By Aasheerwad Dwivedi & Aditya Sinha, Dwivedi is Assistant Professor (Economics), Faculty of Management Studies, Delhi University; Sinha writes on macroeconomic & geopolitical issues
Policy is always made in the shadow of uncertainty. No government, however well-intentioned or technically equipped, can fully anticipate ex ante how millions of dispersed agents, firms, households, states, and markets will interpret, adapt to, or strategically respond to a new rule. Yet the difficulty of anticipation does not absolve governments of responsibility. It shifts the standard from omniscience to responsiveness. Governments fail not because outcomes diverge from intent, but because they mistake intent for impact. They often assume linear causality in a world governed by feedback loops, incentives, and second-order effects.
They model compliance where evasion is rational, predict behavioural change where institutional trust is weak, and expect uniform responses from heterogenous actors embedded in unequal capacities and information. Bureaucratic silos further distort anticipation. Policies are stress-tested against legal form rather than lived practice, against average agents rather than marginal ones. Above all, political systems privilege announcement over adaptation. Treating policy as a statement of resolve rather than a hypothesis to be tested. But this also leads to uncertainty in certain cases.
India’s recent policy experience increasingly reflects this pattern of reversal. In isolation, each rollback may be defended as responsiveness. But in most cases, it also points towards the issue with how government fails in anticipating a fallout from a policy intervention.
The most recent episode concerns the government’s decision to mandate the pre-installation of the Sanchar Saathi cybersecurity app on all mobile phones manufactured or imported into India, an order that was shortly withdrawn after a public outcry on grounds of privacy, consent, and surveillance. Sanchar Saathi itself is not the problem. The deeper concern lies in the process: a sweeping regulatory instruction affecting hundreds of millions of users was issued first and defended later, only to be reversed.
From “Sanchar Saathi” to Angel Tax
There are several other examples. Consider the Union Budget episode on capital gains taxation of real estate. In 2024, the government abruptly withdrew indexation benefits on property transactions, significantly altering the tax liability of millions of homeowners and investors. After widespread criticism, the decision was partially rolled back, offering taxpayers a choice between a higher tax rate with indexation and a lower flat rate without it. The damage, however, had already been done. Expectations were disrupted, transactions were stalled, and credibility was eroded.
In 2023, the Centre attempted to regulate laptop, tablet, and PC imports through a sudden licensing regime affecting companies such as Apple, Dell, and HP. Framed as a national security measure, the policy triggered immediate fears of supply shortages, price spikes, and production bottlenecks. Within months, the measure was quietly relaxed, after firms had already scrambled to restructure supply chains.
Even recruitment policy has not been immune. In 2024, the government floated advertisements for contractual appointments to senior bureaucratic posts, joint secretaries, directors, and deputy secretaries across 24 ministries only to roll them back amid political backlash.
Then there is the angel tax. Introduced in 2012, expanded controversially to non-resident investors in 2023, and abolished entirely in 2024, the tax regime moved in full circles within a single decade.
There are several reasons why this happens. The first possibility could be haste, policy being drafted and notified under compressed timelines. In such cases, regulatory design often runs ahead of institutional readiness, legal vetting, or implementation capacity. Predictably, resistance then forces retreat.
Second, most policies are designed using static, partial-equilibrium logic—assuming a stable mapping between rule leading to behaviour and outcome. In reality, economies and societies are complex adaptive systems characterised by feedback loops, threshold effects, and endogenous responses. A marginal regulatory change can trigger discontinuous reactions once actors coordinate expectations (as in capital gains taxation or import licensing). Governments tend to model first-order compliance effects while systematically underweighting second-order behavioural adaptations such as deferral, arbitrage, lobbying, exit, or legal contestation. The result is predictable policy overshoot followed by corrective rollback.
Third is structural information asymmetry. The state does not observe private adjustment costs, informal sector behaviour, or evasion strategies, nor can it precisely identify the counterfactual baseline absent intervention. Administrative data sets are compliance-focused, backward-looking, and lagged, making them weak instruments for anticipation. Actors with the greatest exposure to policy also have the strongest incentives to misrepresent likely responses during consultation, further degrading foresight.
Fourth, modern political systems reward speed, signalling, and decisiveness, not experimentalism or adaptive learning. Policies are framed as commitments rather than hypotheses, discouraging pilot phases, sunset clauses, or reversible design. Once announced, feedback is treated as opposition rather than information. This produces a paradox: governments become responsive only after reputational damage occurs, converting what could have been ex-ante learning into ex-post reversal, eroding credibility even when intent is defensible.
Policy uncertainty functions as an implicit tax on expectations. It raises discount rates, delays investment, and re-prices risk upward as firms, investors, and innovators shift from commitment to caution, asking not whether a policy is sound but whether it will endure. While democratic course correction is legitimate and often necessary, frequent and abrupt reversals transform flexibility into fragility, eroding credibility rather than improving outcomes.
India’s durable reforms such as the GST and Insolvency and Bankruptcy Code succeeded precisely because they were preceded by consultation, sequencing, and institutional buy-in; recent volatility stands in sharp contrast. The deeper danger is not uncertainty per se but the progressive weakening of reform signalling, when every announcement is discounted for rollback risk, even well-designed policies lose force. Sustainable growth therefore demands not just boldness in policy, but durability in design and execution.