A Neural Network Approach to Estimating the Allowance for Bad Debt

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Fifty-two variables were used, and only 8 variables were not used.

“The study indicates that fifty two variables (see above) were used.Only 8 variables were not used (see Table 3).”

The quote only partly states this claim.

There was no statistically significant difference between NeuroXL Predictor and SPSS residuals in the paired t-test on residuals.

“There was no statistically significant difference between NeuroXL Predictor residuals.”

May combine two claims.

In a second two-sample t-test on absolute differences, NeuroXL Predictor had M=1.4860, SD=2.448 and SPSS had M=2.023, SD=3.486, t(80)=-1.5687, p=.0603 one-tailed, and the null was rejected at the 95% confidence interval.

“NeuroXL Predictor absolute differences (M = 1.4860,SD = 2.448) and SPSS absolute differences (M = 2.023, SD = 3.486), t(80) = -1.5687.The p-value was .0603(one-tailed).”

• Also stated in: § 6.2.1 Prediction Accuracy · plain

  “Since the test statistic was within the reject region, the null was rejected at this significance level.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

Both NeuroXL Clusterizer and NeuroXL Predictor proved to be user-friendly applications that worked seamlessly with Microsoft Excel.

“Both NeuroXL Clusterizer and NeuroXl Predictor proved to be user friendly applications that worked seamlessly with Microsoft Excel.”

The quote only partly states this claim.

In one instance, NeuroXL Predictor provided a more accurate estimate than SPSS regression.

“In this situation, NeuroXL Predictor provided the more accurate estimate.”

May combine two claims.

Bad debt expense and write-off amounts were frequently missing from the WRDS data, leaving the data incomplete.

“The amounts for these accounts were frequently omitted in the WRDS data.Consequently the data was incomplete.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Many predictions from NeuroXL Predictor and SPSS were close.

“A graph of the absolute values derived by NeuroXL Predictor and SPSS demonstrated many of the predictions were close (see Exhibit IX).”

Probably not a substantial claim: may not pass S3 (load-bearing).

The ANN software packages used in this study are only one brand.

“The ANN packages used in this study are just one brand of ANN software.”

May combine two claims.

In a paired t-test on percentage differences, NeuroXL Predictor had M=.4180, SD=.3912 and SPSS had M=.5538, SD=.6033, t(77)=-2.2037, p=.01523 one-tailed, and the null was rejected at the 95% confidence interval.

“There was a statistically significant difference between NeuroXL Predictor percentage differences (M = .4180,SD = .3912)and SPSS Percentage Differences (M = .5538,SD = .6033),t(77) = -2.2037.The p-value was .01523(one-tailed).”

• Also stated in: § 6.2.1 Prediction Accuracy · plain

  “Therefore, for this test, the null hypothesis was rejected at the 95% confidence interval.”

Probably not a substantial claim: may not pass S3 (load-bearing).

NeuroXL Predictor is unable to provide the specific weights it assigns to different independent variables.

“However, like other ANNs, NeuroXL Predictor is unable to provide the specific weights it assigns to different independent variables.”

May combine two claims.

Out of 85 comparisons, NeuroXL Predictor output was more accurate 47 times compared to SPSS, which was more accurate 38 times.

“Out of 85 comparisons, the NeuroXL Predictor output was more accurate 47 times compared to SPSS, which was more accurate 38 times.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

The variables selected by SPSS were easily discernible.

“The variables selected by SPSS were easily discernible (see Table 2).”

Probably not a substantial claim: may not pass S3 (load-bearing).

As the complexity of the event being analyzed increases, training time for backwards propagation increases non-linearly.

“First, as the complexity of the event being analyzed increases, the training time increases non-linearly.”

• cites Costea and Nastac, 2005 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Credit scoring models have been used for decades to classify customers as good or bad credit risks.

“In addition, Zurada and Kunene (2010) also note that credit scoring models are not new phenomena.They have been used for decades to group customers into two categories: good credit and bad credit.”

• cites Zurada and Kunene, 2010Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

Because the allowance estimation process is subjective, artificial neural networks are useful for checking the reliability of management's estimates.

“However, given the subjective nature of the process, ANNs are beneficial for purposes of verifying the reliability of management estimates.”

• AssumptionRests on an assumption: The items incorporated into GAAP-based allowance calculations are very subjective.

May combine two claims.

Artificial neural networks generally outperform conventional techniques because they use a non-linear rather than a linear approach.

“Overall, ANNs perform better than conventional techniques, because these networks employ a non-linear approach as opposed to a linear approach.”

• cites Atiya, 2001Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Certegy's system employs artificial neural network architecture, traceable to United States Patent 6,647,376.

“The origin of Certegy's system can be traced to United States Patent 6,647,376, which clearly shows the system employs ANN architecture.”

• cites United States Patent 6,647,376Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

Artificial neural networks that use backwards propagation often rely heavily on initial starting values.

“Often, ANNs which use BP, rely heavily on initial starting values.”

• cites Costea and Nastac, 2005 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Artificial neural networks use two parameters: learning rate and momentum.

“ANNs use two parameters: learning rate and momentum.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Artificial neural networks are repeatedly tested until an acceptable accuracy level of 80 to 90 percent is reached.

“Ragothaman, et al. (2008) explains that ANNs are subjected to repeat testing until an acceptable level of accuracy is achieved (80 to 90 percent).”

• cites Ragothaman, et al. (2008) sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

There is very little consensus among auditors about evaluating loan collectability.

“There is very little consensus among auditors regarding the evaluation of loan collectability.”

• cites Wright and Willingham (1997)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S2 (checkable).

The importance of accurately evaluating credit risk from financial institutions' loan granting decisions cannot be underestimated.

“Zurada and Kunene (2010, p. 1) note that “accurately evaluating the credit risk posed by financial institutions loan granting decisions cannot be underestimated.””

• cites Zurada and Kunene, 2010Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim. May combine two claims.

The main advantages of neural network classification over traditional approaches are freedom from distributional assumptions, universal approximation, elimination of intercorrelated data problems, and mapping from inputs to outputs without prior knowledge of the function form.

“Costea and Nastac (2005, p. 230) stated that “the main advantages of neural network approaches for classification over the traditional approaches are that ANNs are free of any distributional assumptions, are universal approximations, eliminate problems with intercorrelated data, and provide a mapping function from the input to the outputs without any a priori knowledge about the function form.””

• cites Costea and Nastac, 2005 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Little research has been performed on auditor evaluation of loan portfolios.

“Wright and Willingham (1997) indicated that little research has been performed in the area of auditor evaluation of loan portfolios.”

• cites Wright and Willingham (1997)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

Artificial neural networks can learn by adjusting the weights assigned to relationships based on observed historical patterns.

“ANNs are able to 'learn' because they can adjust the weights assigned to certain relationships based on observed historical patterns.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Auditors use their knowledge and education to evaluate loan collectability using variables such as collateral strength, borrower financial condition, short-term liquidity, business risk, and financial risk.

“Auditors use their knowledge and education to evaluate loan collectability using variables such as: strength of and access to collateral, borrower's financial condition, expected degree of short-term liquidity, expected degree of business risk, and expected degree of financial risk.”

• cites Wright and Willingham (1997)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

Preprocessing the input data can reduce problems associated with backwards propagation.

“As a way to reduce these problems Costea and Nastac (2005, p. 231) suggest preprocessing the input data.”

• cites Costea and Nastac, 2005 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Modern databases may contain a billion records.

“However, this is no longer the case as modern databases may contain billion records.”

• cites Hand and Henley, 1997 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Artificial neural networks are advantageous because they can accurately predict default probabilities with minimal error.

“Hall, et al. (2009) concluded that ANNs are advantageous because with minimal error they are able to accurately predict default probabilities.”

• cites Hall, et al., 2009 sourceNot checked — no full text of the source

Classified as probably not a substantial claim. May combine two claims.

Artificial neural networks are more robust and reliable than conventional analytical procedures.

“Costea and Nastac (2005) find ANNs to be more robust and reliable than conventional analytical procedures.”

• cites Costea and Nastac (2005) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

Applying a rational explanation to an artificial neural network's decision can be very difficult or impossible.

“Since an ANN learns as it is presented more information, and since a computer is able to make faster and stronger statistical inferences than the human brain, applying a rational explanation to an ANN's decision can be very difficult, if not impossible.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The bad debt allowance amount should be based on Generally Accepted Accounting Principles and will involve estimates and subjective judgment.

“All that is known is that: 1) the amount should be based on GAAP and 2) the amount will involve estimates and subjective judgment.”

• cites GAAPNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

In the past, data sets were relatively small, numerical, and clean, which allowed statistics to produce straightforward answers.

“Hand and Henley (1997) noted that in the past, data sets were relatively small, numerical and clean.This enabled the use of statistics to derive relatively straightforward answers.”

• cites Hand and Henley, 1997 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Auditors classify loans as doubtful imprecisely.

“They noted the manner in which auditors classify loans as doubtful is imprecise.”

• cites Wright and Willingham (1997)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Artificial neural networks are capable of both supervised and unsupervised learning.

“Ragothaman, et al. (2008) indicated that ANNs are capable of both supervised and unsupervised learning.”

• cites Ragothaman, et al. (2008) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Approaches to classification can be divided into statistical approaches and human knowledge processing approaches that rely on artificial intelligence.

“Kim, et al. (1993, p. 167) stated that “the various approaches to the classification problem can be grouped into two paradigms: statistical approaches based on statistical assumptions and mathematical processes, and human knowledge processing approaches which rely on artificial intelligence techniques.””

• cites Kim, et al., 1993 sourceNot checked — no full text of the source

Classified as probably not a substantial claim. May combine two claims.

A good credit customer is likely to repay debt, while a bad credit customer is likely to default.

“A good credit customer is likely to repay the debt whereas a bad credit customer is likely to default.”

• cites Zurada and Kunene, 2010Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

Backwards propagation is the most popular artificial neural network method.

“They also note backwards propagation (BP) is the most popular ANN method.”

• cites Costea and Nastac, 2005 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Predictive data mining has two goals: discovering hidden relationships and patterns in data, and building usable prediction models.

“Costea and Nastac (2005) believe predictive data mining has two aims.The first aim is to uncover hidden relationships and patterns in the data.The second aim is to construct usable prediction models.”

• cites Costea and Nastac, 2005 sourceNot checked — no full text of the source

May combine two claims.

A low ratio of allowance for bad debt to bad debt write-offs indicates that the allowance was underestimated, while a high ratio indicates that too much allowance was accumulated.

“Riley and Pasewark (2009, p. 41) state that if the ratio is low, it means that the allowance for bad debt is underestimated, while a high ratio indicates that too much allowance for bad debt was accumulated.”

• cites Riley and Pasewark, 2009Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

If backwards propagation is applied to a fairly simple problem, the training process may take longer.

“Second, if backwards propagation is applied to a fairly simple problem, the training process may take longer.”

• cites Costea and Nastac, 2005 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Even a one percent improvement in the accuracy of a credit scoring model can save millions of dollars in a large loan portfolio.

“Zurada and Kunene (2010, p. 2) underscore the importance of a good credit scoring model: “even a 1 percent improvement in the accuracy of a credit scoring model can save millions of dollars in a large loan portfolio.””

• cites Zurada and Kunene, 2010Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

Artificial neural networks focus on input-output pairs to determine whether generalizable patterns exist.

“Swingler (1996) noted that ANNs focus on input-output pairs in order to determine if generalizable patterns exist.”

• cites Swingler (1996)Not checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The allowance method for bad debt can be estimated in three ways under Generally Accepted Accounting Principles.

“The allowance method for bad debt can be estimated in three different ways according to Generally Accepted Accounting Principles (GAAP).”

• cites GAAPNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

The allowance for bad debt must be estimated using methods prescribed by Generally Accepted Accounting Principles.

“It should be noted that the allowance for bad debt must be estimated using the methods outlined by Generally Accepted Accounting Principles (GAAP).”

• cites GAAPNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

Artificial neural networks are well suited to situations where the structure of the data is poorly understood.

“They are “well suited to situations where we have a poor understanding of that data structure” (Hand and Henley, 1997, p. 12).”

• cites Hand and Henley, 1997 sourceNot checked — no full text of the source

May combine two claims.

Artificial neural networks offer strong pattern recognition capabilities and can examine a company's financial and operating data to indicate possible hazards.

“ANNs offer “tremendous pattern recognition capabilities” and can focus on the financial and operating data of a company and indicate areas where hazards may exist.”

• cites Ramamoorti and Traver (1998)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

An artificial neural network constructs a classification model by identifying existing patterns in input data.

“According to Green and Choi (1997, p. 16) an ANN (artificial neural network) “builds a classification model by finding any existing patterns of input data.””

• cites Green and Choi, 1997 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Artificial neural networks belong to the category of artificial intelligence.

“According to Kim, et al. (1993), ANNs fall into the category of artificial intelligence.”

• cites Kim, et al., 1993 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Auditors examine loan information in a top-down, theory-driven manner.

“Wright and Willingham (1997, p. 105), stated auditors examine loan information “in a top-down (theory-driven) fashion.””

• cites Wright and Willingham (1997)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

Altman (1968) indicated that an overall bankruptcy assessment should be based upon a combination of measures.

“As a result, Altman (1968) indicated that an overall bankruptcy assessment should be based upon a combination of measures.”

• cites Altman (1968) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Setting weights to very small numbers forces the artificial neural network to analyze each variable with no predefined bias.

“If the weights are set to very small numbers, the ANN is forced to analyze each variable with no predefined bias.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The most frequently used method to estimate the allowance for bad debt is the aging of accounts receivable.

“The most frequently used method is the aging of accounts receivable.”

• cites Riley and Pasewark (2009)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

In Fanning and Cogger's study, the artificial neural network produced results that were 63 percent accurate, while none of the other models produced results greater than 50 percent accuracy.

“The findings show that the ANN produced results which were 63 percent accurate, while none of the other models produced results greater than 50 percent accuracy.”

• cites Fanning and Cogger's (1998) sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Exogenous variables such as unemployment rates could reasonably be expected to affect amounts collected.

“This may not be realistic, as exogenous variables, such as unemployment rates, could reasonably be expected to have an effect on amounts collected.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The major difference between artificial neural network results and a statistically-based multiple regression approach is the process in which weights are assigned to the variables.

“However, according to Coakley and Brown (1999), the major difference between ANN results and a statistically-based multiple regression approach is the process in which weights are assigned to the variables.”

• cites Coakley and Brown (1999)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

Accounts receivable and default rates on credit are items managers would want to assess, and executive dashboards would presumably incorporate alerts relating to those issues.

“Since accounts receivable and default rates on credit are items that managers would want to assess the design of executive dashboards would presumably incorporate alerts relating to those issues.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The heart of any dashboard includes revenue forecasts and accounts receivable, both of which have a great effect on allowance for bad debt.

“This is pertinent to the current research as cited by Kirkland (2006, p. 2), the “heart of any dashboard” includes items such as “revenue forecasts” and “accounts receivable,” both of which have a great effect on allowance for bad debt.”

• cites Kirkland (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

In Feroz et al. (2000), the logistic regression model had an accuracy rate of 52 percent while the artificial neural network model had an accuracy rate of 70 percent.

“In comparing the effectiveness of the two models, the logistic regression model had an accuracy rate of 52 percent while the accuracy of the prediction of the ANN model was 70 percent.”

• cites Feroz, et al. (2000) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

A major advantage of an artificial neural network over the human brain is that an artificial neural network can analyze and recognize patterns in large sets of data that would overwhelm a human brain.

“The major advantage an ANN holds over the human brain is that an ANN is capable of analyzing and recognizing patterns in large sets of data.”

• Also stated in: § 3.1 Adaptive Learning Algorithms · hedged

  “The enormity of such large sets of data would overwhelm a human brain.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Information systems can be classified into two paradigms: behavioral and design science.

“Information systems can be classified into two paradigms.”

• Also stated in: § 3.2 Design Science · plain

  “The first information systems paradigm is behavioral based.”

• Also stated in: § 3.2 Design Science · plain

  “The second paradigm is a design science approach.”

• cites Hevner, et al. (2004) sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Credit scoring models are based on linear discriminant analysis.

“West (2000) stated these credit scoring models are based on linear discriminant analysis.”

• cites West (2000) sourceNot checked — no full text of the source

May combine two claims.

The allowance aging method suffers from two weaknesses: the estimate is based on endogenous variables only, and the estimates are ultimately a result of judgment and vulnerable to earnings management manipulation.

“However the allowance aging method suffers from two weaknesses.”

• Also stated in: § 2.3.2 Traditional Approach -Percent of Sales · plain

  “First, the estimate of bad debt is based on endogenous variables only.”

• Also stated in: § 2.3.2 Traditional Approach -Percent of Sales · plain

  “Secondly, the estimates are ultimately a result of judgment and therefore vulnerable to earnings management manipulation.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Artificial neural networks are concurrent; inputs are re-evaluated as new output data is acquired.

“Coakley and Brown (1999) indicate that ANNs are concurrent.”

• Also stated in: § 2.2.2 · plain

  “The recurrent function of ANNs means that inputs are re-evaluated as new output data is acquired.”

• cites Coakley and Brown (1999)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

In a comparison using data from 92 companies and validation on 37, an artificial neural network accurately classified 79 percent of the holdout companies while linear discriminant analysis accurately classified 33 percent.

“Their results showed the ANN is a better predictor and they were able to accurately classify 79 percent of the 37 companies from the holdout sample, while the linear discriminant analysis model was only able to accurately classify 33 percent of the 37 companies.”

• cites Kuldeep and Bhattacharya (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S1 (asserted), S2 (checkable), S3 (load-bearing).

Artificial neural networks would benefit if macroeconomic variables such as interest rates were incorporated into the model.

“Third, they feel that ANNs would benefit if macroeconomic variables such as interest rates were incorporated into the model.”

• cites Kuldeep and Bhattacharya (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Management can easily adjust both the bad debt rate applied to credit sales and the rates applied to an aging of receivables.

“Management can easily adjust both the bad debt rate applied to credit sales as well as the rates applied to an aging of receivables.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

A general rule of thumb is that no layer should have more than five elements; more elements make it harder for the artificial neural network to generalize.

“According to Aragones, et al. (2007) a general rule of thumb is that no layer should have more than five elements.”

• Also stated in: § The ANN method used by Aragones, et al. (2007) is back-propagation (see Exhibit I). · hedged

  “If more were introduced, it would make it harder for the ANN to generalize.”

• cites Aragones, et al. (2007) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Low weights may lead to a slower learning rate but have the benefit of forcing the artificial neural network to consider each independent variable equally.

“Thus, while low weights may lead to a slower learning rate, it has the benefit of forcing the ANN to consider each independent variable equally.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A lower momentum level may be chosen to achieve greater accuracy because it entails more analysis by the artificial neural network.

“A lower momentum level may be chosen in order to achieve greater accuracy because it entails more analysis on the part of the ANN.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Because no judicial injunction has been issued against Certegy and its use of an artificial neural network, using artificial neural networks in credit granting is feasible.

“Thus, since no judicial injunction has been issued against Certegy and its use of an ANN, it can be inferred that usage of ANNs in credit granting is feasible (see Appendix A).”

May combine two claims.

The human brain can discern patterns but cannot process massive amounts of data.

“Ramamoorti and Traver (1998) noted while the human brain is able to discern patterns, it is unable to process the massive amounts of data.”

• cites Ramamoorti and Traver (1998)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

More iteration does not necessarily guarantee better results and less iteration does not necessarily entail poorer results.

“More iteration does not necessarily guarantee better results and less iteration does not necessarily entail poorer results.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

A back-propagation network is good at unsupervised learning, adapting, and generalizing data.

“According to Ding, et al. (2011) a back-propagation network is good at unsupervised learning, adapting, and generalizing data.”

• cites Ding, et al. (2011) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Artificial neural networks are good classifiers but tend to perform better in the short run than in the long run.

“First, while ANNs are good classifiers of items, they tend to perform better in the short run than in the long run.”

• cites Kuldeep and Bhattacharya (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Altman (1968) suggested multiple discriminant analysis as the solution, seeing its major advantage as analyzing the entire variable profile simultaneously rather than sequentially.

“This led him, after careful consideration, to suggest multiple discriminate analyses as the solution.”

• Also stated in: § 2.3.3 Cluster Analysis and Regression Analysis · plain

  “He saw the major advantage of this method as having “the potential of analyzing the entire variable profile of the object simultaneously rather than sequentially examining its individual characteristics.””

• cites Altman (1968) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Using multiple discriminant analysis, Altman (1968) obtained a relevant amount of accuracy in bankruptcy prediction.

“Using multiple discriminant analysis, Altman (1968) was able to obtain a relevant amount of accuracy in bankruptcy prediction.”

• cites Altman (1968) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Artificial neural networks are not static but keep integrating new variables and interactively adjust weights until the error is minimized.

“ANNs are not static, but keep integrating new variables.”

• Also stated in: § 2.3.3 Cluster Analysis and Regression Analysis · plain

  “Unlike the static regression model, where the equation is established and relied upon, an ANN interactively adjusts weights until the error is minimized.”

• cites Coakley and Brown (1999)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

Setting initial weights relatively high could bias the manner in which the artificial neural network analyzes the data.

“If one chooses to set initial weights relatively high, it could bias the manner in which the ANN analyzes the data.”

Probably not a substantial claim: may not pass S2 (checkable).

The speed, accuracy, and capability of computers to handle massive amounts of data provide researchers with a valuation tool for analyzing credit risk.

“Thus, the speed, accuracy, and capability of computers to handle massive amounts of data provide researchers with a valuation tool for analyzing credit risk.”

Classified as probably not a substantial claim.

Artificial neural networks are better at predicting bankruptcy than other models because they are better equipped to deal with missing variables, multi co-linearity, and outliers.

“Second, Kuldeep and Bhattacharya (2006) acknowledge that ANNs are better at predicting bankruptcy than other models because they are better equipped to deal with missing variables, multi co-linearity, and outliers.”

• cites Kuldeep and Bhattacharya (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim. May combine two claims.

If the properties of the data and the distributional assumptions required by a parametric model differ, this indicates non-linearity; if non-linearity exists in the data, use of an artificial neural network is warranted.

“Additionally, Coakley and Brown (1999) note, if the properties of the data and the distributional assumptions required by a parametric model differ, then this is an indication of non-linearity.”

• Also stated in: § 2.2.2 · plain

  “If non-linearity exists in the data, then use of an ANN is warranted.”

• cites Coakley and Brown (1999)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

While artificial neural networks provide more accurate results than traditional statistical models, they offer little rational explanatory capability.

“The authors acknowledge that while ANNs provide more accurate results than traditional statistical models, it offers little rational explanatory capabilities.”

• cites Quek, et al. (2009) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Many companies have large allowances for bad debt reserves, enabling them to absorb very large default rates without a related drop in income.

“Jackson and Liu (2007) found many companies have large allowances for bad debt reserves, which enables them to absorb very large default rates without a related drop in income.”

• cites Jackson and Liu (2007)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Certegy provides little, if any, information to consumers about their credit decision and cannot fully explain the final decisions reached by its artificial neural network system.

“Certegy provides little, if any information to consumers regarding their credit decision.”

• Also stated in: § 2.2.1 Knowledge and Understanding of Problem: · plain

  “Certegy cannot fully explain the final decisions reached by its ANN system.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

In West (2000), logistic regression was the most accurate method, but results might differ if some artificial neural network architectures were combined rather than assessed individually.

“While the results of that study indicated logistic regression was the most accurate method, it also stated the results might differ if some of the ANN architectures were combined rather than assessed individually.”

• cites West (2000) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The neural network approach offers a more general model in bankruptcy prediction problems.

“According to Quek, et al. (2009, p.166), “the neural network approach offers a more general model in bankruptcy prediction problems.””

• cites Quek, et al. (2009) sourceNot checked — no full text of the source

May combine two claims.

The aging method estimates bad debt at specific times but ignores the accuracy of past estimates using the method.

“This method estimates bad debt at specific times; however, the disadvantage of this method is that it ignores the accuracy of past estimates using the method.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

A mixture of expert approaches is more accurate according to West (2000) because this particular artificial neural network was more accurate than any other model in the study.

“However, a mixture of expert approaches (which combines the power of numerous models) is more accurate according to West (2000) because this particular ANN was more accurate than any other model in the study.”

• cites West (2000) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Two key factors, learning rate and momentum, must be considered in arriving at an estimate for the dependent variable.

“In addition two key factors, learning rate and momentum, must be considered in arriving at an estimate for the dependent variable.”

Probably not a substantial claim: may not pass S3 (load-bearing).

An artificial neural network model had a high probability of detecting fraudulent financial statements.

“The study finds an ANN model with a high probability of detecting fraudulent financial statements.”

• cites Fanning and Cogger's (1998) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Data distribution characteristics, size, and additive noise affect how artificial neural networks learn, their reliability, and their predictive accuracy.

“He found that data distribution characteristics, size, and additive noise have an impact on the manner in which ANNs learn, their reliability, and ultimately their predictive accuracy”

• cites Penharkar (2002)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

Bad debt allowance is manipulated up or down to meet earnings expectations.

“Consistent with expectations, they found bad debt allowance is manipulated up or down to meet earnings expectations.”

• cites McNichols and Wilson (1988) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Artificial neural networks are excellent tools because they do not rely on strong assumptions about the distribution of underlying variables and can overcome limitations of more restrictive traditional volatility estimation models.

“Aragones, et al. (2007) reached the conclusion that ANNs are excellent tools because they “do not rely on strong assumptions about the distribution of the underlying variables and, therefore, can overcome some of the limitations of more restrictive traditional volatility estimation models.””

• cites Aragones, et al. (2007) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The artificial neural network approach does not assume any probability distribution or equal dispersion for its data.

“The ANN approach does not assume any probability distribution or equal dispersion for its data.”

• cites Quek, et al. (2009) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Artificial neural network structures should be pyramidal.

“Kuldeep and Bhattacharya (2006) noted ANN structures should be “pyramidal.””

• cites Kuldeep and Bhattacharya (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

The artificial neural network method used by Aragones et al. (2007) is back-propagation.

“The ANN method used by Aragones, et al. (2007) is back-propagation (see Exhibit I).”

• cites Aragones, et al. (2007) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Credit sales do not necessarily represent cash inflow.

“However, credit sales do not necessarily represent cash inflow.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

A write-off expense from a discontinued segment is frequently difficult to determine and can be included with other discontinued segment expenses.

“Such an expense is frequently difficult to determine and can be included with other expenses associated with a discontinued segment.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

As a company becomes more leveraged with debt, an incentive exists to manipulate the allowance for bad debt to meet debt obligations.

“As a company becomes more leveraged with debt, there exists an incentive to manipulate the allowance for bad debt accounts in an effort to meet debt obligations.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A company can manipulate the sales to average total assets ratio through its credit policies.

“Thus, a company can manipulate this ratio through credit policies.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Profitability ratios indicate a company's health to creditors and can influence whether credit is granted.

“from the viewpoint of creditors, profitability ratios are an indication of the health of a company and can influence whether or not they grant credit.”

Probably not a substantial claim: may not pass S3 (load-bearing).

If defaults are high, actual write-offs can be concealed by including bad debt expense in a broader expense category.

“if defaults are high, actual write-offs could be 'concealed' by simply placing it in a broader category of expenses.”

• AssumptionRests on an assumption: Bad debt expense is frequently included within general and administrative expenses and is not separately reported on the income statement.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Financial statement variables reflect the results of operations and company decision-making.

“These financial statement variables reflect the results of operations and company decisionmaking.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

Management is often evaluated on profitability measures and bonuses are frequently tied to end results.

“Ultimately management is often evaluated on such measures and quite frequently their bonuses are tied to the end results.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

If the ratio of market value equity to book value equity is less than one, the stock could be undervalued.

“If this value is less than one, it means the stock could be undervalued.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

To make receivables appear more attractive to a factor, a company may manipulate net receivables through the allowance for bad debt.

“in order to make receivables appear more attractive to a factor, a company may manipulate net receivables through the Allowance for Bad Debt account.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Regardless of which situation a company employs, the allowance for bad debt will be affected.

“Regardless of which situation a company employs, the Allowance for Bad debt will be affected.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

If a company fails to pay its debt obligations, a creditor can take legal action.

“If a company fails to pay its debt obligations, a creditor can take legal action.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

Return on total assets should impact credit policies and eventually the allowance for bad debt.

“As a result this ratio should impact credit policies and eventually the allowance for bad debt.”

Probably not a substantial claim: may not pass S3 (load-bearing).

If a company does not adequately account for bad debts, the number provided to creditors may be inaccurate.

“If a company does not adequately account for bad debts, the number provided to creditors may not be accurate.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Bad debt expense is a component of overall earnings and is related to the allowance for bad debt.

“A component of overall earnings is bad debt expense, which is related to the allowance for bad debt.”

Classified as probably not a substantial claim. May combine two claims.

Because the allowance for bad debt is subjective and net income is measured on an accrual basis, management may manipulate the allowance to meet expectations and realize compensation bonuses.

“Since the allowance for bad debt account is a subjective measure combined with the fact that net income is measured on an accrual basis, there may exist a temptation for management to manipulate the allowance for bad debt account to meet expectations and realize compensation bonuses and incentives.”

• AssumptionRests on an assumption: The allowance for bad debt account contains a subjective component even when GAAP is followed.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Undervalued stock may indicate that the market sees potential problems with the company that are not readily apparent from its financial statements.

“undervalued stock may indicate the market sees potential problems with the company which are not readily apparent from the company's financial statements.”

Probably not a substantial claim: may not pass S3 (load-bearing).

An upward trend in days sales in accounts receivable should impact the allowance for bad debt because it implies longer collection and increased default chance.

“If this number begins to trend upwards, it should impact the allowance for bad debt accounts because it implies that it is taking longer to collect receivables, which increase the chance for defaults.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A higher number of defaults leads to a higher number of accounts a company will have to write off.

“The higher the number of defaults, the higher the number of accounts a company will have to write-off.”

May combine two claims.

A high standard deviation of the ratio of allowance for bad debt to actual write-off implies greater volatility and a company may operate at extremes.

“A high standard deviation would imply greater volatility, which in turn would denote a situation where a company may operate at extremes.”

• cites Riley and Pasewark (2009)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Owners have expectations about their investment returns.

“Owners have expectations in regards to their investment returns.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Ratios enable comparison of companies of differing sizes.

“The usage of ratios enables one to compare companies of differing sizes.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The design science approach focuses on improving and aiding human and organizational capacity through the invention of novel and innovative artifacts.

“The design science approach is not focused on behavioral theories, but instead, is focused on improving and aiding human and organizational capacity through the invention of novel and innovative 'artifacts.'”

• cites Hevner, et al. (2004) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S4 (based).

Companies with relatively high R&D expenses are expected to have more stringent credit policies due to the need for immediate cash.

“One would expect a company with relatively high R&D expenses to have more stringent credit policies due to the need for immediate cash than those companies with relatively low R&D expenses.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

Ratios provide information about a firm that may not be readily apparent on its financial statements.

“Ratios provide information about a firm that may not be readily apparent on a firm's financial statements.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

If more credit accounts are allowed, it will result in increased cash collections in the short term.

“if more credit accounts are allowed, it will result in increased cash collections in the short term.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

A business with a large investment in fixed assets has a higher break-even point than one with a comparatively small investment in such assets.

“A business which has a large investment in fixed assets and equipment (depreciable assets) has a higher break-even point than a company with a comparatively small investment in such assets.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

An artifact can be a computer-based solution that improves the user's experience, creativity, intuition, and problem-solving capabilities.

“An 'artifact' can be a computer-based solution which, according to Hevner, et al. (2004, p. 76), improves the “experience, creativity, intuition, and problem solving capabilities” of the user.”

• cites Hevner, et al. (2004) sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

The design science paradigm requires knowledge of a problem domain, a solution, a design artifact that accomplishes the solution, and an evaluation of the artifact.

“According to Hevner, et al. (2004), the design science paradigm requires the following: 1) knowledge and understanding of a problem domain, 2) a solution of the problem domain and 3) the design artifact which will accomplish the solution and 4) an evaluation of the artifact.”

• cites Hevner, et al. (2004) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

If a company has inadequate cash to pay current debt, its ability to continue into the future becomes a serious concern.

“If a company finds itself with inadequate cash to pay current debt, its ability to continue into the future becomes a serious concern.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Consumer debt is often discharged in bankruptcy and becomes permanently uncollectible.

“Consumer debt is often discharged in bankruptcy (i.e., becomes a bad debt and is permanently uncollectible).”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Management can reasonably be expected to find a mixture of credit policy and bad debt expense that maximizes net cash flow.

“One can reasonably expect management to find a mixture of credit policy and bad debt expense which maximize net cash flow”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Disposing of a segment could entail a very large write-off of accounts receivable.

“if a company decides to dispose of a segment, it could entail a very large write-off of accounts receivable.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Management may lower credit-granting standards to generate cash to meet expectations.

“In order to meet such expectations management may be tempted to lower credit-granting standards in order to generate cash.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

Undervaluing may be due to timing differences, so the market price will eventually adjust to reflect book value.

“This undervaluing may simply be a reflection of timing differences, which means the market price will eventually adjust to reflect book value.”

Classified as probably not a substantial claim.

An unexpected job loss would impact the allowance for bad debt.

“Clearly, this would impact the allowance for bad debt.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). The quote only partly states this claim.

Management can manipulate the ratio of income on assets to sales by under- or overestimating the ending allowance for bad debt.

“Management can very easily manipulate this amount by under or overestimating the amount that should be reflected in the ending allowance for bad debt.”

Probably not a substantial claim: may not pass S3 (load-bearing).

If lenient credit policies cause quick inventory sales, defaults and the allowance for bad debt may be high.

“If inventory is sold quickly (because credit policies are too lenient as opposed to selling at cheaper prices than competitors), this may lead to a high level of defaults and a high allowance for bad debts.”

Probably not a substantial claim: may not pass S4 (based).

When a segment is disposed, its accounts receivable and associated allowance for bad debt are eliminated from the overall company records.

“when a segment is disposed, any accounts receivable and their associated allowance for bad debt is eliminated from the overall company records.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

If a company has major cash flow problems, it may overstate net receivables by understating the allowance for bad debt.

“If a company is suffering major cash flow problems, the temptation would be to overstate net receivables by understating the Allowance for Bad Debt.”

Probably not a substantial claim: may not pass S3 (load-bearing).

To gain credit, a company may understate its allowance for bad debt.

“Thus, in order to gain credit, a company may be tempted to understate their allowance for bad debt.”

Probably not a substantial claim: may not pass S3 (load-bearing).

If bad debts fall far short of the allowance, the allowance account may be excessive.

“Likewise, if bad debts fall far short of the allowance, it may indicate a company's allowance account is excessive.”

• cites Riley and Pasewark (2009)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The allowance exhaustion rate compares beginning allowance to actual write-offs, and ideally it should be close to one.

“The beginning allowance amount is compared to the actual bad debt write-offs.”

• Also stated in: § 4.3 Dependent Variable: Allowance for Bad debt -Year End · plain

  “Ideally, the number should be close to one.”

• cites Riley and Pasewark (2009)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Credit sales are recorded as credit even if the customer fails to pay the associated account.

“credit sales are recorded as credit even if the customer fails to pay the associated account.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The allowance for bad debt is affected whenever credit policies are changed.

“By extension, the Allowance for Bad Debt is affected whenever credit policies are changed.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

If 100% of all accounts receivable were collected, it would indicate an overly strict credit policy.

“If 100% of all accounts receivable were collected, it would indicate an overly strict credit policy.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

If a company lacks sufficient cash, it may understate its allowance for bad debt to appear more solvent.

“If a company does not have sufficient cash balance, it may be tempted to understate the allowance for bad debt accounts in order to appear more solvent.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

A large amount of consumer credit outstanding can indicate that a majority of companies are leaning toward more lenient credit policies.

“This variable can also be an indicator, that a majority of the companies are leaning more towards more lenient credit policies.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Under accrual accounting, there are many ways, including manipulation of the allowance for bad debt, to improve net income.

“under accrual-based accounting, there are a myriad of ways, including the manipulation of the allowance for bad debt, to improve net income.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Lowering credit standards could increase near-term cash collections but lead to higher long-term default rates.

“lowering credit standards could result in more cash collections in the near term, but ultimately lead to higher default rates in the long term.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A lender's decision to grant credit includes an evaluation of the company's allowance for bad debt.

“The decision of a lender to grant a company credit includes an evaluation of the company's allowance for bad debt.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The allowance for bad debt can be manipulated at year end to derive a particular net income.

“the Allowance for Bad Debt account could simply be manipulated at the end of the year in order to derive a particular value for net income.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

If solvency ratios indicate financial problems, the temptation may be to lower credit standards to generate immediate cash when a company cannot meet its cash obligations.

“If the solvency ratios indicate financial problems, the temptation may be to lower credit standards in order to generate immediate cash if a company is faced with an inability to meet its cash obligations.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

If most inventoriable costs have been recaptured through sales, a company may revise credit policies because the danger of a net loss has decreased.

“If the majority of inventoriable costs have already been re-captured through sales, the company may revise credit policies accordingly as the danger of producing a Net Loss has greatly decreased.”

Classified as probably not a substantial claim.

The makeup of credit policies as influenced by R&D will impact the allowance for bad debt.

“Hence, the makeup of the credit polices as influenced by R&D will impact the allowance for bad debt.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

The depreciation and amortization to sales ratio impacts sales goals and ultimately the allowance for bad debt.

“Consequently, this ratio impacts sales goals, which ultimately impacts the Allowance for Bad Debts.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

If inventory takes too long to sell, it may be because credit policies are too strict or competitors sell the same items at lower prices.

“If it takes too long to sell inventory, it may be because its credit policies are too strict or their competitors are selling the same items at lower prices.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Net receivables, total inventory, cost of goods sold, and sales all have a direct impact on the allowance for bad debt.

“variables such as net receivables, total inventory, cost of goods sold and sales all have a direct impact on the allowance for bad debt accounts.”

Probably not a substantial claim: may not pass S3 (load-bearing).

As consumer credit outstanding increases, the likelihood of defaults increases because individuals may have been granted more credit than they can pay.

“As this number increases, the greater the likelihood of defaults as individuals may have been granted more credit than they are capable of paying.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Miscellaneous items such as number of employees, stock prices, dividends paid, and cash flows may not directly impact the allowance for bad debt but reflect overall health and must be considered when calculating the allowance.

“Miscellaneous items, which include variables such as number of employees, stock prices, dividends paid and cash flows may not directly impact the allowance for bad debt accounts, but they do reflect the overall health of a company and as such must be considered when calculating the allowance for bad debt Accounts.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

A higher volume of credit sales increases the probability that some accounts will default.

“the higher the volume of credit sales, the greater the probability that some accounts will default.”

• Also stated in: § 4.4.2 Efficiency Ratios · plain

  “The higher the volume of credit sales the greater the probability the accounts will default.”

Probably not a substantial claim: may not pass S4 (based).

If people are not working, their ability to repay debt is impaired.

“If people are not working, their ability to repay debt is impaired.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A longer time to collect receivables increases the likelihood of defaults.

“the longer it takes to collect receivables, the greater the likelihood of defaults.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

A person who can repay debt today may not be able to repay it tomorrow due to an unexpected job loss.

“a person who has the means to repay debt today may not have the means to repay debt tomorrow due to an unexpected job loss.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

In factoring with recourse, all risk resides with the company factoring its receivables because the factor will demand payment for accounts it cannot collect.

“In some cases (factoring with recourse) all the risk resides with the company factoring its receivables, as the factor will demand payment for accounts it cannot collect.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The price to sales primary ratio can be improved by methods that impact the ending allowance for bad debt, as long as credit policies are not too inappropriate.

“As long as credit policies are not too inappropriate, this ratio can be improved by methods which impact the ending allowance for bad debt.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Once a company reaches break-even, it has more options in setting credit policies.

“once a company reaches a break-even point, it has more options available to it in terms of credit policies.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

The allowance for bad debt may be heavily influenced by many internal and external factors.

“In actuality, the allowance for bad debt may be heavily influenced by any number of factors, both internal and external.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

A company will manipulate the allowance for bad debt if doing so increases the likelihood of securing a loan.

“A company will manipulate the allowance for bad debt if doing so would increase the likelihood of securing a loan.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Because EPS is an important measure, management may use earnings management techniques to make it appear favorable.

“Since EPS is such an important measure, management may have an incentive to use earnings management techniques to make this number appear favorable.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Habitually low liquidity ratios may indicate a cash flow problem.

“If these ratios are habitually low for a company, it may indicate that there is a cash flow problem.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

A high allowance for bad debt could negatively impact stock price.

“a high allowance for bad debt could negatively impact stock price.”

Classified as probably not a substantial claim.

Bad debt expense should closely match actual write-offs over time.

“Ideally the bad debt expense should match closely with actual write-offs.”

• cites Riley and Pasewark (2009)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Creditors can impose penalties on a company or force it into bankruptcy.

“Creditors have the option to impose penalties on the company and/or force it into bankruptcy.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

New accounts can be factored to generate quick cash.

“Also, the new accounts can always be factored to generate quick cash.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

A company may manipulate net receivables to appear financially healthy.

“As with the current ratio, the temptation may be to manipulate net receivables in order to appear more 'healthy' financially.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

Traditional methods of estimating bad debt are subjective.

“The traditional methods of estimating bad debt are subjective.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Credit policies can be changed to increase sales, and this directly affects bad debt.

“Credit policies can be changed to increase sales and this would have a direct effect on bad debt.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Companies typically have the option to factor receivables.

“Companies typically have the option to factor receivables.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Credit policies that produce a large allowance for bad debt could increase cash flow and make higher dividends possible.

“credit policies which produce a large number in the allowance for bad debt account, could increase cash flow and ultimately make higher dividends possible.”

Classified as probably not a substantial claim. May combine two claims.

Because the allowance for bad debt contains a subjective component, management may manipulate it to meet earnings projections.

“Since the allowance for bad debt accounts contains a subjective component (even if GAAP is followed), management may be tempted to manipulate the allowance for bad debt account in order to meet earnings projections and expectations.”

• AssumptionRests on an assumption: The allowance for bad debt account contains a subjective component even when GAAP is followed.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

In some cases only a few variables may be sufficient to estimate a company's ending bad debt allowance, while in other cases many variables may be needed for accurate prediction or classification.

“In some cases, only a few variables may be sufficient to estimate a company's ending bad debt allowance, whereas in other cases, a relatively large number of variables may be needed in order to determine an accurate prediction and/or classification”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

If a company emphasizes the investment associated with generating each sales dollar, it will avoid credit sales to customers who will not pay because such sales waste investment capital.

“If a company emphasizes the investment associated with the generation of each sales dollar, then the company will not want make credit sales to customers who will ultimately not pay because the sales represent a waste of investment capital.”

Probably not a substantial claim: may not pass S4 (based).

As the number of bankruptcies increases, the potential for accounts defaulting increases.

“As the number of bankruptcies increases, the potential for accounts defaulting increases.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Failure to ascertain customer debt information will impact the allowance for bad debt as defaults begin to materialize.

“Failure to ascertain customer debt information will impact the allowance for bad debt as defaults begin to materialize.”

Probably not a substantial claim: may not pass S3 (load-bearing).

It cannot be determined with complete certainty whether NeuroXL Clusterizer's three classifications match the true correct three classifications from other methods.

“There is no means to ascertain with 100% certainty whether the three classifications derived from NeuroXL Clusterizer are the same as the true correct three classifications derived by other means.”

Classified as probably not a substantial claim.

Ignoring exogenous variables can have significant consequences, such as increased default rates when the overall economic environment changes.

“Ignoring exogenous variables can have significant consequences.”

• Also stated in: § 5) Consumer Price Index (All Items Annual Average) -Measures the cost of living. It · hedged

  “If the overall economic environment in which the company operates changes, it is quite likely their default rates will increase.”

Probably not a substantial claim: may not pass S3 (load-bearing).

It is impossible to know how the inclusion of these variables impacted the ANNs used in this study.

“It is impossible to know how the inclusion of these variables impacted the ANNs used in this study.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Inflation can make a person's fixed salary insufficient to pay all their debts, which affects the allowance for bad debts.

“The implication for the allowance for bad debts is clear: A person making a set salary, may suddenly find that their income is no longer sufficient to pay all their debts.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The two-step clustering procedure in SPSS was more appropriate for a relatively large data set with many continuous variables.

“This method was more appropriate for a relatively large data set which possessed many continuous variables.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

If individuals have savings in reserve, they are more likely to have sufficient funds to pay their debts as they become due.

“if individuals have savings in reserve, it increases the likelihood they will have sufficient funds to pay their debts as they become due.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Neural networks can flag suspicious activity almost immediately.

“However, ANNs provide a method to identify suspicious activity almost instantaneously.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

With 10,000 epochs, the ANN's final weights should be nearly as accurate as possible.

“Given that this number was very high, in theory it should mean that the final weights assigned by the ANN are as close to accurate as possible.”

Probably not a substantial claim: may not pass S4 (based).

A large disposable personal income indicates that consumers have more money available to repay debt and buy non-essential or discretionary items.

“A large number indicates that consumers have more money available to repay debt and buy non-essential/discretionary items.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Selecting variables would be the main obstacle in developing such a network.

“The prime difficulty facing the development of such a network would be the selection of variables.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Such a capability would discourage people inclined toward fraud.

“This would act as a deterrent for those who are inclined to engage in fraudulent activity.”

Probably not a substantial claim: may not pass S3 (load-bearing).

This study's problems were not complex enough to require ANNs that could gather data.

“However, the complexity of the problems for this study did not necessitate ANNs capable of data gathering.”

Classified as probably not a substantial claim.