request. Animal feed is a food item that domestic animals consume. Around 35% of animal feed raw materials are still considered imported products in Indonesia. (GMPT, 2021). The animal feed industry is one of the most critical industries in Indonesia. It is even one of the industries that grew in Indonesia's early stages of development. It is said to be important because this industry has a robust forward link with the livestock sector, and livestock also has a related backward link with the need for feed input,mainly corn (Kemenperin, 2019).

The global demand for palm oil is increasing rapidly. A growing trend for palm oil prices relates to petrol prices in global commodity markets (Baffes, 2007). Corn is also an important agricultural production that drives the stock of agricultural companies. Corn plays a critical role in the commodity market, and corn can be transformed into food, animal feed, biofuels, and other raw material ingredients for industrial processes (Nedeljkovic & Maksimovi, 2019). Wheat flour is used primarily for human consumption. The use of wheat flour in animal feed ingredients is typically restricted when wheat is priced aggressively with other grains such as corn. Several of the huge compositions of ingredients needed to produce animal feeds are soybean meal and wheat flour, both of which are the most essential components to produce animal feeds. People have used soybeans as an excellent protein food source and fiber thousands of times. There is also a demand to consider the impact of stocks. Changing the exchange rate in companies differs depending on the tendency to export or import. The exchange rate is the quantum of a particular currency that can change for a specific currency from another country (Joesoef, 2008). When the currency keeps discounting, investors are less consider investing. This condition will surely lead to a drop in the stock price and a lower stock return (Scotti, 2007).

The objective of this study is to know where to go for more information to forecast the best way for stock returns. The best way to gain more profit from investing is not by investing in companies that are considered good companies (profit is already aligned from time to time) but by investing in companies with the potential to have better progress. The background above shows that the agriculture sector's historical stock index tends to be unstable in Indonesia, and the progress of animal feed companies is growing around 35% from 2013 to 2017. The question is how to forecast whether the stock will be better or worse in the following year. First, an investor must understand the critical driver that drives the company's business. This study focuses on two keys that can be considered the drivers of the return of company stock, commodity prices and the USD/IDR exchange rate. By researching several commodities that might impact the company's business process and the USD/IDR exchange rate driving the company's condition, those two key drivers might increase or decrease the company's progress, automatically impacting the stock returns.

This research uses generalized autoregressive conditional heteroscedasticity (GARCH) in the EViews 12 application to know the influence of the independent variable on the dependent variable, using timeline series as monthly and daily data from Jan. 2014 - Dec. 2021. Also, for pandemic information, use monthly series from January 2014 - December 2019 and January 2020 - December 2021 to see if there is a difference between Corona Pandemic before and after. Antoniou et al. (2005) mentioned that the GARCH analysis technique avoids spreading nonstationary returns.

2. LITERATURE REVIEW

Previous research shows a positive effect, (Kang & Vespignani, 2017), and the result also shows a negative effect. (Ildirar & İşcan, 2015). Both pieces of research used more general commodity items. Other pieces of research (e.g., Filis et al., 2013; Chiou et al., 2009; Gorton et al., 2006; Choi et al., 2010) have researched and shown positive signs of commodity prices. Energy and metal types include oil, silver, gold, gas, and more. These types of commodities affect the stock market. In Indonesia, a study also shows a positive influence between oil, gold, silver, and exchange rate against mining companies (Putra & Robiyanto, 2019). Since several previous studies have researched gold, silver, and oil in Indonesia, this current research will focus on other specific commodity items such as palm oil, soybeans, wheat, and corn. Furthermore, the exchange rate of US$ / ID affects the stock market of agricultural companies and animal companies in Indonesia, which is supported by other studies showing that the price of palm oil and the exchange rate significantly influence the stock market (Sabariah, et al. 2014).

3. METHODS

Research on the relationship between stock returns and volatility in South African and China Stock Markets uses the GARCH model methodology.The result of the study revealed persistent volatility in both exchange markets and resembled the same movement in returns. Therefore, financial professionals frequently refer to the GARCH process to approach volatility in financial markets. Furthermore, GARCH provides a more real-world context than other models when trying to predict the rates of financial instruments (Kenton, 2020). Therefore, this study focuses on finding the best GARCH model and the effects of several commodity rates.

This study uses the return of stocks and the ratio of prices to form the required data of the model in the form of the returns of agriculture and animal feed companies in Indonesia. The monthly and daily data must be changed by the return to the value of the shares in the previous month/day. The stock/price return ratio can be calculated using the formula below, where r_tis the return of the ratio on month t, P_tis the closing price on month t, and P_t-1 is the closing price on month t - 1:

r_t = P_t-P_t-1/P_t-1

The population of this research is from agriculture and animal feed companies listed on the Indonesian Stock Exchange (IDX) from 2014 to 2021. Of 29 companies, 23 companies are the best criteria to include (see Table 1). The total data of this study is 96 observations and 2.088 observations. As for easier showing in the result, there is the code for all variables (see Table 2).

Table 1.Sampling DataIDX
Criteria	Number of Companies
Animal Feed companies listed on IDX	4
Agriculture companies listed on IDX and have recording stock returns data from 2014 - 2021	19
Animal Feed companies listed on IDX and have recording stock returns data from 2014 - 2022	4

For GARCH (1,1), the model can be expressed as follows :

σ²_t= ω+ αε²_t-1+ βσ²_t-1>> 1 -----------------(1)

ω= γ ∨_L>> 2 (3) ----------------(2)

Where σ²_t is the estimate of the variance for day t, αε²_t-1= σ²_t-1 Z²_t-1 and σ²_t-1 represent tthe associated squared error and the conditional variance on the previous day, respectively, with α and β being their respective weights. The long-run variance ∨_L is an average level towards which the variances revert through a mean reversion principle, with γ being the weight assigned to such an average level. The main feature of this model is that it captures the clustering of volatility in the data through the persistence parameter + β < 1 to ensure a unique stationary process and positivity of the conditional variance. However, if the persistence parameter + β equals 1, the GARCH model converges to the Integrated GARCH (IGARCH) model, where long-term volatility carries an explosive process IGARCH is the particular version of the GARCH (1,1) model where, as advanced above, the persistence parameter ( α + β ) is equal to 1 and typically imports a unit root under the GARCH process. Thus, the conditional variance in IGARCH (1,1) is :

σ²_t= ω+ αε²_t-1+ (1- α ) σ²_{t-1 -----------------(4)}

Given that β is set equal to (1 – α ) with restrictions ω ≥0, α ≥0, and 1- α ≥ 0. and 1- α ≥0. In the GARCH (1,1) and IGARCH models, the impact of positive and negative news on the conditional variance is symmetric. These models restrict all coefficients to be greater than zero, and thus cannot explain the negative correlation between return and volatility. Therefore, current return and future volatility might have a negative correlation, and the impact of positive and negative shocks on the conditional variance is somewhat asymmetric (Black, 1976). This asymmetrical became known as the "leverage effect," after which more advanced models developed to incorporate its effect. It is essential to distinguish the leverage effect from volatility feedback (Naimy, Haddad, Ferna ́ndez-Avile ́s, & Khoury, 2021). GARCH (1,1) steps are expressed as follows (Figure 2).

Figure 2.GARCH flowchart

4. RESULTS

As mentioned previously, the objective of this study is to show the output by using GARCH modeling with a unit root, indication result, diagnostic test, and discussion.

4.1 Unit Root Test

A series is considered stationary in the broad sense, in the weak sense, or in the second order if it has a fixed mean and constant variance (Yaffee & McGee, 2000). The stationary of a time series has a high possibility of significantly influencing its properties and forecasting behavior. On the contrary, the inability to render a time series in the correct form of stationery can lead to spurious results (Greunen, Heymans, Heerden, & Vuuren, 2014). All variables are stationary. However, for monthly data, MAGP is not suitable to be used due to the time series plot not showing volatility time series (see Figure 1 for a time series plot example). The data creates the singular matrix and is not available to test. That is the opposite in daily data and MAGP showing stationary.

4.2 GARCH Indication

In GARCH (1,1), the sum of the ARCH and GARCH coefficients ( α + β ) is less or close to 1, and this indicates that volatility shocks are pretty persistent (Eviews, 2020) or are equal to 1 for IGARCH model (Naimy, Haddad, Ferna ́ndez-Avile ́s, & Khoury, 2021).

All of the ARCH (RESID) and GARCH models are significant. The monthly data shows 14 output variables using IGARCH, and the total of theARCH and GARCH coefficients equals 1. The rest of the eight GARCH (1,1) models have a total of ARCH and GARCH coefficients of less than 1. The ARCH (RESID) and GARCH models are significant for daily data. Daily data showing all output variables GARCH (1,1) and the total of ARCH and GARCH coefficients show less than 1. All of this data shows that the models are already persistent.

4.3 Diagnostic Test

Using four diagnostic tests, the Q statistic correlogram, the squared residuals correlogram, the normality test, and the ARCH effect test. All of the models for 23 companies show a 24 Lag Correlogram for both Q-Statistic and Squared Residuals significantly higher than 5% that shows the models are correctly specified. ARCH effects also show higher than 5%, meaning the model has no heteroskedasticity in the residuals. As for the normality test, all models support three error distributions, and the Jarque-Bera result shows normality.

4.4 Correlogram and ARCH Effect

The Q-test is used to test for the presence of autocorrelation. The Ljung-Box statistics of the residuals can be used to check the adequacy of a fitted model (Tsay R. S., 2005). In this Q-Test, there are several residual tests: Correlogram Q-Statistic. This test is one of the residual diagnostics to display the correlogram, either autocorrelations or partial autocorrelations. Also, to test the remaining serial correlation in the model's mean equation and check the qualification of the mean equation. If all Q statistics are higher than a 5% significant level, meaning that there is no significance, the mean equation of the model is considered correctly specified (Eviews , 2020) and the correlogram squared residual test. This test is one of the residual diagnostics to display the correlogram (autocorrelations and partial autocorrelations) of the squared standardized residuals to test for the remaining ARCH in the variance equation of the model output and to check the qualification of the variance equation. If all the Q-statistics is higher than the 5% significant level, meaning that there is no significance, the mean equation of the model is considered correctly specified (Eviews , 2020).

The practical problem with this Q-test is to choose the order of lag to use for the test. It is an empirical issue, and there is no a priori guide for the lag's maximum length (Gujarati & Porter, 2009). The number of lags for annual data tends to be minor, with 1 or 2 lags. For quarterly data, 1 to 8 lags are appropriate; for monthly data, 6, 12, or 24 lags can be used given sufficient data points (Wooldridge, 2009). Since this study also uses daily data, there is a way to choose optimal lags in EVIEWS by using VAR (Vector Autoregressive) estimation, advising to use the Akaike Selection Criterion (AIC) in selecting the lag length. The information criterion with the smallest criterion value evidences the ideal lag length to employ (Adeleye, 2018) and the ARCH-LM test to investigate whether the proposed data contain any variation in conditional volatility (Hartman et al. 2012). The desired result is an insignificant statistic, indicating that there is no significant ARCH effect in the standardized residuals (Danielsson, 2011).

The output of the correlogram on monthly and daily data shows mostly 24 lags that are not significant in the Q statistic for the mean equation and squared residual for the variance equation, meaning that all of the models are correctly specified Daily data for AALI, ANJT, BTEK, LSIP, and TBLA must be checked on how many lags are sufficient to use best. Using the VAR lag criteria system identifying AIK and SC that is showing for two lags are enough optimal lag to show insignificant. Therefore, all monthly and daily data are correctly specified. For the ARCH effect, all models are not significant (higher than 5%), meaning that there is no significant ARCH effect in the standardized residuals.

4.5 Jarque-Bera Test

The GARCH model evolved on the basis of standard normal (Gaussian) distribution. Nevertheless, significant evidence suggests that the financial time series is hardly Gaussian but typically leptokurtic and heavy-tailed (see Figure 3). In order to solve this problem, the Maximum Likelihood Estimation Based on Gaussian is extracted from the Student's t distribution and the Generalized Error Distribution (GED) (Lingbing & Yanlin, 2017). Both those two distributions have been used as alternatives in finance research (Chkili, et al. 2012), (Fan, Zhang, Tsai, & Wei, 2008), (Mabrouk & Saadi, 2012).

Figure 3.Normal and Leptokurtic Graph

Several authors have suggested that the time series of asset returns show a very peculiar characteristic since the data distribution is with heavy tails,negative skewness, and asymmetric (Cont, 2001) , (Chu, et al. 1996), and (Ding, et al. 1993). There are also empirical facts on asset returns, the presence of the so-called volatility clustering, conditional heteroskedasticity, and the long-term memory property (Almeida, Basturk, Kaymak, & Sousa, 2014) and (Wan & Si, 2017). All daily data results are non-normal, even using Normal (Gaussian), and for monthly data, mostly showing normal as mentioned before, since this is using data from financial time series showing a very peculiar distribution.

4.6 GARCH (1,1) Output

The GARCH (1,1) model uses five independent variables: CORN, EXC, PLMOL, SYB, and WHT, with sample monthly data and daily data from 2014 to 2021, followed by a significant level of probability of 5%. Table 1 and Table 2 results as CORN rate effects on five companies that are AALI, BISI, BWPT, CPRO and MAIN, in monthly data, and ten companies that are AALI, ANJT, BISI, BTEK, BWPT, CPRO, DSFI, GZCO, IIKP and SGRO in daily data. For US$/IDR the exchange rate affects ten companies which are AALI, BISI, BWPT, JAWA, JPFA, LSIP, MAIN, SIMP, SMARand TBLA, for monthly data, and 14 companies which are AALI, BISI, BTEK, CPIN, CPRO, DSNG, GZCO, IIKP, JPFA, MAGP, MAIN, SIMP, TBLA, and UNSP for daily data. As for Palm Oil Rate effects to 9 companies which are AALI, BWPT, CPRO, DSNG, IIKP, LSIP, SIMP, SMAR, and TBLA for monthly data, and ten companies which are AALI, ANJT, BTEK, BWPT, GZCO, IIKP, JPFA, LSIP, SIMP, and TBLA for daily data. For soybean rate effects, six companies are BISI, BTEK, BWPT, CPRO, DSFI, and JPFA for monthly data, and eight companies are AALI, BTEK, BWPT, CPRO, GZCO, LSIP, SGRO, and SIMP for daily data. Lastly, for Wheat rate effects to 12 companies which are AALI, BISI, BWPT, CPRO, DSFI, IIKP, JAWA, JPFA, MAIN, SGRO, SIPD, and SMAR for monthly data, and nine companies which are AALI, ANJT, BTEK, BWPT, CPRO, GZCO, IIKP, LSIP, and SGRO for daily data.

Table 2.GARCH (1,1) Coefficient Results INotes: * are probability values that are significant 5% level.
Dependent Variable	Monthly Data (Coefficient)
Dependent Variable	CORN	EXC	PLMOL	SYB	WHT

AALI	0.2905*	1.9810*	0.6053*	0.0315	0.2418*
ANJT	0.0808	0.0585	0.1462	0.0297	0.05
BISI	0.3168*	0.5351*	-0.0136	0.5747*	0.6501*
BTEK	-0.0001	-0.0002	-0.0001	0.0000*	0.0001
BWPT	0.7481*	2.7616*	0.8100*	0.7407*	0.6755*
CPIN	0.013	-0.6775	0.1918	0.018	-0.0726
CPRO	2.9496*	2.2452	0.8282*	0.8838*	1.7407*
DSFI	0.2532	0.0912	0.2342	0.9786*	0.7278*
DSNG	0.0689	-0.3098	0.4514*	0.0312	-0.0522
GZCO	-0.0146	-0.0006	0.0848	0.0811	0.005
IIKP	-0.552	0.7699	1.8612*	-0.1653	1.6736*
JAWA	0.1765	0.6246*	0.0062	0.0097	0.1909*
JPFA	-0.3839	2.9359*	0.2987	0.3569*	0.4535*
LSIP	-0.0279	0.3169*	0.8007*	0.1655	0.1614
MAIN	0.7214*	3.2612*	0.2656	0.4029	1.0066*
PALM	-0.3536	-0.397	-0.0033	0.2445	0.2281
SGRO	0.0371	-0.1288	0.1837	0.1308	0.3174*
SIMP	0.0086	1.2709*	0.6654*	0.2714	-0.0871
SIPD	-0.0932	-0.2307	0.0000	0.0126	0.2106*
SMAR	0.1476	0.5591*	0.7289*	0.413	0.5449*
TBLA	0.0537	1.3418*	0.4361*	-0.0185	0.0407
UNSP	0.0137	-0.1145	0.0235	-0.0012	0.0023

Table 3.GARCH (1,1) Coefficient Results IINotes: * are probability values that are significant 5% level.
Dependent Variable	Daily Data (Coefficient)
Dependent Variable	CORN	EXC	PLMOL	SYB	WHT
AALI	0.0747*	0.2272*	0.3107*	0.1135*	0.0762*
ANJT	0.0511*	-0.1107	0.0579*	0.0142	0.0513*
BISI	0.0806*	0.5250*	0.0167	-0.0034	-0.0137
BTEK	0.0046*	0.0138*	0.0004*	0.0086*	0.0018*
BWPT	0.1151*	-0.2746	0.3859*	0.1786*	0.1104*
CPIN	-0.0301	0.6073*	0.0383	-0.0297	0.0088
CPRO	0.0249*	0.0276*	0.0028	0.0811*	0.1208*
DSFI	0.1509*	0.0349	0.0528	0.0792	0.0594
DSNG	-0.0624	0.2542*	0.0387	0.0242	0.0255
GZCO	0.0000*	-0.000*	-0.000*	-0.000*	-0.000*
IIKP	0.0000*	-0.000*	-0.000*	0.0000	-0.000*
JAWA	0.0309	0.0396	0.0405	-0.0197	0.0199
JPFA	0.0285	0.9432*	0.1029*	0.0064	0.0349
LSIP	-0.0391	-0.0291	0.3891*	0.1110*	0.0981*
MAGP	0.0000	0.0000*	0.0000	0.0000	0.0000
MAIN	-0.0214	0.6617*	0.0488	0.0464	0.0416
PALM	0.0011	-0.0179	0.0027	-0.0002	-0.0007
SGRO	0.1170*	0.0064	0.019	0.0921*	0.0907*
SIMP	-0.0291	0.4777*	0.2477*	0.0789*	0.0401
SIPD	0.0003	-0.0039	0.0015	-0.0012	-0.0005
SMAR	-0.0322	-0.1374	0.0132	0.0105	0.0274
TBLA	-0.0126	0.3466*	0.1302*	-0.0283	0.0161
UNSP	0.0074	0.4340*	-0.0091	-0.0077	0.0327

Following evaluation by (Sinbanda & Mlambo, 2014), an evaluation of the variable indicates the coefficient number (see Table 9 for monthly data and Table 10 for daily data). Negative coefficients indicate that increasing exogenous variables led to increasing endogenous variables. On the contrary,positive coefficients indicate that increasing exogenous variables led to decreasing endogenous variables. For example, AALI is affected by the Corn rate significant with a probability of 0.0008 (lower than the 5% level) with a coefficient number negative of 0.029 (monthly data). The result reveals that an increase of a 0.29 percentage point increase in CORN leads to a one percentage point increase in the AALI rate. As AALI is affected by thePalm Oil rate, the results reveal that depreciation of the Palm Oil rate from a 0.6 percentage point increase for one percentage of AALI.

4.7. COVID-19 Pandemic

There is an ongoing COVID-19 pandemic started by the World Health Organization. This pandemic impacted the stock market and reacted with a significant increase in volatility, and in some cases, stock markets plunged by an almost unprecedented amount (Albulescu, 2020), (Zhang, et al. 2020), and (Omari, et al. 2012). Although market volatility was the greatest during March, it remained high compared to prepandemic levels in the month after that (A. Bartolome del Canto, 2021). After showing the volatility results using the GARCH (1,1) model that shows the relevant and suitable model for time series volatility, and with this pandemic situation following different market conditions. This study will also investigate the effect of commodity prices on agriculture and animal feed companies with monthly data on two different timelines, which are before COVID-19 and after COVID 19.

This section covers two different timelines: time series from 2014 until 2019 (before the pandemic) and time series from 2020 until 2021 (after the pandemic). In the previous section, all the data already appear stationary. However, 18 companies can include because five companies have no data return happening on 2020 and 2021, so to have a better comparison, only use 18 companies. All total ARCH and GARCH have less than one and equal to 1 for IGARCH restrictions.

Table 4.Summary of GARCH (1,1) Output (Pandemic) INotes: * are probability values that are significant 5% level.
Dependent Variable	Independent	Monthly Data From 2014 to 2019	Monthly Data From 2020 to 2021
Dependent Variable	Independent	Coefficient	Coefficient
BISI	PLMOL	0.955833*	-0.560454*
CPIN	PLMOL	-0.21182*	0.611*
CPIN	WHT	-0.351102*	-0.565434*
DSNG	PLMOL	0.382183*	0.894686*
LSIP	PLMOL	0.706117*	0.750718*
SIMP	PLMOL	-0.396077*	1.591583*
SIMP	SYB	-0.730591*	-0.680628*

Table 5.Summary of GARCH (1,1) Output (Pandemic) IINotes: * are probability value that are significant 5% level
Dependent Variable	Independent	Monthly Data From 2014 to 2019	Monthly Data From 2020 to 2021
Dependent Variable	Independent	Coefficient	Coefficient
AALI	CORN	-0.389694*	-0.1892
AALI	PLMOL	0.720851*	0.4928
AALI	SYB	0.397687*	0.17327
AALI	WHT	0.459385*	0.1702
ANJT	WHT	0.1974*	0.24685
BISI	CORN	1.352479*	-0.18841
BISI	WHT	-0.79376*	0.21984
BWPT	EXC	-5.556172*	-2.39155
BWPT	PLMOL	1.705463*	-0.25382
DSFI	CORN	-0.494346*	-0.10895
DSFI	EXC	3.109288*	-0.99147
DSFI	PLMOL	0.549959*	0.38243
DSFI	SYB	1.150743*	-0.44251
DSFI	WHT	-0.889046*	0.18892
JAWA	EXC	-1.610287*	-2.41943
JAWA	SYB	1.108163*	0.01696
JAWA	WHT	-0.945967*	0.06211
JPFA	EXC	-3.084604*	-0.85569
SGRO	PLMOL	0.463725*	0.21707
SGRO	SYB	0.274371*	-0.46518
SIPD	WHT	-0.405838*	0.97838
SMAR	SYB	-0.484162*	0.32655
TBLA	EXC	-1.370766*	-1.52008
TBLA	PLMOL	0.326883*	0.47402
UNSP	CORN	-1.845674*	0.03389
UNSP	PLMOL	0.928861*	0.34828
UNSP	SYB	1.25782*	-0.03118

Table 6.Summary of GARCH (1,1) Output (Pandemic) IIINotes: * are probability values that are significant 5% level.
Dependent Variable	Independent	Monthly Data From 2014 to 2019	Monthly Data From 2020 to 2021
Dependent Variable	Independent	Coefficient	Coefficient
AALI	EXC	-0.23074	-2.765064*
ANJT	CORN	0.02154	-0.239238*
BISI	EXC	-0.41721	-2.61069*
BISI	SYB	-0.59801	0.677969*
BWPT	CORN	-0.1622	-1.027809*
CPIN	CORN	0.08988	0.247699*
DSNG	EXC	0.516	-1.370619*
DSNG	SYB	0.05446	0.048874*
JPFA	WHT	0.28566	-2.064958*
MAIN	EXC	-0.44066	-3.277342*
SIMP	CORN	0.24748	0.692395*
SIMP	EXC	-0.87932	-2.390485*
SIMP	WHT	0.32344	-0.810186*
SIPD	EXC	-0.55748	-3.749156*

Several companies have mixed outputs significantly compared to before and after the pandemic. As shown in Table 3, we see that there are still impacted together before the pandemic and after the pandemic. For BISI stock return rate that is affected by the Palm Oil rate, CPIN stock return is affected by the Palm Oil rate and wheat, DSNG stock return is affected by the Palm Oil rate, LSIP stock return is affected by Palm Oil, and SIMP stock return is affected by Palm Oil and soybean. In Table 4, 27 variables are significant on the timeline before the pandemic but not on the timeline after. Also, in Table 5, 14 variables are the opposite, meaning those are significant after the pandemic but not before the pandemic. Surprisingly, both timelines' data showed normal distribution by Jarque-Bera output.

7. DISCUSSION

The result shows that all models show no normal distribution for daily data, which means much fatter tails. Many studies in the empirical finance literature on the volatility relationship of returns primarily use monthly return data. Higher frequency (such as daily) data are frequently not used,since returns are typically rather noisy, rendering the return relationship in high frequency challenging to establish. Moreover, studies (e.g., Goyal et al., 2003; Zhang et al., 2005; Gui et al., 2008; Ludvigson et al., 2007; Jiang et al., 2010; Zhang, 2010) examined the effects of macroeconomic variables on asset pricing, and these variables are only available monthly or quarterly. However, seeing the result of significant effect, there are 51 observation data significant by using daily data and more than 42 observation data significant by using monthly data. Other researchers (e.g., Banerjee, et al., 2006; Dritsaki, 2017; Karmakar, 2006) conducted their research to model volatility by daily frequency data series.

8. CONCLUSION

The companies with more significant effects are AALI, BWPT, and CPRO, using monthly or daily data. These three companies are agriculture companies and have affected almost all dependent variables. PALM has no effect on both monthly or daily data. For the four animal companies, there are two companies, JPFA and MAIN, that have affected most of those independent variables. Both companies are affected by the US$/IDR exchange rate and the wheat rate in the monthly data, and SIPD is affected by the wheat rate in the monthly data. Regarding the impact of the pandemic, the Palm Oil rate still affects several companies both before and after the pandemic, although the pandemic situation has also been impacted with 27 variables significant before but not significant after the pandemic and 14 variables significant after the pandemic but not significant before the pandemic. This research has shown that commodity prices and the US$/IDR exchange rate affect the stock return of agricultural and animal companies. Therefore, the investor can consider these commodity rates to be their key driver for the stock returns of target companies.

9. LIMITATION

There is a problem with finding the parameter estimates for the error distribution both in GED and Student’s t when finding the best GARCH (1,1) model. The best model finding may not present the best probability due to distributions that affect the tail distribution. As a result, daily data are not normal for all model data, and monthly data has a more normal distribution, but daily data have more significant results. Therefore, there is a high possibility that using monthly data is the best. However, it is better to use a more extended period, perhaps more than ten years.

Read Counter : 1198 Download : 247

Share Now

The Effect of Commodity Prices and Exchange Rate on the Stock Return of Agriculture and Animal Feed Companies in Indonesia

Abstract

Full Article

References

Authors

Indexed

ISSN & Website Link

Contact Info

Article Sidebar

Abstract

Full Article

References

Authors

Article Details

Indexed

ISSN & Website Link

Contact Info