GIGO-why data standardisation, cleaning, validation, reconciliation tools, end to end process and SOP are key to digital initiatives

This is the time of the year in India when taxpayers (individuals, corporates) are preparing for the all important annual ritual - filing of income tax returns. I also downloaded my records of AIS TIS 26AS from the Income Tax portal to match with the records that are available with me. Imagine my shock- the capital gains calculated as per AIS TIS is 4 times my actual capital gains (I wish it were true - I wont mind paying the additional taxes if I had really made so much capital gains !) . The mathematician/Computer Scientist in me got to the job to figure out the reasons. Here are the some of the key observations :

• There is possibly lot of mismatch between the data reported by the reporting entities - as reported to me as user/customer and possibly to the regulator.
• If X shares were bought and y sold during the year (y <x), normally the closing stock of x-y should show up as a closing stock. Capital gains should be computed only on the y shares sold during the year. However the tax portal shows capital gains on the complete x shares.
• Shares where the purchase price was not available (possibly because the shares were bought long back and that historical data is not available on the tax portal) - purchase prices of such shares seems to have been taken as ZERO and capital gains computed accordingly. In programming 101 we were taught that NULL value (data not available) is not the same as ZERO - but this fundamental seems to have been forgotten. I would love to get shares at zero cost, sell in the market, pay 15% STCG and pocket the difference - if life were that easy!. Shouldn't there be a logic which catches records with zero cost, validates the reason for zero cost (bonus shares, stock split or similar corporate actions!) and accordingly flag such records in a different bucket (Error bucket ? ). And the logistics of ingesting the historical records from the reporting entities can always be explored going back say n no of years (apply 80:20 rule or similar) or if too cumbersome at least flag such records in a different bucket instead of short term capital gains. Blind reporting leads to loss of faith in the reports itself and can cause more harm than good.
• Likewise in borderline cases where the dates are close enough (367 days) Long term capital gains has become short term.
• Many reporting entities have reported the same data multiple times. Some have taken the trouble to rectify the multiple entries by corresponding reversal entries - others have not taken the trouble. One large private bank has reported and reversed the same transaction 25 times - yes 25 times - of course they have been kind enough to reverse the duplicates. Another large Housing finance company which merged with a large Private bank - dividend paid once has been reported twice by each of the entities -how I wish I had actually received the money twice.
• While I could export the securities transactions data to an excel file for further analysis, there is no way to export dividend data to compare with our own bank records . One needs to compare manually record by record - a painful process - I found out the necessity of comparison when my Chartered Accountant pointed out duplicate reporting when the dividend has been paid only once!.
• First time I am looking at the AIS TIS 26AS data with so much detail - maybe I am making a mistake and not reading the data correctly/properly - would be grateful if any of the readers of this article can point out my mistake. But if you are also experiencing similar data issues, please do share your views.
• We are also expected to look at the data presented by AIS TIS and report back for any erroneous records. Without any easy data export, reconciliation mechanisms how is it even manually possibly to look at each data item and report for errors ? Normally one would first look for 20% of the items with 80% of the impact - or the A B C rule. However since one cannot export and analyze easily across the spectrum of AIS TIS data this is not possible to do - at least I could not figure out ! There is a need for tools (duly supported by proper data export from AIS TIS portal) so that tax filers / tax practitioners can reconcile the data with a 360 degree view in a data driven way.
• If the correctness of the data reporting by the reporting agencies and the ingestion by the tax portal can be made more robust and accurate, the annual tax filing exercise will become so much less painful and possibly will also further enhance the revenues.
• The data reporting by various data reporting entities and the data presentation to the end users need to be more organized and amenable to analysis - online and offline. At present when I was looking at our own data it was in pretty much random order - no logical organization of the data could be inferred by me. And there was no way I could do it myself - only for the securities transactions once I managed to export to excel I could sort, organize, group etc - however again reconciliation with data from the reporting entities was manual since every entity has decided to organize and present data in their own way - with ZERO standardization.
• Considering the importance of AIS TIS 26AS both for the revenue authorities as well as the taxpayers is it not incumbent to have some discipline in the data reporting - and should the reporting entities not be penalized for careless and incorrect reporting ! On top of it I am given to understand that the reporting entities can keep revising the data (possibly by paying some penalties if the due date is passed!) - not sure if this is correct not being a tax expert. I am supposed to file my tax returns which is expected to be in line with AIS TIS 26AS. How do I do that if the reporting entities keep changing - seen major updates in AIS TIS 26AS every few days myself.
• Any comments or suggestions welcome - is it just me or is a common issue ?
• Needless to say the fundamental principles of Information Sciences around ensuring proper data validation while ingestion (formats, ranges, business rules) seem to require strengthening here as in any large system design and rollout !

印孚瑟斯 Axis Bank HDFC Bank Income Tax Department Of India Indian Revenue Service - Income Tax Institute of Chartered Accountants of India The Institute of Chartered Accountants of India (ICAI) ClearTax Tax2win ?? Taxmann KDK Softwares India Private Limited Indian Administrative Service (IAS) - Government of India nasscom Debjani Ghosh Alumni Cell, IIT Kharagpur Prof.Ravikumar Bhaskaran Anand Deshpande Persistent Systems Salil Parekh N. Chandrasekaran Kris Lakshmikanth K Krithivasan Abhay Karandikar Subhasis Chaudhuri Indian Institute of Management Ahmedabad Indian Institute of Management, Calcutta Indian Institute of Management Bangalore

Ranjeet Kumar Agarwal FICCI Confederation of Indian Industry Nishith Desai Associates Rohit Jawa Sanjiv Puri Amitabh Chaudhry Neil Borate