Effendi, Magdalena (2023) Pemodelan dan Peramalan Volatilitas Return Antar Indeks Pasar Modal dengan Menggunakan Fractional Cointegration Model. Masters thesis, Institut Teknologi Sepuluh Nopember.
Text
6003212014-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 September 2025. Download (5MB) | Request a copy |
Abstract
Salah satu bentuk investasi yang banyak dipilih oleh investor adalah saham. Pada dasarnya harga saham bersifat fluktuatif sehingga dalam menyikapi hal tersebut, investor perlu untuk memonitor pergerakan harga saham. Adanya perubahan harga saham mengakibatkan indeks harga saham berubah sehingga banyak digunakan untuk menganalisis return indeks saham. Semakin besar return yang diharapkan maka semakin besar pula risiko yang akan dihadapi. Faktanya, hampir semua investasi mengandung risiko. Oleh karena itu, penting untuk meramalkan risiko dengan melihat fluktuasi return indeks atau volatilitas. Volatilitas dapat menyebabkan return indeks mengalami heteroskedastisitas. Oleh karena itu, untuk menanggulanginya dapat menggunakan model GARCH. Selain efek heteroskedastisitas, pada sebagian besar volatilitas juga ditemukan adanya fenomena long memory. Dewasa ini, pasar modal menjadi lebih terkointegrasi dari yang sebelumnya hanya tersegmentasi di dalam suatu negara. Akibatnya, perubahan pada pasar modal suatu negara dapat mempengaruhi pasar modal negara lainya. Atas dasar pertimbangan tersebut, penelitian ini bertujuan untuk melakukan pemodelan dan peramalan volatilitas antar indeks pasar modal IHSG, S&P 500, N225, SZSC, FTSE 100, dan BSESN 30 dengan membandingkan model tanpa adanya asumsi cointegration terhadap model yang ada asumsi cointegration pada data volatility return kuadrat dan realized volatility dengan periode pengamatan terbagi menjadi dua yakni data in-sample periode 1 Januari 2012-31 Desember 2022 dan data out-sample periode 1 Januari 2023-30 April 2023. Dari hasil penelitian menunjukkan bahwa model fractional cointegration dapat diterapkan untuk peramalan baik pada data volatility return kuadrat maupun realized volatility. Penelitian ini juga menunjukkan kecenderungan bahwa model terbaik untuk meramalkan volatility return kuadrat adalah model dengan adanya asumsi cointegration, sedangkan untuk meramalkan realized volatility adalah model tanpa asumsi cointegration. Namun, jika data volatility return kuadrat dan realized volatility terindikasi memiliki proses long memory non stasioner maka model tanpa adanya asumsi cointegration cenderung lebih baik untuk digunakan.
=================================================================================================================================
One form of investment that many investors choose is stocks. Stock prices fluctuate, so investors need to monitor stock price movements in response. A change in stock prices causes the stock price index to change, which is widely used to analyze stock index returns. The greater the expected return, the greater the risk that will be faced. The fact is almost all investments involve risk. Therefore, predicting risk by looking at index return fluctuations or volatility is essential. Volatility can cause index returns to experience heteroscedasticity. Therefore, to overcome it, you can use the GARCH model. In addition to the heteroscedasticity effect, most of the volatility also found the phenomenon of long memory. Capital markets are becoming more cointegrated than previously only segmented within a country. As a result, one country`s capital market changes can affect other countries` capital markets. Based on these considerations, this study aims to model and predict volatility between capital market indices IHSG, S&P 500, N225, SZSC, FTSE 100, and BSESN 30 by comparing models without cointegration assumptions to models with cointegration assumptions on quadratic volatility returns and realized volatility with observation periods divided into two, namely in-sample data for January 1st 2012-December 31st 2022 and out-sample data for January 1st 2023-March 31st 2023. The research results show that the fractional cointegration model can be applied to better forecast quadratic volatility return data and realized volatility. This study also shows a tendency that the best model for predicting quadratic volatility returns is a model with cointegration assumptions while predicting realized volatility is a model without cointegration assumptions. However, if the quadratic volatility return data and realized volatility are indicated to have non-stationary long memory processes, the model without cointegration assumptions tends to be better.
Item Type: | Thesis (Masters) |
---|---|
Uncontrolled Keywords: | fractional cointegration, GARCH, pasar modal, return, volatilitas; fractional cointegration, GARCH, capital market, return, volatility |
Subjects: | H Social Sciences > HA Statistics > HA30.3 Time-series analysis H Social Sciences > HB Economic Theory > Economic forecasting--Mathematical models. H Social Sciences > HG Finance > HG4529 Investment analysis Q Science > QA Mathematics > QA280 Box-Jenkins forecasting Q Science > QA Mathematics > QA401 Mathematical models. |
Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis |
Depositing User: | Magdalena Effendi |
Date Deposited: | 26 Sep 2023 02:23 |
Last Modified: | 26 Sep 2023 02:23 |
URI: | http://repository.its.ac.id/id/eprint/103558 |
Actions (login required)
View Item |