The 1991 turn

In June 1991 the Indian government could pay for about two weeks of imports. By the first week of August it had airlifted gold to London and Zurich, devalued the rupee twice, and published a budget that removed industrial licensing from most of the economy. The speed is what people remember. The more useful question is what actually changed on the ground, because the answer is narrower and more durable than the drama suggests.

What changed was not the amount of state in the economy. Public spending did not collapse in 1991 and the public sector did not shrink. What changed was who held the right to say no. Before July, a proposal to build a factory went to a licensing office and waited for a clerk to decide whether the capacity was allowed to exist. After July, for most industries, it went to a bank.

How a financing problem became a liquidity problem

The crisis built up over two years rather than arriving in a week. Through 1990 the fiscal deficit widened, inflation ran high, and the government borrowed short-term at high rates to cover a growing current-account gap. Then the Gulf war pushed the oil bill up and cut remittances from Indian workers in the region, at the moment when credit was already short. Reserves fell while the obligations against them got shorter, which is the mechanism that turns a financing problem into a sudden stop.

By June 1991 usable reserves were around a billion dollars against imports of roughly two billion a month. The government restricted foreign exchange, slowed letters of credit for imports, and then did the thing that made the news: it sent gold to be pledged with the Bank of England and with Swiss banks, raising a few hundred million dollars against it. A stand-by arrangement with the IMF followed in October, which is the part that mattered for the following decade, because the conditions attached to it set the direction of the reforms.

Two things are usually folded together here and should be separated. The gold pledge was a bridge: it bought weeks, not years. The IMF programme was a framework: it committed the government to a set of changes it might not have made on its own timetable. The budget of 24 July sits between them, written by people who assumed the programme was coming.

Two devaluations in three days

On 1 July the rupee was devalued, and on 3 July it was devalued again. The two steps together moved it by close to a fifth against the dollar. The reason for doing it twice is instructive: the first step was meant to be enough, the market disagreed, and rather than defend a rate the authorities moved again. It was the first clear signal that the exchange rate would be set by the market's opinion rather than by the government's preference.

A devaluation is a transfer, and it is worth saying from whom to whom. It makes exports cheaper and imports dearer, which helps producers who sell abroad and hurts anyone whose costs are denominated in dollars — importers, travellers, anyone buying fuel or machinery. It also reduces the real value of domestic debts. None of this is a free lunch; it is a decision about who absorbs the adjustment, taken in a hurry because the alternative was not paying for oil.

The budget that removed the licence

Industrial licensing began with the 1951 Industries (Development and Regulation) Act. Its logic was physical: a planning authority decided how much capacity each industry should have, and a licence was permission to build that capacity. Over four decades the list of what required permission grew, and so did the discretion of the officials administering it. The cost was not mainly corruption, although there was that. The cost was time and scale: a plant sized to a licence rather than to a market, and a queue measured in years.

The policy of 24 July 1991 abolished licensing for most industries, kept it for a short list, cut the number of sectors reserved for the public sector from seventeen to eight, and opened most of the rest to foreign investment through an automatic route. The budget the same day cut import duties and began dismantling the tariff structure that had made domestic production profitable regardless of its efficiency.

What the reform did not touch

The reforms of 1991 and the years after were concentrated on industrial licensing, trade, and capital flows. They left the parts of the economy that determine whether a factory can actually be built largely alone: land title and acquisition, labour law, electricity pricing and distribution losses, and agricultural marketing. Those are the constraints a manufacturer meets after the licence is gone, and they were not in the package.

This is not a footnote. It is the explanation for the shape the growth took. The sectors that grew fastest after 1991 were the ones that needed skills and capital and little land: software services, telecoms, finance, pharmaceuticals. The sector that needed land, power and flexible labour — manufacturing — grew more slowly than the planners of 1991 expected, and employed far fewer people per unit of output than the East Asian pattern had.

The bill arrives on a schedule

The short-run result was a recovery. Reserves rebuilt, the current account stabilised, and growth moved to a higher average through the 1990s. The long-run result is a set of deferred costs that are easy to list and hard to price. Public investment in agriculture and irrigation fell as a share of output while subsidies rose; state electricity boards accumulated losses that were later converted into debt; and the fiscal deficit, after a brief improvement, returned to a level that absorbs a large part of domestic savings.

The honest way to describe 1991 is not as a decision to shrink the state. It is as a decision to stop deciding one particular thing — whether a given capacity may exist — and to keep deciding many others. That is why the reform was both larger than its critics said and smaller than its admirers said.

Why this is a good case for the pillar

Every entry in this pillar has a stated reason and an actual one. Here the stated reason was the balance of payments: the country was weeks from default and had to do something. The actual reason the changes lasted is different. Once the licensing office no longer stood between a firm and its investment, the constituency that had an interest in the old system lost its gate, and the constituency that benefited from the new one grew every year. Policies survive when they create their own supporters, and this one did, in a decade.

The crisis was real, the response was improvised, and the outcome was neither the collapse that was predicted nor the transformation that was promised. It was a change in who says no — which is the sort of change that outlasts the government that makes it.

This is one of the Modern India entries in About India. The pillar's argument — that each of these systems has a stated reason and an actual one — is on the pillar page.