Mukesh Arya* SAI - India
* The author was involved in conceiving, designing, implementing and monitoring of the unique project of auditing a tax amnesty scheme, Voluntary Disclosure of Income scheme-1997, under the direction of the Comptroller & Auditor General of India. The work involved marshalling huge resources across the country with a very strictly defined time frame and the need to produce an exhaustive report for the Parliament of India on audit findings.
This paper traces a unique project of database creation from widely dispersed documents in a non-computerized environment facilitating an audit investigation with serious forensic implications. The comprehensive audit findings were possible due to the innovative use of the information technology. "Voluntary Disclosure of Income Scheme -1997", a tax amnesty scheme, introduced by the Government of India to unearth undisclosed income generated 475,000 declarations having 1.2 million pages. These could be audited through using imaging technology for capturing the images of the data contained in manually filled up declaration forms, conversion of images into RDBMS database and application of macro and micro analysis of the database after establishing linkages to other data-warehouses. The project reckoned computer non-literates to fill up the specially designed input sheets which were used for imaging. Conclusions as to widespread misuse, frauds, issues of public accountability for manipulation of legislative intent, unexplained lacunae in the amnesty scheme and complicity of the Executive that enabled large-scale evasion of taxes through the very amnesty scheme that sought elimination of such evasion could be precisely identified. The findings provided the Government with the alternatives for corrective actions in the tax assessments and tax reforms on a large scale. The project provides a framework to developing and developed countries to avoid pitfalls in the framing of legislation and implementation of similar schemes.
Spurred by the need to unearth undisclosed income and to bring into tax-net the evaders that dot the economy more often than not with holes large enough to enable, enhance and enthuse a parallel economy, the Government of India offered, the fifth time in 20 years, a tax amnesty scheme to the tax defaulters. The initiative was to provide one more opportunity to the tax evaders to declare their undisclosed income and to adopt the path of rectitude and civic responsibility and also to mobilize resources for the country to invest funds into priority sectors of the economy. The declarants were assured of immunity from prosecution and penalty and more importantly, complete secrecy about their declarations.
The scheme was operational for six months of July to December 1997. The tax defaulters were required to fill up a very simple declaration form and some basic documents in support of the assets declared. The entire internal revenue department was geared up for the success of the scheme reinforced by sustained media campaign that sought to instill mixed feelings of fear, patriotism and relief. However, the executives made a number of concessions in the scheme in the days the scheme was operational at the press conferences. Certain circulars and notifications that ripped apart the shield of the scheme were major concerns for the Auditor General as these had exposed drafting lacunae in the legislation for major abuses by the unscrupulous tax dodgers. A number of circulars were clear cases of manipulation of legislative intent. Some other events definitely pointed towards fraud like situation and misuse of the scheme.
The declarations were secret but accessible by the Comptroller & Auditor General of India. The number of declarations totaled up to 475,000 and the declared amount was Rs. 3336 millions. Sampling technique, normally used for obtaining audit assurance and confidence, is always irrelevant in a situation of forensic audit of presumptive fraud. The state of affairs called for 100% audit checks.
The Commissioners of Income Tax were under instructions to keep the declaration forms in the safe vaults at 121 locations. Any leakage of information could result in contempt petitions in the court apart from ridicule to the Parliament for breach of promise on secrecy. The Declaration forms were so designed that these did not reveal much, as they were very simple and devoid of detailed information. Other records were necessary to be linked and analyzed to obtain meaningful information.
A project for creation of a computerized database was so designed that the security concerns were addressed without compromising the essential needs of audit. Audit access to the secret forms was restricted to authorized personnel for the minimum period of time to avoid any adverse possibilities. Innovative application of IT enabled audit in handling highly sensitive and well guarded information in the most secured way without compromising the confidentiality and process them with utmost speed and competence.
Sensitivity of data also necessitated a single focal point, diverse database structure, generic software and coded information. A project of such complexity and dimension was unique for any Supreme Audit Institution and indeed was an experiment not found undertaken earlier anywhere.
At the end, audit concluded that the Scheme was not in the interests of revenue and in fact it provided an uncalled for opportunity to the dishonest assessees to pay tax on their concealed income at the preferred rate and then retire to the old habit of concealing income.
The report generated intense debate in the media and the official circles and is presently under examination by the Public Accounts Committee of the Parliament of India. Several Public Interest Litigations have also been filed in the Supreme Court / High Courts.
1.0 Why create a computerized database?
Creation of computerized database (RDBMS) was decided upon for several reasons besides the security concerns.
1.1 Non-computerized environment
In the non-computerized environment, inevitably, the transactions have to be recorded only in the manual files, These are not susceptible for easy manipulation or accessible in a variety of formats for analysis. The Indian income tax department presented a non-computerized scenario wherein the details contained in the declaration forms and other traditionally filed information relating to the tax payer were dispersed all over the country and in an array that presented difficulties of record linkages.
1.2 Existing database may not be relevant
Given that non-computerized environment may also be possible in developed countries where the need for data is for specific purposes and not capable of easy inference from the existing database, a structured questionnaire was used to evaluate the Indian scenario. The exercise confirmed that the records relating to assessments of taxpayers were located at several places and files for different types of taxes for the same taxpayer, such as wealth tax, estate duty, income tax, corporation tax were dealt by different assessing officers, not easily amenable to linkages in the manual format.
1.3 Need to link with other database
Typically, success of a taxman to detect the tax dodgers depends to a large extent on how well he correlates other financial and non-financial information. An auditor has to be a notch above that if he has to proactively track down the exceptions in fraud like situation. Several provisions in the Scheme required that the taxman look at other information before granting immunity to the declarants. For example, the scheme was not available to those who were involved in narcotics, drugs, economic offences, etc. Specific circular from the Executive during the period of the Scheme prevented the assessing officers from looking into the antecedents of the declarants. Moreover, no database of such persons was provided to the assessing officers to check such disabilities. These were major concerns for the Audit. A computerized database of declarations turned into a necessity for audit to establish such linkages prompted by the absence of such exercise by the tax department in several areas of concern.
1.4 Financial data alone is not sufficient
Assessment of returns of income and particularly of a tax amnesty scheme cannot depend wholly on the financial data alone. Questions such as whether the same person availed of the earlier amnesty schemes or whether the declarant is a large business house subjected to thorough tax scrutiny every year or whether the assets declared are fairly valued and tax offered or whether the declarant is a criminal not entitled to the scheme, are relevant for appraisal of the implementation of the scheme. Unarguably, non-financial information is as important to understand the success or otherwise of the scheme as the financial information. Computerized database provides the most vital linkages to disparate records swiftly and effectively for audit analysis and conclusions.
1.5 IT platform necessary for quick results
Assessment records for last five years in each case, linkages with other national and local data-warehouse, 475,000 declaration forms and many attendant information, were needed to be cross-checked and pin pointed as audit evidence for drawing conclusions. A traditional audit would involve painstaking exercise to isolate objectionable cases from the large population of declaration forms and then collect evidence thereon. Given the enormous amount of information spread over several parts of the country, this entailed massive exercise involving large human resources and finances over a long period of time. Cent percent audit check in a manual audit presents one with the possibility of missing the wood for a tree. Computerized database could provide the solution for these dilemmas, given the budget and time allocated for the purpose. A common IT platform was quintessential for quick results.
1.6 Data integrity, validation and conformity
Predictably, there were similar or same names with details that appeared to be matching in all respects and yet one would make a mistake if these were taken at their face values. It was essential that we performed a number of checks on parameters that conclusively proved the identity of an assessee for evidence purposes. Data integrity, validation and conformity tests assumed greater dimension in fraud like situations and had implications that were serious in each individual case. In the normal manual checks, it would have taken ages for just a preliminary list of exceptions on certain audit parameters. A computerized database could provide this preliminary list in an amazingly short time frame.
2.0 Project Outline
The project could be broadly grouped under following phases.
Phase I Creation of structure of input sheet for data capture.
Phase II Filling up the input sheets
Phase III Scanning of input sheets, conversion into database and linkages with other data-warehouses.
Phase IV Macro analysis of the data and substantive audit tests on the aggregated and disaggregated data to support conclusions through micro analysis, writing of the Audit Report and presentation to the Parliament of India
2.1 Human Resources Dynamics
Human resource is the key to the success of the project. There was a paramount concern for the secrecy of the information contained in the declaration forms, particularly of the details regarding the declarants. In fact, the declaration forms were made available to the Auditor General on the specific understanding that the secrecy will be maintained. One of the options for creation of a database was to provide each of the 800 odd auditors with a laptop for direct data entry of information from the files. This was not practicable for three reasons, (a) most of the auditors were computer non-literates (b) heavy investment in the hardware and software for a single project was unviable from cost angle and, (c) there was risk of losing the soft data in transit or otherwise, impacting on the perception of threat to security of the information.
The project as conceived and designed for implementation took care of the strengths and weaknesses of the audit personnel. Pre-project training-cum-workshops for the auditors at 19 locations throughout the country within 15 days and continuous monitoring of their work after the launching of the project simultaneously at 60 sites was found essential for improved results consistent with the audit objectives. Human dynamics are as important in a project of this nature as the synergy takes a variety of forms to lead to the desired goals.
The division of work ensured that each auditor could access information in parts only and that too within a tight schedule for each declaration form. As a result, one would be as wise as one of the blind men expressing opinion about the shape of an elephant.
Additionally, each auditor was required to acquaint himself of the government instructions on maintaining of confidentiality of information and the implications under the conduct rules for the government employees.
2.2 Selection Of Technology
The project was information technology centric. Imaging technology was found available in the market with several variants, essentially for capturing images from the paper media. Some specific applications were also found using imaging technology in the conduct of objective type examinations and compilation of results quickly. These are very simple applications. Nevertheless, this information was a vital cue in the selection of software. None of the software available, including the one finally procured, was ever tested for a complex project of this nature. Several pre-pilot tests, re-configuration of software with the help of the vendors, customization of the software for audit needs and the testing of the technology in its entirety on a pilot project had to be undergone before the project was fully launched.
Figure 1 - Format of Input Sheet
3.0 How was database created?
PHASE I
3.1 Creation Of Structure Of Input Sheet For Data Capture
Specially designed input sheets containing about 125 fields were used for collection of information. Imaging technology was customized and adopted to capture the specially designed input sheets.
The most critical component of the imaging technology was the designing of the input sheet (Figure 1). The boxes were provided for each field and the information was captured through shading of the boxes. These sheets were scanned and images executed through special application software (Batch Scanning system from Newgen Software Technologies limited) in batches and the resultant data was converted into RDBMS, in this case the generic Microsoft Access software.
3.1.1 Structure Of Input Sheet
Extraction of information from the declaration forms, the returns of income tax and wealth tax for last five years and the relevant folios in respect of each assessee was to be done by the computer non-literates and therefore the input sheet had to be designed in a way that facilitated easy transcription and the information automatically codified. It was essential that the structure of the input sheet is kept as simple as possible for those transcribing the information by shading the relevant boxes. However, it had to be complex one for someone to decipher it for quick comprehensive information. Again, it had to be scanner friendly and be able to afford checks for internal consistency of information. These were complex issues.
What finally emerged was a combination of coded information on a two-page form uniquely numbered. This unique identification was used for all the other linked databases, for example that of names and addresses database. The Sheet provided for ‘registration’ marks on both the sides and the software for conversion of images into database was set for relative distances from the boxes for default values. Relevant information and the fields were identified on the input sheet for capture of information quickly and efficiently. A specimen of the finalized input sheet is displayed below.
3.1.2 Fields for data input and data validation
Certain parameters for audit evidence required collection of details that were either numerical and or characters. Logical fields (Yes/No) were preceded by questions whereas numerical fields had to be marked in the boxes assigned for the values expected. To assure data validation, enforced handwritten figures just above the boxed markings were provided for.
Whereas we could provide for numerical fields with boxes within a particular length on the page, it became impossible to contain information as to names, addresses, designation and the verifier of the declaration form. Pilot study proved that the handwritten characters were not capable of error free conversion through the Optical Character recognition (OCR) component of the software. This was a major limitation of the imaging technology.
Recognizing this drawback, manual input of the information in a separate database was planned for with the same unique identity for each record in the RDBMS. This OCR limitation was effectively converted into strength of the project whereby manual input of names and addresses in the computers was centrally executed through a select highly trusted individuals in a time bound programme. This data was kept out of bounds for all except two in the audit department thereby effectively combating the threat to the secrecy of names of the declarants
3.1.3 Coding of fields
Figure 2 - example of coding
Length of fields could not be manipulated beyond a point and given the two-page limit; Input sheet was designed with maximum coding that was possible. Separate booklets were printed to indicate codes for different fields. This ideally facilitated the security concerns too.
An extract of one field is shown below as an example of the coding done to extract information relating to “buildings” declared under the scheme that may relate to any of the assessment years beginning 1962-63 to 1997-98, a span of 36 years!
B1,B2 etc, for example, indicated the different building properties declared whereas X2 codified the assessment year 1972-73. ‘Sp’ signified the grouped years.
3.1.4 Numerical fields
Figure 3 - example of numerical field
Essential information as to the amount declared by the assessee, amount of tax paid, dates of payment, Permanent Account Number, the charge of the Commissioner of Income tax, the circle and the ward number and the status of the assessee could be captured through the numerical fields. The shading of the boxes was both easy and mechanical for most computer non-literates. In fact they were not required to touch the computer at all.
Following fields, for example, relating to the date of declaration and the amount declared have the audit parameters built in. The information effectively calculated, for example, the interest payable for delayed payment of tax or the amount of tax payable at particular rates of income tax.
It may be seen that boxes were not provided for the ‘year’ under the Date of Declaration since the year was only 1997. Dates can never be more than 31 and therefore no box beyond 3 in the first column. These measures effectively reduced the chances of transcribing information wrongly
3.1.5 Logical fields
Certain information was gathered from the folios of each individual assessee on a test check basis. For example, it was prudent to ascertain the sectors of the economy that generated maximum unaccounted income or similarly the professions that have remained outside the detailed scrutiny of the income tax department. All these fields were coded for saving the space on the input sheet and for easy conversion into database. Logical fields were also used for checking the internal consistency of data collected.
Figure 4 - fields for logical questions
Field No.13 was devised to provide several logical questions. Either shading the box or leaving it blank could indicate a Yes or No answer to each field. This composite field is reproduced below for appreciation of the leading questions that an auditor would pose for his usual probe.
PHASE II
3.2 Filling Up The Input Sheet
3.2.1 Input sheets were printed centrally with numbers to identify them uniquely. All 1.2 million pages were distributed to the 19 field offices and further transported to 121 locations. The logistics for distribution of forms was decided upon in advance. The filled forms were re-transported to the headquarters at New Delhi with adequate security for every trunk box.
3.2.2. The whole exercise of filling up of forms was planned in such a way as to ensure minimum quantum of work with highest standards of quality. Instruction booklets were issued to each auditor that illustrated “How to fill the Input Sheet” with examples. The booklet also prescribed the codes for different fields. Supervisory staff was instructed on how to review the work of subordinates for quality assurance.
3.2.3 The declaration forms were obtained from the Commissioners of Income tax daily and returned to the safe vaults at the end of the day. No telephones were permitted in the room in which this exercise was carried out. No member of the team was allowed to carry papers into or out of the room. Entry of outsiders including that of the staff of the income tax offices was prohibited. The room was sealed at the end of the day and reopened by the same officer next day. This was done at each of the 121 locations to ensure secrecy of information. The work of collecting the information on input sheets was completed within a period of two months.
PHASE III
3.3 Scanning Of Input Sheets, Conversion Into Database And Linkages With Other Data-Warehouses
3.3.1 Scanning of input sheets
Heavy-duty scanners were deployed at the secured centralized location in headquarters at New Delhi where all the input sheets were transported and then scanned using the customized Software for a full-scale application.
The software could have been customized either for batch processing or continuous processing. Batch processing for scanning was preferred to the continuous mode for reasons of privacy, safety, security and simultaneous conversion of images to database on other computers for validation and for increased interface for data integrity.
3.3.2 Batch scanning software
Batch Scanning System was so integrated as to scan a large volume of pages, perform page management operations, collation of the batch pages into documents and associating indexes to the pages. Six high-speed scanners from Fujitsu with a speed of 15 pages per minute were used. The scanned images could then be exported as .exp or .exd files. The system provided following features for the whole operation.
3.3.2.1 Job and Batch definition
Several batches are contained in a job that needed to be defined. Jobs were created for each state having several batches in the order in which the input sheets were received from field audit parties. Keeping tab on batches for security and privacy reasons was accomplished through specially designed protocol.
3.3.2.2 High speed scanning
This is the crux of the technique used. The software enabled simultaneous viewing of the images on the monitor to check quality of scan and the interface between the scanner and the computer.
3.3.2.3 Document separation
Figure 5 - Flow Chart of the Project
Information for each assessee was marked on a two-page input sheet. This, then, constituted a document for one record. The scanning was defined with the separator after every two pages and for additional assurance; identifiers were incorporated for linkages in the RDBMS.
3.3.2.4 Page management
Page management system enabled insertion, deletion, interchange, rotation and autodeskew the pages scanned. This required the operator to be alert to rectify the exceptions, as they occur to avoid loss of time subsequently.
3.3.2.5 Document marking
Marking of document for its bad quality was always an integral part of the project to ensure highest level of quality assurance. Slippages in marking documents as bad led to avoidable additional scanning work in some instances.
3.3.2.6 Index entry and association
Though the application provided for manual entry of indexes to each document, we relied on the zone execution for automatic generation of indexes.
3.3.2.7 Zone definition and execution
Defining zones in a scanned image is the most important task in the project for picking up values from the marked locations. Zones could be defined as OCR, Bar code, MICR or the OMR. Once having clearly defined and coordinated the zones, the entire batch could be executed for results.
3.3.3 Statewise grouping of images
All the images were grouped by the 19 major states of the Union of India that facilitated controls over the receipt of input sheets in batches from the field offices. The grouping provided a series of trails that were unique for a particular region or class of assessees. It also enabled the back office to continuously check on the quality of work done by the field audit parties. The internal consistency parameters were invoked at several levels in order to effectively upgrade the quality of output from field offices. This provided the highest possible data integrity.
3.3.4 Secrecy issues
Scanning operations were executed on six machines installed at Delhi headquarters for eight hours a day. Evidently, no operator could work beyond five hours at a stretch and two shifts were the necessity. Batch processing enabled masking the work done by operators from each other. The software additionally ensured camouflaging of identities through assigning secret codes and passwords for each operator. The final execution of the batches was entrusted to two specific persons on the project, thereby ensuring that sensitive data had highly restrictive access.
3.3.5 Accuracy of the data at scanning stage
Having scanned and executed the batch, the application enabled view of all the pages. Since we had incorporated handwritten figures above the boxes for marking on the Input sheets, visual checking of figures with that of scanned results provided additional assurance on the accuracy of the data.
3.4 Conversion of images to database
Zone creation and execution of the batch produced the default values assigned to each marked location. These documents in batches were then exported along with the associated data to an office application. These were saved in ASCII format. Thereafter, Microsoft Access was used to convert the ASCII data into database.
Each record in the database had the unique identifier. An All-India database of all information extracted from the declaration forms was captured electronically from a non-computerized environment. This database did not have the names and addresses and therefore privacy of the individual Tax assessee was ensured.
3.5 Database of Names and personal details of the declarants
A separate exercise was carried out to create a database (475,000 records) of names and addresses and personal details of the declarants by direct input into specially prepared software in MS-Access. Each record had the unique identifier to link with the All-India database mentioned para above. This database was securely kept away from the other databases and used selectively to identify individuals for evidence purposes and for performing other substantive audit tests.
This database proved very useful while providing linkages with national data warehouse of different agencies such as police, National Crimes Record Bureau, Central Economic Offences Wing, etc.
3.6 Auditors’ notepad entries
Profession of auditing dictates one to be inquisitive and one must provide for this duty in any audit project. Input sheet obviously could not provide for all conceivable facts in all cases. An auditor as a professional would apply his acumen to follow through the leads that may be vital for audit conclusions. Each of the 800-odd auditors was given a 100-page notepad to jot down their findings. Certain exceptions could be foreseen for certain population of the declarants, whereas some could not be and were left to the ingenuity of the auditors. The auditors were instructed to invariably use the unique identifier for record linkages in other databases. These notepad entries were subsequently converted into a separate computerized database for exception reporting on certain individuals, by category of assesses or by class. It is relevant to mention that this information was manually culled out from folios, local authenticated knowledge, etc., but always from a non-computerized source.
3.7 Valuers’ information
The Scheme enjoined upon the declarants to value their assets, particularly jewellery, gold, silver and precious and semi-precious stones supported by authorized Valuers’ certificates. Media reports suggested misuse of the scheme through these certificates of valuation. A separate computerized database was prepared from the information contained in the records of Income tax department relating to approvals granted by the tax authorities at various locations to various jewelers over a number of years. These were cross-checked with relevant fields of the All-India database to establish whether the certificates of valuation in each case were issued by the approved valuers. Detection of few cases of money laundering through jewelers’ stock-in-trade of gold, silver, etc., was the spin off of this exercise.
3.8 Past Assessment records information
The department maintains a folio on every tax payer at least for last 10 years. Conclusions as to the track record of taxpayers are as interesting as declarations under the Tax Amnesty Scheme. A database of the important financial and non-financial information was created in respect of the top 12,000 declarants. Data linkages on these parameters expectedly contributed a number of audit conclusions for corrective actions, for example non-filing of wealth tax returns in subsequent assessment years for the wealth declared under the scheme, multiple declarations by the same tax payers, ineligible declarants, tax certificates issued without payment of tax or interest.
3.9 Other National database
The scheme had laid down a number of conditions for validity of the declarations. Most importantly, those involved in criminal cases, economic offences, drugs and narcotics and similar offences were not eligible for the tax amnesty scheme. Preliminary audit enquiry revealed that the tax department had not cross-checked on these aspects before issuing the certificate of amnesty on declared amount. Collecting these databases from various investigating agencies required consent for sharing the data and provision for their secrecy.
Accessing of information by tax officers from investigating agencies, though implicit in the provisions of the Scheme, was never undertaken in the absence of any enabling mechanism. Secrecy assurances from Audit and effective liaison for data download from different types of database formats resulted in forming audit conclusions that were both revealing and shocking. For example, a number of high volume declarations from those involved in large financial scams under police investigation could be established beyond doubt with clinching evidences and that made out the case for immediate withdrawal of immunity from prosecution.
Internet information and commercially available databanks on companies, corporations and partnership firms were obtained. This contained excellent quality of information that is normally available in the public domain sources. Stock prices, annual financial accounts, production and inventory details, schedule of assets, etc., in respect of companies listed on the stock exchanges of the country proved immensely useful for record linkages.
Strategically, an essential audit check in regard to any third party databases is to test them as to completeness, correctness and integrity of data to provide for high quality of reliable evidence. Procedures were established to assess these parameters and results obtained through linkages of records from suspect sources were ignored for evidence purposes.
PHASE IV
4.0 Macro Analysis Of The Data And Substantive Audit Tests, Micro Analysis, Writing Of The Audit Report
4.1 Macro Analysis Of Database
A decision was taken to deploy Microsoft Access as a generic tool for macro-analysis of database and for linkages of records in other database. Options of Oracle or SQL were not cost effective and user friendly for the stand-alone project of limited duration.
A few important conclusions that could be drawn through macro-analysis are briefly stated below.
4.2 Micro Analysis Of Database
Unquestionably, database analysis cannot completely replace substantive audit tests or traditional auditing techniques. Having obtained major results from macro analysis of database, it was inevitable to apply sampling technique for in-depth audit. Considering the size of the population of 475,000 declarations and geographical locations; segmental sampling technique was applied for the 121 Commissioners of Income Tax. Additional details such as wealth tax paid by the assessee during the last 10 years, income returned during last five years, ghost declarations, money laundering, track record of these assesses, were probed into for evidence. Had there been no database, focusing on to these high value assesses with supporting broad based analysis and leads would not have been possible except perhaps in respect of a very small sample. The main argument of the affected adversaries to the audit findings through small sample would have been that the conclusions are not representative of the entire population. Use of the information technology effectively countered this argument.
Main audit conclusions based on substantive audit checks and subsequent findings through micro-analysis of data are as follows.
5. Bottlenecks
Every project throws up challenges. This project was not an exception.
Input sheet preparation was most challenging since all conceivable information had to be visualized beforehand and provided for. There was no provision for insertion of additional fields subsequently. A single box missed in one field necessitated reprinting of 1000 input sheets at the printing stage.
Pilot project could throw up the possibility of generation of incorrect values if an inappropriate marker pencil was used for shading the boxes. Pre-project training took care of this small technology specific problem, which had larger dimensions as to data integrity.
Person specific bottlenecks could be found in the spelling of different kinds of names from different regions of the country. Auditors’ notepad entries on the other hand had entries that could not be standardized beyond a point.
6.0 Conclusions
Audit of a tax amnesty scheme that involved high security and public concerns for privacy in a non-computerized environment was made possible through the innovative use of information technology. Data extraction or mining could be launched through the computer non-literates and analysis of a national database could be attempted successfully by linking other data-warehouses, whether government or those in the public domain, for arriving at conclusions. These provided a framework for launching nation-wide reforms in tax administration. The report is also the beacon to other developing and developed countries and forewarns about the pitfalls in the framing of legislations and implementation of similar future schemes.
Awareness of the potential of the information technology for e-governance and more specifically for the public sector auditor in securing public accountability needs to be promoted. This pioneering project highlights the use of IT for time and cost-effective audit methodology in a fraud like situation and is relevant for all countries.
Significantly, this methodology is also relevant for the Executive, Legislature and the Judiciary in countries that have lagged behind computerization but wish to critically appreciate the outcomes of their important decisions and functioning through database analysis.
7.0 Acknowledgments
I am grateful to the C&AG of India for entrusting this project and for continuous encouragement in the audit endeavours with effective guidance at critical times. Assistance of several officers and audit staff all over India and the software vendors contributed to the successful completion of the project.
8.0 References
1. Report of the Comptroller & Auditor General of India for the year ended March 1999 “Voluntary Disclosure of Income Scheme, 1997” Union Government (Direct Taxes) No.12A of 2000
2. Batch Scanning System, User’s Guide M/s., Newgen Software, Technologies Ltd, New Delhi