Predicting the survival or failure of click-and-mortar corporations: A knowledge discovery approach

doi:10.1016/j.ejor.2005.05.009

European Journal of Operational Research

Volume 174, Issue 2, 16 October 2006, Pages 959-982

https://doi.org/10.1016/j.ejor.2005.05.009 Get rights and content

Abstract

With the boom in e-business, several corporations have emerged in the late 1990s that have primarily conducted their business through the Internet and the Web. They have come to be known as the dotcoms or click-and-mortar corporations. The success of these companies has been short lived. This research is an investigation of the burst of the dotcom bubble from a financial perspective. Data from the financial statements of several survived and failed dotcom companies is used to compute financial ratios, which are analyzed using three classification techniques—discriminant analysis, neural networks, and support vector machines to find out whether they can predict the financial fate of companies. Neural networks perform the task better than other techniques. Using discriminant analysis and neural networks, the key financial ratios that play a major role in the process of prediction are identified. Statistical tests are conducted to validate the findings.

Introduction

The late 1990s have witnessed a phenomenal revolution—the growth of the Internet based companies. In a short span of about 10 years, the Internet has taken the form of an alternate ‘shopping mall’. Companies have started selling merchandise on the Web ranging from pet supplies to garden tools to cosmetics. Some of these companies have a physical presence in retailing like Barnes and Nobles, Wal-Mart, etc. These are the brick-and-mortar corporations. Many of these companies also have online subsidiaries, which produce their own financial reports separate from the parent company. There are others who are essentially pure play Internet and only have an online existence. These are known as the dotcoms or the click-and-mortar corporations, examples of which are Amazon.com, expedia.com, etc. As the number of Internet users across the world increases at a steady rate, retailing over the Internet continues to gain popularity. It is stated in a report released by Jupiter Research Corporation, (Jupiter Research Corporation, 2003) the online retail sector will experience an average annual growth rate of 21% between 2002 and 2007. The report also states that, by 2007, more than 5% of US retail sales will be transacted online. On a similar vein, Forrester Research has also predicted that e-commerce sales, growing at a steady rate of 19% per year, will increase to $229.0 billion in 2008 from $95.7 billion in 2003, with the total number of online shopping households in the US reaching 63 million by 2008 (Rush, 2003).

However, the phenomenal growth in Internet related e-tailing has suffered a major setback in the early years of the new millennium. In 2001, Forrester Research had indicated that weak financial strength, increased competition, and investor flight would drive most click-and-mortar companies out of business by 2001 (Grenier, 2003). At the same time, the Gartner Group had predicted that as many as 95% of all click-and-mortar companies will fail by 2002 (Cane, 2000). Sure enough, the prediction of the financial pundits has come true and the dotcom bubble has burst. Within a very short time, several corporations that have seen phenomenal growth in their stock prices in the late 1990s have gone out of business. In 2000, when CNNfn.com had asked the market data and research firm Birinyi Associates of Westport, Connecticut, to calculate the market value of the stocks in the Bloomberg US Internet Index, they had found that the combined market values of all stocks had fallen to $1.193 trillion from $2.948 trillion at their peak, sustaining a loss of $1.755 trillion in a span of only seven months (Kleinbard, 2000).

Several factors are thought to be responsible for this demise and as indicated by Sharma (2001), these have included among others, the dotcoms’ inability to improve revenues and earnings, failure to post-profits, attempt to capture a major market share in the smallest possible time, and tendency to operate in limited geographical areas. The objective of this research is to conduct a rigorous study of the financial statements of dotcom companies and to discover the factors responsible for the survival or attrition of these companies. Classification techniques are used to find out how good financial ratios are for predicting performance of click-and-mortar corporations and to identify those financial ratios that are most predictive of financial health of click-and-mortar companies based on historical financial data.

The results of this research may offer directives to emerging companies that plan to do business on the Internet in the coming years. It will be able to provide guidelines about which financial factors to monitor closely for long-term competitive advantage in the market. It will be extremely useful for new as well as seasoned investors who plan to invest in similar companies. It can act as a guide to investors on what factors to look for to ascertain the financial health of a click-and-mortar company before they make the decision to purchase stocks of such a corporation.

Section snippets

Literature review

Due to the importance of predicting the financial health of an organization, this is a widely researched area. Different approaches and techniques have been used for forecasting the likelihood of failures as indicated by Wong et al. (2000) and it is quite common to use financial ratios. The interest is this area has been spurred by Altman (1968) with the use of multiple discriminant analysis with five financial ratios, to predict the risk of failure. Another notable example of the use of

Research methodology

In this paper the click-and-mortar corporations and online subsidiaries of brick-and-mortar corporations are studied to find out whether they will survive or cease to exist. We follow an approach based on knowledge discovery to conduct this study. The process of knowledge discovery can be divided into four distinct steps. They are shown in Fig. 1(a), which is adapted from the literature (Piatetsky-Shapiro and Frawley, 1991; Fayyad et al., 1996; Cios et al., 1998). The path of research followed

Numerical experimentation

The first set of experiments involves the use of the DA technique for the classification problem. DA is conducted using SAS System for Windows Version 8.2. DA models can be set up easily and can be applied quickly on the testing data for classification. Initially, all 24 financial ratios are considered as input variables. Initial experimentation indicates that better results are likely using a subset of the 24 input variables. Hence, stepwise DA is used to identify the subset of input

Discussion of results

From the stated numerical results in Section 4, it can be concluded that in addition to the method of data analysis, the four factors—training to testing ratio, size of the training sample, size of the testing sample, and balance of sample play an important role in determining the performance of a model. Accordingly, an ANOVA experiment is conducted to study which of these factors are statistically significant in determining the total testing classification accuracy for the ‘best’ model. A

Conclusion

In this paper DA, NN and SVM are used to find out whether it is possible to predict the survival or failure of click-and-mortar corporations based on financial ratios. On the whole, NN seems to perform better than the other methods but a hybrid model that consists of initial choice of financial ratios by DA and is followed by analysis using NN yields the best result (77.5%). This indicates that hybrid systems can improve performance in classification in some situations. It is observed that in

Acknowledgments

The authors want to thank the two anonymous reviewers whose comments have significantly improved the quality of the paper. Thanks are due to Mr. James Pang for his help with the numerical experiments. The first author gratefully acknowledges the support received as a seed grant from The University of Hong Kong and as the Central Earmarked Research Grant from the Research Grants Council of Hong Kong. The second author acknowledges support received as Faculty Research Grant from College of

References (55)

E.I. Altman
The success of business failure prediction models: An international survey
Journal of Banking and Finance
(1984)
A.I. Dimitras et al.
Business failure prediction using rough sets
European Journal of Operational Research
(1999)
A.I. Dimitras et al.
A survey of business failures with an emphasis on prediction methods and industrial applications
European Journal of Operational Research
(1996)
D. Fletcher et al.
Forecasting with neural networks: An application using bankruptcy data
Information and Management
(1993)
R.C. Lacher et al.
A neural network for classifying the financial health of a firm
European Journal of Operational Research
(1995)
M. Leshno et al.
Neural network prediction analysis: The bankruptcy case
Neurocomputing
(1996)
M. Luoma et al.
Survival analysis as a tool for company failure prediction
Omega
(1991)
S. Piramuthu et al.
A classification approach using multi-layered neural networks
Decision Support Systems
(1994)
J. Scott
The probability of bankruptcy: A comparison of empirical predictions and theoretical models
Journal of Banking and Finance
(1981)
F.E.H. Tay et al.
Application of support vector machines in financial time series forecasting
Omega
(2001)

R. Tsaih et al.

Forecasting S&P 500 stock index futures with a hybrid AI system

Decision Support Systems

(1998)

R.L. Wilson et al.

Bankruptcy prediction using neural networks

Decision Support Systems

(1994)

B.K. Wong et al.

A bibliography of neural network business applications research: 1994–1998

Computers and Operations Research

(2000)

Y. Alici

Neural networks in corporate failure prediction: The UK experience

E.I. Altman

Financial ratios, discriminant analysis and the prediction of corporate bankruptcy

The Journal of Finance

(1968)

C.J.C. Burges

A tutorial on support vector machines for pattern recognition

Data Mining and Knowledge Discovery

(1998)

L.J. Cao et al.

Financial forecasting using support vector machines

Neural Computing Applications

(2001)

Cane, E., 2000. Online laughing stocks: Betting on dotcom failures takes off. (Downloadable from website:...

Chang, C.-C., Lin, C.-J. 2001. LIBSVM: A library for support vector machines. (Downloadable from website:...

K.J. Cios et al.

Data Mining Methods for Knowledge Discovery

(1998)

Dahlberg, T., Horluck, J. 2004. Internet hype overreaction—and what we can learn from it. Working paper. Department of...

E.B. Deakin

Discriminant analysis of predictors of business failure

Journal of Accounting Research

(1976)

R.A. Eisenbeis

Pitfalls in the application of discriminant analysis in business and economics

The Journal of Finance

(1977)

Fan, A., Palaniswami, M., 2000. Selecting bankruptcy predictors using a support vector machine approach. In:...

U. Fayyad et al.

From data mining to knowledge discovery in databases

AI Magazine

(1996)

H. Frydman et al.

Introducing recursive partitioning for financial classification: The case of financial distress

The Journal of Finance

(1985)

Grenier, M.P., 2003. High-tech downturn is no surprise to startups. (Downloadable from website:...

Cited by (36)

Survival of e-commerce entrepreneurs: The importance of brick-and-click and internationalization strategies
2021, Electronic Commerce Research and Applications
Citation Excerpt :
Our research thus aims to help entrepreneurs and investors in their decisions by identifying the factors that increase the survival rate. Most studies examine e-commerce success, whereas few deal with survival (Banerjee et al., 2007; Bose and Pal, 2006; Janda and Moreira, 2017; Kauffman and Wang, 2008; Nikolaeva, 2007; Sung, 2006). Some of the determining factors that explain short-term survival may not serve to ensure long-term survival; in this sense, a goal of short-term profitability can be drawn from the investments required to ensure future growth (Kogut and Zander, 1992).
E-commerce is a fast-growing industry that attracts many entrepreneurs; however, the survival rate is lower than that of other industries. Entrepreneurs take many strategic decisions that have a significant bearing on the success of their e-commerce ventures. Two are particularly salient for our research: Is it better to start as a pure-click or open a physical store? To what extent should e-commerce ventures internationalize? We also wonder if it is worth reviewing the first annual accounts available. Grounding on organizational ecology theory, we develop a model hypothesizing that (1) the brick-and-click strategy favours survival; (2) internationalization favours survival; (3) firm size and financial health matter; (4) earnings management mediates the relationship between financial health and survival; and (5) the brick-and-click and internationalization strategies mediate the relationship between size and survival. The empirical study is performed by analysing seven years of data on 632 new e-commerce ventures using non-parametric means tests, a Cox regression, and a generalized structural equation model. The study tests the strategies followed by e-commerce entrepreneurs and firm characteristics that influence survival. We find that the risk of bankruptcy is 1.437 times greater for pure-click than for brick-and-click retailers, 2.778 times greater for local players than for internationalized firms, and 1.787 times greater for unprofitable firms than for profitable firms. Hence, it is worth analysing the financial statements provided by entrepreneurs; however, signs of earnings management should be checked. All these factors affect the probability of early bankruptcy as well as explain survival several years later.
Deep learning models for bankruptcy prediction using textual disclosures
2019, European Journal of Operational Research
Citation Excerpt :
Given the high impact of corporate bankruptcy events, researchers in operational research (OR) and artificial intelligence (AI) further propose intelligent models to forecast bankruptcy. New modeling techniques include boosting, discriminant analysis, support vector machine, and neural networks (Alfaro, García, Gámez, & Elizondo, 2008; Bose & Pal, 2006; Geng, Bose, & Chen, 2015), data envelopment analysis (Li, Crook, & Andreeva, 2014), least absolute shrinkage and selection operator (Tian, Yu, & Guo, 2015), dynamic slacks based model (Wanke, Barros, & Faria, 2015), two-stage classification (du Jardin, 2016), to name a few. A common element of those models is the application of market-based and accounting-based variables, which are usually constructed using numeric data in a well-structured format.
This study introduces deep learning models for corporate bankruptcy forecasting using textual disclosures. Although textual data are common, it is rarely considered in the financial decision support models. Deep learning uses layers of neural networks to extract features from textual data for prediction. We construct a comprehensive bankruptcy database of 11,827 U.S. public companies and show that deep learning models yield superior prediction performance in forecasting bankruptcy using textual disclosures. When textual data are used in conjunction with traditional accounting-based ratio and market-based variables, deep learning models can further improve the prediction accuracy. We also investigate the effectiveness of two deep learning architectures. Interestingly, our empirical results show that simpler models such as averaging embedding are more effective than convolutional neural networks. Our results provide the first large-sample evidence for the predictive power of textual disclosures.
Prediction of financial distress: An empirical study of listed Chinese companies using data mining
2015, European Journal of Operational Research
The deterioration in profitability of listed companies not only threatens the interests of the enterprise and internal staff, but also makes investors face significant financial loss. It is important to establish an effective early warning system for prediction of financial crisis for better corporate governance. This paper studies the phenomenon of financial distress for 107 Chinese companies that received the label ‘special treatment’ from 2001 to 2008 by the Shanghai Stock Exchange and the Shenzhen Stock Exchange. We use data mining techniques to build financial distress warning models based on 31 financial indicators and three different time windows by comparing these 107 firms to a control group of firms. We observe that the performance of neural networks is more accurate than other classifiers, such as decision trees and support vector machines, as well as an ensemble of multiple classifiers combined using majority voting. An important contribution of the paper is to discover that financial indicators, such as net profit margin of total assets, return on total assets, earnings per share, and cash flow per share, play an important role in prediction of deterioration in profitability. This paper provides a suitable method for prediction of financial distress for listed companies in China.
Predicting financial distress and corporate failure: A review from the state-of-the-art definitions, modeling, sampling, and featuring approaches
2014, Knowledge-Based Systems
As a hot topic, financial distress prediction (FDP), or called as corporate failure prediction, bankruptcy prediction, acts as an important role in decision-making of various areas, including: accounting, finance, business, and engineering. Since academic research on FDP has gone on for nearly eighty years, there are abundant literatures on this topic, which may appear chaotic to the researchers of the field and make them feel confused. This paper contributes to the current review researches by making a full summary, analysis and evaluation on the current literatures of FDP. The current literatures of FDP are reviewed from the following four unique aspects: definition of financial distress in the new century, FDP modeling, sampling approaches for FDP, and featuring approaches for FDP. By considering the new state-of-the-art techniques in this area, FDP modeling are classified and reviewed by the following groups: namely, modeling with pure single classifier, modeling with hybrid single classifier, modeling by ensemble techniques, dynamic FDP modeling, and modeling with group decision-making techniques. Sampling methods for FDP are classified and reviewed by the following paired groups, namely: training sampling and testing sampling, single industry sampling and cross-industry sampling, balanced sampling and imbalanced sampling. Featuring methods for FDP are categorized and reviewed by qualitative selection and combination of qualitative and quantitative selection. We comment on the current researches from the view of each category and propose further research topics. The review paper is valuable to guide research and application of the area.
Forecasting business failure: The use of nearest-neighbour support vectors and correcting imbalanced samples - Evidence from the Chinese hotel industry
2012, Tourism Management
Previous studies on firm failure prediction (FFP) have chiefly addressed predictions based on balanced datasets without considering that the real-world target population consists of imbalanced data. The current study investigates tourism FFP based on the imbalanced data of Chinese listed companies in the hotel industry. The imbalanced dataset was collected and represented in terms of significant financial ratios, and a new up-sampling approach and forecasting method were proposed to correct imbalanced samples. To balance the imbalanced dataset, the up-sampling method generates new minority samples according to random percentage distances from each minority sample to its nearest neighbour (NN). The NNs of unlabelled samples are retrieved from the balanced dataset to produce a knowledge base of nearest-neighbour support vectors, from which base support vector machines (SVMs) are generated and assembled. Empirical results indicate that the proposed sampling approach helped models produce more accurate performance on minority samples, with accuracy rates in excess of 90 per cent. This method of using nearest-neighbour support vectors and correcting imbalanced samples is useful in controlling risk in tourism management.
Dynamic financial distress prediction using instance selection for the disposal of concept drift
2011, Expert Systems with Applications
Prior studies of financial distress prediction (FDP) all focus on static modeling and ignore whether the model is still suitable with time passing on. This paper devotes to the first investigation on what the concept of financial distress concept drift (FDCD) is, whether FDCD exists and how to dispose FDCD. We construct a dynamic FDP modeling based on instance selection for the disposal of FDCD. Dynamic FDP consists of instance selection, FDP modeling and future prediction. Instance selection methods including full memory window, no memory window, window of fixed size, window of adaptable size, and batch selection are used to tackle FDCD. For feature selection, we construct a wrapper by integrating forward and backward selections on Mahalanobis distance. Empirical results indicate that gradual and constant virtual concept drift does exist in FDP, and dynamic FDP models perform much better than static models. Meanwhile, window of fixed size and batch selection are more suitable for Chinese listed companies’ dynamic FDP.

View all citing articles on Scopus

Indranil Bose is an Associate Professor of Information Systems at School of Business, The University of Hong Kong. His degrees include B.Tech. (Hons.) in Electrical Engineering from the Indian Institute of Technology, M.S. in Electrical and Computer Engineering from University of Iowa, M.S. in Industrial Engineering and Ph.D. in Management Information Systems from Purdue University. He has research interests in telecommunications and networking, data mining and artificial intelligence, electronic commerce, applied operations research, and supply chain management. His teaching interests are in telecommunications, database management, systems analysis and design, and data mining. His publications have appeared in Communications of the ACM, Computers and Operations Research, Decision Support Systems, Ergonomics, European Journal of Operational Research, Information and Management, Operations Research Letters, and in the proceedings of numerous international and national conferences.

Raktim Pal is an Assistant Professor in the Department of Computer Information Systems and Management Science at the College of Business, James Madison University. He has more than five years of industry experience in the area of supply chain management and logistics. He holds a Ph.D. in Transportation Systems from Purdue University. Also, he has M.S. in Civil Engineering and M.S. in Industrial Engineering from Purdue University, and B.Tech. (Hons.) in Civil Engineering from the Indian Institute of Technology. His research interests are in the areas of supply chain management and logistics, electronic commerce, transportation systems modeling, and applied operations research. His teaching interests are in supply chain management, operations management, management science, and statistics. His publications have appeared in Computers and Operations Research, ASCE Journal of Transportation Engineering, and Transportation Research Record.

¹: Tel.: +1 540 568 7094; fax: +1 540 568 3273.

View full text

Computing, Artificial Intelligence and Information ManagementPredicting the survival or failure of click-and-mortar corporations: A knowledge discovery approach

Abstract

Introduction

Section snippets

Literature review

Research methodology

Numerical experimentation

Discussion of results

Conclusion

Acknowledgments

Journal of Banking and Finance

European Journal of Operational Research

European Journal of Operational Research

Information and Management

European Journal of Operational Research

Neurocomputing

Omega

Decision Support Systems

Journal of Banking and Finance

Omega

Decision Support Systems

Decision Support Systems

Computers and Operations Research

Neural networks in corporate failure prediction: The UK experience

Financial ratios, discriminant analysis and the prediction of corporate bankruptcy

The Journal of Finance

A tutorial on support vector machines for pattern recognition

Data Mining and Knowledge Discovery

Financial forecasting using support vector machines

Neural Computing Applications

Data Mining Methods for Knowledge Discovery

Discriminant analysis of predictors of business failure

Journal of Accounting Research

Pitfalls in the application of discriminant analysis in business and economics

The Journal of Finance

From data mining to knowledge discovery in databases

AI Magazine

Introducing recursive partitioning for financial classification: The case of financial distress

The Journal of Finance

Computing, Artificial Intelligence and Information Management
Predicting the survival or failure of click-and-mortar corporations: A knowledge discovery approach