Organizational Unit: Department of Financial Engineering
Loading...
Date established
City
Country
ID
12 results
Publication Search Results
Now showing 1 - 10 of 12
Master ThesisPublication Metadata only Return prediction in turkish stock market via machine learningBabayakalı, Selen; Ahi, Emrah; Ahi, Emrah; Güntay, Levent; Akyıldırım, E.; Department of Financial Engineering; Babayakalı, SelenIn this study I compare machine learning methods for predicting the stock returns of individual Turkish stocks listed in the Istanbul Stock Exchange (Borsa Istanbul). As the main machine learning model I use the Instrumented Principal Component Analysis (IPCA) and as a benchmark model I use Fama-French Factor Model. The IPCA model generates the stock-level expected returns based on observable stock-level and firm-level characteristics and latent common factors estimated within the model. Within the model stock-level characteristics determine the factor betas, namely the covariances of stock returns with the latent common factors. I estimate versions of the benchmark Fama-French models between 3 to 5 factors. The versions of the IPCA models use between 3 and 6 factors and use 10 characteristics. The sample covers all stocks in the XUTUM Index and the sample period includes forecasts between 2010 and 2022. Using a panel data of 252 firms listed in the Borsa Istanbul XUTUM Index and I analyze the comparative performance of the IPCA and Fama-French models. More specifically, I look at the in sample and out of sample performances of the models by comparing the realized and predicted series of returns for each individual stock. I find that the IPCA model significantly outperforms the Fama-French model by obtaining significantly higher out of sample R-squared levels and correlation of return forecasts and realized returns. The performance difference between Fama-French and IPCA models is more pronounced in the Turkish stock market compared to results of (Kelly, Pruitt and Su 2018) for the US stock market. Therefore, my results imply that the use of asset pricing models based on machine learning techniques may provide better results in emerging stock markets.Master ThesisPublication Metadata only Do pension funds provide liquidity: Lessons from em countries(2017-01) Peksevim, Seda; Yalçın, Atakan; Yalçın, Atakan; Ozsoylev, H. N.; Özsoy, Satı Mehmet; Department of Financial Engineering; Peksevim, SedaIn response to the financial crisis, there has been increased attention on the importance of the long-term investors for financial stability. We contribute to this literature by investigating the role of pension funds on stock market liquidity, using a data set covering 23 EM countries over the period 2004-2014. In particular, we focus on whether liquidity supply by pension funds are more pronounced in periods of market turbulence. We find strong evidence that pension funds, i) supply liquidity in stock markets, and ii) liquidity provision effect is stronger in financial crisis times. Our findings have key implications for both policymakers and global fund managers. As liquidity provision is an important function of financial markets both in general and periods of market stress, EM economies that are relying heavily on foreign investors should increase domestic and long-term investor base (pension funds) in their financial markets. Auto-enrollment reform may provide a viable solution to increase pension fund size in EM countries.Master ThesisPublication Metadata only Determinants of liquidity adequacy ratio: An empirical study on Turkish banksÇokaklı, Osman Serhan; Ahi, Emrah; Ahi, Emrah; Güntay, Levent; Öztürk, G.; Department of Financial EngineeringLiquidity and liquidity risk is a phenomenon that has important consequences in the management of banks. In addition, liquidity risk has an important role in banking crises. With the new regulations created, liquidity indicators are taken into account in the evaluation of the financial strength of the banking sector and banks. This study examines the variables that affect the liquidity risk of the Turkish banking sector. As a result of the analysis using simple linear regression, it was found that the factors affecting the liquidity management in the Turkish banking sector, although their effectiveness changes periodically, are the ratio of cash values to total assets, the ratio of demand deposits to total deposits, the ratio of non-performing loans to total cash loans, and the ratio of funding from repo transactions to total liabilities.Master ThesisPublication Metadata only FX markets co-movement dynamics and global factors: Evidence from emerging marketsYurttaş, Gizem Ece; Ahi, Emrah; Ahi, Emrah; Güntay, Levent; Danışman, G. Ö.; Department of Financial Engineering; Yurttaş, Gizem EceEmerging markets are particularly different from developed markets as they possess vulnerability, easily disturbed by financial crises, and not completely liberalized. Some global shocks such as Covid-19 and 2018 global financial crisis easily affect the emerging markets in especially currency-based systems. Therefore, it is important to investigate the impacts of global factors on their currencies in the last few decades. The main purpose of this thesis is to investigate the effects of global factors on emerging markets between the years of 2009-2021. The cluster was narrowed by PCA analysis and regression analysis was performed on the factors obtained as a result of the analysis. Then, the analysis was made to minimize the average loss that will occur due to restructuring with the factors included in the PCA analysis results and to compare the movements of the currencies of developing countries within themselves with the regression results. Empirical tests are implemented into three different time periods to understand the effects of global factors: • High Liquidity Market Regime (January 2009 – June 2013) • Monetary Tightening Market Regime (July 2013 – December 2019) • The Emerging Markets in Pandemic (January 2020 – August 2021) The findings show that the currencies of the developing countries are highly correlated with the Dollar Index, MSCI World Index and EURUSD parity, but also it is observed that they are correlated with the commodity markets in times of crisis. On the other hand, even the emerging markets countries do not diverge from each other much during the global financial crisis, it is analyzed that the currencies of the countries that have difficulties in their domestic monetary and fiscal policies such as Turkey and Argentina are decomposed from the emerging markets currencies in all periods regardless of the type of crises.Master ThesisPublication Metadata only Valuation of fixed income securities and estimation of term structure on international bond market using machine learning techniquesDartanel, Ali; Ahi, Emrah; Ahi, Emrah; Güntay, Levent; Akyıldırım, E.; Department of Financial Engineering; Dartanel, AliIn this study, I focus on predicting bond risk premia in Turkish Eurobonds market using machine learning methods. Machine learning uses statistical learning techniques to gather useful structures of a data set without being explicitly programmed. In recent years machine learning has become a very popular topic and shown very good results in a wide variety of fields, but there is a lack of research in the field of term structure modeling. In order to predict Turkish Eurobond returns, I implemented several machine learning models such as OLS, PCA, Ridge, Lasso, Elastic net and neural networks. The raw data set I used comprises of Turkey Government Eurobond yields between 2005 and 2020, inclusive. Both monthly and yearly returns are estimated separately. Zero-coupon rates and forward rates are calculated from the raw data and used as left-hand site elements for machine learning predictions. Macroeconomic variables are also added to forward rates as factors. I compared the out-of-sample performance of the models and I found that Penalized linear regression yields the best results for excess bond return prediction, providing nearly 10% out-of-sample R2. Neural networks are the second-best performer yielding around 3-4% out-of-sample R2. Plus, adding macroeconomic variables to the models slightly improved the results by 2-3%. Also, yearly returns estimation performed better than monthly returns for OLS, Ridge, Lasso and Elastic net regressions, but not for neural networks.Master ThesisPublication Metadata only A analysis of the low-volatility anomaly on the Borsa Istanbul(2021-06-04) Yaman, Erkin Levent; Akat, Muzaffer; Akat, Muzaffer; Ahi, Emrah; Özuğurlu, E.; Department of Financial Engineering; Yaman, Erkin LeventIn financial markets, obtaining high returns against high systematic risk is basically the expected situation. It is regarded as a financial anomaly that high-risk stocks yield lower returns than low-risk stocks. This situation is called as low volatility anomaly or beta anomaly in financial literature. In this study, it was investigated whether there was a low volatility anomaly in BIST 100 between 2010-2019. 80 companies listed on the stock exchange between 2010 and 2019 were included in the study. Within the scope of the study, the presence of beta anomaly was investigated for periods of semiannual, 1, 2, 5 and 10 years. In addition, it was examined whether there was a beta anomaly for the years 2010-2013 and 2014-2019. The results obtained show that there is a beta anomaly within a certain beta value range in each period examined. Especially periodically, the presence of a general beta anomaly was detected in the period 2014-2019. In addition, it has been determined that beta anomalies generally do not occur in cases where the beta value is 1 and above in the short term, and very low and very high beta values reduce the return rate in the long term.Master ThesisPublication Metadata only Skipping across the Bosphorus: Post-jump returns at ultra-high frequencyPayze, Halil Bilgin; Güntay, Levent; Güntay, Levent; Ahi, Emrah; Yıldızhan, Ç.; Department of Financial EngineeringThis study rigorously investigates intraday jumps in highly liquid stocks listed on Borsa Istanbul (BIST) using two common jump test methods and ultra-high frequency data. Additionally, three purely price-based jump methods, referred to as Price Skipping, were developed to improve jump detection accuracy. The findings reveal the presence of reversal post-jump returns, irrespective of the jump direction. A profitable trading strategy was established, exploiting this market behavior by taking long positions after down jumps and short positions after up jumps. The strategy's performance was compared across different jump detection methods and confidence intervals. Logistic regression analyses demonstrate a significant negative correlation between relative tick size and jump occurrence. Additional analyses using Ordinary Least Squares (OLS) and Locally Estimated Scatterplot Smoothing (LOESS) consider factors such as relative tick size, volume, pre-jump returns, sector affiliation, jump timing, index jumps, and alternative jump detection methods. Importantly, a positive correlation was discovered between relative tick size and the magnitude of reversal post-jump returns. Another significant finding suggests a positive correlation between trading volume during jumps and the magnitude of reversal post-jump returns. These results indicate that high volumes, potentially resulting from large market orders, induce stock price jumps involving multiple ticks, and when the relative tick size is large, these movements trigger an overreaction, subsequently leading to a reversal within a defined time period. These findings suggest that the tick size is excessive for the small priced stocks and suboptimal for the middle range. Therefore, it is recommended that Borsa Istanbul's management consider revising their tick size policy to enhance market liquidity, considering the implications of this study.Master ThesisPublication Metadata only Effectiveness of technical indicators in predicting BIST100 index returnsKarabudak, Burak; Gökçen, Umut; Gökçen, Umut; Ahi, Emrah; Ertan, A. S.; Department of Financial EngineeringThis paper presents a logistic regression-based approach for predicting the direction of the Borsa Istanbul (BIST) index's price movements using technical indicators. Historical price data from 2005 to 2021 is utilized, and five thresholds are determined from -2% to 2% based on the daily or weekly returns of the BIST index. Binary variables which utilized in logistic regression as dependent variable are defined according to the thresholds, and logistic regression is applied using annual or semiannual training data. Daily models perform better than weekly models in logistic regression, and On Balance Volume (OBV) and Average Directional Index (ADX) are the indicators within the best statistical results. 0% threshold models are best in accuracy of prediction (approximately half of prediction is accurate). Weekly models are better than daily models, and annual models are better than semiannual models in accuracy. Some of the models outperformed the BIST index in cumulative return (daily cumulative return is 4,84; weekly is 4,48), with two models having approximately three times the cumulative return of the BIST index namely Daily-1% threshold-6 month's cumulative return is 13,35; Weekly-2% threshold-12 month's is 12,96. In summary, the results show that technical indicators can be successful in predicting stock or index returns depending on the training period, time period, and the combination of technical indicators used.Master ThesisPublication Metadata only Modelling prepayment in mortgages with a bank exerciseÖzdil, Ayşin; Güntay, Levent; Güntay, Levent; Ahi, Emrah; Danışman, G. Ö.; Department of Financial Engineering; Özdil, AyşinHome loans are one of the longest-term products in the Banking sector and are exposed to multiple macroeconomic cycles throughout their maturity. Each home loan contract includes the right of the loan to pay at any time during the term of the loan which causes the risk of changes in the contractual cash flows of the Banks. When the literature for the Turkish market is analyzed, studies are done on calculating home loan option prices and most of the calculations are based on classical option techniques. In previous studies, market spot interest rate and house price levels are used as variables. Aside from other studies in the literature, this thesis is based on prepayment probability of the home loans. Borrower-specific, loan-specific and macro economic specific factors are chosen as the variables affecting prepayment probability. To analyze the cyclical effects, the study is carried out in one of the top 10 banks with asset sizes by selecting a 12-year data observation interval between 01.01.2010 and 31.12.2021. The study aims to model customer prepayment behavior by estimating a logistic regression using 694.778 fixed-rate home loans and 247.572 prepayment events. The data set has about 30 million observations where each home loan has monthly observations until each home loan is closed. The logistic regression model can accurately predict the prepayment behavior of contracts with an Area-under-the-Curve (AUC) statistic of 0.921 and Gini coefficient of 0.843. In studies investigating home loans out of Turkey, interest rate, borrower income, loan to value ratio, borrower age, loan age and region are the variables that affect prepayment in the loan portfolios. This thesis shows that the most effective variables in the prepayment behaviors are the interest rate level changes, reference market interest rates, current risk of home loan and customer total debt which affect home loans payment schedule. Interestingly, original and current Loan to Value ratio, loan maturity, loan age, customer age, customer education status and customer income have limited impact on the prepayments as indicated by the model. The most plausible for this observation is the fact that people in Turkey buy their houses for residential purposes, not for trade and do not sell their houses in a short time unless it is compulsory can be interpreted.Master ThesisPublication Metadata only Comparison of different customer segmentation models for the financial sector versus the Turkish factoring sector and a clustering model proposal for a sme focused factoring companyAktuna, Mehmet; Güntay, Levent; Güntay, Levent; Ahi, Emrah; Özlük, Ö.; Department of Financial EngineeringFactoring is the process of transferring a forward receivable arising from commercial transactions to a factoring company by assignment. Companies that prefer to do factoring, instead of waiting for the due date of the collection of their receivables arising from their trade, transfer their receivables to the factoring company with a certain discount, with documents such as post-dated cheque, trade invoices and factoring contracts. The use of cheque in Turkey is different from that in the world due to the post-dated maturity written on the cheque and SMEs in Turkey frequently use cheques in their trade with maturity. In terms of segmentation, there are many studies for banks in the literature, but there is no study in the SME focused factoring segment. This study is the first, where a two-stage customer segmentation is implemented using the data of a SME focused Turkish factoring company. When the clusters generated by the K-Means and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) models are compared, it is observed that the clusters of the K-Means algorithm are more balanced in terms of distribution of the number of customers across clusters. On the other hand, the clusters generated by the DBSCAN model are more homogeneous, especially for outlier clusters. Homogeneity means an effective segmentation, which is there are few outliers within a cluster as defined by the Mahalanobis distance of each observation. The DBSCAN model can generate several homogeneous clusters containing a small number of customers if model parameters are appropriately calibrated. On the contrary, if few number of clusters are preferred, the homogeneity of the DBSCAN algorithm gets worse and the K-Means model gives more balanced clustering results.