本研究以Panel Data模型探討台灣中小企業信用風險，依企業產業別分別建立新發生逾期比率模型、違約機率模型二套迴歸模型，檢定總體經濟及金融環境相關重要指標對於中小企業違約情形之間的關係。實證結果分述二點為：1.新發生逾期比率模型解釋能力較沈中華、黃博怡（2006）研究台灣中小企業產業別信用風險之檢定，明顯不足，僅對私部門放款金額年增率及退票張數比例年增率有顯著影響，亦不同於沈中華、黃博怡強調公司之ROA、資產週轉率以及GDP成長率、通貨膨脹率之影響因子。2.違約機率模型，解釋能力較新發生逾期比率模型來的高，其中景氣同時指標綜合指數、工業生產指數、對私部門放款金額、消費者物價指數，其年增率與違約機率皆呈現負向關係，年增率越大代表景氣越佳，違約機率越低。另銀行牌告基準放款利率年增率、退票張數比例年增率與違約機率皆呈現正向關係，當放款利率增加對於企業融資成本將提高，導致企業違約機率增加，而退票張數的增加也代表企業在財務上出現問題，違約機率自然攀升。 This paper applies the Panel Data models to describe credit risk of SMEs in Taiwan. We establish two sets of regression models：rate of default model and probability of default model by industry sectors. In addition, this study also investigates the relationships among SMEs default, economic and financial environment factors. According to our results, we found that first, the rate of default model’s results are worse than Shen and Huang （2006）which stress ROA, Asset Turnover, GDP growth rate and Inflation rate . The influence of annual growth rate of loan amount and dishonored ratio of checks & bills - ratio in number are significant. Second, the result of probability of default model’s are better than the rate of default model. As the annual growth rate of coincident indicators, industrial production index, loan amount and consumer prices increase, the probability of default declines. Besides, as the annual growth rate of prime lending rates and dishonored ratio of checks & bills - ratio in number increase the probability of default increase too.
