Karnataka was seen as a pioneer when it deployed an algorithm to make sure public money goes to the right person despite one’s name not matching with documents such as Aadhaar for direct benefit transfer (DBT).
Name mismatch was a teething trouble citizens faced in getting government benefits. The algorithm negated the need to manually check names.
Now, at least six years after it was first put to use, the Comptroller and Auditor-General (CAG) has found “significant gaps” in the name-matching tool, raising questions over its “reliability and
integrity”.
The algorithm written by the Centre for Development of Advanced Computing (C-DAC) assigns a name match score between 0 and 100 comparing the name of the beneficiary as per Aadhaar with the name of the person whose account is actually credited with money under various government schemes. This is done post payment.
The CAG found that different name-match scores were awarded for the same beneficiary on different transactions “indicating inconsistency in the process”.
Sample this: A beneficiary named G Satyavathi received Rs 900 under the government’s milk incentive scheme. Her account name was Gubbal Sathyavathi Suryachandrarao. The name match score was 19. However, when she received Rs 1,100 under the same scheme, the name match score was 0.
The CAG audit analysed 22.91 lakh such transactions where the name match score was between 0 and 20 where chances of cash benefits going to a wrong beneficiary is higher.
Of these 22.91 lakh transactions, 77,148 records did not have beneficiary name, 3,210 did not have account holder’s name, 6,567 had no bank name, 3,211 did not have the bank account number and 5,610 did not have masked Aadhaar number.
“The above gaps indicate the lack of validation input controls and impacts the reliability and integrity of the name-matching exercise,” the CAG said. The government told the auditors that the mismatch could be “due to software updates and issues”, that the name match score is being constantly reviewed and changes are suggested to C-DAC. “Software updates affecting a process indicates the lacunae in the application,” the CAG said.
Senior IAS officer Munish Moudgil, an IIT Bombay alumnus, first wrote the name-matching algorithm in 2014, which was used by the Rajiv Gandhi Rural Housing Corporation Ltd and for the Parihara drought relief compensation.
According to sources, when the e-Governance department wanted the source code, Moudgil refused. So, the e-Governance department got it done from C-DAC, which the CAG says needs “corrective action”.