Study, Based On Indicators Of The Health Of The Financial Early Warning System

The 2004 annual reports of the listed companies in China showed that the total loss of the 20companies which performed worst has reached 45.35% of the loss of overall listed companies,while the percentage in 2003 was 52.79%. Since 2001, a series financial scandals like Enronand WorldCom has accelerated the legislation of the Sarbanes-Oxley (SOX) Act whichindicates that the US government are trying to prevent and obviate the financial crisis throughthe reinforcement of internal control. However, it is still far from enough without a perfectfinancial crisis prediction system which provides the financial controlling system and investmentstrategy with adequate acuity and reactivity.On the basis of the existing theories, this dissertation argued about four points of view asfollows: firstly, this study focuses on the integrative indexes which reflect the relations betweenevery two statements and build up an overall financial health inspection system. Secondly, thisdissertation builds up a financial crisis prediction system in different year before the crisis whichis another way of resisting the affect of the "financial number game". Thirdly, the crisiscompanies defined in this study are not only the companies which named by ST but also the 50worst performed companies. Fourthly, this dissertation makes it clear that the sample crisiscompanies could not experience any financial crisis as identified before until the supposedcrisis year.31 crisis companies and 31 non-crisis companies are respectively picked out from themanufacturing industry. It is found that the prediction indexes are not up to the hypothesis ofnormal distribution. So the nonparametric tests are adopted and it is found that 9 indexes in 11ones possess significant prediction abilities. The single variable models in this dissertationshow quite a good prediction effect but lack of stability and adaptability. Then, the multipleprediction models are built up by the step wise regression method. And finally, the dissertationmakes a conclusion as follows: in the year before the crisis, the logistic multiple modelcomposed by AZ (percentage of the other account receivable in the current assets), BD (ratioof the operational cash inflow to the outflow) and BC (percentage of the cash inflow from sellingproducts and providing services in the total operational cash inflows) shows an errordifferentiation ratio at 12.90%. While in two year before the crisis, the model composed by AZ,BC and BB (current liability coverage ratio by the operational cash) shows an errordifferentiation ratio at 14.52%. It is obvious that the model shows a satisfactory effect in the twoyears before the financial crisis.