• Selected
• Entire List

Abstract:
Financial academics and practitioners have recognized that average stock returns are related to past performance and cross-section of stock returns is that predictable based on past returns. Number of researchers report that past losers (negative or lowest return-stocks) outperform past winners (positive or highest return-stocks) or vice versa over the subsequent three to five years not only in US markets but also in other stock markets. This study examines the momentum and contrarian effects on stock returns in one of the leading emerging markets, Istanbul Stock Exchange (ISE) between years 1991 and 2000 by using the same empirical methodology in Jegadeesh and Titman (1993). It also investigates the weak-form efficiency of the stock market by examining the profitability of a number of contrarian strategies based on past prices, size, price, book-to-market, earnings-to-price ratios of stocks. Prior loser-stocks are found to outperform prior winner-stocks consistent with the predictions of the overreaction hypothesis. Compounded annual return difference between the top-winners and top-losers is around 15% in favor of loser-stocks since the average return difference during the 10-year period is 1.14% per month. Empirical findings for the longer-term average returns up to 36 month holding periods reveal a reversal of returns from 15 months to 36 months. On the other hand, we find that average abnormal returns and the average abnormal return difference per month between losers and winners increase as the holding period extends. Results also indicate that there is a downward trend in average returns for the winner stocks based upon the length of past returns that is used for portfolio formation but upward trend for the losers. Profitability of the strategies is robust to changes in the size of the portfolios. We also find that contrarian profits in January are significantly higher than those in non-January months, particularly for the short-term holding periods such as one and three months, however, losers outperform the winners in most of the months of the year. The overreaction is significantly stronger for smaller firms than for larger firms. Losers portfolio are typically smaller, lower priced, high-B/M and high-E/P stocks (as distressed stocks) than stocks in the winners portfolio. Our evidence indicates that there is significant price, size, B/M and E/P effects in stock returns in ISE, consistent to previous empirical work. After we analyzed 80 different strategies based on five different factors such as past-return, size, price, B/M and E/P, in various length of formation and holding periods, we find that stocks that have lower price, smaller size, lower past-return, higher-B/M and E/P are significantly provide higher returns than others. Price, size and E/P-based portfolios earn a larger return than loser-winner portfolios suggested by the overreaction hypothesis. Large profits of winners&losers portfolios might be subsumed or caused by the other factors such as size, E/P and B/M effects. It might most probably be size-based phenomenon with the contribution of other factors. On the other hand, findings show that losers are riskier than the winners because they are more sensitive to all three Fama-French factors, however, significant contrarian profits are partially related with risk factors and not captured by the three-factor model of Fama and French. Large abnormal returns of winner and loser portfolios and overreaction effects might be explained only in part by the risk factors. Finally, our results show that contrarian effects or more specifically "winners and losers effect" are existing on stock returns in ISE and the contrarian trading strategies that is buying past losers and selling past winners realize significant abnormal profits to the investors consistent to DeBondt and Thaler (1985). However the evidence is also consistent with the overreaction and partially with the behavioral hypothesis and risk, reasons behind the profitability of the contrarian strategies needs to be scrutinized by using new explanatory risk factors and hypothesis.

Contrarian strategies, momentum effects, winners-losers effect, overreaction, emerging markets

Abstract:
Previous evidence of US stock markets has shown that stocks included in (excluded from) index exhibit significant positive (negative) abnormal returns on the announcement day and trading volume is effected by the event. However the existing literature has not provided a unique explanation of the results. This is the first study that examines the price and volume effects on stocks associated with the changes in value-weighted index composition of two separate indices, ISE-100 and ISE-30, of the leading emerging market - Istanbul Stock Exchange in the period between years 1995 and 2000 and also in several sub-periods. 204 additions and 180 deletions in 24 quarterly periodical index revision intervals are analyzed by using event-study and standardized cross-sectional test methodology shown by Boehmer, Musumeci and Poulsen (1991).

Consistent to previous evidence of US stock markets, we find that stocks included in (excluded from) index, particularly for the ISE-30, tend to generate positive (negative) abnormal returns in the event period until effective change date and trading volume is effected by the event significantly. Price decreases after the change date both for included and excluded stocks. Behavior of stock prices has interesting patterns around the index revision period, however the statistical and economical significance of the cumulative abnormal returns are weak due to lack and weakness of index funds, institutional investors, mutual funds invest in stocks and non-existing of derivatives trading. Therefore, both the price pressure and imperfect substitutes hypotheses are unlikely explanations for the findings for ISE, consistent to the results of Beneish and Whaley (1995).

The effects of changes in index composition on stock market have been explored not only by cumulative abnormal returns but more strongly by the volume ratios. Abnormal volume behavior exists both for added and excluded stocks which are greater than one and strongly significant. There is significant increase in abnormal volume after announcement and on change date for excluded stocks, whereas it falls after effective change date. Volume variability of the added-deleted stocks indicates a significant increase after the announcement and then hits to the highest level on the effective change day. This evidence confirms our previous findings that market reacts to changes in index composition, particularly for the stocks subject to deletion from index. One factor in explaining this behavior might be the regulation of the ISE that requires only the stocks in ISE-100 can be subject to margin trading, lending and borrowing and also to short selling which causes the closing of margin trading accounts on deleted stocks after the deletion is effective on change date and also rejecting of those stocks as collateral by brokerage houses.

Decreasing price and volume for excluded stocks after announcement is also consistent with the information cost and liquidity explanation. Additionally, long-run abnormal volume in event period can not be attributed to the use of index stocks in arbitrage with derivative instruments as Lynch and Mendenhall (1997) did, since there is no derivatives trading in Turkish capital markets yet. Abnormal trading on or just prior to announcement day may partly be attributed to speculative trading based on the estimations of the possible included-excluded stocks that are spreaded out by rumors. This is consistent to our expectation that the revisions can be estimated by the traders since the criteria and the data used for index revisions are publicly available like Italian Stock Market shown by Rigamonti and Barontini (2000) recently.

Finally, all evidence show that changes in index composition does matter for the market participants and have impact on return and volume of stocks added to and particularly deleted from the index in Istanbul Stock Exchange. One implication for future research is that we expect that the effects of changes in index composition on stock market will be clear and stronger both statistically and economically in ISE in the future by the growth of index funds, institutional investors, mutual funds invest in stocks and also by existing of the derivatives trading in Turkish capital markets.

Stock index revisions, Index composition, Price and Volume effects

Abstract:
Various seasonal patterns in stock markets have been shown that disappeared, or at least substantially weakened, since they were first documented in 1980s. To detect whether calendar anomalies are still alive or not, this paper investigates all types of seasonalities such as the day-of-the week, turn-of-the-year and January, turn-of-the-month, intra-month, and holiday effects in stock returns as well as in trading volume in one of the leading emerging markets-Istanbul Stock Exchange (ISE).

Results indicate that calendar anomalies are still significantly existed in the ISE both in stock returns and trading volume consistent to international evidence. In addition, I find that a change in settlement period significantly effects the distribution of daily returns across the days of the week. And, low and negative Monday effect disappears when the return of the last trading day of the previous week and/or return of the previous week is positive. Moreover, another turn-of-the-month effect in the middle of the month (around 15th of each month) has been discovered which seems to be related to standardization in payment systems in public sector. Most of the anomalies seem to be related to the combination of various factors such as settlement procedures, window-dressing, information processing, inventory adjustments, risk, standardization in payment systems, regularities in dividend payments and earnings announcements, market closures, and measurement errors. Evidence also implies that regulators of markets and institutional practices have an important impact on the existence of the seasonalities in stock markets. However, no explanation yet is sufficient completely for this phenomenon by providing direct evidence. Finally, there is still necessity to explore alternative explanations and refine the existing factors that keep most of the calendar anomalies still alive in stock markets.

Market Efficiency, Anomalies, Seasonalities, Day-of-the-Week Effect, January Effect, Turn-of-the-Year Effect, Turn-of-the-Month Effect, Intra-Month Effect, Holiday Effect

Abstract:
Financial academics and practitioners have recognized that average stock returns are related to past performance and cross-section of stock returns is that predictable based on past returns. Number of researchers report that past losers (negative or lowest return-stocks) outperform past winners (positive or highest return-stocks) or vice versa over the subsequent three to five years not only in US markets but also in other stock markets.

This study examines the momentum and contrarian effects on stock returns in one of the leading emerging markets, Istanbul Stock Exchange between years 1991 and 2000 by using the same empirical methodology in Jegadeesh and Titman (1993). It also investigates the weak-form efficiency of the stock market by examining the profitability of a number of contrarian strategies based on past prices of stocks.

Prior loser-stocks are found to outperform prior winner-stocks consistent with the predictions of the overreaction hypothesis. Compounded annual return difference between the top-winners and top-losers is around 15% in favor of loser-stocks since the average return difference during the 10-year period is 1.14% per month. Empirical findings for the longer-term average returns such as 24 and 36 month holding periods reveal a reversal of returns from one to three years. On the other hand, we find that average abnormal returns per month and the average abnormal return difference between losers and winners increase as the holding period extends. Results also indicate that there is a downward trend in average returns based upon the length of past returns that is used in the strategy for the winner stocks but upward trend for the losers. The gap between losers and winners increases in favor of losers increases as goes back and the effect of length of past returns on holding period return seems to be more important than that of the length of holding period (K). Profitability of the strategies is robust to changes in the size of the portfolios. We also find that contrarian profits in January are significantly higher than those in non-January months, particularly for the strategies which are based on relatively shorter holding periods such as one and three months. Persistently, losers earn large January returns which are significantly higher than those of winners.

Finally, our results show that contrarian effects or more specifically 'winners and losers effect' are existing on stock returns in the Istanbul Stock Exchange and the contrarian trading strategies that is buying past losers and selling past winners realize significant abnormal profits to the investors. The evidence is consistent with the overreaction and partially with the behavioral hypotheses.

Momentum effects, contrarian strategies, winners-losers effect, overreaction, market efficiency, emerging markets

Abstract:
One of the interesting findings among the seasonalities in stock markets is that the return, volume and volatility of the stock prices and bid-ask spreads all broadly follow a U-shaped pattern over the trading day. This study examines the intra-daily seasonalities of the stock returns in the emerging Turkish Stock Market which is an order-driven market using electronic trading without market makers, in the period from 1996 to 1999, by using 15-minute (and also 5- and 1-minute) interval data. Results show that stock returns follow a U-shaped or more precisely a W-shaped pattern over the trading day at the Istanbul Stock Exchange (ISE) since there are two trading sessions in a day. This result is consistent with the previous findings in the literature. Opening (Overnight) and closing returns are significantly large and positive. Volatility is higher at the openings and follows an L-shape pattern during the both sessions. Interestingly, the daily average close-to-close returns are generated only during the opening and closing intervals and the average intra-day return is negative when the returns at the opening and/or closing intervals (even the first and the last minutes of the day) are excluded from the analyses. Relatively higher mean return and standard deviation at the openings of the trading sessions seem to be significantly generated by the accumulated overnight information and the closed-market effect (halt of trade). Large day-end returns are strongly affected by the activities of fund managers and speculators for the window-dressing around the close. Finally, intra-day seasonalities exist significantly also in the Turkish Stock Market as consistent with those of the international stock markets.

Anomalies, Seasonalities, Microstructure, Intraday Effects, Day-end Effect

Abstract:
One of the interesting findings among the seasonalities in stock markets is that the return, volume and volatility of the stock prices and bid-ask spreads all broadly follow a U-shaped pattern over the trading day. This study examines the intra-daily seasonalities of the stock returns in the emerging Turkish Stock Market which is an order-driven continuous auction market using electronic trading without market makers in the period from January 1, 1996 through January 15, 1999 by using 15-minute (and also 5 minute and 1 minute) interval data.

Results show that stock returns follow a U-shaped or more precisely a W-shaped pattern over the trading day at the Istanbul Stock Exchange (ISE) since there are two trading sessions in a day. This result is consistent with the previous findings in the literature. Opening (Overnight) and closing returns are significantly large and positive. In addition, volatility is higher at the openings and follows an L-shape pattern during the both sessions. Interestingly, the daily average close-to-close returns are generated only during the opening and closing intervals and the average intra-day return is negative when the returns at the opening and/or closing intervals (even the first and the last minutes of the day) are excluded from the analyses. Thus, the rest of the trading day provides no gains (losses) to close-to-close overall returns.

Findings suggest that relatively higher closing prices are not corrected by the market at the opening of the next trading day. Relatively higher mean return and standard deviation at the openings of the trading sessions seem to be significantly generated by the accumulated overnight information and the closed-market effect (halt of trade). Mondays have the highest opening mean return and volatility among the days of the week supports this explanation. On the other hand, large day-end returns are strongly affected by the activities of fund managers and speculators who, at the end of the day, boost their portfolios' asset value by bidding higher and accepting ask prices that result to higher closing prices with the contribution of relatively higher minimum tick sizes determined by the Stock Exchange.

Intra-day seasonalities that also exist significantly in the Turkish Stock Market, are consistent with those of the international stock markets. This conclusion implies that large profits can be realized by using a simple trading rule, based on the strong intra-day seasonalities in stock returns at the ISE, such as buying and selling stocks at a particular time of the day.

Anomalies, seasonalities, microstructure, intraday effects, day-end effect

Abstract:
In spite of the strong existence of price limits in financial markets, there is not much agreement and information on the effects of price limits on volatility and price discovery, which has important policy implications for the investors and regulators. This study examines the effects of price limits on stock return volatility by testing the overreaction and information hypotheses for the Istanbul Stock Exchange. We implement structural break tests as well as compherensive GARCH framework to estimate the impact of price limits on volatility, controlling for structural breaks, financial and economic crises, trading activity, and business cycle fluctuations. Our results do not support the information hypothesis. The fundamental conclusion of this paper is that the two-hour break between the two daily sessions reduces volatility by acting as a circuit breaker, which facilitates the dissemination of valuable information, thus preventing severe overreactions to news events, which are consistent with the overreaction hypothesis.

Price limits, volatility, emerging markets, ARCH-GARCH modelling

Abstract:
There has been much discussion among the regulators, investors and academics in policy circles to control the increasing volatility by using the price limits on financial markets. In spite of the strong existence of price limits worldwide, there is no much information regarding the effects of price limits on volatility and price discovery. Most of the previous studies find no evidence for the price limits that reduce the volatility.

This study examines the effects of price limits on volatility in stock returns through testing the overreaction and information hypotheses by using the same methodology in Phylaktis et al. (1999) for Athens Stock Exchange in one of the leading emerging markets - Istanbul Stock Exchange (ISE) in the period between years 1990 and 2001 for a larger sample. More specifically, we investigate the effects of increase in price limits on volatility in ISE in the period following the structural change in July 14, 1994 since daily cumulative price limit is doubled as a result of transition from one to two sessions in a trading day by using the econometric techniques such as serial correlation and GARCH models.

Our results do not support the information hypothesis in contrast to findings of Phylaktis et al. (1999). Serial correlation analysis gives us no strong evidence to reject or confirm the information hypothesis and inconclusive. Therefore, this inconclusive result motivates us to further analysis in future. GARCH estimation on daily and monthly stock returns controlling for structural breaks, financial and economic crises, trading activity and macroeconomic factors point out in a direction that volatility on stock returns has reduced despite the increase in daily price limits in the period following the structural change in ISE on 14 July, 1994. The majority of stock exhibit a negative and significant sign for the coefficient on the dummy variable. The results are robust to data frequency, leverage effect, financial crises and macroeconomic indicators.

In other words, with double sessions and despite broader implicit daily price limit ranges, volatility seems to decline. Findings imply that transition from one to two sessions in a trading day with lunch-break makes positive impact in reducing the volatility in a environment where the price limit is almost doubled. It seems that the two hour lunch-break between the daily sessions has the effect of a circuit breaker, thus facilitating the dissemination of information and preventing severe overreaction to news events which is consistent to overreaction hypothesis.

Finally, we find that volatility has decreased after the increase in price limits both for cross-section of stocks and overall index as well in ISE and thus price limits have no impact on volatility in stock market by the positive contribution of trading halt in the middle of the trading day.

Price limits, volatility, emerging markets, ARCH-GARCH modelling

Abstract:
This study examines the daily seasonalities in emerging Turkish Stock and Money Markets. Within this framework, day-of-the-week-effects in overnight interest rate changes both in the Central Bank Interbank Money Market and the Istanbul Stock Exchange (ISE) Repo Market as well as in stock returns and liquidity at the ISE Stock Market are investigated.

Results have shown the existence of significant day-of-the-week-effects both in overnight interest rate changes and stock returns. Overnight interest rates significantly fall on Wednesdays and increase on Mondays relative to the previous days. At the stock market, returns are significantly higher in the second part of the week and lower in the first two days of the week. Seasonalities in volatility and liquidity across the week confirmed the prior findings. Dynamic trading strategies based on the daily seasonalities in stock returns are able to produce significantly higher returns than market return of a simple "buy and hold" strategy in some periods neglecting transaction costs.

Some evidences are documented for the relationship between the existence of day effect in overnight interest rates and Treasury Bill Auctions, institutional practices and the other factors which effect the liquidity conditions of the market creating the seasonality in liquidity.

Findings support the "settlement procedure" explanation for the day-of-the-week effect in stock returns due to the fact that significant differences and movements in stock returns, return/risk, and liquidity indicators are observed between different periods in which different settlement procedures are employed in the stock market. On the other hand, negative Monday returns disappear when the mean return of previous Friday is positive.

Finally, these results have shown the existence of significant day-of-the-week-effects both in overnight interest rate changes and stock returns which is promising excess return for portfolio managers, and indicate that regulatory bodies of the markets, institutional practices and public authorities have an important creative impact on the seasonalities in stock and money markets.

Abstract:
This study examines the long-standing IPO puzzle in the Istanbul Stock Exchange (ISE) by using new factors affecting the performance of IPOs such as source of shares in the IPO (new issue or sale of large shareholders), allocation of shares and dispersion of investors (ownership of foreign investors, number of investors, and breadth), as well as existing factors such as market conditions (hot/cold), underwriters' reputation, and firm characteristics (issue and firm size, E/P, and B/M ratios).

Our results differ from previous studies at least three ways. First, magnitude of underpricing is significantly lower, while underperformance is higher than other studies. Our strong evidence supports the existence of the underpricing by positive initial excess returns (5.94%) and the long-term underperformance up to three-year holding period (-84.5%) in the ISE. We explain this with heavy competition among investment banks to mandate the IPOs in a market, where the number of IPOs is very limited, which leads to overvaluation and so underperformance in the long run, while limiting abnormal initial returns. Second, underperformance starts much earlier than in other markets, and has even been known to commence at the end of first month following the IPO because of myopic behavior of investors, who mostly seek short-term returns in a very volatile and inflationary environment. Third, we show that underperformance of IPOs disappears for IPOs made in a cold market, and those made through the sale of large shareholders. Allocation of shares in an IPO and size also impact after-market performance of shares. The broader and the more even allocation of shares, the smaller both the long-run underperformance and the underpricing are. Finally, a temporary large and positive initial return experienced by firms issuing stock to the public for the first time turns out to be hazardous to the wealth of their shareholders in the long-run. However, the mystery of these phenomena requires further investigation by researchers in the future until a complete and satisfying explanation is found.

Initial Public Offerings, underperformance, underpricing, market efficiency, emerging markets.

Abstract:
There has been considerable discussion in policy circles about controlling volatility by imposing price limits on asset prices. This study examines the effects of price limits on a stock market by testing volatility spillover, delayed price discovery, and trading interference hypotheses in a leading emerging market, Istanbul Stock Exchange (ISE), which has a unique microstructure with related to price limits. Our results support volatility spillover, delayed price discovery, and trading interference hypotheses. We also show price locks at limits measured by Volume-Weighted Average Prices provide significantly stronger evidence regarding the effects of price limits than measurement of limit moves only. Finally, price limits have a significant impact on stock market, casting doubt on their effectiveness in financial markets.

Price limits, volatility, microstructure, overreaction, price discovery

Abstract:
Firm-specific trading halts have become a common practice in many international stock markets during the last two decades. However, the effects and effectiveness of trading halts remain controversial among academics and regulators. In this debate, it seems crucial to understand how the trading behavior of institutional and individual investors, the market microstructure and the duration of the halts are related to the effects of the trading halts. By considering these factors, this paper assesses the efficiency of trading halts by examining the return, volatility and volume behavior around news-initiated trading halts through the unique microstructure and trade-by-trade data of the Istanbul Stock Exchange (ISE). It also investigates, for the first time, the trading behavior of different types of investors such as individuals, mutual funds and brokerage houses around trading halts.

Findings show that most of the new information is absorbed by prices within fifteen minutes (almost completely in an hour) following the resume of trading after a halt. Reaction of investors to bad news is slower and stronger than good news. Our results are robust to time-of-halt and duration-of-halt effects. Price discovery mechanisms based on fully computerized trading, non-existence of monopolist specialists and opening batch mechanisms, and restrictions on order cancellation during trading are some of the factors that accelerate the speed of adjustment in prices. In spite of halts, institutional investors would take the price advantage of new information during the halt period ahead of the individual investors by doing better timing in trading after halts. Institutional investors systematically buy and sell at more favorable prices around halts than individual investors do. Finally, overall evidence suggests that trading halts are effective in dissemination of valuable information and play an important role in enhancing the efficiency of the price discovery mechanism.

trading halts, microstructure, price discovery, regulatory effectiveness, institutional investors,

Abstract:
This study examines the momentum and contrarian effects on stock returns in one of the leading emerging markets, which has a unique market structure, with record-high inflation, high volatility, high turnover, low correlation of returns with other exchanges and myopic investors: the Istanbul Stock Exchange (ISE). It also investigates the weak-form efficiency of the stock market by examining the profitability of a number of contrarian strategies based on past returns, size, price, book-to-market and earnings-to-price ratios of stocks in various lengths of formation and holding periods. Our findings show that a self-financing trading strategy, buying past loser stocks and selling past winner stocks generate significant abnormal returns (approximately 15% annually) in ISE. However, these large contrarian profits are for bearing the extra risk of loser stocks similar to the US results. We also find significant price, size, and B/M effects in stock returns. Finally, our results show the continuous profitability of contrarian strategies both in very short (starting from 1 month) and in long holding periods (up to 36 months), which appears to be related to country-specific factors.