In August 2014, the newly-elected National Democratic Alliance (NDA) administration announced the reintroduction of the previous government’s then-suspended ‘Direct Benefit Transfer for LPG’ (DBTL) programme to distribute LPG subsidies by bank transfer, while retaining (against the recommendations of both the petroleum and finance ministries, and despite extensive media speculation) the annual cylinder cap per household first instituted (but twice raised) by the United Progressive Alliance (UPA) administration at its existing level. 

There is a long-standing consensus, both within and outside government, on the need to reform India’s cooking gas subsidies. In the last financial year, the Indian government spent Rs. 48,378 crore (US$7.8bn) subsidizing Liquefied Petroleum Gas (LPG) – Rs. 8,756 crore (US$1.4bn) more than the equivalent central budget allocation for primary education, and Rs. 15,378 crore (US$2.5bn) more than the allocation to the flagship National Rural Employment Generation Scheme (NREGS) public employment programme. The bulk of this subsidy transfer accrued to wealthier households in urban areas, with the majority of the population in the bottom two-thirds of the income distribution scale (who currently use highly-polluting and inefficient solid fuels) receiving little or no direct benefit.

The stated objectives of LPG subsidy reform are typically threefold: to contain total subsidy outlay, to distribute existing subsidy transfers more progressively, and to increase clean fuel access and use among poorer households. Any policy intervention with the specified purpose of reforming LPG subsidies should therefore be judged against three interrelated criteria: the quantity of fiscal consolidation delivered; the quality of any fiscal consolidation; and the impact on clean energy access and use. 

As currently designed, DBTL meets none of these criteria. 

DBTL is not a ‘cash transfer’ program

Understanding the current mechanism of LPG subsidy entitlement and disbursement, and the government’s proposed changes, is key to understanding why DBTL does not constitute subsidy reform. Under the existing system of LPG subsidy provision, households who possess a registered connection at an LPG distributor (to which every household in the country is theoretically entitled, but half of the population do not possess) order a subsidized LPG cylinder and pay the subsidized price in cash on collection or delivery. Under the revised (DBTL) system, households order an LPG cylinder from their LPG distributor, receive a payment equivalent to the current subsidy amount via electronic transfer to a bank account registered with their gas distributor, withdraw this subsidy amount, then pay the full (unsubsidized) market price for the cylinder in cash. DBTL therefore simply represents a shift in the modality of subsidy payment  – the scheme does not decouple receipt of subsidy from fuel consumption (subsidy receipt is contingent on purchase of LPG), nor does it apply any form of targeting in selecting beneficiaries (retaining the highly regressive distribution of the existing subsidy). 

Despite this, DBTL is frequently mis-characterised as a ‘cash transfer’ programme, effectively conflating it (intentionally or otherwise) with Latin American social programmes such as Bolsa Familia or Oportunidades. DBTL is emphatically not a ‘cash transfer’ program — in technical terms, DBTL is a ‘cash-assisted in-kind transfer’, with subsidy receipt entirely dependent on fuel purchase. In its current form, the program is in effect a social transfer contingent on purchasing power; the wealthier the household (and therefore the higher the propensity to consume LPG) the greater the amount received, with non-consuming households — including the majority of the poorest half of the population — receiving nothing.

DBTL will not deliver fiscal consolidation

In addition to retaining the existing (highly regressive) structure of subsidy entitlements, the expansion of the DBTL program is likely to result in little or no net fiscal gain relative to the current system, and will introduce a series of previously non-existent costs into the process of subsidy disbursement that will generate a net loss both to individual beneficiaries and to the wider economy. Projections of a notional fiscal gain resulting from the introduction of DBTL are largely based on an extrapolation of short-run reductions in consumption immediately following the scheme’s imposition in pilot districts, with the entirety of the difference between prior and post-DBTL subsidized LPG consumption levels conceptualised as ‘leakage’. This is highly misleading, both because it fails to accurately account for (and offset) the direct and indirect fiscal costs of introducing the program, and because it misrepresents the nature of any reduction in subsidized LPG consumption achieved.

Alongside the extensive (and largely unquantified) expenditure on system infrastructure and administration, the government is now providing all consumers with an increased fixed cash advance upon registration for DBTL in response to the widespread disruption of household finances witnessed during the initial launch of the scheme. This cash advance, currently set at Rs. 568 per household, is equivalent to three and a half months of subsidized consumption at present rates (Rs 161.81 per cylinder as of 1st Feb 2015), thereby offsetting any possible short-term fiscal gain through disruption of subsidized consumption. All subsidy transfers under DBTL are also subject to a 2% commission payment, further increasing the cost of delivery relative to the previous embedded subsidy approach.

In addition to the direct fiscal costs associated with DBTL, the program also imposes a series of discrete and recurring costs on beneficiaries. Beyond the initial (and often sizeable) costs to consumers associated with scheme registration, a significant proportion of the estimated 120-140 million LPG-consuming households now have to undertake recurring — and previously unnecessary — physical bank transactions; as frequently as once a month for withdrawals alone, with additional visits for transfer confirmation. This represents a substantial (and unrecognised) net cost both to the consumer and to the economy overall, and one that disproportionately affects more marginalised and poorer consumers (as they are typically those with the most limited ability to access formal banking services). The government’s apparent reliance on fuel subsidy transfers to effectively cross-subsidize expenditure by public sector banks on the flagship Jan Dhan Yojana (JDY) financial inclusion program potentially introduces an additional layer of dysfunction, obscuring the actual cost of providing banking services and creating a further constituency for the continuation of LPG subsidies in their current form.

DBTL primarily disrupts eligible consumption, not ’leakage’ 

The substantial direct and indirect costs of introducing DBTL, where recognised, are frequently justified on the grounds that the scheme delivers an equivalent or greater reduction in LPG subsidy ‘leakage’ (often undefined). There is, however, no evidence that the majority, or even a substantial percentage, of the initial 10-15% reduction in subsidized LPG consumption associated with the introduction of DBTL in pilot districts is the result of the targeted removal of illegitimate consumption (in the sense of over-quota household consumption, or direct diversion to non-household use). Instead, experience from pilot districts indicates that the introduction of DBTL causes widespread disruption of legitimate household consumption through a variety of mechanisms.

DBTL’s inability to deliver substantial savings through targeted reductions in undesired subsidized LPG consumption is understandable. As with the existing method of disbursement, DBTL has no effect on post-purchase diversion (which would be rapidly and effectively curtailed by a lower cylinder cap) – under the direct transfer system, a tea-stall owner who currently consumes seven subsidized cylinders per annum for household use will continue to be free to use the remaining five cylinders of her subsidized allocation for ‘commercial’ use. In addition, DBTL will not yield significant savings through the removal of invalid connections and associated (illegitimate) consumption, for two reasons. Firstly, all domestic LPG consumption was subsidized prior to the introduction of the per-household cylinder cap in September 2012, and there was no effective limit on the number of cylinders that could be purchased per registered connection. Households therefore possessed little or no incentive to obtain and maintain multiple connections, and consequently the large majority of functioning (i.e. non-dormant) residential connections was essentially legitimate. Secondly, any de-duplication exercise is inherently reliant on existing systems of connection registration, and the public-sector Oil Marketing Companies (OMCs) have already undertaken extensive (and administratively simple) validation exercises on the existing connection databases both before and after the institution of the cylinder cap to identify and verify inactive or duplicate domestic LPG connections, resulting in the suspension of a reported 13.3 million connections by November 2012. 

Aadhaar is irrelevant to LPG subsidy rationalization

Importantly, DBTL’s marginal scope for reducing ineligible consumption through de-duplication is unrelated to the invalidation of the program’s mandatory linkage with the UPA administration’s ill-conceived Aadhaar scheme (constrained by the  Supreme Court’s order that enrolment in the scheme could not be made mandatory for the provision of public services, the NDA administration has modified DBTL to remove the compulsory requirement for Aadhaar registration, while retaining Aadhaar linkage as ‘the primary option’). 

Linkage with Aadhaar — a program established by executive order, with no formal legislative standing, which was entirely absent from the new government’s pre-election manifesto — was always irrelevant to the functioning of the DBT mechanism (even for the narrow purpose of removing duplicate connections) due to the household-based nature of the entitlement, and instead represented an attempt to instrumentalize the existing subsidy transfer in order to compel enrolment. The removal of mandatory registration simply reduces DBTL’s administrative costs for both government and beneficiaries, however the retention of any association with the program remains highly problematic, as the threat of benefit provision or withdrawal continues to be habitually used at the field level to compel ‘voluntary’ Aadhaar registration (in direct contravention of the Supreme Court’s guidance).

DBTL will inhibit access to cleaner cooking fuel by the poor

The disruption to eligible consumption caused by introducing DBTL has two implications: that a significant proportion of any immediate post-implementation drop in subsidized consumption is temporary, and will rebound as excluded households recover access; and that a proportion of any remaining reduction will relate to consumption by eligible households whose ability to access subsidized fuel has been durably disrupted or removed – a burden that is likely to fall disproportionately on lower and middle-income households. 

In addition to inhibiting access to subsidized LPG among connected households, DBTL also impedes LPG uptake by non-connected households — alongside the pre-existing financial and administrative barriers to access, particularly for the poor (beyond fuel and equipment costs, registration for an LPG connection is itself often difficult, requiring significant time and effort and, in some cases, informal payment), prospective beneficiaries now require a bank account, physical access to banking services on a regular basis, and a level of financial literacy necessary to operate the account. 

Beneficiaries in pilot areas also report that the switch to direct payment has specific (negative) implications for women unrelated to connection disruption: as the large majority of household gas connections and associated bank accounts are in the name of the male ‘head of household’, female household members (who are typically responsible for managing household budget expenditure) are reliant on timely and complete withdrawal and transfer of the subsidy amount in order to avoid destabilising household finances – a problem that persists even with the provision of an advance payment.

Reforming LPG subsidies: build on success, not failure

As in the later years of the UPA administration, the new government’s misguided emphasis on DBTL (and associated initiatives such as Aadhaar) has diverted valuable time, financial and institutional resources, and political capital from simple, equitable and cost-effective reforms to LPG subsidies — such as reinstatement of a realistic per-household cylinder cap, and institution of a comprehensive program of access extension — and come at a huge opportunity cost both to the poor and to the wider economy.

There is undoubtedly a case for a progressive shift towards electronic transfer of payments for many cash-based entitlement programmes including pensions, scholarships, and public employment programmes such as NREGS (which have been operating such transfers for several years through existing payment mechanisms), and for the creation of new or expanded (cash-based) social welfare programs. 

This case does not extend to the use of cash payments for unreformed in-kind subsidy transfers such as LPG, whether on the grounds of equity, administrative efficiency or fiscal responsibility. On the contrary, DBTL increases the administrative cost of delivering the subsidy, reduces the net value of the subsidy to the consumer, exacerbates the regressive distribution of subsidy expenditure, and will not meaningfully reduce total fiscal outlay. The NDA’s newly-appointed economic policy establishment is therefore faced with a clear choice: between continuing to support a costly and regressive program which entrenches the existing subsidy system, or charting an alternative path towards genuine subsidy reform.