Over the last decade, welfare programmes in India have proliferated with a particular eye to making delivery more efficient. This essentially means that politicians seek to improve last-mile delivery by using technology-enabled solutions that reduce leakages, discretion, and favoritism.

The Economist estimates that this “high-tech welfare safety-net” includes nearly 300 schemes that have benefited 950 million people, most of them poor, and accounts for 3% of the gross domestic product, or GDP, in government spending. Political observers believe this “new welfarism”, with its emphasis on rule-based, direct transfers and use of “JAM” (jan dhan bank accounts, adhaar biometric identification, and mobile phones), has won votes and political support for ruling parties.

But does efficient welfarism bring with it the political dividends that many assume it does? The impact of welfare programmes hinges on two interlinked questions: do voters reward politicians for delivering benefits? And do voters care about process, namely how those benefits are delivered? In other words, do voters reward efficient implementation that reduces corruption and favoritism?

Go­­­­­vernance reforms often attempt to reduce stealing or corruption by minimising discretion and favoritism. Such efforts assume public disdain for corruption. Yet, our lived experience reveals widespread expectations of favoritism from family, kin, or coethnics, and occasional stealing from the state.

A humorous portrayal of the latter is shown in films like Katiyabaaz. The former is unmistakable in our caste-based politics. In my two years of fieldwork, journalists, politicians, and ordinary voters frequently described the expectation, “hamara hai hamare liye karega” (she or he is one of us and will do something for us). As one politician summarised, “jati svarth se judi hai” (caste is linked to material self-interest).

So where, then, do voters stand on the tradeoff between efficient implementation and favoritism along ethnic, political, or geographic lines? I argue, based on evidence from my research, that while voters in India reward good performance, they are more focused on outcomes than efficient implementation. Voters place a modest premium on efficiency. Far from the picture painted in our newspapers, voter appetite for governance reform (specifically, better implementation) seems limited.

Experiments involving distributive policy

To get at these questions, I fielded an online survey (1,047 voters) and a nationally representative telephone survey conducted in 12 different languages by CVoter India (5,350 voters). Within each survey, voters were randomly assigned to read either negative performance information on the incumbent government, positive information about one of its welfare programmes, or positive information about a welfare program that emphasised measures taken by the government to reduce corruption.

The negative condition emphasised that “India’s economy is not doing well”, pointing to historic unemployment, fuel inflation, and growing income inequality. The positive condition stated the number of people who have benefited from a welfare programme (Ujjwala or Pradhan Mantri Awas Yojana). The efficiency condition added to this by mentioning specific steps taken by the government to reduce fake claim-making (biometric authentication), corruption and discretion (direct benefit transfer to bank accounts, geotagging of assets and others).

The promise of such designs is that if the researcher randomly assigns people to read or listen to different pieces of information, the resulting groups of people are identical, on average, with the only difference being the type of information they read or listen to. This allows the researcher to say with greater confidence that differences in their political preferences are caused by the type of information, and not some other factors.

In the study, I measured political preferences in two ways: I asked voters to rate the Narendra Modi-led government’s performance on a 0 to 10 scale (higher values indicate more favorable assessments) and to hypothetically donate 10 rupees to major political parties in their state (where people could choose to “donate” nothing to any of them or split up to 10 rupees among the parties). I was interested in the amount notionally “donated” to the Bharatiya Janata Party.

What the survey shows

I found that voters’ preferences do move in response to positive performance information. Specifically, voters, on average, gave the Modi government’s performance a 6 out 10 when primed to think about historic unemployment, growing inequality, and fuel inflation.

They evaluated the government more favorably (just over seven out of 10) after being told about welfare programmes; the government’s rating improved by 1.14 points or roughly 11 percentage points of the full scale. When it comes to donating, voters gave, on average, 3.9 out of 10 rupees to the BJP after encountering negative information. They donated 28 paisa (0.28 rupees) more on learning about welfare programmes. This is approximately 2.8 percent of the available budget.

What jumps out from these results? First, voters seem to favorably evaluate the BJP government even when their attention is focused on bad macroeconomic performance. This tells us something about the limits of performance-based appeals to voters: they work but can’t be the only thing going for a party. Second, attitudes stated in a survey can overstate support for the BJP but behavioral measures give a more accurate picture.

Turning to efficient implementation, do voters care how benefits are delivered? To get at this, I compared the impact of additional information that mentions steps taken by the government to reduce corruption and discretion. What does the additional information do? Not all that much. Evaluations of government performance increased by 0.19 points, or a meagre 2 percentage points of the full scale. Donations to the ruling party increased by 29 paisa, or approximately 3 percentage points of the available budget.

Put together, there is a clear political pay-off to delivering welfare benefits but only a modest premium for doing it “efficiently,” at least to the bystander. Indeed, in separate research, I find this is the case for actual beneficiaries of the Pradhan Mantri Awas Yojana, for affordable housing.

Governance, policy implications

The upshot of this analysis is that a lot of governance reforms aimed at reducing leakages, discretion, and favoritism are politically viable if voters value process (that is, how a benefit is delivered to people) to a greater extent. Currently, public discourse on corruption is limited to high profile cases and is focused on making an example out of influential individuals engaging in corruption or nepotism.

There is, however, a need to deepen this discourse. Donors and experts focused on policy design and last-mile delivery need to qualitatively assess how citizens view efficiency-enhancing measures and whether they see any value in them.

Public information and awareness campaigns need to highlight the deleterious impacts of favoritism and corruption more explicitly and point to the gains from efficient implementation. A case for efficiency needs to be made, not assumed. Until there is pressure from voters, the political incentives to improve governance and undertake governance reforms will remain weak.

Shikhar Singh is a Center for the Advanced Study of India Postdoctoral Research Fellow. His Twitter handle is @shikhar_46.

This article was first published on India in Transition, a publication of the Center for the Advanced Study of India, University of Pennsylvania.