پيش بيني ورشكستگي و راهبري شركت ها: ديدگاه نسبت هاي مالي

Bankruptcy prediction and Corporate Governance: Financial Ratio Approach

ﭘﯿﺶﺑﯿﻨﯽ ورﺷﮑﺴﺘﮕﯽ در ﻣﻄﺎﻟﻌﺎت و ﻣﻘﺎﻻت ﻣﻮﺟﻮد در ﺣﻮز ﻫﺎي ﺣﺴﺎﺑﺪاري و ﻣﺪﯾﺮﯾﺖ ﺑﺴﯿﺎر ﻣﻮرد ﺑﺤﺚ واﻗﻊ ﺷﺪهاﺳﺖ و ﻣﻄﺎﻟﻌﺎت ﻓﺮاواﻧﯽ در راﺑﻄﻪ ﺑﺎ روشﻫﺎي ﺗﺠﺮﺑﯽ ﺑﻬﺘﺮ ﺑﺮاي ﭘﯿﺶﺑﯿﻨﯽ ورﺷﮑﺴﺘﮕﯽ اﻧﺠﺎم ﺷﺪهاﺳﺖ. ﻫﺪف ﺗﺤﻘﯿﻖ ﺣﺎﺿﺮ اﺳﺘﻔﺎده از ﻧﺴﺒﺖﻫﺎي ﻣﺎﻟﯽ و ﺷﺎﺧﺺﻫﺎي راﻫﺒﺮي ﺷﺮﮐﺘﯽ ﺑﺮاي ﭘﯿﺶﺑﯿﻨﯽ ورﺷﮑﺴﺘﮕﯽ ﺷﺮﮐﺖﻫﺎي ﭘﺬﯾﺮﻓﺘﻪ ﺷﺪه در ﺑﻮرس اوراق ﺑﻬﺎدار ﺗﻬﺮان اﺳﺖ. ﭘﮋوﻫﺶ ﺣﺎﺿﺮ ﺑﻪ ﻟﺤﺎظ ﻫﺪف، ﺑﻨﯿﺎدي و از ﻧﻈﺮ روش ﺗﺤﻘﯿﻖ ﺗﻮﺻﯿﻔﯽ از ﻧﻮع ﻫﻤﺒﺴﺘﮕﯽ ﻣﯽﺑﺎﺷﺪ. در ﭘﮋوﻫﺶ ﺣﺎﺿﺮ ورﺷﮑﺴﺘﮕﯽ ﺷﺮﮐﺖﻫﺎ ﺑﻪ ﻋﻨﻮان ﻣﺘﻐﯿﺮ واﺑﺴﺘﻪ و ﺗﻌﺪاد40 ﺷﺎﺧﺺ در ﭘﯿﺶﺑﯿﻨﯽ ورﺷﮑﺴﺘﮕﯽ در دو ﮔﺮوه 31 ﺗﺎﯾﯽ ﻧﺴﺒﺖﻫﺎي ﻣﺎﻟﯽ و 9 ﺗﺎﯾﯽ ﺷﺎﺧﺺﻫﺎي راﻫﺒﺮي ﺷﺮﮐﺖ ﺑﻪ ﻋﻨﻮان ﻣﺘﻐﯿﺮ ﻣﺴﺘﻘﻞ ﻣﻮرد اﺳﺘﻔﺎده ﻗﺮار ﮔﺮﻓﺘﻪﺷﺪهاﺳﺖ. در اﯾﻦ ﭘﮋوﻫﺶ ﻧﺴﺒﺖ ﺑﻪ ﻣﻘﺎﯾﺴﻪ ﭼﻬﺎر ﻣﺪل ﭘﯿﺶﺑﯿﻨﯽ ﻣﻌﺮوف ﺷﺎﻣﻞ ﻣﺪل ﻣﺎﺷﯿﻦﺑﺮدار، ﺷﺒﮑﻪﻫﺎي ﻋﺼﺒﯽ ﻣﺼﻨﻮﻋﯽ، ﺷﺒﮑﻪ ﻫﺎي ﻋﺼﺒﯽ ﻣﺼﻨﻮﻋﯽ ﺑﻬﯿﻨﻪﺳﺎزي ﺷﺪه ﺑﺎ اﻟﮕﻮرﯾﺘﻢ ژﻧﺘﯿﮏ و رﮔﺮﺳﯿﻮن ﻻﺟﯿﺖ اﻗﺪام ﺷﺪه اﺳﺖ ﮐﻪ ﻧﻬﺎﯾﺘﺎ ﺷﺒﮑﻪ ﻋﺼﺒﯽ ﻣﺼﻨﻮﻋﯽ ﺑﻬﯿﻨﻪﺷﺪه ﺑﺎ اﻟﮕﻮرﯾﺘﻢ ژﻧﺘﯿﮏ ﺑﻬﺘﺮﯾﻦ ﮐﺎراﯾﯽ را ﻧﺴﺒﺖ ﺑﻪ ﺳﺎﯾﺮ ﻣﺪلﻫﺎ از ﺧﻮد ﻧﺸﺎن داد. ﻫﻤﭽﻨﯿﻦ ﺑﺎ ﻣﻘﺎﯾﺴﻪ وﯾﮋﮔﯽ ﻧﺴﺒﺖﻫﺎي ﻣﺎﻟﯽ و ﺷﺎﺧﺺﻫﺎي راﻫﺒﺮي، ﻧﺴﺒﺖﻫﺎي ﻣﺎﻟﯽ ﺧﻮد را ﺑﻪ ﻋﻨﻮان وﯾﮋﮔﯽﻫﺎي ﺗﺎﺛﯿﺮﮔﺬار و ارزﺷﻤﻨﺪﺗﺮي ﺑﺮاي ﭘﯿﺶﺑﯿﻨﯽ ورﺷﮑﺴﺘﮕﯽ ﻧﺸﺎندادﻧﺪ. ﭼﻨﺎﻧﭽﻪ دﻗﺖ ﺗﺨﻤﯿﻦ ﺑﻪ ازاي ﻧﺴﺒﺖﻫﺎي ﻣﺎﻟﯽ در ﺑﺎﻻﺗﺮﯾﻦ ﺳﻄﺢ ﺧﻮد ﻗﺮار دارﻧﺪ ﮐﻪ در ﭘﺎﯾﺎن ﻣﯽﺗﻮان ﻧﺘﯿﺠﻪﮔﺮﻓﺖ ﮐﻪ ﺑﻬﺘﺮﯾﻦ ﻣﺪل ﺑﺮاي ﭘﯿﺶﺑﯿﻨﯽ ورﺷﮑﺴﺘﮕﯽ ﻣﺪل اﺳﺘﻔﺎده از ﻧﺴﺒﺖﻫﺎي ﻣﺎﻟﯽ در ﺷﺒﮑﻪ ﻋﺼﺒﯽ ﻣﺼﻨﻮﻋﯽ ﺑﻬﯿﻨﻪﺷﺪه ﺑﺎ اﻟﮕﻮرﯾﺘﻢ ژﻧﺘﯿﮏ ﻣﯽ ﺑﺎﺷﺪ. اﯾﻦ اﻟﮕﻮرﯾﺘﻢ ﺑﯿﺸﺘﺮﯾﻦ دﻗﺖ را ﺑﺪﺳﺖ آوردهاﺳﺖ و ﺧﻄﺎي آن ﮐﻤﯿﻨﻪ اﺳﺖ و ﻣﯽﺗﻮان آن را ﺑﻌﻨﻮان ﯾﮏ ﻣﺪل ﻗﺎﺑﻞ اﻋﺘﻤﺎد، ﭘﺎﯾﺪار و ﻋﻤﻠﯽ در ﻧﻈﺮ ﮔﺮﻓﺖ.

Bankruptcy prediction in studies and articles in the areas of Accounting and Management are discussed and many studies on the experimental method is more effective for bankruptcy prediction was carried out. The aim of this study is to compare the financial and indicators of corporate governance for bankruptcy prediction of companies listed on Tehran stock exchange. As the sample were selected variables into two categories that bankruptcy as the dependent variable and the number of 40 indicators or factors affecting predicted the crisis or financial distress in two groups of 31 rats financial ratios and 9-indices corporate governance as an independent variable used is taken. In this study we compare the 4 methods famous prediction models vector machines, artificial neural networks, artificial neural networks optimized by genetic algorithm and logit regression action. Which ultimately artificial neural network optimized by the genetic algorithm works best compared to other models showed. It also has a feature comparison ratios and financial indices of governance, Ratios your finances as characteristics of effective and valuable for predicting bankruptcy showed. The precision of the estimates for properties Ratios Financial is the highest level. At the end it can be concluded that the best model for bankruptcy prediction is the use of ratios financial artificial neural network optimized algorithms Genetics is. This algorithm has the highest accuracy achieved and error is minimal. Therefore it could make it as a model of reliable, sustainable and practical.