Lessons from GST for eliminating the Rs1 trillion lying around as float

The government of India today spends around Rs3 trillion each year on centrally sponsored schemes. Though many explanations such as lack of political will, poor implementation capacity, etc. have been offered as reasons for poor outcomes, there is a larger structural issue: the government fund flow and payments architecture have got very little attention. 

The government’s current fund management system pushes funds down a hierarchical chain from the centre to the panchayat. Each level of government must first produce utilization certificates verifying that a majority of its previously disbursed funds have been used. When this is done, the level above the one requesting the funds releases a portion of the annual allocation down the chain. The effects of delayed fund transfers in one part of the chain therefore cascade through the entire system and in the process, amplify failure. For instance, if a district fails to receive its second instalment in time, all panchayats under it will be refused further instalments even though individually many of them may have fulfilled all eligibility conditions. 

This has led to the paradox of “unspent funds and unpaid dues”. According to an Accountability Initiative report on the National Health Mission (NHM), in 2015-16, only 68% of the total funds approved for NHM were actually spent. The story in India’s education sector is similar. A school-level expenditure-tracking exercise conducted in 300 schools across 10 districts in five states in December 2015 revealed that 31% of the schools were yet to receive their annual grants at the time of the survey, well into the academic year. 

Since funds are disbursed in bulk and in advance, unspent funds accrue in various government accounts. These idle funds, or float, put considerable strain on government finances. At present, there is no systematic tracking of float; however, estimates place this at over Rs1 trillion, i.e. Rs1 trillion of avoidable fiscal deficit. 

Ultimately, the burden is borne by beneficiaries who fail to receive entitlements on time, programme managers who spend a disproportionate amount of time chasing funds rather than ensuring implementation, government suppliers who cannot predict when their payments will come (and therefore increase prices) and contractual workers who do not know when they will receive payments. Interestingly, senior officers and staff ensure that their salaries are never delayed. 

The ministry of rural development, however, is an exception. In 2014-15, 70% of wages of MGNREGS (Mahatma Gandhi National Rural Employment Guarantee Scheme) workers were delayed by over 15 days. Of these, 13% encountered a delay of over 90 days. Delays occurred even though various levels of central and state governments had parked unspent funds. In August 2015, the ministry took cabinet approval to introduce direct transfer of wages from the Consolidated Fund of India to the accounts of MGNREGS workers based on an electronic fund transfer order generated at the panchayat or block level.

The cabinet proposal was informed by a 2012-13 study of MGNREGS in Bihar by MIT, Harvard and Oxford economists Abhijit Banerjee, Esther Duflo, Rohini Pande and Clément Imbert affiliated to J-PAL, a global research network of professors from around the world. The study found that linking the flow of MGNREGS funds to actual expenditure and reducing the number of intermediaries in the payment structure brought down programme expenditure by 24% without compromising workers’ employment or wages. The authors attribute this to a reduction in corruption. The Bihar intervention, if scaled nationally, was estimated to save MGNREGS Rs5,744 crore in 2012-13. 

The August 2015 change in architecture of MGNREGS was possible because it had a complete transaction-based, workflow-driven IT platform. The data created by such a platform significantly reduces the burden that programme managers face in maintaining accounts and pursuing requests for funds. Very few schemes that use IT have these features. Most use computerized MIS (management information system) that are largely used for data upload. Though many schemes have transitioned to the Public Financial Management System, it functions largely as a payment gateway. 

Unclogging the government’s financial pipeline requires expanding workflow- driven IT platforms akin to MGNREGS for all central sector, centrally sponsored and state government schemes. At their ideal state, such platforms would automate payment, releasing money once all conditions precedent have been fulfilled, allowing programme managers to pull funds daily and directly from the Consolidated Fund of India and/or the State Consolidated Fund and pay directly to worker/beneficiary/vendor bank account. Criteria for payment eligibility would be precoded into the platform, thereby eliminating discretion on when funds are released. 

This would mean zero advance release, implying zero float. Since payments would be made in real time on a voucher-to-voucher basis, departments, auditors and the public will be able to track expenditure on a real-time basis. 

In 2010, the government of Mexico initiated a centralized e-payment system for federal salaries, pensions and social transfers, resulting in $1.27 billion in savings. The goods and services tax (GST) in India is achieving a similar feat. GST relies on the GST platform that was built by GST Network (GSTN), a not-for-profit section 8 company jointly owned by the centre, the states and financial institutions. 

If we can do it for taxes, we should do it for expenditure. Utilizing the expertise that GSTN has gained in rolling out GST may be our best bet. But for this to happen, the department of expenditure should task GSTN with the job of building a modular workflow driven-IT platform for the government. Individual government departments at the centre and states can then customize and use it in the same way that businesses customize and implement enterprise cloud and on-premise software from entities such as Oracle, Ramco or Microsoft.

A. Santhosh Mathew is a former 1985 batch IAS officer from the Bihar cadre.