Scribd
Upload
54 views7 pages

Course: Multivariate Data Analysis and Forecasting Techniques Course Coordinator: Dr. PK Das Assignment On Case Analysis

Principal component analysis (PCA) and cluster analysis were used to identify groups of financial ratios that convey similar information. PCA identified 8 components accounting for 90% of variation. Box plots showed significant differences in medians and intervals for ratios measuring solvency, liquidity, profitability, and leverage between bankrupt and non-bankrupt firms. Specifically, lower values of ratios like cash/debt, current ratio, cash flow/assets, and assets/debts increased the probability of bankruptcy.

Uploaded by

Vaibhav Bajpai
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
54 views7 pages

Course: Multivariate Data Analysis and Forecasting Techniques Course Coordinator: Dr. PK Das Assignment On Case Analysis

Principal component analysis (PCA) and cluster analysis were used to identify groups of financial ratios that convey similar information. PCA identified 8 components accounting for 90% of variation. Box plots showed significant differences in medians and intervals for ratios measuring solvency, liquidity, profitability, and leverage between bankrupt and non-bankrupt firms. Specifically, lower values of ratios like cash/debt, current ratio, cash flow/assets, and assets/debts increased the probability of bankruptcy.

Uploaded by

Vaibhav Bajpai
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 7

Indian Institute of Foreign Trade

New Delhi | Kolkata


Course: Multivariate Data Analysis and Forecasting Techniques
Course Coordinator: Dr. PK Das
Assignment on Case Analysis
Case Study: Predicting Corporate Bankruptcy

Group 8: Abhishek Batni – 57KB;

Overview
The aim of bankruptcy prediction is to help the enterprise stakeholders to get the comprehensive
information of the enterprise. Much bankruptcy prediction has relied on statistical models and
got low prediction accuracy. However, with the advent of the AI (Artificial Intelligence),
machine learning methods have been extensively used in many industries (e.g., medical,
archaeological and so on).

A1. Big data analytics technique(s) that would be appropriate in assessing whether there are
groups of variables/ratios that convey the same information and how important that information
is are: -
 Principal Component Analysis(PCA)
 Factor Analysis
 Cluster Analysis
Here, in the case we have performed to PCA to arrive at such groups or componennts
summary(PCA)

As shown in total 8 components are account for reducing 90% variation


plot(PCA, type="l")
biplot(PCA)

The above figure gives the classifications or groups for the ratios which have similar loadings for
the two major components as shown. These set of ratios in each groups have similar effects on
the firms and can be studied for profiling of a firm as bankrupt or not.
A2. Distinct goals of profiling the characteristics of bankrupt firms is to predict the possibility of
bankruptcy using financial ratios. The models hence derived were based on classical statistical
classification techniques such
 linear discriminant analysis (LDA),
 logistic regression
 probit model
 These methods gained great popularity because of simplicity of the interpretations.The price
paid for this simplicity, is possible misrepresentation of the real relation between the predictor
variables and the response, which is likely nonlinear and too complex to be expressed
analytically.
 New methodologies for bankruptcy prediction are based on machine learning and artificial
intelligence methods. These algorithms are simply predicting (black box style) by design-
oriented more toward prediction performance using a set of financial ratios. Among the most
popular of these new classification models are
 artificial neural networks (NN)
 support vector machines (SVM)
 different ensemble methods
 deep learning
 To obtain high prediction performance, it is necessary to allow for a flexible (non-parametric)
relation between the predictors and the response. Therefore, behind the NN, SVM, deep
learning and ensembles classification success, there is a multivariate nonlinear complex
discriminant function that is estimated nonparametrically.
 This flexibility, however, entails a difficult interpretation of the model results. These black box
predictive algorithms are not able to explain the mechanism by which the predictors affect the
mean response. Consequently, although they work very well to predict when bankruptcy will
occur, they fail to account for the cause.
The point of departure for our comparative analysis is the practical finding that black-box-type
data-mining methods typically show better classification performance than models whose results
are easy to interpret, i.e. there seems to be a kind of trade-off between the interpretability and
predictive power of bankruptcy models. Most of the efforts are done on extracting
comprehensible rules from the black box type of function estimated in a machine learning
algorithm. However, these methodologies in general only represent an approximation of
the original model and therefore might lose accuracy
List of 132 firms

A3. Comparing Box plots for each ratios between bankrupt and non-bankrupt firms
We used the below code for each ratio and captured box plot snippets and merged them.
#Box plot for R1, similar code for other ratios changing column values
boxplot(PCA_Data[1:66,1:1],PCA_Data[66:132,1:1])
Here 1 – Bankrupt 2 – Non-Bankrupt
From the above box plots, the following ratios are observed with significant difference between
the medians and the intervals of bankrupt firms and non-bankrupt firms,
R4, R6, R9, R13, R14, R15, R16, R17, R19, R20, R21, R22, R23, R24

Solvency Ratio
R4 – Cash/Debt This ratio measures the financial strength of a company, if it is smaller than 1,
it means the company has more debt than the cash in hands.

Liquidity ratios:
1) R9 – Current Ratio = Current Assets/Current Debt
2) R23 – Working Capital/Assets
When liquidity ratios decrease, the probability of bankruptcy increases confirming that these
ratios are signals of financial health of the firms.

Profitability ratios:
1) R6 – Cash Flow/Assets
2) R14 – Income/Assets
3) R15 – Income/Debt
The probability of bankruptcy increases when the profitability ratios decrease.

Leverage ratio inverse


R21 – Assets/Debts
A low ratio of total assets to total debts indicates that a firm with high indebtedness increases
its probability of bankruptcy.
So, a firm with reduced liquidity, low profitability or excessive leverage has a high
probability of bankruptcy.
Decrease in profitability (net income to sales) and efficiency (sales to total assets) have an
increasing effect on bankruptcy’s probability.
1) R13 – Income/Sales
2) R20 – Sales/Assets
Similarly, lower values of the following ratios indicate a high probability of bankruptcy.
1) R16 - Income plus Depreciation/Sales
2) R17 - Income plus Depreciation/Assets
3) R19 – Sales/Receivables
4) R22 – Working Capital/Sales
5) R24 – Working Capital/Debts

Key Takeaways:
 Excess of liquidity is also not good.
 Low proportions of current assets or quick assets could also decrease the probability of
bankruptcy, if the number is too high, it could mean the company is not leveraging its
assets.
 Both high current liabilities and sales, have a decreasing effect.
 Very low levels of money, quick assets or total assets increase the probability of
bankruptcy.

You might also like