KnE Social Sciences | International Conference on Economics, Business and Economic Education 2018 (ICE-BEES 2018) | pages: 860–868

, , and

1. Introduction

The availability of basic needs is important for human lives. People need basic need in their lives therefore, the price that circles around the basic needs fluctuates and shakes the economic situation. Generally, the correlation is influenced by the supply and demand of the situated market. When the supply is increased, usually, what happens in the market's demand is decreased with all other things being equal and vice versa. Basic needs that people consume are from agriculture and farm commodities. Those commodities are one of many staple goods that have irresponsive demand elasticity to price changes hence, it is called inelastic. If there is a price change to certain basic needs in a massive amount of change, the price of the good is considered stable (inelastic) to price change [1]. Since, there is a low price elasticity of demand mixed with the change of seasons in the production causes high price fluctuation [2]. The effect of that phenomenon is when there is a increased price change on certain basic need commodities will cause a shock effect, such as higher price change when there is a change on supply side.

The aim of this paper is to observe and do technical analysis to price fluctuation to certain basic food commodities in Semarang City using ARIMA (Autoregressive Integrated Moving Average) model as the analysis tool. Food price analysis is used to formulate price stability policy and increased production. Accurate public information can form an accurate policy by forecasters, also to reduce price variance in the market [3]. The application of ARIMA as an analysis tool of short-term prediction, hopefully, can anticipate and give an early warning message to food price fluctuation in the future. The more accurate policy formulation, the more it can prevent the shock of price to happen in the future. So, the shock could be prevented to either happen to consumers or producers. Consumers can still consume basic needs with reasonable price, and also for the farmers and stock farmers still can sell their commodities to the market with accurate policy. Same idea [4] that forecasting can help people on making accurate decision timely to face the uncertainty of price in the future.

In this paper, the ARIMA-forecasted commodities are chicken meat, eggs, chili, and shallots. This is because within the period of analysis starts from 1 st of August 2015 to 31 st of May 2018, all four of the commodities showed that the price fluctuated different than other commodities, especially it showed on the special periods such as at the end of the year. The analysis is meant as an early detection of price change, especially those basic needs for people. Price control is one of the effort to decrease the risk for producers (farmers). The form of price control is the way to forecast the price in the future. The result of this forecasting analysis can be used as part of the farmers' decision also for government as the policy maker.

Figure 1

Food Price Growth In Jawa Tengah.

fig-1.jpg

In previous research, [5] stated that forecasting using ARIMA model can predict the price of staple goods on a national scale with an error rate of 2.22%. The study [6] compares the use of the Winter Smoothing model with the ARIMA model on Bulog's rice stock and concluded that the use of Winter Smoothing further yields Mean Square Error (MSE) smaller than ARIMA model. Also, [7] concluded that the deficiency in the ARIMA model cannot predict a downward or sharply moving data movement in the Consumer Price Index (CPI) due to external factor interventions such as the financial crisis.

Recently, the determination of price commodity involves the trial-error system so, the stability of chili commodity price becomes unattended and shows unstable price fluctuation. This is because the system still embraces fundamental analysis. Fundamental analysis of commodity involves each of product of commodity, generally, fundamental analysis of commodity includes something very staple and basic such as, supply and demand analysis, seasons, plagues, and natural disasters. This research is conducted in order to develop complementary models, especially in terms of data processing of food commodity price technically. In this study, the ARIMA model is used to predict the daily price of staple goods in Semarang City in the short term. ARIMA is suitable for time-series forecasting studies especially for short-term prediction [8,9].

2. Method

The data used in this study used secondary data on Semarang City Food Commodity Price Daily from April 2015 to June 2018 obtained through the Central Java Inflation Control Team through pricejateng.org website. Data processing tools used in this study using E-Views.

This research uses Autoregressive Moving Average (ARIMA) model. The development and design of the ARIMA model as the forecasting tool of the known financial-economy variables is ARIMA Box-Jenkins (1976). Box-Jenkins finds the ARIMA model (p, d, q) that meets the stochastic procedure in which the sample originated. The Box-Jenkins procedure focuses on installing a mixed Autoregressive Integrated Moving Average (ARIMA) model into a set of data [10]. The Box-Jenkins methodology consists of four repetitive steps of model identification, parameter estimation, diagnostic test and model forecasting [4]. In the study [11], ARIMA model applications through three phases of model identification, parameter estimation, and diagnostic examination. The Box-Jenkins ARIMA model requires stationary procedures, as long as the data is not stationary, the time-series data identification step needs to be done in order to become stationary. A stationary procedure is a necessary condition in building an ARIMA model.

Figure 2

ARIMA Approaching Model Scheme.

fig-2.jpg

3. Result and Discussion

Results of data plots on all four food products analyzed showed a recurrent pattern in time series data. This indicates the presence of seasonal elements in the data. The nature of food prices that depend on nature such as temperature and availability of fertilizer and food business intensity such as peak harvest and planting season and national holiday [12]. In further analysis, the stationary assumption of variance and means is necessary as both are the main requirements of forecasting in this ARIMA method. The test results on the stationary of the data show that the four commodities studied have been stationary, the test results show that the Augmented Dickey Fuller (ADF) test statistic, it is known that the CHICKENMEAT, REDCHILI, EGGS, and SHALLOTS variables at a real level of 1% have stationary at the level (table 1).

Table 1

Stationary Test.


variabel ADF Test Statistic KETERANGAN
CHICKENMEAT -16,38809* Stationer
REDCHILI -27,38118* Stationer
EGGS -25,42486* Stationer
SHALLOTS -25,58626* Stationer
*) test critical values α 1%: -3,436462

By considering the stationary test, the next step can be identified plot Correlogram for estimation ARIMA Box Jenkins model of each commodity studied. Here is the Correlogram plot:

Figure 3

Correlogram Plot.

fig-3.jpg

Considering the correlogram, the ARIMA model identification for each commodity is as follows:

Table 2

ARIMA Model Identification.


Variabel Pola ACF dan PACF Model terpilih
CHICKENMEAT AR (1,2,3) MA (1,2,9) AR1, AR2, MA1, MA2, MA9
REDCHILI AR (1,2,4) MA(1,2,7) AR1, AR2, MA1
EGGS AR(1,2) MA (1) AR1, AR2, MA1
SHALLOTS AR (1), MA (1,7) AR1, MA1, MA7

The estimation results from the four commodities are as follows:

Table 3

ARIMA Model Estimation Results.


Variabel Hasil Estimasi R 2 RMSE MAPE
CHICKENMEAT CHIKENMEAT = 0 + [AR(1)=0.746128202189, AR(2)=0.253687714401, MA(1)=0.448089479807, MA(2)=-0.135154031822,MA(9)=-0.112326367488 0.94 683,82 0,97%
REDCHILI REDCHILI = 27628.0745422 + [AR(1)=1.52546016094, AR(2)=-0.534442706153, MA(1)=-0.384879131553 0.97 1539,23 3,21%
EGGS EGGS = 20167.3834621 + [AR(1)=1.81617177758, AR(2)=-0.821601427631, MA(1)=-0.672717913042 0,98 260,02 0,68%
SHALLOTS SHALLOTS = 25578.1160093 + [AR(1)=0.992525677226, MA(1)=0.186374751894, MA(7)=0.0849543289561 0,99 715,28 1,53%

Based on the estimation results used can be used to forecast for each food commodity in the period to come. Root Means Square Error (RMSE) and Mean Absolute Percent Error (MAPE) show significant results. MAPE is the most important part of statistics and can be used on observations and has the smallest variability from sample to sample [10]. MAPE can often be understood in user research, so MAPE is often used to report research results [13,14]. The results of the model's ability to be used as forecasts for each commodity show that the resulting model can be used to make predictions, which is shown by the actual value pattern identical to the fitted value pattern.

Figure 4

ARIMA Residual Model, Actual Pattern And Fitted Pattern.

fig-4.jpg

4. Conclusion

From the results of the research, for prediction data of chicken, egg, chili and red onion price using ARIMA time series model, based on Root Means Square Error (RMSE) and Mean Absolute Percent Error (MAPE) values, the four ARIMA models can be used as early warning model of commodity prices. The R-Squared values of all four models greater than 90% also indicate that ARIMA's overall model has a good performance to predict future prices, in anticipation of fluctuating market demand. The benefit of using this prediction model is that at the farmer / breeder level the relative will choose the starting time of production at the high price, so it is expected to help increase the income of farmers. For the government, it can be used as guides in determining floor price and ceiling price so that it can be maintained conducive price in terms of demand and supply.

References

1 

Fridaus, M. (2012). Manajemen Agribisnis. Jakarta: Bumi Aksara. W433W9360

2 

Brandt, J.A and Bessler, D.A. (1983). “Price Forcasting and Evaluation: An Application in Agriculture,” Journal of Forecasting vol. 2, pp. pp. 237-248.W433W9360

3 

Manfredo, Sanders, D.R, Mark, R. (2002). “USDA Production Forecasts for Pork, Beef, and Broilers: An Evaluation,” Journal of Agricultural and Resource Economics 27(1), pp. pp. 114-127. W433W9360

4 

Gujarati, D.N and Porter, D.C. (2010). Essentials of Econometric 4th Edition. New York: McGraw-Hill. W433W9360

5 

Rasyidi, M.A. (2017). “Prediksi Harga Bahan Pokok Nasional Jangka Pendek Menggunakan ARIMA,” Journal of Information Systems Engineering and Business Intelligence Vol. 3, No. 2, pp. 107-112. W433W9360

6 

Sugiarto, S, Gamal, M.D.H, Sanjaya, A. (2015). “Peramalan Stok Beras Bulog Pekanbaru Dengan Menggunakan Pemulusan Winter Dan ARMA (p,q),”. Prosiding Semirata 2015 bidang MIPA BKS-PTN Barat. Semirata, Pontianak: Universitas Tanjungpura Pontianak.W433W9360

7 

Pimpi, L. (2013). “Penerapan Metode ARIMA Dalam Meramalkan Indeks Harga Konsumen (IHK) Indonesia Tahun 2013,” Paradigma, pp. pp 35-46. W433W9360

8 

Kyungjoo, L, Schwan, Y, John, J. (2007). “Neural Network Model vs. SARIMA Model In Forecasting Korean Stock Price Index (KOSPI),” Information System vol. 8 no. 2, pp. pp 372-378. W433W9360

9 

Merh, N, Saxena, V.P, Pardasani, K.R. (2010). “A Comparison Between Hybrid Approaches of ANN and ARIMA For Indian Stock Trend Forecasting,” Journal of Business Intelligence, vol. 3, no. 2, pp. pp 123-131. W433W9360

10 

Jadhav, V, Chinnappa Reddy, B.V, Gaddi, G.M. (2017). “Application of ARIMA Model for Forecasting Agricultural Prices,” J. Agri. Sci. Tech., vol. 19, pp. 981-992.W433W9360

11 

Shahwan, T and Odening, M. (2007). “Forecasting Agricultural Commodity Prices using Hybrid Neural Networks,” Computational Intelligence in Economics and Finance, vol. 2, pp. 63-74.W433W9360

12 

Weron, R. (2014). “Electricity price forecasting: A review of the state-of-the-art with a look into the future,” International Journal of Forecasting, pp. pp. 1030-1081.W433W9360

13 

Farjam, A, Omid, M, Akram, A (et al). (2014). “A Neural Network Based Modeling and Sensitivity Analysis of Energy Inputs for Predicting Seed and Grain Corn Yields,” J. Agr. Sci. Tech, vol. 16, pp. 767-778.W433W9360

14 

Jadhav, V, Chinnappa Reddy, B.V, Gaddi, G.M (et al). (2016). “ Exploration of Different Functional Forms of Growth Models: A Censorious Analysis with Reference to Horticultural Sector in Karanataka,” Int. J. Tropica. Agr, vol. 34, no. 4, pp. 1107-1116. W433W9360 W433 W433

FULL TEXT

Statistics

  • Downloads 9
  • Views 33

Navigation

Refbacks



ISSN: 2518-668X