CN112200647A - Audit quality determination method and system - Google Patents
Audit quality determination method and system Download PDFInfo
- Publication number
- CN112200647A CN112200647A CN202011086884.9A CN202011086884A CN112200647A CN 112200647 A CN112200647 A CN 112200647A CN 202011086884 A CN202011086884 A CN 202011086884A CN 112200647 A CN112200647 A CN 112200647A
- Authority
- CN
- China
- Prior art keywords
- audit
- company
- probability
- quality
- type
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/12—Accounting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/80—Management or planning
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- Data Mining & Analysis (AREA)
- Entrepreneurship & Innovation (AREA)
- Accounting & Taxation (AREA)
- Mathematical Physics (AREA)
- Finance (AREA)
- General Business, Economics & Management (AREA)
- Mathematical Analysis (AREA)
- Educational Administration (AREA)
- Pure & Applied Mathematics (AREA)
- Operations Research (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Marketing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Life Sciences & Earth Sciences (AREA)
- Algebra (AREA)
- Tourism & Hospitality (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Technology Law (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention discloses a method and a system for measuring audit quality, wherein the method comprises the following steps: extracting the deduction type detail data and the associated transaction type detail data based on the company financial statement and the supplementary notes thereof; based on the accounting measurement attribute of the company, the balance sheet item and the profit sheet item are divided into a historical cost item and a to-be-verified item again; calculating the audit quality of the company level; predicting the audit quality of a company level based on an Mscore model to generate a probability estimation result of financial fraud occurrence; generating an audit failure probability estimation result of company level audit quality based on an audit failure prediction model; and generating an audit quality determination report based on the probability estimation result of the financial fraud and the audit failure probability estimation result. The embodiment of the invention processes the related financial statement data to generate a corresponding audit quality measurement result so as to meet the audit risk assessment, audit industry quality detection and government supervision requirements of audit practitioners and realize accurate positioning prediction service.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a method and a system for auditing quality determination.
Background
The current definition of audit quality includes: (1) finding the financial statement by an auditor to have bugs, and reporting the possibility of the bugs; (2) the probability that the auditor does not provide a non-standard audit opinion report for the financial statement containing the major error is obtained; (3) the faithful expression degree of the data reported by the auditor; (4) the ability to measure audits to reduce noise and variation and improve accountant information fidelity. Although the definition of the audit quality is different, the influence of the audit on the information quality of the financial statement is emphasized to a certain extent. In consideration of the importance of the financial information quality, the effective measure of the audit quality undoubtedly has important value on capital market and economic boundary, has a fundamental effect on improving the audit quality, and has important practical significance on the supervision of an accountant affairs institute and accounting supervision.
However, since the audit investment and the audit process cannot be directly observed from the outside world, and the audit report is standardized, the audit quality is difficult to measure directly (Liu Feng et al, 2007). The measurement of each boundary on the audit quality problem generally designs a substitute index through some observable audit output or behaviors, and indirectly measures the audit quality according to the substitute index. Currently, the more common substitute indexes for the audit quality measure include the reputation of the office, the scale, the audit cost, the non-standard audit opinions and the surplus quality. It should be noted that, the indirect alternative measurement of the audit quality often mixes the influence of many other factors, which cannot be effectively eliminated, and the measurement of the audit quality is inaccurate (liu feng, etc., 2007), which is also the reason why a contradiction conclusion may be drawn when different alternative measurements are used to verify the audit quality of the same service. On the other hand, the auditing quality is a complex concept, and the alternative measure variables have certain reason but are not comprehensive, and also deviate from the characteristics of the auditing main body. Therefore, the measurement of the audit quality needs to tightly surround the attribute of the audit, and from the essential problem, important factors reflecting the audit quality are mined.
According to the general definition of auditing, auditing quality relates to the ability of an auditor to exert auditing functions so as to ensure that generated accounting information is consistent with recognized accounting principles. The audit quality is defined in a large and small way, the audit quality is recognized to contain various abilities of an auditor, and the abilities are sensed through the information quality of financial statements. In other words, the financial statement of the listed company is an important output of the auditing work, and the information quality reflects various abilities of the auditor and becomes an important presentation object of the auditing quality. Therefore, the measure of the audit quality can be based on the measurement attribute of the accounting information from the financial statement of the listed company, namely the output level, and the audit quality is regarded as a structure with a plurality of mutually related dimensions, so that a comprehensive measure framework of the audit quality is constructed, and a reliability measuring mode needs to be provided from an informatization angle so as to meet the related requirements of related accountants.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides an audit quality determination method and an audit quality determination system.
In order to solve the above technical problem, an embodiment of the present invention provides a method for auditing quality measurement, where the method includes:
extracting the deduction type detail data and the associated transaction type detail data based on the company financial statement and the supplementary notes thereof;
based on the accounting measurement attribute of the company, the balance sheet item and the profit sheet item are divided into a historical cost item and a to-be-verified item again;
calculating company level audit quality based on the reduced value type detail data, the associated transaction type detail data, the historical cost type project and the to-be-verified type project;
predicting the company level audit quality based on an Mscore model through an R language to generate a probability estimation result of financial fraud occurrence;
generating an audit failure probability estimation result by using an R language to base the audit quality of the company level on an audit failure prediction model;
and generating an audit quality measurement report through the R language based on the probability estimation result of the financial fraud and the audit failure probability estimation result.
The calculating of the company-level audit quality based on the reduced value type detail data, the associated transaction type detail data, the historical cost type project and the to-be-verified type project comprises the following steps:
calculating a historical cost item ratio and a reasonable part ratio of the to-be-verified item based on the historical cost item and the to-be-verified item, wherein the sum of the historical cost item ratio and the reasonable part ratio of the to-be-verified item is the auditing work quality;
calculating the occupation ratio of the items to be verified based on the items to be verified, and calculating the Euclidean distance of the items to be verified of the sample company in a multi-dimensional space, wherein the Euclidean distance of the occupation ratio of each item to be verified is the judgment level of the auditing specialty;
and calculating the auditing effort degree, wherein the auditing effort degree is | management layer performance forecasting deviation |/| the audited actual profit |, and the management layer performance forecasting deviation is the difference between the forecasted net profit and the actual net profit.
The method for predicting the company level audit quality based on the Mscore model through the R language to generate the probability estimation result of the financial fraud comprises the following steps:
constructing a financial fraud early warning model based on the Mscore model, and increasing auditing work quality, auditing professional judgment level and auditing effort degree on the basis of the Mscore index;
calculating the probability of making the first type of errors and the probability of making the second type of errors through Logistic regression, selecting a value with the minimum expected cost as a threshold value according to the cost of making the first type of errors and the cost of making the second type of errors, and predicting the probability of financial fraud of a company on the market, wherein: the probability of making a first type of error is the probability of predicting a fraud company as a normal company, and the probability of making a second type of error is the probability of predicting a normal company as a fraud company.
The generating of the audit failure probability estimation result by the company level audit quality based on the audit failure prediction model through the R language comprises the following steps:
on the basis of a pure financial index audit failure prediction model, the audit work quality, the audit professional judgment level and the audit effort degree are increased on the basis, the probability of making a first type of errors and the probability of making a second type of errors are calculated through Logistic regression, and then a value with the minimum expected cost is selected as a threshold value according to the cost of making the first type of errors and the cost of making the second type of errors, so that the probability of audit failure of a listed company is predicted, wherein: the probability of making a first type of error is the probability of predicting a fraud company as a normal company, and the probability of making a second type of error is the probability of predicting a normal company as a fraud company.
Correspondingly, the embodiment of the invention also provides a system for auditing quality measurement, which comprises:
the data processing module is used for extracting the reduced value type detail data and the associated transaction type detail data based on the company financial statement and the supplementary notes thereof;
the project classification module is used for reclassifying the asset liability statement project and the profit statement project into a historical cost project and a to-be-verified project based on the accounting measurement attribute of the company;
the audit quality calculation module is used for calculating the company level audit quality based on the subtractive detail data, the associated transaction detail data, the historical cost items and the items to be verified;
the fraud prediction module is used for predicting the company level audit quality based on the Mscore model to generate a probability estimation result of financial fraud occurrence;
the audit failure prediction module is used for generating an audit failure probability estimation result based on the audit failure prediction model on the company level audit quality;
and the processing module is used for generating an audit quality measurement report based on the probability estimation result of financial fraud and the audit failure probability estimation result.
The audit quality calculation module calculates the ratio of the historical cost items and the reasonable part ratio of the items to be verified based on the historical cost items and the items to be verified, and the sum of the ratio of the historical cost items and the reasonable part ratio of the items to be verified is the audit working quality; calculating the ratio of the items to be verified based on the items to be verified, and calculating the Euclidean distance of the items to be verified of the sample company in the multi-dimensional space, wherein the Euclidean distance of each item ratio of the items to be verified is the judgment level of the auditing specialty; and calculating the auditing effort degree, wherein the auditing effort degree is | management layer performance forecasting deviation |/| the audited actual profit |, and the management layer performance forecasting deviation is the difference between the forecasted net profit and the actual net profit.
The fraud prediction module constructs a financial fraud early warning model based on the Mscore model, and increases the auditing work quality, the auditing professional judgment level and the auditing effort degree on the basis of the Mscore indexes; calculating the probability of making the first type of errors and the probability of making the second type of errors through Logistic regression, selecting a value with the minimum expected cost as a threshold value according to the cost of making the first type of errors and the cost of making the second type of errors, and predicting the probability of financial fraud of a company on the market, wherein: the probability of making a first type of error is the probability of predicting an audit-failed company as a non-audit-failed company, and the probability of making a second type of error is the probability of predicting a non-audit-failed company as an audit company.
The audit failure prediction module is based on a pure financial index audit failure prediction model, increases audit work quality, audit professional judgment level and audit effort degree on the basis, calculates the probability of making a first type of error and the probability of making a second type of error through Logistic regression, selects a value which enables the expected cost to be minimum as a threshold value according to the cost of making the first type of error and the cost of making the second type of error, and predicts the probability of audit failure of a company on the market, wherein: the probability of making a first type of error is the probability of predicting an audit-failed company as a non-audit-failed company, and the probability of making a second type of error is the probability of predicting a non-audit-failed company as an audit company.
The embodiment of the invention constructs a measuring index system of the audit quality in a multi-level and multi-dimensional manner from the audit output angle. Therefore, compared with a single audit quality substitute index, the method has the advantages that the substantial attribute of the audit quality is more directly and comprehensively disclosed from a wider visual angle and returns to the characteristics of the audit subject. In addition, the method improves the accuracy of the existing financial fraud prediction model and the audit failure prediction model by extracting the financial statement information, and can form a corresponding decision report for monitoring and related supervision reference of the audit industry. According to the embodiment of the invention, the relevant financial statement data is processed to generate the corresponding audit quality measurement result, so that the audit risk assessment, the audit industry quality detection and the government supervision requirements of audit practitioners are met, and the accurate positioning prediction service is realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow diagram of a method of auditing quality determinations in an embodiment of the invention;
fig. 2 is a system diagram of audit quality determination in an embodiment of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
FIG. 1 shows a flow diagram of a method of auditing quality determinations in an embodiment of the invention, the method comprising the steps of:
s101, extracting the reduced-value type detail data and the associated transaction type detail data based on the company financial statement and the supplementary notes thereof;
s102, based on the accounting measurement attributes of the company, reclassifying the asset liability statement item and the profit statement item into a historical cost type item and a to-be-verified type item;
the data used in the embodiment of the invention relates to an A stock listed company asset liability statement, a profit sheet, a financial statement remark and a management layer performance forecast, all the information is from listed company public information, the data interval is 2011-:
1.1, extracting detail data of the value-reduced type (eight items such as bad account preparation and the like) according to the financial statement of the company and the remarks thereof. The related items are as follows: ten statement items such as receivables.
And 1.2, extracting associated transaction detail data according to the company financial statement and the remarks thereof. The related items are as follows: nine statement items such as receivable bills.
1.3, according to the accounting measurement attribute, the balance sheet item and the profit sheet item are divided into two categories again: items not related to estimation, judgment and selection (called historical cost items for short) and items related to estimation, judgment and selection (called verified items for short) are classified according to the following relevant classification results:
(1) historical cost category items: a. asset class: thirty-five reporting items such as monetary funds; b. liability category: twenty-nine report items such as short-term borrowing and the like; c. the equity class: actual income of three reporting items such as capital (or stock) and the like; d. income-profit category: exchange revenue; e. cost and expense loss class: eleven statement items such as a pay-off net amount and the like;
(2) the items to be verified are: A. the assets comprise twenty-four report items such as transactional financial assets and the like; B. liability category: the twelve statement items such as the obtained tax and debt are postponed; C. the equity class: nine reporting items such as the price accumulation and the like; D. income-profit category: nine reporting items such as premium have been earned; E. cost and expense loss class: business costs, etc.
And 1.4, selecting performance forecast data which is released by a management layer for the first time. If the net profit value is missing, the method is implemented according to the following steps: step one, replacing annual prediction net profit of the performance report; and secondly, replacing the annual net profit predicted by the net profit announced by the third quarter financial statement for the remaining missing value.
In S101-S102, the requirement for auditing quality measurement can be met through data processing, and the corresponding model calculation process can be better met.
S103, calculating company level audit quality based on the reduced value type detail data, the associated transaction type detail data, the historical cost type project and the to-be-verified type project;
it should be noted that, based on the data processing performed in S101-S102, company-level audit quality calculation, audit team-level audit quality calculation, accounting firm-level audit quality calculation, and calculation of internal quality control level can be implemented.
2.1, calculating the audit quality at the company level, wherein the audit quality at the company level comprises the following three steps:
(1) an audit Work Quality (WQ) is calculated.
The corporate financial statement is the output of the audit work, the fair expression degree of the corporate financial statement is the differentiated attribute of the information quality, and the corporate financial statement is the basic embodiment of the audit work quality. Based on the historical cost class and the data to be verified, respectively calculating the historical cost class item occupation ratio and the reasonable part occupation ratio of the project to be verified according to formulas (1) and (2). And the sum of the historical cost type project occupation ratio and the reasonable part occupation ratio of the to-be-verified type project is the auditing work quality.
Audit work quality is the ratio of the historical cost item and the reasonable part of the item to be verified (3)
(2) And calculating the professional judgment level (JP) of the audit. Because the financial statement has a discretionary part, an auditor is required to exert the professional judgment level capability of the auditor. The invention uses the part related to estimation, judgment and selection in the financial statement of the company, namely the deviation degree of the items to be verified from the industry to represent the judgment level of the auditing specialty. And (3) calculating the proportion of the items to be verified according to the formula (4), and then calculating the Euclidean distance of the items to be verified of the sample company in the multidimensional space (every two companies in the industry), namely the judgment level of the auditing specialty.
Audit professional judgment level (5) is equal to Euclidean distance of each item ratio to be verified
Note: in order to ensure that the sum of the historical cost item ratio and the item ratio to be verified is 1, the formulas (1), (2) and (4) are subjected to standardized adjustment.
(3) An audit Effort Level (EL) is calculated. The performance forecast issued by the management layer for the first time is usually not audited by a registered accountant, the invention measures the auditing effort degree by the deviation of the performance forecast of the management layer, and the calculation method is a formula (6):
audit effort | management layer performance prediction deviation |/| actual profit audited | (6)
Wherein, the management layer performance forecast deviation is the difference between the forecast net profit and the actual net profit.
Calculating the ratio of the historical cost items and the reasonable part ratio of the items to be verified based on the historical cost items and the items to be verified in the three steps (1), (2) and (3), wherein the sum of the ratio of the historical cost items and the reasonable part ratio of the items to be verified is the auditing working quality; calculating the occupation ratio of the items to be verified based on the items to be verified, and calculating the Euclidean distance of the items to be verified of the sample company in a multi-dimensional space, wherein the Euclidean distance of the occupation ratio of each item to be verified is the judgment level of the auditing specialty; and calculating the auditing effort degree, wherein the auditing effort degree is | management layer performance forecasting deviation |/| the audited actual profit |, and the management layer performance forecasting deviation is the difference between the forecasted net profit and the actual net profit.
And 2.2, calculating the auditing quality of the auditor team.
The method specifically comprises three indexes, specifically as follows:
the auditing work quality of the auditing teachers and teams is ═ Σ (the auditing work quality of the auditing teachers and teams in the company level) or/the number of companies of the auditing teachers and teams participating in the auditing;
the auditor team level auditing professional judgment level is equal to Σ (company level auditing professional judgment level of the auditor team participating in auditing)/the number of companies of the auditor team participating in auditing;
and the auditor team level auditing effort degree is equal to Σ (company level auditing effort degree of the auditor team participating in the auditing)/the number of companies of the auditor team participating in the auditing.
And 2.3, calculating the audit quality of the accounting firm level.
The auditing work quality of the accounting firm is equal to Σ (auditing work quality of the company participating in auditing of the accounting firm)/the number of companies participating in auditing of the accounting firm;
the level of audit professional judgment of the accounting firm is ═ Σ (the level of audit professional judgment of the company of the accounting firm participating in the audit)/the number of companies of the accounting firm participating in the audit;
the auditing effort degree of the accounting firm level is ═ Σ (the auditing effort degree of the accounting firm in the company level)/the number of companies in which the accounting firm participates in the auditing.
The level of quality control inside the accounting firm is the dispersion degree of the audit quality of different audit teams of the accounting firm.
S104, predicting the company level audit quality based on an Mscore model through an R language to generate a probability estimation result of financial fraud occurrence;
the financial fraud early warning model is built based on the Mscore model, and the auditing work quality, the auditing professional judgment level and the auditing effort degree are increased on the basis of the Mscore indexes; calculating the probability of making the first type of errors and the probability of making the second type of errors through Logistic regression, selecting a value with the minimum expected cost as a threshold value according to the cost of making the first type of errors and the cost of making the second type of errors, and predicting the probability of financial fraud of a company on the market, wherein: the probability of making a first type of error is the probability of predicting a fraud company as a normal company, and the probability of making a second type of error is the probability of predicting a normal company as a fraud company.
Firstly, a financial fraud early warning model based on an Mscore model is perfected. The Mscore model is a financial fraud early warning model constructed based on pure financial indexes and named after the financial fraud behaviors of the safety company are successfully predicted. Wherein the Mscore model is:
Mscorei＝-4.84+0.92DSRIi+0.528GMIi+0.404AQIi+0.892SGIi+0.115DEPIi-0.172SGAIi-0.327LVGIi+4.679TATAi (M1)
table 1 Mscore variable definition table
The financial fraud early warning model is built based on the Mscore model, the auditing working quality, the auditing professional judgment level and the auditing effort degree are increased on the basis of the Mscore indexes, the probability of making a first type of errors and the probability of making a second type of errors are calculated through Logistic regression, then the value with the minimum expected cost is selected as a threshold value according to the cost of making the first type of errors and the cost of making the second type of errors, the probability of financial fraud occurring in a company on the market is predicted, and whether three sub-variables contained in the auditing quality play an optimizing role in the Mscore model is checked. Wherein FrauditAnd (4) indicating whether financial fraud behaviors occur in the t year of the company i, if so, assigning the value to be 1, and otherwise, assigning the value to be 0. The invention judges the types of financial fraud as follows: fictitious profits, fictitious assets, false records, significant omissions, insubstantial disclosure, and general accounting mishandling. The improved financial fraud prediction model comprises the following steps:
Fraudit＝1/{1+exp(β0+β1DSRIit+β2GMIit+β3AQIit+β4SGIit+β5DEPit+β6SGAit+β7LVGIit+β8TATAit+β9WQit+β10JPit+β11ELit)} (M2)
wherein beta is0Is a constant term, β1、β2、β3、β4、β5、β6、β7、β8、β9、β10、β11Are coefficients.
And finally, constructing a financial fraud prediction model only based on the audit work quality, the audit professional judgment level and the audit effort degree:
Fraudit＝1/{1+exp(β0+β1WQit+β2JPit+β3ELit)} (M3)
wherein beta is0Is a constant term, β1、β2、β3Are coefficients.
Based on the models, the financial fraud probability is predicted by taking data in 2011-2016 as a training set and data in 2017-2018 as a prediction set. Since the cost of a financial fraud company predicting a normal company is much higher than the cost of a normal company predicting a financial fraud company, the present study is primarily focused on the probability of a financial fraud offender being a first type of error (the probability of a financial fraud company being misinterpreted as a normal company), where: the probability of making a first type of error is the probability of predicting a fraud company as a normal company, and the probability of making a second type of error is the probability of predicting a normal company as a fraud company.
As can be seen from table 2, compared with the Mscore model (M1), the probability of making the first type of error when predicting financial fraud is significantly improved by adding the financial fraud prediction model (M2) constructed based on the audit work quality, the audit professional judgment level and the audit effort level and the financial fraud prediction model (M3) constructed based on only the audit work quality, the audit professional judgment level and the audit effort level on the basis of the Mscore indexes.
TABLE 2 financial fraud prediction Effect
S105, generating an audit failure probability estimation result by using the R language to base the company level audit quality on an audit failure prediction model;
the method is based on a pure financial index audit failure prediction model, increases audit work quality, audit professional judgment level and audit effort degree on the basis, calculates the probability of making a first type of error and the probability of making a second type of error through Logistic regression, selects a value which enables the expected cost to be minimum as a threshold value according to the cost of making the first type of error and the cost of making the second type of error, and predicts the probability of audit failure of a listed company, wherein: the probability of making a first type of error is the probability of predicting a fraud company as a normal company, and the probability of making a second type of error is the probability of predicting a normal company as a fraud company.
The Logistic regression involved in the embodiment of the invention is a probabilistic nonlinear regression model and is used for solving the problem of binary classification. Vector x ═ of n independent variables (x)0,x1,x2,...,xn) The interpreted variable y is a dummy variable. Assuming that the conditional probability P (y ═ 1| x) ═ P is the probability that y would be 1 under the x condition, we can derive:
wherein g (x) ═ ω0+ω1x1+ω2x2+ω3x3+...+ωnxn. Meanwhile, a probability that y is 0 under the condition of x can be derived.
The probability ratio of occurrence to non-occurrence of an event is:
taking the logarithm of formula (3) can obtain
Assuming that m samples are in total in the t year, the observed values are respectively y1t,y2t,...,ymtThen the likelihood function is:
taking the logarithm of the maximum likelihood function, and solving the partial derivatives for n +1 parameters can obtain n +1 equations, and solving the n +1 equations can obtain the parameter ω ═ (ω ═ c)0，ω1，ω2，ω3，...，ωn) The value of (c). In the embodiment, the regression model is used for judging the samples according to the threshold value, the samples higher than the threshold value are classified into one class, the samples lower than the threshold value are classified into one class, namely, the related values of the audit work quality, the audit professional judgment level, the audit effort degree and the like are used as vectors of n independent variables to perform the calculation process under the regression model, and the probability of making the first class errors and the probability of making the second class errors are calculated.
Part of financial indexes in the audit failure model can be selected, and a pure financial index prediction model is constructed:
Audit Failure＝1/{1+exp(β0+β1epsit+β2ldblit+β3resinit+β4zhggit+β5growthit+β6reczzit+β7rewageit+β8retaxit+β9stockzzit)}(M4)
wherein beta is0Is a constant term, β1、β2、β3、β4、β5、β6、β7、β8、β9It is a coefficient, which is calculated by exp function in R language.
Among them, the Audit FailureitIs to measure whether the company i has the trial in the t yearAnd (4) defining a variable (with the value of 0 or 1) for accounting failure as audit failure, wherein the company on the market punishs due to the financial statement but the auditor issues no-reservation audit opinions in the current year. If the audit failure happens in the current year in company i, the value is 1; otherwise it is 0. The invention removes the penalty of non-financial report-level reasons, and the specific financial report-level penalty reasons comprise fictitious profits, virtual column self-check, false records (misleading statement), delayed disclosure, significant omission, disclosure incompleteness (other) and the like.
Table 3 details the definition of the financial indicators:
TABLE 3 financial index variable definition Table
Based on a pure financial index audit failure prediction model, the invention increases the audit work quality, the audit professional judgment level and the audit effort degree on the basis, calculates the probability of making a first type of error and the probability of making a second type of error through Logistic regression, selects a value which enables the expected cost to be the minimum as a threshold value according to the cost of making the first type of error and the cost of making the second type of error, predicts the probability of the market company to have audit failure, and checks whether three sub-variables contained in the audit quality play an optimization role in the pure financial index audit failure prediction model, wherein the improved audit failure prediction model is as follows:
AuditFailureit＝1/{1+exp(β0+β1epsit+β2ldblit+β3resignit+β4zhggit+β5growthit+β6reczzit+β7rewageit+β8retaxit+β9WQit+β10JPit+β11ELit)} (M5)
wherein beta is0Is a constant term, β1、β2、β3、β4、β5、β6、β7、β8、β9、β10、β11Are coefficients.
The probability of audit failure of a company on the market is directly influenced by the audit quality, so that the invention only constructs an audit failure prediction model based on the audit work quality, the audit professional judgment level and the audit effort degree:
AuditFailureit＝1/{1+exp(β0+β1WQit+β2JPit+β3ELit)} (M6)
wherein beta is0Is a constant term, β1、β2、β3Are coefficients.
According to the method, data in 2011 and 2016 years are used as a training set, data in 2017 and 2018 years are used as a prediction set, and auditing failure is predicted. Since the cost of predicting a company with an audit failure as a normal company is much higher than the cost of predicting a normal company as an audit failure company, the present study is mainly focused on the probability of the audit failure making a first type of error (the probability of misjudging an audit failure company as a normal company).
As can be seen from Table 4, compared with the pure financial index model (M4), the audit failure prediction model (M5) which is constructed by increasing the audit work quality, the audit professional judgment level and the audit effort degree on the pure financial indexes has the significant improvement on the probability of making the first type of errors when the audit failure behavior is predicted.
TABLE 4 Audit failure prediction Effect
And S106, generating an audit quality measurement report through an R language based on the probability estimation result of financial fraud and the audit failure probability estimation result.
It should be noted that, here, the visualization of the report can be realized through the R language, a Web interface based on D3 can be directly generated through the rCharts package based on the probability estimation result of the occurrence of financial fraud and the audit failure probability estimation result, in the specific implementation process, the data source and the drawing mode are specified through the formula and the data, the type of the corresponding type is specified, and finally, the visualization of the audit quality measurement report is realized, which is convenient for the user to look up. In the visualization process, functions of filtering, paging, sequencing and the like under a Web interface can be realized based on the R language, so that the audit quality measurement report meets different requirements of users.
The embodiment of the invention constructs a measuring index system of the audit quality in a multi-level and multi-dimensional manner from the audit output angle. Therefore, compared with a single audit quality substitute index, the method has the advantages that the substantial attribute of the audit quality is more directly and comprehensively disclosed from a wider visual angle and returns to the characteristics of the audit subject. In addition, the accuracy of the conventional financial fraud prediction model and the accuracy of the audit failure prediction model are improved by extracting the financial statement information, and a corresponding decision report can be formed for use as a common audit reference. According to the embodiment of the invention, the relevant financial statement data is processed to generate the corresponding audit quality measurement result so as to meet the relevant requirements of relevant accountants and realize accurate positioning prediction service.
Accordingly, fig. 2 shows a schematic structural diagram of a system for auditing quality measurements in an embodiment of the present invention, the system comprising:
the data processing module is used for extracting the reduced value type detail data and the associated transaction type detail data based on the company financial statement and the supplementary notes thereof;
the project classification module is used for reclassifying the asset liability statement project and the profit statement project into a historical cost project and a to-be-verified project based on the accounting measurement attribute of the company;
the audit quality calculation module is used for calculating company level audit quality based on the subtractive detail data, the associated transaction detail data, the historical cost items and the items to be verified through the R language;
the fraud prediction module is used for predicting the company level audit quality based on the Mscore model through the R language to generate a probability estimation result of financial fraud occurrence;
the audit failure prediction module is used for generating an audit failure probability estimation result based on the audit failure prediction model on the company level audit quality;
and the processing module is used for generating an audit quality determination report through the R language based on the probability estimation result of financial fraud and the audit failure probability estimation result.
Specifically, the audit quality calculation module calculates the historical cost item ratio and the reasonable part ratio of the items to be verified based on the historical cost items and the items to be verified, and the sum of the historical cost item ratio and the reasonable part ratio of the items to be verified is the audit work quality; calculating the ratio of the items to be verified based on the items to be verified, and calculating the Euclidean distance of the items to be verified of the sample company in the multi-dimensional space, wherein the Euclidean distance of each item ratio of the items to be verified is the judgment level of the auditing specialty; and calculating the auditing effort degree, wherein the auditing effort degree is | management layer performance forecasting deviation |/| the audited actual profit |, and the management layer performance forecasting deviation is the difference between the forecasted net profit and the actual net profit.
Specifically, the fraud prediction module constructs a financial fraud early warning model based on the Mscore model, and increases auditing work quality, auditing professional judgment level and auditing effort degree on the basis of the Mscore index; calculating the probability of making the first type of errors and the probability of making the second type of errors through Logistic regression, selecting a value with the minimum expected cost as a threshold value according to the cost of making the first type of errors and the cost of making the second type of errors, and predicting the probability of financial fraud of a company on the market, wherein: the probability of making a first type of error is the probability of predicting a fraud company as a normal company, and the probability of making a second type of error is the probability of predicting a normal company as a fraud company.
Specifically, the audit failure prediction module is based on a pure financial index audit failure prediction model, on the basis, the audit working quality, the audit professional judgment level and the audit effort degree are increased, the probability of making a first type of errors and the probability of making a second type of errors are calculated through Logistic regression, and then according to the cost of making the first type of errors and the cost of making the second type of errors, a value which enables the expected cost to be the minimum is selected as a threshold value to predict the probability of the company on the market that the audit failure occurs, wherein: the probability of making a first type of error is the probability of predicting a fraud company as a normal company, and the probability of making a second type of error is the probability of predicting a normal company as a fraud company.
The system of the embodiment of the invention constructs a measuring index system of the audit quality in a multi-level and multi-dimensional manner from the perspective of the audit output. Therefore, compared with a single audit quality substitute index, the method has the advantages that the substantial attribute of the audit quality is more directly and comprehensively disclosed from a wider visual angle and returns to the characteristics of the audit subject. In addition, the accuracy of the conventional financial fraud prediction model and the accuracy of the audit failure prediction model are improved by extracting the financial statement information, and a corresponding decision report can be formed for use as a common audit reference.
Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable storage medium, and the storage medium may include: a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic or optical disk, or the like.
In addition, the above embodiments of the present invention are described in detail, and the principle and the implementation manner of the present invention should be described herein by using specific examples, and the above description of the embodiments is only used to help understanding the method and the core idea of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (8)
1. A method of auditing a quality determination, the method comprising:
extracting the deduction type detail data and the associated transaction type detail data based on the company financial statement and the supplementary notes thereof;
based on the accounting measurement attribute of the company, the balance sheet item and the profit sheet item are divided into a historical cost item and a to-be-verified item again;
calculating company level audit quality based on the reduced value type detail data, the associated transaction type detail data, the historical cost type project and the to-be-verified type project;
predicting the company level audit quality based on an Mscore model through an R language to generate a probability estimation result of financial fraud occurrence;
generating an audit failure probability estimation result by using an R language to base the audit quality of the company level on an audit failure prediction model;
and generating an audit quality measurement report through the R language based on the probability estimation result of the financial fraud and the audit failure probability estimation result.
2. The method of auditing quality determinations of claim 1, wherein said calculating a company level audit quality based on a reduced value class detail data, an associated transaction class detail data, a historical cost class item, and a class item to be verified comprises:
calculating a historical cost item ratio and a reasonable part ratio of the to-be-verified item based on the historical cost item and the to-be-verified item, wherein the sum of the historical cost item ratio and the reasonable part ratio of the to-be-verified item is the auditing work quality;
calculating the occupation ratio of the items to be verified based on the items to be verified, and calculating the Euclidean distance of the items to be verified of the sample company in a multi-dimensional space, wherein the Euclidean distance of the occupation ratio of each item to be verified is the judgment level of the auditing specialty;
and calculating the auditing effort degree, wherein the auditing effort degree is | management layer performance forecasting deviation |/| the audited actual profit |, and the management layer performance forecasting deviation is the difference between the forecasted net profit and the actual net profit.
3. A method of auditing quality determinations as claimed in claim 2, wherein said generating a probability estimate of financial fraud occurring based on an Mscore model prediction of company level audit quality in the R language comprises:
constructing a financial fraud early warning model based on the Mscore model, and increasing auditing work quality, auditing professional judgment level and auditing effort degree on the basis of the Mscore index;
calculating the probability of making the first type of errors and the probability of making the second type of errors through Logistic regression, selecting a value with the minimum expected cost as a threshold value according to the cost of making the first type of errors and the cost of making the second type of errors, and predicting the probability of financial fraud of a company on the market, wherein: the probability of making a first type of error is the probability of predicting a fraud company as a normal company, and the probability of making a second type of error is the probability of predicting a normal company as a fraud company.
4. A method of auditing quality determinations as in claim 2, wherein said generating an audit failure probability estimate based on a company level audit quality by R language based on an audit failure prediction model comprises:
on the basis of a pure financial index audit failure prediction model, the audit work quality, the audit professional judgment level and the audit effort degree are increased on the basis, the probability of making a first type of errors and the probability of making a second type of errors are calculated through Logistic regression, and then a value with the minimum expected cost is selected as a threshold value according to the cost of making the first type of errors and the cost of making the second type of errors, so that the probability of audit failure of a listed company is predicted, wherein: the probability of making a first type of error is the probability of predicting an audit-failed company as a non-audit-failed company, and the probability of making a second type of error is the probability of predicting a non-audit-failed company as an audit company.
5. A system for auditing quality determinations, the system comprising:
the data processing module is used for extracting the reduced value type detail data and the associated transaction type detail data based on the company financial statement and the supplementary notes thereof;
the project classification module is used for reclassifying the asset liability statement project and the profit statement project into a historical cost project and a to-be-verified project based on the accounting measurement attribute of the company;
the audit quality calculation module is used for calculating the company level audit quality based on the subtractive detail data, the associated transaction detail data, the historical cost items and the items to be verified;
the fraud prediction module is used for predicting the company level audit quality based on the Mscore model to generate a probability estimation result of financial fraud occurrence;
the audit failure prediction module is used for generating an audit failure probability estimation result based on the audit failure prediction model on the company level audit quality;
and the processing module is used for generating an audit quality measurement report based on the probability estimation result of financial fraud and the audit failure probability estimation result.
6. The system for auditing quality determinations of claim 5, where the audit quality calculation module calculates a historical cost class item duty ratio and a reasonable part duty ratio for the class item to be verified based on the historical cost class item and the class item to be verified, the sum of the historical cost class item duty ratio and the reasonable part duty ratio for the class item to be verified being an audit work quality; calculating the ratio of the items to be verified based on the items to be verified, and calculating the Euclidean distance of the items to be verified of the sample company in the multi-dimensional space, wherein the Euclidean distance of each item ratio of the items to be verified is the judgment level of the auditing specialty; and calculating the auditing effort degree, wherein the auditing effort degree is | management layer performance forecasting deviation |/| the audited actual profit |, and the management layer performance forecasting deviation is the difference between the forecasted net profit and the actual net profit.
7. The system for auditing quality determinations of claim 6, wherein the fraud prediction module constructs a financial fraud early warning model based on an Mscore model, and increases auditing work quality, auditing professional judgment level and auditing effort level on the basis of Mscore indicators; calculating the probability of making the first type of errors and the probability of making the second type of errors through Logistic regression, selecting a value with the minimum expected cost as a threshold value according to the cost of making the first type of errors and the cost of making the second type of errors, and predicting the probability of financial fraud of a company on the market, wherein: the probability of making a first type of error is the probability of predicting an audit-failed company as a non-audit-failed company, and the probability of making a second type of error is the probability of predicting a non-audit-failed company as an audit company.
8. An audit quality determination system according to claim 6 wherein the audit failure prediction module increases the audit work quality, the level of professional judgment of the audit and the level of the audit effort based on a purely financial index audit failure prediction model, calculates the probability of the first type of error and the probability of the second type of error by Logistic regression, and predicts the probability of the company on market that will have an audit failure based on the cost of the first type of error and the cost of the second type of error by selecting the value that minimizes the expected cost as the threshold, wherein: the probability of making a first type of error is the probability of predicting an audit-failed company as a non-audit-failed company, and the probability of making a second type of error is the probability of predicting a non-audit-failed company as an audit company.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011086884.9A CN112200647A (en) | 2020-10-12 | 2020-10-12 | Audit quality determination method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011086884.9A CN112200647A (en) | 2020-10-12 | 2020-10-12 | Audit quality determination method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112200647A true CN112200647A (en) | 2021-01-08 |
Family
ID=74008554
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011086884.9A Pending CN112200647A (en) | 2020-10-12 | 2020-10-12 | Audit quality determination method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112200647A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107492032A (en) * | 2017-09-18 | 2017-12-19 | 发贵科技（贵州）有限公司 | A kind of business financial auditing reports management system and its database platform |
| CN108764689A (en) * | 2018-05-22 | 2018-11-06 | 深圳供电局有限公司 | A kind of audit project verification appraisal procedure and system |
| CN109214907A (en) * | 2018-11-16 | 2019-01-15 | 合肥大能信息科技有限公司 | A kind of financial audit system Internet-based |
| CN110796335A (en) * | 2019-09-18 | 2020-02-14 | 广东电网有限责任公司审计中心 | Audit project implementation quality monitoring system and equipment based on data analysis |
| CN111652426A (en) * | 2020-05-29 | 2020-09-11 | 广州大学 | Accounting information comparability determination method |
| CN111695721A (en) * | 2020-04-28 | 2020-09-22 | 胡立昂 | Accounting information comparability determination method, application method and device |
-
2020
- 2020-10-12 CN CN202011086884.9A patent/CN112200647A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107492032A (en) * | 2017-09-18 | 2017-12-19 | 发贵科技（贵州）有限公司 | A kind of business financial auditing reports management system and its database platform |
| CN108764689A (en) * | 2018-05-22 | 2018-11-06 | 深圳供电局有限公司 | A kind of audit project verification appraisal procedure and system |
| CN109214907A (en) * | 2018-11-16 | 2019-01-15 | 合肥大能信息科技有限公司 | A kind of financial audit system Internet-based |
| CN110796335A (en) * | 2019-09-18 | 2020-02-14 | 广东电网有限责任公司审计中心 | Audit project implementation quality monitoring system and equipment based on data analysis |
| CN111695721A (en) * | 2020-04-28 | 2020-09-22 | 胡立昂 | Accounting information comparability determination method, application method and device |
| CN111652426A (en) * | 2020-05-29 | 2020-09-11 | 广州大学 | Accounting information comparability determination method |
Non-Patent Citations (1)
| Title |
|---|
| 王杏芬等: "会计师事务所内外部治理与审计风险——来自中国上市公司的经验分析", 《山西财经大学学报》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chi et al. | The effects of audit partner pre-client and client-specific experience on audit quality and on perceptions of audit quality | |
| Bhat et al. | The implications of credit risk modeling for banks’ loan loss provisions and loan-origination procyclicality | |
| Amiram et al. | Financial statement errors: Evidence from the distributional properties of financial statement numbers | |
| Huang et al. | Value-at-risk analysis for Taiwan stock index futures: fat tails and conditional asymmetries in return innovations | |
| Engelmann et al. | The Basel II risk parameters: estimation, validation, and stress testing | |
| Dias | Market capitalization and Value-at-Risk | |
| Stefanescu et al. | The credit rating process and estimation of transition probabilities: A Bayesian approach | |
| Li et al. | Capital reallocation and adverse selection | |
| Alexander | Statistical models of operational loss | |
| Hutauruk et al. | The Factors Impact On Earnings Management On Indonesia Mining Company | |
| Veronica et al. | Using cash flow ratios to establish a manufacturing bankruptcy prediction model | |
| Rahim et al. | Re-thinking: risk governance? | |
| Adasi Manu et al. | Climate policy uncertainty and corporate leverage policies | |
| Siagian et al. | The determinants of financial report restatement with audit quality as moderating variable: Manufacturing companies listed in Indonesia Stock Exchange: Financial restatement | |
| Askarzadeh et al. | Managerial sentiment and predicted and opportunistic special items | |
| CN112200647A (en) | Audit quality determination method and system | |
| Daromes et al. | Firm Value Effect of Managerial Ability and External Control Mechanism: Mediation Role of Financial Reporting Quality | |
| Yi | Corporate governance, information disclosure and investment-Cash flow sensitivity | |
| Melastiani et al. | The Effect of Cash Flow Volatility, Sales Volatility, and the Operating Cycle on Earnings Persistence | |
| Jitmaneeroj et al. | Time matters less: Variance partitioning of return on equity for banks in Uganda | |
| CN112163942B (en) | Method and system for measuring reliability of accounting information | |
| ALBAYRAK | The Firm Specific Capital Structure Determinants of Turkish Firms Listed in Istanbul Stock Exchange: A Structural Equation Approach | |
| Fadjar et al. | Analysis calculation of allowance for impairment losses of credit before and after the implementation of Psak 50 & 55 on profit at bank X | |
| Carvalho et al. | A stochastic approach for measuring the uncertainty of claims reserves | |
| Kosasih et al. | Effects of Profitability, Financial Leverage, and Dividend Policy on Income Smoothing Practice with Firm Size as a Moderator among Listed Manufacturing Companies in Indonesia |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210108 |