استفاده از مدل های کمک تصمیم گیری چندمعیاری در پیش بینی خرید مجدد سهم بازار آزاد
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|14451||2012||9 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 8424 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
|شرح||تعرفه ترجمه||زمان تحویل||جمع هزینه|
|ترجمه تخصصی - سرعت عادی||هر کلمه 90 تومان||13 روز بعد از پرداخت||758,160 تومان|
|ترجمه تخصصی - سرعت فوری||هر کلمه 180 تومان||7 روز بعد از پرداخت||1,516,320 تومان|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Omega, Volume 40, Issue 6, December 2012, Pages 882–890
This study presents the first attempt to develop classification models for the prediction of share repurchase announcements using multicriteria decision aid (MCDA) techniques. We use three samples consisting of 434 UK firms, 330 French firms, and 296 German firms, to develop country-specific models. The MCDA techniques that are applied for the development of the models are the UTilités Additives DIScriminantes (UTADIS) and the ELimination and Choice Expressing REality (ELECTRE) TRI. We adopt a 10-fold cross validation approach, a re-sampling technique that allows us to split the datasets in training and validation sub-samples. Thus, at the first stage of the analysis the aim is the development of a model capable of reproducing the classification of the firms considered in the training samples. Once this stage is completed, the model can be used for the classification of new firms not included in the training samples (i.e. validation stage). The results show that both MCDA models achieve quite satisfactory classification accuracies in the validation sample and they outperform both logistic regression and chance predictions. The developed models could provide the basis for a decision tool for various stakeholders such as managers, shareholders, and investment analysts.
The last two decades have witnessed a dramatic increase in the use of share repurchases. For example, as Grullon and Michaely  highlight, expenditures on share repurchase programs (relative to total earnings) increased from 4.8% in 1980 to 41.8% in 2000, while more recent data from Standard and Poor's show that share repurchases among companies that comprise the S&P 500 reached a record 172 billion US dollars during the third quarter of 2007. In the EU-15, the value of share repurchases of industrial companies increased from 6.15 billion Euros in 1989 to 58.84 billion Euros in 2005, with their value over the entire period reaching 252.94 billion Euros . Given the growth in the importance and popularity of share repurchases, it is not surprising that this topic has attracted considerable attention in the literature, with numerous studies examining the short-and long-run valuation effects  and  as well as the determinants and motives of share repurchases  and . In the present paper we deviate from existing studies by proposing the application of multicriteria decision aid (MCDA) techniques for the prediction of firms' announcements of open market share repurchases.1 While past studies have employed MCDA techniques in other finance and accounting problems such as bankruptcy prediction, mergers and acquisitions, auditing, etc. with promising results ,  and , there is a lack of studies focusing on share repurchases announcements, and we aim to close this gap in the literature.2 The development of such a model is necessary because it is not possible to use models built for other important business events (e.g. bankruptcy) or to draw any conclusions from their application. There are two reasons for this. First, the decision makers (e.g. analysts, investors, etc.) have different objectives, and the models are built with different goals in mind. Second, different business events are being driven by different factors and theoretical reasoning, and as such the underlying variables (criteria) also differ. As we discuss in more detail in Section 4, a model with the ability to predict share repurchases could have practical implications for various decision makers (e.g. existing shareholders, prospective investors and peer firm managers), and especially for investment managers who could use it as the basis for an investment strategy. While some studies have tried to explain the determinants of share repurchases , to the best of our knowledge, up to date only Andriosopoulos and Hoque  test the out-of-sample prediction accuracy of their model using logistic regression.3 However, the MCDA methods proposed in the present study pose various advantages over traditional statistical and econometric methods such as discriminant analysis and logistic regression. For example: (i) they do not make any assumptions about the normality of the variables or the group dispersion matrices, (ii) they are not sensitive to multicollinearity or outliers, (iii) they can easily incorporate qualitative data, and (iv) they are also very flexible in terms of incorporating any preferences of the decision maker. We use a sample of 530 open market share repurchases that were announced in France, Germany and the UK between 1997 and 2006 and an equally matched control group. There are a number of reasons for which we focus on these three countries. First, they are the three largest economies in the EU, in terms of GDP, number of listed companies, etc. Therefore, some of the largest and most important European firms operate in these three countries. Second, data from von Eije and Megginson  indicate that over the period 1989–2005, these three countries accounted for a combined 76.16% of the total value of share repurchases by industrial firms in the EU-15 (UK: 49.38%, France: 19.95%, Germany: 6.82%). Thus, our study provides an extensive coverage in terms of open market share repurchases in the EU. Third, there are important differences between these countries. For instance, the majority of UK firms are widely held companies whereas France and Germany have a more concentrated ownership structure . Hence, differences in the level of shareholder protection can potentially lead to different managerial attitudes towards shareholder value maximization. Consequently, this could result in different attitudes on firms' cash utilization and the choice of firm payout decisions. For example, in France firms tend to be family owned, and in Germany firms are less widely held than UK firms. Thus, it is likely that share repurchases in Germany, and especially in France, would not be such a popular payout mechanism as it is in the UK. Moreover, managers have different attitudes and priorities in different countries regarding the management of their firms. For instance, Brounen et al.  find that shareholder wealth maximization is one of the most significant priorities for managers in the UK. In contrast, managers in Germany and France place more emphasis on other factors such as leverage optimization. Finally, the magnitude of the market reaction to the announcement of the intention to repurchase shares in the open market differs significantly among these countries  and . Thus, differences in the operating environment can have a significant impact on the markets' perception and reaction to such announcements, as well as the managerial incentives and implications for making such announcements. Consequently, the simultaneous application of the MCDA techniques in these three countries, allows us to test their usefulness in different institutional and regulatory settings, and in countries with potentially different managerial attitudes.4 We develop two MCDA models for each country, using the UTilités Additives DIScriminantes (UTADIS) and ELimination and Choice Expressing REality (ELECTRE) TRI methods. These two methods use different modeling forms (i.e. value functions and outranking relations), thus enabling the investigation of the generalizing ability of different MCDA models in the prediction of share repurchases. For benchmarking purposes we compare the classification accuracies of the MCDA models with the ones obtained by logistic regression. Thus, we develop a total of nine models. All the models are estimated and tested using a ten-fold cross-validation approach. Our results show that the MCDA models classify correctly around 70% of the firms in the validation sample, and they outperform logistic regression in all the cases. The remaining of this research study is organized as follows. Section 2 presents the data, variables and methodology. Section 3 provides a discussion of the empirical results. Section 4 discusses the practical usefulness of the developed models, along with some differences between the UTADIS and the ELECTRE TRI models. The conclusions are in Section 5.
نتیجه گیری انگلیسی
This research study contributes to the literature by providing an analysis of the ability of MCDA techniques to predict the likelihood of an open market share repurchase announcement. To examine the effectiveness of the models, we used three samples consisting of 434 UK firms, 330 French firms, and 296 German firms, half of which announced a share repurchase between 1997 and 2006. The models were developed using UTADIS and ELECTRE TRI, through a ten-fold cross-validation approach. Logistic regression was also employed for benchmarking purposes. The variables were selected on the basis of theoretical reasons and past studies in the repurchasing literature. To account for differences across countries we developed country specific models. Thus, three models for each country are developed, resulting in a total of 9 models. Our results indicated that the firm characteristics that can be useful in discriminating between the two groups of firms vary among countries which may be related to country-specific attributes that influence the managerial decisions with regards to share repurchases. We also found that these may differ across the methods used to develop the models. However, this is not surprising and it has been the case in past studies from other disciplines as well (e.g. prediction of acquisitions, bankruptcy prediction, etc.). For example, firm size appeared to be the most important variable in the three models developed through the ELECTRE TRI method as well as in the UTADIS-UK model, while at the same time it was one of the most important variables in the UTADIS models developed for Germany and France. In contrast, ROA was quite important in the UTADIS-UK model, while being considerably less important in the remaining models. As it concerns the classification ability of the models, the average results over the 10 replications in the validation set illustrated that all the models achieve quite balanced accuracies between the two groups and they performed better than a naïve model based on random assignment to outcomes based on prior probabilities (i.e. 50% in an equal sample). The highest overall accuracy among all the three sample countries was achieved in France. In this case, UTADIS managed to classify correctly approximately 8 out of the 10 firms, a performance that was slightly better than that of ELECTRE TRI. In contrast, in the case of the UK, it was ELECTRE TRI that performed slightly better than UTADIS. In general, the lowest accuracies were observed in the case of Germany, with the MCDA models achieving quite similar accuracies. The satisfactory performance of the MCDA models in the validation dataset illustrates that they could be used for the classification of any new firm not included in the training sample. For example, the developed models could be of use to various decision makers such as investment managers, firm managers, and stockholders, by providing objective information that can be prove useful in an initial screening of the firms. This could result in important savings in terms of time and money. Future research could extend the present study towards various directions such as the testing of the usefulness of the models in other countries, the employment of and comparison with alternative methods (i.e. support vector machines, neural networks, etc.), and the combination of MCDA and other methods into integrated models. It could also consider the use of non-financial variables (e.g. corporate governance) and the development of decision support systems.