Where We Left Off
This is the largest item in the series and it has no scandal attached to it. Nobody took a bribe to write the Second Plan. That is exactly why it is worth eight decades of everybody's attention and has never had it.
Three parts of chronology. Now one part that steps out of it, because forty-one years of a single system cannot be seen a decade at a time.
Part Three covered the first decade of the republic: a constitution amended within fifteen months, the largest election ever held, land reform that passed everywhere and happened almost nowhere, the Hindu Code, the redrawing of the map, and the first elected state government dismissed from Delhi.
Four things from it carry into this part.
A minority of votes produced the power to change the rules. Congress never crossed half the vote in a general election, and first-past-the-post converted about 45 per cent of votes into about three-quarters of the seats — which is more than the two-thirds needed to amend the constitution.
The internal check disappeared between 1950 and 1954. Patel died, Tandon was forced out, Ambedkar resigned, Kripalani left. Nothing institutional replaced them, because nothing institutional had been doing the work. Chapter Four of this part is what that meant when the warnings started arriving.
The strongest test in this series is a government’s own stated target. Nobody can call the standard unfair when they set it. This part is built on that test more than any other, because the Five-Year Plans state targets in numbers and the outturns are published.
The binding obstacle to land reform was the record, not the landlord. Part Three found that a state which located 173 million voters could not locate its tenants. Hold that finding. It has a twin in this part.
One thing about the shape of this part before we start. It leaves the chronology. Parts Two and Three moved year by year; this one takes 1950 to 1991 as a single object, because the licence system is not visible in any single decade. It looks reasonable in 1955, defensible in 1965, questionable in 1975 and indefensible in 1985, and the only way to see that arc is to stand back from it.
How to read this book
The Six Boxes
Six coloured boxes run through the series, each doing one job. Here they are with examples from the material ahead.
A Word Box explains a hard word the moment it first appears.
Import substitution: a policy of making things at home instead of buying them from abroad, by taxing or banning the imported version until a domestic producer exists.
The logic is not stupid. A country that buys everything from abroad never learns to make anything, and needs foreign currency it may not have. Protect an industry while it grows, and eventually it can stand on its own.
The difficulty is the word “eventually”. A producer protected from competition has every reason to stay protected and no reason to become efficient, and he is a much better-organised lobby than the millions of consumers paying more.
Why it matters here: this was the strategy for forty years, and the question the chapters ask is not whether it was reasonable in 1950 but what happened when it stopped working.
An In Real Terms box turns a number too big to picture into something with a body.
India’s economy grew at about 3.5 per cent a year from 1950 to 1980. Population grew at over 2 per cent. So income per person grew at somewhere around 1.3 per cent a year.
At 1.3 per cent, an income takes about fifty-four years to double. A man who started work in 1950 would reach the end of his working life with his household about half as well off again as when he began.
At 7 per cent — the rate of several comparable countries in the same decades — income doubles in about ten years. The same working life covers four doublings.
That is the whole of this part in one comparison. Not a difference in percentage points. A difference between one doubling and four, inside one lifetime.
A How We Actually Know This box shows the physical evidence and then says what it cannot prove.
The best evidence in this part comes from committees the government appointed to investigate itself. The Monopolies Inquiry Commission of 1965. The Hazari committee on industrial licensing. The Industrial Licensing Policy Inquiry Committee under Subimal Dutt, which reported in 1969.
These are strong sources for a specific reason: they were commissioned by the government running the system, staffed by people inside the establishment, and reported findings that embarrassed it. A hostile critic’s estimate can be dismissed. A committee your own finance minister appointed cannot.
What they cannot show: what happened to their recommendations. A committee reports; it does not implement. Tracing whether a finding changed anything requires following the policy afterwards, which is Chapter Four’s job and is much harder than reading the report.
An Argument box appears where serious people disagree, each side at its strongest.
Part One set the rule: a charge needs an alternative that somebody proposed at the time, that was politically possible, and that somebody comparable actually did. Apply it to the founding choice.
In 1950 the export-led model did not exist as a template. The countries later held up as counter-examples had not yet done it, and when they did they did it under conditions India did not have: military governments that could impose unpopular decisions, small homogeneous populations, and enormous American aid and guaranteed market access granted for Cold War reasons. India chose non-alignment and got neither. Judging 1950 against a model that was demonstrated in the 1970s is exactly the hindsight trap.
Nobody serious argues the 1950 choice was wicked or stupid, and this series does not. The charge is about what happened after the evidence arrived. By 1967 the government’s own committees were reporting that the system was concentrating economic power rather than dispersing it. By the mid-1970s comparable countries were visibly pulling ahead. The response was not correction but extension: more controls, more nationalisation, tighter permissions. A bet you refuse to settle for a further twenty-five years has stopped being a bet.
What would settle it: the internal record — which warnings arrived in which year, and what happened to the people who wrote them. A good deal of it is public. Chapter Four uses it.
Why people care so much: because this is the argument about whether India’s poverty in 1990 was inherited from the British or manufactured after 1947, and both answers assign an enormous debt to somebody.
A Hidden Assumption box digs out a belief sitting underneath an argument that both sides accept without discussing.
Everyone who argues about the Indian economy before 1991 — defenders and critics alike — assumes that the country was too poor to invest.
The defenders’ version: you cannot build industry without capital, and a country with an income of a few hundred rupees a head has none, so slow growth was arithmetic rather than policy. The critics’ version is the mirror image: the state absorbed the little capital there was.
India’s gross domestic savings rose from around 9.6 per cent of national income in the 1950s to somewhere in the range of 20 to 25 per cent by the 1970s and 1980s. That is not the profile of a country that could not save. By the later part of this period India was saving at rates comparable to countries growing three times as fast.
So the question was never how much was put in. It was what came out. A rupee saved and invested in a plant that runs at half its capacity, making something at twice the world price, for a market that cannot afford it, has been spent, not invested.
This changes what the whole argument is about. If the constraint was capital, the story is a tragedy about poverty. If the constraint was the return on capital, the story is about allocation — about who decided where the money went and on what basis — and that is a story with decisions in it, and people who made them.
Both camps prefer the first version. It lets the defenders say nothing could have been done, and it lets the critics blame an abstraction called socialism rather than doing the work of finding out where specific rupees went.
And every chapter closes with a Remember This box: the chapter in the plainest words available.
One last note on how to read this part. It is long on argument boxes and short on villains, and that is deliberate. There is no bribe at the centre of this story and no moment where somebody knowingly did the wrong thing for gain. That is exactly what makes it the largest item in the series and the least discussed.
This part leaves the chronology and takes forty-one years as one object, because the system looks reasonable in 1955 and indefensible in 1985 and the arc is invisible a decade at a time.
Growth of about 3.5 per cent against population growth over 2 per cent gave income per head about 1.3 per cent a year — one doubling in a working life, against four in comparable countries.
India was not too poor to invest. Savings rose to around a fifth of national income. The question was never how much went in. It was what came out.
1The Bet
In 1950 a group of serious people made a serious decision for serious reasons. Almost everything in this part follows from what they built instead of what they intended.
1.1 — What was chosen
The problem in front of the government in 1950 was not complicated to state. India was poor, it made almost nothing, and it had very little foreign currency with which to buy what it could not make.
The answer was a plan. India would build, at state expense, the industries that make other industries possible — steel, heavy machinery, power, chemicals, cement. Private business would continue, but inside a framework where the state decided what got built, where, and how much.
The intellectual architecture came largely from P.C. Mahalanobis, a statistician, and it shaped the Second Five-Year Plan from 1956. Its logic is worth stating properly rather than caricatured, because it is coherent.
A poor country can spend its scarce investment on consumer goods, which makes people better off now. Or it can spend it on the machines that make machines, which makes nobody better off now but raises how much the country can build in twenty years. Mahalanobis argued for the second. Accept a generation of scarcity and you buy the capacity to grow permanently afterwards.
That is not a foolish argument. It is a recognisable argument about deferred gratification, made at national scale, by people who genuinely expected the payment to come due.
1.2 — Why it was reasonable at the time
Four things made this look like the sensible course in 1950, and a prosecution brief that hides them is not worth reading.
The intellectual consensus pointed this way. Not just in India. Planning was mainstream development economics. The Soviet Union appeared to have industrialised from peasant poverty in one generation, and the figures showing what that had actually cost were not available.
Britain was doing a version of it. Coal, rail, steel and health had just been nationalised by a Labour government. This was not an exotic choice.
The available example of an unregulated market was two centuries of colonial extraction, ending in a famine in Bengal within living memory. The people making these choices had watched it.
Private capital was genuinely thin. There were large Indian business houses, but not on a scale that could build a steel plant and a heavy engineering complex simultaneously, and they had shown limited appetite for trying.
And the First Five-Year Plan, from 1951 to 1956, largely worked. It concentrated on agriculture and irrigation and it exceeded its own growth target. That is a fact and it belongs here, because it is the reason confidence was so high going into the Second.
1.3 — The third thing that got built
Now the observation this chapter exists for.
India did not build a planned economy. It did not build a market economy. It built a third thing, and confusing it with either of the other two is why most arguments about it go nowhere.
In a genuinely planned economy — the Soviet model — the state owns the factories and directs production. It decides how much steel to make and instructs the plants to make it. The state is the producer.
In a market economy, private firms decide what to make, and succeed or fail according to whether anybody buys it.
India built neither. It left most production in private hands, and then required permission for the decisions those private hands would otherwise make. What to produce, how much, where, with what machinery, using how much foreign currency — each of these became something you applied for.
The state was not the producer and it was not a referee. It was a gatekeeper, and this is a genuinely different arrangement with genuinely different consequences.
Licence Raj: the popular name for the system of industrial permissions that ran in India from 1951 to the early 1990s. “Raj” here means rule or regime.
The term is used loosely and often as an insult, which obscures what it actually described. It was not state ownership; most of the firms were private. It was a requirement that private firms obtain official permission for the ordinary decisions of running a business.
Why it matters here: nearly every argument about this period is conducted as though the choice were state versus market. The thing that existed was neither, and Chapter Six shows that it produced outcomes that neither of those systems would have produced.
Once you see this clearly, the whole chapter’s argument follows, and it points at something both camps have always agreed on without discussing it.
Every argument about India’s economy before 1991 — from the left and from the right — assumes that the choice was between planning and markets.
The left says: planning was correct for a poor country and the reformers dismantled it. The right says: planning failed and markets were the answer. They disagree completely about which was better and agree entirely that those were the options.
India had neither. Under real planning, the state owns the plant and bears the loss when the plant is wrong — the decision-maker and the loser are the same body, which is a harsh but coherent arrangement. Under a market, the firm decides and bears its own loss. Both systems, whatever else is true of them, put the decision and the consequence in the same place.
A permission system separates them. The official who grants or refuses a licence bears no consequence for being wrong. The firm bears the consequence but did not make the decision. Nobody in the arrangement both chooses and pays.
That single structural fact predicts almost everything in the chapters ahead. It predicts why capacity was licensed and never built, why plants ran below capacity for decades without anybody being answerable, why the highest-return activity in the Indian economy became obtaining permission rather than producing anything, and why the system survived thirty years of evidence that it was not working.
It also explains why the standard argument is so sterile. Defending “planning” defends something India did not have. Attacking “socialism” attacks something India did not have. Both sides are arguing about a system neither of them lived under, and the actual system — the file, the queue, the sanctioned capacity — is discussed by almost nobody, because it is boring and has no ideology attached to it.
Everything in the next eight chapters is an elaboration of that one point. The system was a gatekeeper rather than a producer or a referee, and a gatekeeper is the only one of the three that can be wrong at no cost to itself.
The 1950 choice was reasonable on the information available: it was mainstream economics, Britain was doing a version of it, the available example of markets was colonial extraction, and private capital really was thin.
The First Plan concentrated on agriculture and exceeded its own target, which is why confidence going into the Second was so high.
India built neither planning nor markets but a permission system. Under both of the others, whoever decides also pays. Under this one, nobody does both — and that single fact predicts most of what follows.
2How a Permission Actually Worked
Not the theory. The desk, the file, the queue, and the man in Delhi with a rubber stamp. This is what forty-one years of Indian economic life actually consisted of.
2.1 — The Act
The legal foundation was the Industries (Development and Regulation) Act of 1951. It brought a long list of industries under central control and required a licence to establish a new undertaking in any of them.
The Industrial Policy Resolution of 1956 then divided industry into three groups: those reserved exclusively for the state, those where the state would lead and private firms might participate, and everything else. The list reserved for the state was long and included most of what a growing economy needs.
Over the following two decades the framework was extended. The Monopolies and Restrictive Trade Practices Act of 1969 added a further layer of clearance for large business houses. The Foreign Exchange Regulation Act of 1973 tightened control over foreign currency and foreign-owned companies, and led some international firms to leave India rather than dilute their shareholdings.
2.2 — What you actually had to obtain
Abstractions are useless here, so here is the concrete version. Suppose it is 1975 and you want to make bicycles.
You need a licence to set up the undertaking at all. The licence specifies your capacity — the number of bicycles you may make in a year. Making more than your licensed capacity is an offence, even if people want to buy them.
You need permission for where the factory goes. Location was a policy instrument, used to push industry towards backward areas, so your commercial judgment about proximity to suppliers and markets is subordinate to somebody else’s regional policy.
You need a separate capital goods licence to import machinery, which requires establishing that no Indian machine will do — and if an Indian producer says his machine will do, it generally will do, whatever you think of it.
You need an allocation of foreign exchange to pay for anything imported, from the machine to the components to the specialised steel. Foreign currency was rationed by the state and applied for.
If you want to change what you make — bicycles this year, mopeds next — that is a new product and needs new permission.
If you want to expand, that is an amendment to your licensed capacity, and is applied for.
If you are a large business house, you additionally need MRTP clearance, which is another authority with another queue.
And from the mid-1970s, if the business fails, you need permission to close it. Firms above a size threshold required prior government approval to retrench workers or shut down. A permission to exist and a permission to stop existing, from the same government, and in practice the second was harder to get than the first.
Imagine you want to open a tea stall. You apply for permission to open it, and the permission specifies that you may sell four hundred cups a day. On a hot afternoon the queue is longer. You may not serve them; four hundred is what you are licensed for.
You want a bigger kettle. The kettle is imported, so you apply for permission to import a kettle, and you must first prove that no Indian kettle will do. An Indian kettle-maker writes in to say his kettle will do. It is a worse kettle. You get the worse kettle.
You want to sell samosas as well. Samosas are a different product; that is a fresh application.
Now notice where your energy goes. Not into better tea. Into the application, the follow-up, the man who knows the officer, the second application. The tea is almost incidental to the business, and the business is the file.
2.3 — What this did to time
The most-cited feature of the system is delay, and delay is usually described as an inconvenience. It was worse than that, and the reason is worth understanding.
Applications took years. A businessman who saw an opportunity in 1972 might have permission to pursue it in 1976, by which time the opportunity had a different shape or belonged to somebody else.
But the deeper effect was on what kind of business could survive at all. An enterprise that must wait four years for permission needs capital to survive four years of waiting, and needs someone in Delhi who can move the file. A large established house has both. A new entrant with a better idea and no money has neither.
So the system selected. Not for good ideas or efficient production, but for staying power and access. Over forty years, selecting for those two things every day, an economy becomes composed of firms that are good at them.
Everything in section 2.2 is statutory. The Industries (Development and Regulation) Act, the Industrial Policy Resolution of 1956, the MRTP Act, the Foreign Exchange Regulation Act and the industrial disputes provisions on closure are all published law, and the licensing procedures were published as rules.
This is the highest grade of evidence available anywhere in this series. There is no need to rely on anybody’s recollection of how difficult it was; the requirements are on the statute book and anyone can read them.
What it cannot show: how the system was actually administered. A statute tells you what was required, not how long it took, what was asked for informally, or which applications moved and which sat. For that you need the committee reports in Chapter Four, and they are the only systematic evidence that exists.
Before drawing the conclusion, the defence deserves its strongest hearing, because there is a version of it that survives everything in this chapter.
Chapter Two has described the system as an obstacle course. Its defenders say the obstacle course is why there is anything to obstruct.
In 1950 India made almost nothing. An Indian firm attempting to make steel, machine tools or chemicals would have been undercut immediately by established foreign producers with decades of scale behind them, and would never have reached the size at which it could compete. Protection bought the time. By 1991 India had a broad industrial base, a large engineering workforce and companies that went on to compete internationally — and every one of them was raised behind that wall. You cannot admire Indian industry today and condemn the thing that let it survive its childhood.
The argument for protecting an infant industry from foreign competition is respectable and this series does not dispute it. But that is a case for tariffs, not for licences. A tariff shields a domestic industry from outside while leaving Indian firms competing furiously with each other, which is what forces improvement. India did both: it shut out foreign competitors and then licensed how many domestic ones there could be. The second removed the only remaining pressure to get better, and it is not required by the first.
What would settle it: comparing sectors where Indian firms faced many domestic competitors with sectors where licensing left two or three. Some of this has been done and it points the same way, but not systematically.
Why people care so much: because “we needed protection” and “we needed licensing” have been treated as one argument in India for seventy years, and separating them costs one side its best defence.
2.4 — The scarce thing
Every economy has something scarce, and the scarce thing is what everybody competes for.
In a market economy the scarce thing is customers. Firms compete by trying to be chosen by buyers, which is why they spend their effort on price and quality.
In India between 1951 and 1991 the scarce thing was permission. Capacity was licensed, so it could not be competed into existence. Foreign exchange was allocated, so it could not be earned into your hands. Imports were restricted, so a foreign competitor could not take your customers.
Once a licence was granted, the holder faced a market with limited supply, restricted imports and licensed competitors. Being chosen by a buyer was not the problem. Being chosen by an official was.
That is the single most important sentence in this part, and Chapter Six is about who understood it best.
A firm needed permission to exist, expand, relocate, change product, import a machine, obtain foreign currency — and from the mid-1970s, permission to close.
Delay was not merely inconvenient. It selected for firms with the capital to wait years and the contacts to move a file, and against anyone with a better idea and no money.
The scarce thing was permission, not customers. Being chosen by a buyer was not the problem. Being chosen by an official was.
3The Farm That Was Not Funded
Seven Indians in ten lived off the land. The plan that mattered most gave the money to something else, and within a decade the country was buying its bread from a foreign government with conditions attached.
3.1 — The First Plan, which worked
The First Five-Year Plan, from 1951 to 1956, put its weight on agriculture and irrigation. Big dams, canals, community development. It was unglamorous and it exceeded its own growth target.
Some of that was luck — the monsoons were kind, and a country coming off the disruption of Partition had recovery growth available. But the emphasis was right for the country it was aimed at, and it delivered.
3.2 — The Second Plan, which turned away
From 1956 the weight moved decisively to heavy industry. Steel plants, machine tools, capital goods. Agriculture’s share of plan investment fell.
The reasoning was the Mahalanobis argument from Chapter One, and it had an agricultural component that is often left out: the planners did not think farming was unimportant. They thought it did not need capital. Their view was that Indian agriculture could be improved by better organisation — cooperatives, extension services, land reform — rather than by investment, because the binding constraint was social rather than technical.
Two of the three parts of that answer had already failed. Part Three showed what happened to land reform. Cooperatives were pushed and mostly did not take. The extension services worked in places and not in others.
So agriculture got a diagnosis that assumed it needed no money, and a set of remedies that did not work, at the moment the population began growing at over 2 per cent a year.
Foodgrain buffer stock: grain the government buys and stores so that it can release supplies when harvests fail and prices spike.
A country with a buffer stock can survive a bad monsoon by opening its own godowns. A country without one has to buy on the world market at exactly the moment its need is visible to every seller — or accept whatever a donor offers, on whatever terms.
Why it matters here: India in the mid-1960s had no meaningful buffer, and the terms it then accepted are the subject of section 3.4.
3.3 — What happened to food
Output rose through the 1950s and early 1960s. It did not rise fast enough.
The reason is arithmetic. Population was growing at over 2 per cent a year, which meant food output had to grow faster than 2 per cent every year merely to hold the position, and faster still to improve it. Indian agricultural output growth in this period was not reliably above that line, and it depended on the monsoon, which is not a policy.
Then came the droughts of the mid-1960s, two bad years close together, and the position went from tight to critical.
3.4 — Ship to mouth
PL-480: an American law, Public Law 480, under which the United States sold surplus grain to poor countries and accepted payment in the buyer’s own currency rather than in dollars.
For India this was genuinely valuable. Dollars were the scarcest thing the country had, and buying food with rupees meant the dollars could go on machinery instead.
It also meant that a large share of what Indians ate arrived by decision of a foreign government, which could slow the flow at any time without doing anything a lawyer could object to.
Why it matters here: this is the difference between a purchase and a dependency, and India crossed it somewhere in the early 1960s without noticing.
India bought grain from the United States under a programme known as PL-480, which allowed purchase in rupees rather than scarce dollars. Through the 1960s the quantities became very large. The phrase used at the time was that the country lived “from ship to mouth” — grain going from the dock into distribution with no stock in between.
Then the terms changed. During the food crisis of the mid-1960s the American President moved to releasing shipments in small tranches rather than committing to a year’s supply — a practice described afterwards as keeping India on a short tether. Continuation was linked to India’s positions on matters including its agricultural policy and its criticism of American conduct in Vietnam.
Whatever one thinks about the rights of that, its meaning for India was unambiguous. A country that cannot feed itself does not have an independent foreign policy, whatever its constitution says. Non-alignment, which was the central plank of Indian external policy, was being conducted by a government whose people ate at the discretion of one of the two blocs it was declining to align with.
Picture a household that spends every rupee it earns and keeps nothing in reserve. It is not poor exactly; it eats every day.
Then a month comes when the earnings do not arrive. The household goes to a neighbour, who lends generously — and then, the following month, mentions that he has views about how the household is being run, and about the household’s friendship with a man he dislikes.
Nothing has been threatened. Nothing needs to be. The conversation has changed permanently, and it changed the moment the reserve ran out.
That is what the absence of a buffer stock meant in 1966. Not hunger alone. A permanent alteration in what India could say.
Which raises the question this chapter has been circling, and it is not a simple one.
Everyone agrees the Second Plan moved resources to industry. What that means is contested.
The planners did not think agriculture unimportant; they thought it did not need capital. On the evidence available in 1956 that was a defensible reading: Indian yields were far below what the same land produced elsewhere with the same technology, which pointed at organisation rather than investment. And the technology that eventually transformed Indian yields — the high-yielding seed varieties — did not exist in usable form in 1956. You cannot fund a solution that has not been invented.
The diagnosis may have been defensible in 1956. It was not defensible after the first bad harvest, or the second. Food output relative to population is measured every single year and published, and the line was visible from the early 1960s. What the government did during those years was continue the industrial emphasis and cover the gap with imported grain — which is to say, it treated a structural problem as a series of emergencies, for about a decade, until the emergencies became a crisis with a foreign government’s hand on the tap.
What would settle it: the internal record of the Planning Commission and the Ministry of Food through 1960 to 1965 — who raised the food position, in which year, and what was said back. Some of this is published in plan documents and parliamentary answers; the internal papers are not open.
Why people care so much: because if agriculture was starved, then the humiliation of 1966 and everything that followed from it — including a devaluation forced from outside — was self-inflicted, and the whole planning framework carries the blame rather than the weather.
One last point, and it will matter in Chapter Seven. The response to this crisis, when it finally came, was not more planning. It was a technology programme — new seed, fertiliser, water and a guaranteed price — that largely bypassed the permission system altogether. That contrast is the most useful natural experiment in this whole part.
The First Plan put its money on agriculture and irrigation and beat its target. The Second moved the weight to heavy industry, on a diagnosis that farming needed organisation rather than capital.
Two of the three organisational remedies — land reform and cooperatives — had already failed. Population was growing at over 2 per cent, so output had to beat that line every year just to stand still, and it did not reliably do so.
By the mid-1960s India lived ship to mouth on American grain, released in small tranches with conditions attached. A country that cannot feed itself does not have an independent foreign policy, whatever its constitution says.
4The Warnings
Part One said a bad outcome only becomes a blunder if somebody saw it at the time. Somebody did. Three committees, appointed by the government, reporting to the government, in five years.
4.1 — The test, and why this chapter decides the part
Chapter One conceded the 1950 decision. Chapter Three conceded that the 1956 agricultural diagnosis was defensible on the information then available. Both concessions were real and both have the same structure: you cannot blame people for not knowing what nobody knew.
That defence has a precise expiry date, and this chapter establishes it. If the correct diagnosis was produced, in writing, by people the government itself appointed, and nothing followed, then everything after that date is not a mistake. It is a refusal.
4.2 — 1965: the Monopolies Inquiry Commission
The first was the Monopolies Inquiry Commission, which reported in 1965. It examined the concentration of economic power in India and proposed measures to limit it.
Note the framing, because it matters. This was a body looking at whether wealth was becoming concentrated — a socialist’s question, asked by a socialist government, in the terms that government would have chosen. And it found that concentration was happening.
That is important because it means what follows cannot be dismissed as an argument imported from outside. The critique of the licence system in the 1960s did not come from free-market economists. It came from the people running the system, using their own criteria, and finding that the system was failing by its own stated purpose.
4.3 — 1967: Hazari
The Planning Commission then appointed a committee under R.K. Hazari to review the working of industrial licensing under the 1951 Act.
Its conclusion was that the licensing system had produced disproportionate growth for some of the large business houses; that big business had succeeded in turning the policy to its own purposes; and that the granting of licences had become biased in favour of the large houses.
Read that once more and hold what it says. The instrument created to prevent the concentration of economic power was producing the concentration of economic power. Not failing to prevent it. Producing it.
4.4 — 1969: Dutt
The government then appointed the Industrial Licensing Policy Inquiry Committee under Subimal Dutt, a senior civil servant and diplomat. It reported in 1969. Its members included H.K. Paranjape and S. Mohan Kumaramangalam — the latter a figure of the left, not a critic of state control.
Its findings were specific and they are the most damaging documents in this part.
Large houses obtained licences and did not implement them. The committee identified companies doing this. Taking out a licence you never intend to use is not a failure to invest; it is an investment — in keeping the field empty.
This pre-emption forestalled others from entering production. The committee said so directly. A licence held and unused is a wall around a market.
No specific instructions had ever been given to licensing authorities to prevent concentration or monopoly. The system’s stated purpose had never been translated into an instruction to the people operating it.
Large houses had penetrated areas reserved for the public sector, and obtained licences for non-essential items and for items that should have gone to the small and medium sector.
The objectives of the 1956 Industrial Policy Resolution had not been fulfilled. That is the government’s own committee, on the government’s own founding economic document, thirteen years in.
These are published reports of committees appointed by the Government of India. The Dutt committee’s report is a public document and can be read today.
This is the strongest possible evidential position for a charge of this kind, and it is worth being explicit about why. The usual defence against economic criticism of this period is that it comes from people with an ideological objection to state action. That defence cannot touch these documents. The commissioning authority was the government running the system. The members were establishment figures, in one case a prominent man of the left. The criteria applied were the government’s own stated objectives.
What it cannot show: what happened to the reports inside the machine. Who read them, what was argued, and why the response took the form it did are not in the reports themselves, and the Cabinet and party papers that would answer it are not open.
4.5 — What was done about it
Now the part that makes this a chapter rather than a footnote.
The response to a finding that the permission system was concentrating economic power was the Monopolies and Restrictive Trade Practices Act of 1969: a further layer of permission, requiring large houses to obtain an additional clearance from an additional authority.
The licensing system itself was not dismantled. It was not substantially loosened. The diagnosis — that requiring permission to produce had created a market in permissions which the largest players were best equipped to win — was accepted, and the remedy was another permission.
Four years later, the Foreign Exchange Regulation Act of 1973 tightened control again.
You have seen this shape twice already in this series. In Part Two, an emergency justified detention without trial and the power was made permanent rather than temporary. In Part Three, a blockage justified the Ninth Schedule and the schedule acquired no limit. Here, a control fails and the response is more control. In each case the instrument’s failure is treated as evidence that there was not enough of it.
Everyone who argues about India’s economic failure — the people who blame socialism and the people who defend it — assumes that the failure was intellectual. That people believed wrong things, and that if they had believed the right things they would have done differently.
The right’s version: they were captured by a bad idea. The left’s version: the idea was sound and the critics misunderstand it. Both are arguments about beliefs, conducted as though the decisive question were what was in people’s heads.
But the correct diagnosis existed. It was produced in 1965, refined in 1967, and set out in detail in 1969, by three bodies the government itself appointed, using the government’s own criteria, in language nobody had to translate. Nobody needed to be persuaded of anything. It was on the file.
So the operative question is not what anybody believed. It is: what happens to a correct diagnosis inside an organisation?
Part Three answered that for this particular organisation, four years before the first of these reports. Between 1950 and 1954 every figure with an independent base who could tell the leadership it was wrong had died, resigned or been defeated, and nothing institutional replaced them. A finding that contradicts the leadership needs somebody with standing to carry it, and by 1965 there was nobody in the Congress party whose objection had weight.
That is why this chapter matters more than any economic argument in the part. The failure was not that India chose planning. It was that India built a machine with no way of being told it was wrong, and then ran that machine for forty years. The economics is downstream of the politics, and the politics happened in Part Three.
Both camps avoid this, because it is not an argument either can win. It offers the right no vindication of markets and the left no defence of planning. It says the system’s defect was that correct information could not travel upwards — which is a criticism of a structure, and structures have no ideology to attack.
One clarification before leaving this chapter, because it is the most likely place for this part to be misread. Nothing here says the people running the system were stupid or dishonest. Several of them commissioned the reports that condemned their own policy, which is not the behaviour of people avoiding the truth. The finding is narrower and worse: they produced the correct diagnosis and could not act on it, because the organisation had no route by which a finding that contradicted the leadership could travel.
Three committees — 1965, 1967 and 1969 — all appointed by the government, all finding that the licensing system was producing the concentration of economic power it existed to prevent.
Dutt found that large houses took out licences they never used, to keep others out, and that licensing authorities had never been instructed to prevent concentration at all.
The response was another permission — the MRTP Act. The correct diagnosis was on the file by 1969. Everything after that is not a mistake.
5The Comparison
Most comparisons made in this argument are rigged, and they are rigged by both sides. Here is the honest set, and the honest set is worse.
5.1 — The comparison everybody makes, and why it is weak
The standard charge is: look at South Korea. In 1950 they were as poor as us or poorer. Now look.
The defence’s answer to this is strong and this chapter is going to concede it in full.
South Korea and Taiwan were small — populations a fraction of India’s, which makes almost every administrative problem different in kind rather than degree. They were ethnically and linguistically uniform, so they had no equivalent of the reorganisation described in Part Three. They spent much of the period under authoritarian government, which means unpopular decisions could simply be imposed. They received enormous American aid and guaranteed access to the American market, granted for Cold War reasons that had nothing to do with economics. Their land reforms were carried out under military occupation, which is a method not available to a democracy.
India had none of these. It chose non-alignment and got no bloc patron. It was enormous, plural and democratic, which meant every decision had to survive an election.
So: point conceded. The South Korea comparison is not a clean test, and a prosecution brief that leans on it is leaning on the weakest available support.
Compound growth: growth that builds on itself, because each year’s increase is calculated on the total including all the previous years’ increases.
This is why small differences in a growth rate matter enormously over long periods and matter almost not at all in any single year. At 3.5 per cent an economy doubles in about twenty years. At 7 per cent it doubles in about ten. Over forty years the first has doubled twice and the second four times — which means the second is four times the size of the first, from the same start.
Why it matters here: nobody experiences a growth rate. People experience a year, and in any single year the difference between 3.5 and 7 per cent is almost invisible. The cost only becomes visible when it is far too late to have prevented it.
5.2 — The honest comparison set
A fair comparison needs countries that were large, poor, plural and post-colonial. That set is small but it is not empty, and here is what happened in it over roughly the same decades.
| Country | Approximate annual growth | Why the comparison is fair or not |
|---|---|---|
| India | about 3.5% | — |
| Pakistan | about 5% | Identical starting date, same colonial inheritance, same partition, comparable poverty, comparable plurality. The closest comparison that exists anywhere. |
| Indonesia | about 6% | Large, poor, plural, post-colonial, thousands of islands and hundreds of languages. Fair, with the caveat of oil. |
| Thailand | about 7% | Smaller and never formally colonised. Partly fair. |
| Taiwan | about 8% | Small, uniform, authoritarian, American-backed. Not a fair test. |
| South Korea | about 9% | The same objections, more strongly. Not a fair test. |
Now look at what happens when you throw out the unfair comparisons.
India still comes last.
Pakistan is the comparison that ought to be uncomfortable, and it is almost never made, by anybody, in either country. Same starting line, same year, same inherited administration, same partition, comparable poverty. Pakistan spent much of the period under military rule and produced political outcomes India did not, which is a point in India’s favour that this series will make elsewhere. But on the narrow economic question over these decades, it grew faster.
Removing the rigged comparisons does not rescue the record. It makes the charge harder to answer, because every objection available against South Korea fails against Pakistan.
Two brothers inherit neighbouring shops on the same day in 1947. Both shops are in poor condition. Both brothers work hard.
Over thirty years the first brother’s takings grow by about 3.5 per cent a year, the second’s by about 5 per cent. It does not sound like much of a gap; on any given day you could not tell the shops apart.
After thirty years the second shop is taking about fifty per cent more than the first, from the same street, the same customers and the same start. Neither brother did anything dramatic. The difference is entirely in a rate that was never visible on any single day.
That is why this part exists, and why it is placed early. Almost nothing in this series is as expensive as a growth rate, and almost nothing is as difficult to notice while it is happening.
5.3 — The savings puzzle
Here is the fact that removes the last available defence.
India’s gross domestic savings rose from around 9.6 per cent of national income in the 1950s to somewhere in the region of 20 to 25 per cent by the 1970s and 1980s.
That is a remarkable achievement in a poor country and it deserves saying. Indian households saved. The rate reached levels comparable to countries that were growing at two and three times India’s speed.
Which forecloses the most common defence of this period — that a poor country simply lacks the capital to grow. India had the capital. It went in.
What it bought was plants running below capacity, protected products made at above world prices, and licensed monopolies with no reason to improve. The money was not missing. The return on it was.
And that turns the whole subject from an argument about resources into an argument about allocation — which means an argument about who decided, on what basis, with what consequence for being wrong. Chapter One’s hidden assumption said nobody in a permission system both chooses and pays. This is the bill for that.
Comparison is the only tool available for a counterfactual, and it is a contested tool. Both positions are serious.
No two countries are alike, and every difference you can name may be the one doing the work. India was uniquely large, uniquely plural and uniquely democratic among poor countries in this period, and democracy is expensive: it means you cannot suppress consumption, cannot move populations, cannot impose costs on a region that votes. Comparing India with places that did all three is comparing outcomes achieved under different rules. And India’s political record over the same decades — no coup, no famine on the colonial scale, an unbroken sequence of elections — does not appear in any growth table.
Every objection above is answerable by choosing the comparison better, and the answer survives the choosing. Drop the small, uniform and authoritarian cases. Keep only the large, poor, plural, post-colonial ones. India is still last. Pakistan in particular defeats the democracy objection from the other direction: it was less democratic and grew faster, so if democracy explains India’s rate then democracy is a cost worth naming and defending openly rather than mentioning only when the figures are bad.
What would settle it: a comparison across all large poor democracies in the period. The sample is very small, which is itself the finding: there were almost no others, which makes India’s case genuinely close to unique and genuinely hard to test.
Why people care so much: because if the comparisons hold, roughly forty years of Indian poverty were avoidable, and the number of people involved is in the hundreds of millions.
All of which rests on a class of figure that deserves inspection before this much weight is put on it.
Growth rates come from national accounts — the estimate a government makes of everything produced in a year. India’s have been compiled since the 1950s and revised several times, with new base years and improved methods.
They are good enough for this chapter’s purpose for a specific reason: the argument here does not turn on a decimal point. A gap between 3.5 and 5 per cent, sustained for thirty years, survives every plausible revision to either figure. If the argument needed 3.5 rather than 3.8, it would not be worth making.
What they cannot show: anything about who received the growth. National accounts are a total. A country can grow steadily while a majority of its people do not, and India’s poverty rate stayed somewhere between 40 and 50 per cent through most of this period — a fact that sits alongside the growth figures rather than being explained by them. They are also weakest exactly where most Indians worked: output in small farms and unregistered workshops is estimated rather than counted.
With those limits stated, the comparison in this chapter is about as solid as economic history gets, and it goes one way.
The South Korea comparison is weak and this part concedes it: small, uniform, authoritarian, American-backed. That objection is real.
Use only the fair comparisons — large, poor, plural, post-colonial — and India is still last. Pakistan is the closest comparison that exists and it grew faster.
India was not short of capital. Savings rose to around a fifth of national income. What was missing was the return, and a return is a question about allocation, which means a question about who decided.
6Who It Actually Benefited
A system built to restrain large business was the best thing that ever happened to large business, and the government’s own committee said so in 1969.
6.1 — A permission is an asset
Start with a piece of arithmetic that most descriptions of this period miss.
If making bicycles requires a licence, then a bicycle licence is worth money. Not because it lets you make bicycles — anybody can make bicycles — but because it stops other people making them.
The value of the licence is the value of everybody else’s exclusion. And that value goes up the more attractive the market is, which means the better the opportunity, the more valuable it is to hold the permission and the more it costs anybody else to enter.
This is not a subtle effect and it does not require anybody to be corrupt. It follows from the design.
6.2 — Pre-emption
Once you see that a licence is an asset, one behaviour becomes obviously rational, and the Dutt committee found it happening.
Apply for licences you have no intention of using.
Chapter Four quoted the finding: large houses obtained licences and did not implement them, and this pre-emption of capacity forestalled others from entering the field. In the language of the committee, the practice was against the objective of rapid industrial growth.
Consider what a large house got for the cost of an application. It did not have to build the plant. It did not have to raise the capital, hire the workers, or take the risk. It simply had to hold the paper — and for as long as it held it, the licensed capacity for that product was accounted for, and a rival applying to make the same thing could be told the capacity already existed.
The state had, without intending to, created a cheap and legal way of buying a market and never entering it.
Imagine a town where the council decides how many bakeries there may be, and settles on three. Applications open.
You are the richest man in town. You do not particularly want to run a bakery. But you apply for all three permissions, and you get them, because you have the lawyers and the contacts and the patience for the process.
You then open one bakery, badly, and sit on the other two permissions.
Nobody in that town will ever eat better bread. Not because you baked well, and not because the council was corrupt, but because the council created a thing that was worth owning and you owned it. And you did it entirely within the rules, which is why nobody can point at a moment when the wrong was done.
6.3 — The market in files
The second consequence is the one everybody knows about, and it is worth putting in its proper place — which is second, not first.
When permission is the scarce thing, permission acquires a price. Sometimes that price was paid in money. Often it was paid in more durable currencies: political contributions, jobs for relatives, favours held in reserve.
The government examined this too. A committee on the prevention of corruption reported in the mid-1960s and connected corruption directly to the system of licences, permits and controls — the machinery, not the men. It led to the creation of a central vigilance body.
The important analytical point, which is easy to miss in the moralising this subject attracts, is the direction of causation. A system that makes permission valuable will produce a market in permission. You can prosecute individual officials indefinitely and the market will persist, because the thing being traded is created by the rules rather than by the character of the people administering them.
Rent-seeking: spending effort to capture a bigger share of existing wealth rather than to create new wealth.
A firm that invents a cheaper process has created something; the wealth did not exist before. A firm that spends the same money persuading an official to refuse its rival’s licence has moved wealth from one pocket to another and produced nothing.
Both are rational for the firm. Only one is useful to anybody else. Which of the two a business pursues depends on which pays better, and that depends entirely on how the rules are arranged.
Why it matters here: the licence system made the second reliably more profitable than the first, for forty years, and businesses responded exactly as anybody would.
6.4 — Who was on both sides
Which produces the arrangement that this chapter exists to name.
By the late 1960s the licence system had three groups of people with a stake in its continuation, and they were not the groups anybody would have predicted in 1951.
The officials who issued the permissions, whose importance was entirely a function of the permissions existing.
The politicians, for whom a discretionary permission is the most useful thing in politics: it can be granted, withheld, delayed, and remembered.
And the large business houses — the very people the system was built to restrain — for whom it was a barrier around their markets, paid for by the state and enforced by law.
Between them, these three groups contained everybody with the organisation and the access to change the policy. The people harmed by it were the firms that were never founded, the products never made, and the consumers who paid more. None of those is a constituency. A business that does not exist cannot lobby.
Everyone arguing about the licence system assumes that it restrained big business.
Its defenders say: it prevented private monopoly and kept the commanding heights in public hands. Its critics say: it strangled enterprise and held Indian business back. Both are describing a state pressing down on capital. They disagree only about whether the pressing was justified.
It did the opposite, and the government’s own inquiry established this in 1969. A permission requirement is a barrier to entry, and a barrier to entry is worth far more to whoever is already inside than to the authority that erected it. Every rule making it harder to start a bicycle factory is a gift to whoever already has a bicycle factory.
So the system’s real distribution of benefit ran opposite to its stated purpose. It restrained the entrant, not the incumbent. It protected the largest houses from competition more effectively than any of them could have managed privately, and it did it with the moral authority of a socialist state and at no cost to the beneficiaries.
This is why the assumption survives so comfortably. It flatters everybody. It lets the left remember a state standing up to capital. It lets the right remember heroic entrepreneurs strangled by bureaucrats. What it hides is that the biggest houses of that era and the officials regulating them were, on this specific question, on the same side — and that the losers were people who never appeared in the argument because they never got to exist.
Watch for this shape in later parts. Whenever a regulation’s stated purpose is to restrain the powerful, ask who bears the cost of complying with it, and whether that cost is larger for the powerful or for whoever is trying to become powerful.
Which leads to the accusation most commonly made about this period, and the one most often made carelessly.
India’s reputation for administrative corruption dates substantially from this period. What that means is disputed.
Corruption requires something worth buying. A permission regime manufactures thousands of valuable, discretionary, individually-granted decisions, each of which is worth more to the applicant than the official earns in a year. No moral instruction can survive that arithmetic at scale. Change the rules and the market disappears, which is broadly what happened to industrial licensing corruption after the permissions were abolished.
Corruption existed before 1951 and persists in sectors that were never licensed — land records, policing, contracts, admissions. Blaming the licence system is convenient because it locates the problem in a policy that has since been abolished, rather than in enduring features of Indian administration: low official pay, weak accountability, and social norms about obligation to kin and community that do not switch off at an office door.
What would settle it: sector-by-sector measurement before and after delicensing. Partial evidence exists; nobody has assembled it systematically.
Why people care so much: because one account says the problem was a policy and can be fixed by policy, and the other says it is a property of Indian public life, and those two beliefs imply completely different politics.
Set this chapter beside Chapter Four and the shape of the whole part is visible. The government appointed committees which found the system producing the opposite of its purpose. The committees were right. And the three groups with the organisation to change the policy were the three groups the policy was making better off. That is not a conspiracy and nobody had to arrange it. It is what happens when the people harmed by an arrangement are firms that were never founded.
A licence is an asset, and its value is the value of everybody else’s exclusion. That follows from the design and requires nobody to be corrupt.
The Dutt committee found large houses taking out licences they never used — buying a market and never entering it, legally, for the price of an application.
Three groups had a stake in the system continuing: officials, politicians, and the largest business houses. The losers were firms that were never founded, and a business that does not exist cannot lobby.
7The One Thing That Worked
In the middle of forty years of failure, one programme succeeded completely. Understanding why is worth more than all the rest of this part.
7.1 — What was actually done
By 1965 the food position described in Chapter Three had become intolerable. What followed was the most successful thing any Indian government did to the economy in the whole period covered by this part.
It had four components and none of them was a plan target.
Seed. New wheat varieties, bred in Mexico, which produced far more grain per acre if given enough water and fertiliser. India imported them in enormous quantity — the largest seed import anywhere up to that time — and Indian scientists adapted them to local conditions.
Fertiliser and water. The new seeds only work with heavy inputs. Fertiliser supply was expanded and subsidised, and tube-well irrigation spread rapidly with subsidised power and credit.
A guaranteed price. A minimum support price was established, along with the machinery to announce it, and a state agency to buy grain at that price and hold it. A farmer adopting an unfamiliar seed was taking a risk with his family’s food, and the guaranteed price removed the downside of that risk.
Credit. Someone had to pay for the seed, the fertiliser and the pump before the harvest came in.
Within about seven years India was self-sufficient in wheat. The ship-to-mouth dependence described in Chapter Three ended, and it has not returned.
7.2 — Why it worked when nothing else did
Now the question that matters, and the answer is not “because the state acted”.
The state acted constantly through this period. It acted in steel, in machine tools, in textiles, in chemicals. What was different here?
No farmer had to apply for permission to plant the new seed.
Read the four components again with Chapter Two in mind. Every one of them is the state acting as a supplier or a buyer. It supplied seed, fertiliser, water, power and credit. It bought grain at a guaranteed price. At no point did it decide which farmer might grow how much wheat, or require an application, or specify a licensed capacity.
The decision stayed with the person who bore the consequence. A farmer who adopted the package and made it work kept the gain. A farmer who did not, did not. Millions of separate people made millions of separate decisions with their own land at stake, and the state’s job was to make the good decision possible rather than to make it on their behalf.
That is the exact inverse of the permission system. And it is the only part of the Indian economy in these forty years that performed like the countries in Chapter Five’s comparison table.
Agricultural output is among the best-measured things in India, for an unglamorous reason: the state buys grain, and anything the state buys must be weighed and paid for.
Procurement records, support prices, area under high-yielding varieties, fertiliser offtake and stock levels are all published, annually, by agencies whose job was operational rather than promotional. The wheat transformation is visible in several independent series that were collected for different purposes and agree with each other.
What it cannot show: what happened to people. Output per acre is not income per household, and neither is a measure of who ate. A district can double its wheat and contain more malnourished children than before if the gains go to owners and the losers are labourers. The production series is excellent and it is silent on distribution, which is why Chapter Seven’s success and India’s persistent malnutrition are both true at once.
The technical term for what was imported deserves unpacking, because the whole geography of the programme follows from what these seeds require.
High-yielding variety: a crop bred to produce far more grain per plant than traditional varieties — mainly by growing a shorter, stiffer stalk that puts its energy into grain rather than straw, and does not fall over under a heavy head.
The catch is that the extra grain has to come from somewhere. These varieties need much more water and much more fertiliser than the seeds they replaced. Given both, they transform a harvest. Given neither, they can do worse than what the farmer was already growing.
Why it matters here: that requirement decided the geography of the whole programme, and therefore which parts of India would be prosperous for the next fifty years.
7.3 — The districts that were chosen
The programme was not applied evenly across India, and it could not have been.
The new seeds needed assured water. That meant the effort concentrated where irrigation already existed or could be created quickly: Punjab, Haryana and western Uttar Pradesh above all, and later parts of the coastal south.
The choice was technically correct. Putting water-hungry seed into rain-fed districts would have produced failure and discredited the whole programme in its first season.
It also decided, in a few years, which parts of India would be prosperous for the next half-century. A national programme that could only be delivered in some districts was, unavoidably, a decision about which districts. Part Five of this series is about exactly this — how central decisions made some Indian states rich and others poor — and this is one of the three or four largest entries in it.
7.4 — What was mortgaged
The bill came later and it is still being paid.
The package tied farmers to a specific crop pattern — wheat and, in the paddy areas, rice — because those were the crops the guaranteed price covered. Once the procurement machinery, the credit and the whole rural economy were built around two crops, switching became individually irrational even where it was collectively necessary.
Those crops are thirsty. Free or near-free electricity for pumping, introduced as an input subsidy and impossible to withdraw afterwards, made pumping groundwater effectively costless to the farmer while it remained very costly to the aquifer.
The result, over fifty years, is a groundwater position in Punjab and Haryana that is among the most serious resource problems in India, and a crop pattern that nobody now defends and no government has been able to change.
None of this makes the Green Revolution a mistake. A country facing famine and foreign leverage over its food supply took the option available and it worked. But the way it was paid for — a permanent input subsidy and a permanent procurement guarantee, neither of which any government has since been able to withdraw — is the same pattern this series keeps meeting: an emergency instrument that acquires a constituency and never expires.
Fifty years on, the same programme is cited as India’s greatest policy success and as the origin of its worst resource crisis. Both cases are serious.
India in 1966 faced genuine hunger and a foreign government’s hand on its grain supply. Within seven years that was over permanently, and it has never returned despite the population roughly tripling since. Whatever the later costs, they are the costs of a country that has food. Judging the decision by problems that only exist because it succeeded is the clearest possible case of the hindsight trap Part One warned about.
Nobody disputes the emergency or the achievement. The objection is to what was signed alongside it: a procurement guarantee and an input subsidy, both permanent, neither reviewed. Those two commitments locked a whole region into two thirsty crops and made pumping groundwater free to the farmer while it remained very costly to the aquifer. Fifty years later no government of any party has been able to withdraw either, and Punjab’s water position is among the most serious in India. An emergency measure that acquires a constituency and never expires is the exact pattern this series keeps finding.
What would settle it: whether any government seriously attempted to restructure the support and was defeated, or whether none tried. The record suggests very few attempts and no sustained one.
Why people care so much: because the region that benefited most is now paying most, which makes every proposal to change the arrangement look like a punishment aimed at the people who fed the country.
Both of those positions, though, share something they never state.
Both sides of the Indian economic argument claim the Green Revolution, and both do it by assuming that it shows what state action can achieve.
The left’s version: here is proof that a determined state can transform an economy, and the reformers who dismantled state capacity forgot it. The right’s version: it worked because it was about technology and prices rather than planning, which proves markets. Both are arguing about how much state there should be.
Neither is looking at the specific thing that was different. The state in this programme was a supplier and a buyer, not a gatekeeper. It did not decide who could farm, how much they could produce, or what capacity they were licensed for. It removed constraints — seed, water, fertiliser, credit, price risk — and left the decision with the person whose land it was.
That distinction is not on either side’s map, and it explains something both sides find awkward. The same government, in the same years, with the same ideology and often the same ministers, produced the best economic outcome of the period in agriculture and the worst in industry. If the variable were the amount of state, that could not happen. The variable was not the amount. It was where the state stood relative to the decision.
Which suggests the useful question about any government programme is not “how interventionist is it” but “does it add a permission, or remove a constraint?” Those two things wear the same clothes in a budget document and behave in opposite directions in the world.
Carry that question into Part Five, where it does most of its work. The programme that succeeded removed constraints, and it removed them in some districts and not others, and that is why Part Five is about a map.
Four components: seed, fertiliser and water, a guaranteed price, and credit. Within about seven years India was self-sufficient in wheat and the ship-to-mouth dependence ended for good.
It worked because no farmer had to apply for permission to plant. The state was a supplier and a buyer, not a gatekeeper, and the decision stayed with the person who bore the consequence.
It also chose which districts would be rich for the next fifty years, and it was paid for with a permanent subsidy and a permanent procurement guarantee that nobody has since been able to withdraw.
8The Two Moments
Twice before 1991 the direction could have changed. The first attempt was punished so severely that it settled the question for twenty years. The second happened without anybody announcing it.
8.1 — June 1966
Indira Gandhi had been Prime Minister for a few months. The food crisis was at its worst, foreign exchange had run down, and aid donors — principally the United States and the World Bank — were pressing for a change in economic direction as a condition of resuming support.
On 6 June 1966 the rupee was devalued, from about 4.76 to the American dollar to 7.50. That is a cut of over a third in the external value of the currency, and it was accompanied by promises of import liberalisation.
Overvalued currency: a currency whose official exchange rate is set higher than what it would fetch if freely traded.
An overvalued rupee makes foreign goods look cheap to Indians and Indian goods look expensive to foreigners. That sounds pleasant and is not: it means imports are attractive exactly when you cannot afford them, and exports are uncompetitive exactly when you need to earn foreign currency.
Why it matters here: India held an overvalued rate for years because devaluing is politically painful and doing nothing is not — until the reserves run out, at which point the choice is made for you. That happened in 1966 and again in 1991.
The economics of it were defensible. An overvalued currency makes imports artificially cheap and exports artificially expensive, which is precisely the wrong arrangement for a country short of foreign exchange. Devaluation makes exports competitive and imports dear, which is what a country in that position needs.
Politically it was a catastrophe, and the reasons were not economic.
It was seen as capitulation — a national humiliation performed under foreign instruction, by a new Prime Minister widely regarded as weak, at the direction of the country whose grain shipments were already being released month by month. The aid that was expected did not arrive in the volumes anticipated. The immediate effect on prices was felt long before any export benefit.
Congress lost heavily in the general election that followed in 1967, losing power in a large number of states for the first time.
8.2 — What was learned from it
Here is why this moment matters more than the devaluation itself.
The lesson taken from 1966, by the government and by the political class generally, was not that the devaluation had been badly executed or badly explained. It was that opening the economy is politically fatal and is done at the instruction of foreigners.
What followed was a sharp turn in the opposite direction. Banks were nationalised in 1969. The MRTP Act arrived the same year, adding the permission layer described in Chapter Four. The Foreign Exchange Regulation Act followed in 1973 and tightened control over foreign companies and foreign currency; several international firms left India rather than comply. Privy purses were abolished. The rhetoric moved decisively left and so did the policy.
It is worth being careful about the causation here. The turn after 1966 was not only a reaction to the devaluation; there was a genuine political struggle inside Congress in these years, and the leftward move served a purpose in that struggle that had nothing to do with economics. Part Eight of this series takes that story properly.
But the effect on the economic argument is not in serious doubt. After 1966, any Indian politician proposing to open the economy was proposing something that had been tried once, had failed visibly, and had been widely understood as surrender. The single attempt at liberalisation before 1991 discredited liberalisation for a generation.
This is argued about in India to this day and the two cases run past each other, because they are about different things.
The currency was overvalued and everybody serious knew it. Devaluation was correct and overdue, and the reason it failed to deliver was not the devaluation but the two failed monsoons either side of it, which wrecked output and prices at exactly the wrong moment. It was also undersold: no serious attempt was made to explain to the country why a weaker rupee was in its interest, so the only available interpretation was the one the opposition offered.
A devaluation works by making exports competitive. India in 1966 had almost nothing to export, because forty years of import substitution had built an industry aimed entirely at the domestic market behind a tariff wall. So the benefit could not arrive quickly, while the cost — dearer imports of food, oil and machinery, in a famine year — arrived immediately. Doing the right thing at a moment when its benefits cannot appear and its costs are maximal is not merely bad timing; it is a policy error, and it burned the option for twenty-five years.
What would settle it: export performance in the following three to five years, controlled for the monsoon. The data exists and the answer is contested among economists rather than settled.
Why people care so much: because the political reading of 1966 — that opening the economy is surrender to foreigners — is still live in Indian politics, and it was formed in these months.
8.3 — The second moment, which nobody announced
The other thing that happened before 1991 is much less discussed and it complicates the standard story considerably.
Indian growth accelerated in the 1980s, from around 3.5 per cent to somewhere in the region of 5.5 per cent — before the reforms of 1991, and under Congress governments led by Indira Gandhi after 1980 and then Rajiv Gandhi.
There was no announcement and no doctrine. What happened was a series of unglamorous relaxations: some industries were delicensed, capacity limits were loosened, restrictions on expansion by existing firms were eased, imports of components and technology became somewhat freer, and a computer policy in 1984 opened the door for what became the software industry.
This belongs on the credit side and it is routinely omitted by both camps. The reformers omit it because it complicates the story that everything began in 1991. The defenders of the old model omit it because the acceleration came from loosening the model.
But it has a second half, and Chapter Nine is about that half. The growth of the 1980s was financed in part by external borrowing, and the borrowing is what turned the shock of 1990 into a crisis. The decade got its growth partly on credit, and the bill fell due in a single summer.
A household that has lived carefully for thirty years starts spending more freely. The house looks better. The children eat better. This is real: the family genuinely is better off, and anyone visiting can see it.
Some of it is a pay rise. Some of it is an overdraft. From the inside, and from the street, the two are indistinguishable — the difference only shows up when the bank calls.
India in the 1980s had both. The delicensing was a pay rise and it was real. The external borrowing was an overdraft. The bank called in June 1991.
Two moments, then, and they teach opposite lessons that the country learned in the wrong order. In 1966 the right economic decision was taken at the worst possible moment and was punished, which taught a generation that the decision itself was wrong. In the 1980s a series of small right decisions were taken quietly and worked, and almost nobody noticed, because nothing was announced and no argument was won.
The rupee was devalued in June 1966 — economically defensible, politically catastrophic, and followed by heavy Congress losses in 1967.
The lesson taken was that opening the economy is fatal and foreign-directed. What followed was bank nationalisation, the MRTP Act and FERA. One failed attempt discredited liberalisation for twenty-five years.
Growth accelerated in the 1980s, before 1991, under Congress, through quiet delicensing. That is a credit item both camps omit — and it was financed partly by borrowing, which is what made 1991 a crisis rather than a shock.
9Why It Took Until 1991
The correct diagnosis was on the file in 1969. The system ended twenty-two years later. What changed in between was not the argument.
9.1 — The constituency
Chapter Six named the three groups with a stake in the permission system continuing: the officials who issued permissions, the politicians for whom a discretionary permission is the most useful object in politics, and the large houses for whom it was a state-funded wall around their markets.
Add two more that the 1970s created.
Organised labour in the public sector and in protected industry. Employment in a licensed firm behind a tariff wall was secure and comparatively well paid, and after the mid-1970s a firm above a certain size could not retrench or close without permission. That is a genuine benefit to the people holding those jobs, and they were organised.
Farmers in the procurement system. Chapter Seven’s guaranteed price created a large, geographically concentrated group with a direct interest in a permanent arrangement.
Now count who was left. The people paying for the system were consumers paying above world prices, workers in the unorganised sector who had none of the protections, and firms that were never founded. Not one of those is a constituency. They have no association, no strike, no delegation, and in the third case no existence.
So the system was not sustained by belief. By 1980 there was very little belief left; the people running it could read the growth figures. It was sustained by the ordinary arithmetic of who could organise.
9.2 — What actually happened in 1991
The end, when it came, was not an argument being won.
Three things hit at once. The Gulf War pushed oil prices up sharply and India imports oil. The Soviet Union — India’s largest trading partner, with bilateral trade of over five billion dollars a year — collapsed, and with it a large share of Indian exports. And non-resident Indians, watching both, withdrew their deposits.
Underneath those shocks sat the structural position: after forty years of building for the domestic market behind a tariff wall, India had a small export base and therefore no reliable way of earning the foreign currency it needed. That is the same finding as Chapter Eight’s verdict on 1966, arriving twenty-five years later with interest.
Foreign exchange reserves stood at about 1.2 billion dollars in January 1991 and had roughly halved by June — enough for something in the region of two to three weeks of essential imports. The International Monetary Fund suspended its programme and the World Bank stopped disbursing. India was weeks from defaulting on its external obligations.
In May 1991, under a caretaker government, twenty tonnes of gold were flown to Zurich as collateral. In July, a further forty-seven tonnes went to the Bank of England. Sixty-seven tonnes in total, raising a few hundred million dollars, against an emergency loan from the Fund of about 2.2 billion.
The gold left the country in secret and the public learned about it roughly a week later.
Two to three weeks of imports means exactly what it says. Not a recession. Not a slowdown.
It means that in about a fortnight, the country cannot pay for the next tanker of crude oil. No oil means no diesel, which means no lorries, which means grain does not move from the surplus states to the deficit ones. It also means no fertiliser, no imported medicines, and no components for factories.
And it means a household’s gold — the thing an Indian family holds precisely because it is the last thing you sell — being flown out of the country by the state, in secret, for the same reason a family sells it.
The thing being avoided at that cost has a name, and it is worth knowing exactly what it would have meant.
Sovereign default: a country failing to pay what it owes to foreign lenders on time.
It is not the same as a business going bankrupt, because there is no court that can seize a country. What happens instead is that lending stops — for the government, and for every bank and company in that country — and for a poor country that imports its oil and much of its machinery, lending stopping means the machinery stopping.
Why it matters here: in June 1991 India was weeks away from this, and the government knew it before anybody else did.
9.3 — Who ended it
The reforms of July 1991 abolished industrial licensing for most industries, devalued the rupee in two steps, and began dismantling the trade restrictions.
They were carried out by a Congress government, under Prime Minister P.V. Narasimha Rao, with Manmohan Singh as Finance Minister.
This is a credit item of the first order and it goes in without qualification. The party that built the permission system dismantled it, quickly, competently, and against its own thirty-year rhetoric, at a moment when it held a minority in Parliament. Part Fourteen of this series covers that government properly and gives it its due.
What it does not do is settle the question this chapter asks, which is why it took until 1991.
This is contested in Indian politics every year, usually by people with an interest in the answer.
A minority government with no mandate for any of this dismantled its own party’s forty-year framework within weeks of taking office, and made it stick. Rao provided the political cover and Singh the design. Plenty of governments have faced a balance-of-payments crisis and responded with emergency borrowing and nothing else; this one used the crisis to change the system, which required a judgment about what the crisis meant and the nerve to act on it.
The reforms arrived at the precise moment there was no alternative, in substantially the form an external lender required, from a party that had rejected all of it for four decades and had been in power for most of them. Give credit for competent execution under duress, certainly. But a government that changes course only when it has run out of money is not being credited with a decision; it is being credited with recognising that the decision had already been made for it.
What would settle it: whether the same people would have done the same thing without the crisis. There is some evidence — the quiet liberalisation of the 1980s ran in the same direction — but the pace and scope of 1991 have no precedent in it.
Why people care so much: because the credit for 1991 is now a live claim in Indian politics, made by parties that were not in office and denied to a party that no longer campaigns on it.
Underneath that argument, though, both sides accept something that is not true.
Everyone telling the story of 1991 — the reformers who celebrate it and the critics who resent it — assumes that a bad policy ends when it is shown to be bad.
The celebratory version: the evidence accumulated, the argument was won, and in 1991 the right people finally had the courage to act on it. The resentful version: a crisis was used as a pretext to impose an argument that had not been won. Both are stories about an argument, differing only over whether it was won honestly.
The argument was over in 1969. Chapter Four established it: the government’s own committees, using the government’s own criteria, found the system producing the opposite of its purpose. Nothing of substance was added to that diagnosis in the following twenty-two years. Nobody’s mind needed changing.
What changed in 1991 was not knowledge. It was the bank balance. The system ended when the country ran out of money to run it, and not one day earlier — and the specific form the ending took was determined by what an external lender required, which is the least dignified possible way for a sovereign country to change its economic policy.
Draw the general rule, because this series will meet it again. A policy with an organised constituency does not end when it is refuted. It ends when the money runs out, when the constituency loses its grip, or when the policy is destroyed by an outside event. Being wrong is not on that list.
And note what this implies about the reforms themselves. Because they were made under duress rather than won by argument, the argument was never actually had in public. India changed its economic direction without any government ever persuading its own electorate that the change was right — which is why, thirty-five years later, the case still has to be made every time, and why the 1966 reading of liberalisation as foreign-imposed surrender has never entirely gone away.
One thing this part cannot do, and it should be said before the Remember box rather than left to Chapter Ten. It cannot tell you what those forty-one years cost in human terms, because that requires a counterfactual and Part One explained why counterfactuals cannot be measured. What it can tell you is that the diagnosis was complete in 1969, that the money was there, that the fair comparisons all run one way, and that the change came from a bank balance rather than an argument. Those four are enough.
The system was sustained not by belief but by who could organise: officials, politicians, large houses, protected labour and procurement farmers. Those paying for it — consumers, unorganised workers, and firms never founded — are not constituencies.
In June 1991 reserves covered about two to three weeks of imports. Sixty-seven tonnes of gold were flown out as collateral, in secret.
Licensing was abolished by a Congress government, and that is a credit item without qualification. But the argument had been over since 1969. What changed in 1991 was the bank balance.
10An Honest List of What We Do Not Know
The largest charge in this series rests on a comparison with a world that never happened. Here is exactly how much weight that can bear.
10.1 — Genuinely unknown
This part makes the biggest claim in the series — that a policy framework cost India decades of growth — and it is the claim that can least be proved. That has to be said plainly at the front of this chapter rather than buried.
What India’s economy would have been worth in 1991 under a different policy. Every estimate you will ever see of this is a model, and a model’s answer is decided by its assumptions.
Why it is unknown: it is a counterfactual across forty years and a billion people. Part One’s Chapter Three explained why these cannot be measured, and this is the largest one in the series. What can be done is comparison, which is Chapter Five, and comparison is weaker than measurement and better than opinion.
How much of the growth gap is the policy and how much is everything else. India differed from every comparison country in ways beyond economic policy: size, plurality, democracy, non-alignment, the absence of a bloc patron, geography, the monsoon.
Why it is unknown: you cannot hold those constant. Any statement about the share attributable to policy is an estimate with a wide band around it, and anybody quoting a precise figure is overstating what the method can deliver.
How much of the 1980s acceleration came from delicensing and how much from borrowing. Both happened. Both plausibly contributed.
Why it is unknown: disentangling them requires assumptions about what the borrowed money was spent on, and the accounts do not answer that at the necessary level of detail.
What happened to the committee reports inside the government. Chapter Four’s three reports are public. What was said about them in Cabinet, who resisted acting, and why the response was another permission rather than fewer, is not.
Why it is unknown: Cabinet papers and Congress party records for this period are not open. This is the decided category, and it is the single closed file that would do most to settle this part.
The scale of the market in permissions. Chapter Six argued that a permission system creates a price for permission. The size of that market — what proportion of licences involved a payment, and how large — is not established.
Why it is unknown: nobody keeps records of it, by design. The government’s own committee on corruption identified the mechanism and could not quantify it either. Anybody who gives you a figure has invented it.
10.2 — Solid
What the law required. The Industries (Development and Regulation) Act 1951, the Industrial Policy Resolution 1956, the MRTP Act 1969, the Foreign Exchange Regulation Act 1973, and the provisions requiring permission to retrench or close. All published statute. Nothing in Chapter Two rests on recollection.
What the government’s own committees found. The Monopolies Inquiry Commission in 1965, the Hazari committee, and the Dutt committee reporting in 1969. Published reports, commissioned by the government, staffed by establishment figures, applying the government’s own criteria. This is the strongest evidence in the part.
The growth rates. About 3.5 per cent a year from 1950 to 1980, against population growth over 2 per cent, giving per capita growth of roughly 1.3 per cent. Acceleration to somewhere around 5.5 per cent in the 1980s. These are standard national accounts figures and the disputes about them are at the margins.
The savings rate. From around 9.6 per cent of national income in the 1950s to roughly 20–25 per cent by the 1970s and 1980s. This is what forecloses the “too poor to invest” defence.
The 1966 devaluation. 6 June 1966, from about 4.76 to 7.50 rupees to the dollar. A matter of record, as is the electoral result that followed in 1967.
The Green Revolution outcome. Self-sufficiency in wheat within about seven years, and the end of ship-to-mouth dependence. Visible in several independent published series collected for different purposes.
The 1991 position. Reserves of about 1.2 billion dollars in January 1991, roughly halved by June, covering something in the range of two to three weeks of essential imports. Sixty-seven tonnes of gold pledged — twenty to Zurich in May under a caretaker government, forty-seven to the Bank of England in July. An IMF loan of about 2.2 billion dollars. Licensing abolished for most industries in July 1991 by a Congress government.
10.3 — What can actually be tested
Since the central counterfactual cannot be run, it is worth being clear about the four things in this part that can be checked by anybody who wants to.
Whether the diagnosis existed before the policy changed. Read the Dutt report. It is public. Its date is 1969 and the licensing system was abolished in 1991. That gap is not an interpretation.
Whether the system’s stated purpose was achieved. Its purpose was to prevent the concentration of economic power. Its own inquiry committee found it producing concentration. That is not an outside critic’s judgment.
Whether India was short of capital. The savings series answers it. It was not.
Whether the fair comparisons still go against India. Drop every country to which a legitimate objection exists. Pakistan and Indonesia remain, and both grew faster.
Those four are the load-bearing structure of this part. The counterfactual arithmetic — how many hundreds of millions of people would have been how much better off — is not, and I have deliberately not given you a number for it, because the honest answer is that nobody has one.
Genuinely unknown: what the economy would have been worth under different policy, how much of the gap is policy rather than everything else, how the 1980s growth divides between reform and borrowing, what was said about the committee reports inside government, and the size of the market in permissions.
Solid: what the law required, what the government’s own committees found, the growth rates, the savings rate, the 1966 devaluation, the wheat transformation, and the 1991 position.
The counterfactual cannot be run and I have not given you a number for it. Four things can be checked by anybody: the diagnosis predates the change by twenty-two years, the system produced the opposite of its purpose, India was not short of capital, and the fair comparisons still go against it.
Sources & further reading — Part 4
Timeline, 1950–1991
Dates only. The arguments are in the chapters.
| Date | What happened |
|---|---|
| 1951 | The Industries (Development and Regulation) Act. A licence is required to establish an undertaking in a listed industry. |
| 1951–56 | The First Five-Year Plan. Weight on agriculture and irrigation. Exceeds its own growth target. |
| 1956 | The Industrial Policy Resolution divides industry into three schedules. The Second Plan moves the weight to heavy industry. |
| 1957–58 | A foreign exchange crisis. Import controls tightened. |
| Early 1960s | Food output fails to reliably outpace population growth. Grain imports under PL-480 rise. |
| 1965 | The Monopolies Inquiry Commission reports on the concentration of economic power. The Agricultural Prices Commission and the Food Corporation of India are created. |
| 1965–67 | Two failed monsoons. Severe food crisis. Grain released in short tranches, with conditions attached. |
| 1966 | Mexican high-yielding wheat seed imported in bulk. The Green Revolution package begins. |
| 6 Jun 1966 | The rupee is devalued from about 4.76 to 7.50 to the dollar. Politically catastrophic. |
| 1967 | The Hazari committee finds licensing has produced disproportionate growth for large houses. Congress loses power in many states in the general election. |
| 1969 | The Dutt committee finds large houses taking licences and not implementing them, pre-empting capacity. Banks are nationalised. The MRTP Act adds a further permission layer. |
| Early 1970s | India reaches self-sufficiency in wheat. Ship-to-mouth dependence ends. |
| 1973 | The Foreign Exchange Regulation Act tightens control over foreign currency and foreign-owned firms. |
| Mid-1970s | Firms above a size threshold require government permission to retrench workers or close. |
| 1977 | Several international companies leave India rather than dilute shareholdings under FERA. |
| 1980s | Growth accelerates to around 5.5 per cent. Quiet delicensing, easier component imports, a computer policy in 1984. Financed partly by external borrowing. |
| 1990 | The Gulf War raises oil prices. Soviet trade collapses. Non-resident deposits are withdrawn. |
| Jan 1991 | Foreign exchange reserves at about 1.2 billion dollars. |
| May 1991 | Twenty tonnes of gold flown to Zurich as collateral, under a caretaker government. The public learns about a week later. |
| Jun 1991 | Reserves cover roughly two to three weeks of essential imports. India is weeks from default. |
| Jul 1991 | Forty-seven tonnes of gold to the Bank of England. An IMF loan of about 2.2 billion dollars. Industrial licensing abolished for most industries by a Congress government. |
Part Five starts here. It asks a question this part could not: the permission system operated on the whole country, so why did it leave some Indian states rich and others hollowed out? The answer is a set of rules almost nobody has heard of.
Glossary
Every hard word used in this part, in plain English.
| Word | What it means |
|---|---|
| Balance of payments | The record of a country’s money coming in from abroad against money going out. A crisis means it cannot pay for what it must buy from other countries. |
| Buffer stock | Grain the government buys and stores, so supplies can be released when harvests fail. Without one, a country must buy at the moment its need is visible to every seller. |
| Capital goods | Machines used to make other things, as opposed to things people consume. The centre of the Second Plan’s strategy. |
| Devaluation | Reducing the official value of a currency against others. Makes exports cheaper abroad and imports dearer at home. |
| FERA | The Foreign Exchange Regulation Act of 1973, controlling foreign currency and the operations of foreign-owned companies in India. |
| Foreign exchange | Foreign currency. Needed to buy anything from abroad. Rationed by the Indian state throughout this period. |
| Import substitution | Making things at home instead of buying them abroad, by taxing or banning the imported version until a domestic producer exists. |
| Licence Raj | The system of industrial permissions running from 1951 to the early 1990s. Not state ownership: private firms requiring official permission for ordinary business decisions. |
| Licensed capacity | The maximum quantity a firm was permitted to produce. Exceeding it was an offence, whatever the demand. |
| MRTP Act | The Monopolies and Restrictive Trade Practices Act of 1969, adding a further clearance requirement for large business houses. |
| Minimum support price | A guaranteed price at which the state will buy a crop. Removes the downside risk of adopting a new method. |
| PL-480 | The American food aid programme under which India bought grain for rupees rather than scarce dollars. |
| Pre-emption | Obtaining a licence you do not intend to use, so that nobody else can obtain one. Identified by the Dutt committee in 1969. |
| Rent-seeking | Spending effort to capture a larger share of existing wealth rather than to create new wealth. Rational for the firm, useless to everybody else. |
| Reserves (foreign exchange) | The stock of foreign currency a country holds. Usually measured in how many months of imports it would cover. |
| Savings rate | The share of national income not consumed. India’s rose from under 10 per cent in the 1950s to around a fifth by the 1970s. |
What Comes Next
Part Five — The Map of Who Got What
Part Five asks the question this part could not. The permission system applied to the whole country, so why did it leave some Indian states rich and others hollowed out? It asks:
- From 1952 to 1993 the central government equalised the price of coal, iron ore, steel and cement across India, so a factory anywhere paid what a factory at the mine paid. Who proposed it, what was it for, and what did it do to the states sitting on the minerals?
- Industrial licences were granted for particular places. Where did they physically go, decade by decade, and does the pattern match any stated policy?
- The Green Revolution went where the water was. Was any attempt made to extend it, and what happened to the districts that were never in it?
- River water and canal capacity were allocated by agreements and awards. Who won, who lost, and which disputes are still unresolved sixty years later?
- Plan transfers to states ran on formulas. What did the formulas reward, and did they narrow the gap between states or widen it?
- Bihar’s mineral belt was separated into Jharkhand in 2000, not by Congress. West Bengal’s sharpest decline runs from 1977 under the Left Front. What is left of the charge once both concessions are made?
Questions, not topics. If any of them is answered badly, you will be able to tell.