PMO centralisation is the charge most often levelled at the government Narendra Modi has led since 26 May 2014, and the argument usually proceeds by analogy. The Prime Minister’s Office under Modi, on the familiar telling, either repeats the office Indira Gandhi built between 1966 and 1977 with new branding, or it marks a complete rupture from every earlier pattern of rule in New Delhi. This article argues that both analogies fail, and that the failure can be demonstrated rather than merely asserted. Read through its institutional mechanisms and its structural constraints, the Modi PMO shares four structural features with the Indira PMO, and differs from it in nearly every mechanism by which those features are produced and nearly every constraint against which they operate.

Two Analogies, and the Claim This Article Defends

The simple-analogue reading treats the post-2014 Prime Minister’s Office as the Indira-era secretariat restored. Its holders point to a recognisable pattern: a leader who campaigns as the government, a cabinet that ratifies rather than deliberates, a parallel centre of authority beside the formal one, and regulatory bodies whose autonomy narrows under pressure. Opposition politicians state the analogy as an accusation. Some scholars state it as a diagnosis of continuity, arguing that the Indian executive was personalised long before 2014 and that Modi inherited an office whose dominance Indira Gandhi had already established between 1969 and 1977. On this reading, nothing structural separates the two periods; only the personnel and the party colour have changed.

The rupture reading runs in the opposite direction. Its holders, who include the government’s defenders and a strand of commentary that treats 2014 as a founding election, describe an office rebuilt from first principles: a think tank where a planning commission stood, committees and groups of secretaries where cabinet notes once travelled, a personnel system run through structured appraisal where personal loyalty once decided postings. On this reading the comparison with Indira Gandhi is a category error, because the instruments on the desk are different instruments, built for a different state.

Both readings fail on the same point. Each selects one layer of the institution and treats that layer as the whole. The analogue reading is correct about the structural features and silent about the mechanisms. The rupture reading is correct about several mechanisms and silent about the features those mechanisms serve. The claim defended here integrates the layers: Modi-PMO and Indira-PMO share structural features but operate through different institutional mechanisms and within different structural constraints, and the differences are not minor. The features are four. Both offices work as extensions of the Prime Minister’s personal authority. Both reduce the cabinet meeting to ratification. Both generate a parallel centre of power beside the formal architecture. Both press against the autonomy of regulatory and constitutional bodies. What produces those features, and what resists them, is where the two eras separate.

The evidence for the comparison is documentary before it is impressionistic. The Cabinet Secretariat’s procedural notifications on cabinet structure, in the 1969 to 1984 set and the 2014 to 2024 set, record how the committee architecture was formally arranged in each period. The founding resolution and annual reports of NITI Aayog record the post-2014 replacement for the Planning Commission. The Department of Personnel and Training notifications record the post-2014 pattern of senior appointments. Against those records stand the scholarly treatments: Inder Malhotra’s biographical work on the Indira-era office, Pratap Bhanu Mehta’s essays on institutional decline written between 2014 and 2024, and Christophe Jaffrelot’s Modi’s India (2021), which reads the post-2014 order as a new kind of regime. The article tests the three readings against the documents and adjudicates between them.

The comparison also belongs inside the larger frame of the series. The Modi era renegotiates the three founding tensions of the republic, and the integrated three-tensions framework is the frame within which the PMO question sits: an office is one of the places where the tension between personal authority and institutional form gets settled, government by government. The PMO is where that settlement is drafted.

Each analogy also has a political use, and the use explains its survival better than its accuracy does. The analogue reading serves an opposition that lacks a parliamentary majority: to call the Modi PMO a restoration of the Indira office is to accuse it of a known offence with a settled public verdict attached, the verdict of March 1977. The rupture reading serves a government elected on discontinuity: to call the office unprecedented is to claim that its methods cannot be judged by standards formed under earlier governments, because the standards belong to the order the election displaced. Both uses reward imprecision. The scholarly versions of the two readings are more careful than their political versions, and the adjudication below engages the scholarly versions at their strongest, Mehta for the continuity case and Jaffrelot for the rupture case, with Malhotra’s record of the earlier office as the baseline both must answer.

The method matters to the verdict. Claims about an office can rest on three kinds of material: the memoirs and testimony of participants, the journalism of the period, and the procedural record the institution itself generates. This article weights the third highest, because notifications, resolutions and appointment orders survive checking in a way recollections do not. The Cabinet Secretariat’s procedural notifications are the least cited of these sources; partially available through Right to Information requests and parliamentary answers, the 2014 to 2019 set records the committee compositions, the Groups of Secretaries and the reporting lines through which the post-2014 office was actually assembled, decision by decision. Where the procedural record and the memoirs conflict, the article follows the record and notes the conflict.

What PMO Centralisation Means as a Measurable Condition

Centralisation is a loose word in political argument, so the article fixes a working definition before using it. PMO centralisation means the concentration, inside the Prime Minister’s Office, of four functions that the formal architecture distributes elsewhere: the initiation of policy, the movement and clearance of files, the approval of senior appointments, and the review of implementation. A PMO is centralising to the degree that ministries originate less and clear more, that decisions are settled before the cabinet meets, that the Appointments Committee of the Cabinet and the personnel ministry process choices the PMO has framed, and that implementation is monitored from the PMO rather than from the line ministry or the Cabinet Secretariat.

Each function leaves evidence. Initiation shows in where a proposal is first drafted and whose draft reaches the cabinet note. Clearance shows in the routing of files and in the record of decisions announced before a cabinet discussion took place. Appointments show in the Department of Personnel and Training notifications and in the tenure pattern of the officers notified. Review shows in the meeting architecture: who summons secretaries, who sets the agenda of their meetings, and who receives their reports. Measured on those four traces, centralisation is an empirical condition, not a mood, and the condition can be compared across decades because the traces survive in the record of both eras.

The legal position of the office makes the measurement necessary. The Prime Minister’s Office has no statutory charter and no constitutional text of its own. The Constitution distributes executive power through Articles 74 and 77: the President acts on the aid and advice of the Council of Ministers, and the President makes rules for the transaction of government business and its allocation among ministries. The office of the Prime Minister draws its procedural force from those rules and from the Prime Minister’s position inside them, not from any Act of Parliament. An institution with no legal personality of its own can grow very large without any formal amendment occurring, which is precisely why impressions mislead and traces matter.

What do the Transaction of Business Rules require before a decision binds the government?

The Transaction of Business Rules, framed in 1961 under Article 77 of the Constitution, list the classes of case that must reach the Cabinet or a Cabinet committee before a decision binds the government: proposals for legislation, changes in taxation, and cases the Prime Minister calls for. Ministries may dispose of other cases themselves.

The same rules contain the hinge on which centralisation swings. The Prime Minister may permit a departure from the rules in a given case, and a ministry may act on a decision communicated through the PMO with the assurance that formal approval will follow. The Cabinet Secretariat administers the rules, maintains the schedules of cases reserved for collective decision, and services the cabinet and its committees. The text therefore describes two governments at once: a formal government in which reserved classes of business require collective approval, and a working government in which the Prime Minister’s permission can route a decision around the collective stage. Every era of PMO strength has lived in the gap between those two descriptions. The Indira era widened the gap by practice. The post-2014 era, the sections below argue, widened it by design, building standing machinery inside the gap itself.

The Cabinet Secretariat deserves a place in the measurement, because the formal design gives it the role of the cabinet system’s memory and referee. Its secretary, the Cabinet Secretary, is the head of the civil service, services the cabinet and its committees, and administers the Transaction of Business Rules. In a distributed executive the Cabinet Secretary’s office is where inter-ministry disputes are composed before they reach ministers. The centralisation trace appears here as a shift of gravity: where the PMO’s own review platforms summon secretaries directly, set the agenda of their collective meetings and receive their reports, the Cabinet Secretariat’s composing function attenuates, and the Cabinet Secretary’s role adjusts from referee of the system to coordinator of the centre’s programme. The adjustment is visible in the post-2014 record, where the Cabinet Secretary presented the Groups of Secretaries’ work to the Prime Minister at the meeting of 27 October 2016 and the groups themselves reported past the secretariat to the PMO. No rule was amended to produce this shift. It is the kind of change the four traces exist to catch, invisible to a reader of the formal chart and unmistakable in the meeting record.

The Office Before 2014: A Secretariat That Grew by Accretion

The institution began small. The Prime Minister’s Secretariat was created in 1947 to service Jawaharlal Nehru, and through the Nehru years it remained a modest office, because the cabinet contained figures of independent standing, Vallabhbhai Patel until his death in December 1950 foremost among them, and because policy travelled through ministries headed by politicians with their own followings. The secretariat coordinated; it did not command. Lal Bahadur Shastri, who succeeded Nehru in June 1964, worked through L.K. Jha, an Indian Civil Service officer who served as his secretary and gave the office a professional weight it had not previously carried, and Jha continued briefly under Indira Gandhi after January 1966.

The founding generation argued about the office’s proper size in terms that set the baseline for every later argument. The correspondence between Nehru and Vallabhbhai Patel in the government’s first years, over Kashmir, over Hyderabad and over the handling of the services, records two principals negotiating the boundary between a Prime Minister’s secretariat and a cabinet of colleagues, with Patel insisting that decisions belonged to the government collectively and the secretariat’s task was to service that collective process. Practice matched the doctrine: a small staff handling correspondence, tours and the Prime Minister’s engagements, its officers drawn from the Indian Civil Service and the personal staff in roughly equal informality, its files carrying no authority beyond the Prime Minister’s signature at the end of them, and policy advice arriving through ministries whose heads Nehru treated as principals in their own right. The baseline matters because both the eras compared in this article define themselves against it. Indira Gandhi’s rebuilding was a deliberate departure from the Nehruvian secretariat, understood at the time as such. The post-2014 office presents itself as a restoration of decisiveness against a later drift, but its architects inherited the same founding chart and the same unanswered question the chart poses: an office with no legal personality grows exactly as large as the Prime Minister’s method requires, and no larger rule restrains it.

The expansion came with Indira Gandhi, and it came in stages. P.N. Haksar, who became secretary to the Prime Minister in 1967 and principal secretary in 1971, rebuilt the secretariat into the government’s policy centre: Haksar’s office vetted appointments, drafted the economic programme that produced bank nationalisation in July 1969, and ran the political secretariat of the 1971 war from inside the Prime Minister’s house. The secretariat’s staff and reach grew with each crisis, and the growth was personal in the sense that it depended on Haksar’s standing with the Prime Minister rather than on any notified redesign of the institution. When Haksar moved out in 1973 and P.N. Dhar succeeded him, the office kept its centrality and lost none of its habits.

When did the Prime Minister’s Secretariat become the Prime Minister’s Office?

The Janata government of Morarji Desai renamed the Prime Minister’s Secretariat as the Prime Minister’s Office in 1977, in the months after the March 1977 election ended the Emergency. The renaming announced an intention: the office would serve a cabinet government again rather than command one, and its staff was cut.

The Janata interlude demonstrated both the intention and its limits. Morarji Desai governed through his cabinet to a degree no Prime Minister since Nehru had attempted, and the office shrank accordingly between 1977 and 1979. Yet the formal architecture the Janata government inherited, the Transaction of Business Rules, the committee schedules, the personnel ministry inside the Prime Minister’s charge, was left intact, and the contraction of the office lasted only as long as the Prime Minister who willed it. Indira Gandhi’s return in January 1980 brought the office back to the centre, and the reconstruction of those years, the post-1980 office-rebuilding completed between 1980 and 1984, installed P.C. Alexander as principal secretary in 1981 and restored the secretariat’s vetting of appointments and files. Rajiv Gandhi, who succeeded his mother in October 1984, gave the office a different texture, a younger circle of advisors and an impatience with file procedure, and the period’s controversies, from the legislative reversal of the Shah Bano judgment in 1986 to the Bofors purchases, turned repeatedly on how decisions had really been taken inside that circle. P.V. Narasimha Rao, heading a minority government from June 1991, then ran the office as an instrument of quiet management: his principal secretary A.N. Verma coordinated the economic reform programme across ministries, and Rao’s method, deciding in committee and letting ministers front decisions the PMO had framed, showed that a weak parliamentary position could coexist with a strong office where the Prime Minister’s political skill supplied the authority the arithmetic denied.

The decades after 1989 then supplied a natural experiment. The coalition governments from 1989 to 2014, surveyed in the pillar study of the coalition era, constrained the office by arithmetic: a Prime Minister dependent on allies could not run the government from a single secretariat, and the office under H.D. Deve Gowda and I.K. Gujral, whose governments rested on outside support, contracted to a coordinating staff. The exception inside the coalition decades was Atal Bihari Vajpayee, whose principal secretary Brajesh Mishra held the additional charge of National Security Advisor from November 1998 and gave the Vajpayee PMO a strategic weight its coalition base would not have predicted, though the allies retained vetoes that the PMO had to price into every decision.

The Manmohan Singh decade, from 2004 to 2014, inverted the Indira pattern. The PMO under T.K.A. Nair and then Pulok Chatterjee was administratively competent and politically subordinate, because a second centre of authority stood outside the government altogether: the Congress presidency, institutionalised in the National Advisory Council that Sonia Gandhi chaired from 2004 to 2006 and again from 2010 to 2014, which drafted legislation the cabinet then processed. Sanjaya Baru, who served as the Prime Minister’s media advisor from 2004 to 2008, recorded in The Accidental Prime Minister (2014) how files and appointments bent toward that second centre. The lesson the post-2014 designers drew from the decade was available to any reader of Baru’s memoir: an office vacated at the top would be occupied from outside the government, and the remedy was to fill the office completely.

Feature One: The Personality-as-Institution Dynamic

The first shared feature is the oldest. In both eras the office functions as an architectural extension of the Prime Minister’s personal authority rather than as an institution that constrains or balances that authority. The analytical baseline for this feature is the personality-as-institution framework, which describes how Indira Gandhi converted a party and a government into instruments of a single political person. Her secretariat embodied the conversion. Haksar’s authority rested on his access to her, the officials who rose were those she trusted, and the discourse of a committed bureaucracy, defended in print by her minister Mohan Kumaramangalam, held that officials should serve the programme of the elected leader rather than a neutral ideal of administration. The office did not survive its occupant in the same shape: the institution was the person, staffed.

The post-2014 office reproduces the pattern with precision. Narendra Modi won the 2014 general election on a campaign that named the leader rather than the party as the offer before the voter, and the Bharatiya Janata Party’s own campaign committee architecture was subordinated to that offer after July 2013, when Modi took charge of the campaign. The 2019 campaign repeated the design and returned 303 seats for the party on the strength of it. Inside the government, the pattern shows in the branding of programmes to the Prime Minister’s person, in the direct communication channels that run from the Prime Minister to the citizen without ministerial mediation, and in a PMO staffed by officers whose standing derives from the Prime Minister’s confidence, among them Principal Secretary Nripendra Misra from 2014 to 2019 and National Security Advisor Ajit Doval from 30 May 2014.

Two consequences follow from the shared feature, and both eras display them. The first is succession fragility: an office built as a personal extension has no settled second line, so authority gathers in a small number of confidants and the institution’s memory lives in their heads. The second is accountability direction: officials inside such an office answer upward to the person rather than outward to the cabinet, the Parliament or the record, and their written advice adjusts accordingly. Inder Malhotra’s reporting on the Indira-era office documented the upward answerability in detail, and the post-2014 record, from the appointment notifications to the review architecture described below, displays the same direction of answerability. The feature is genuinely shared. What differs is the machinery through which personal authority is converted into decisions, and that machinery occupies the mechanism sections of this article.

The personal mandates were constructed in parallel ways forty-three years apart. Indira Gandhi fought the 1971 election on a plebiscitary appeal over the heads of her party’s organisation, which had split in 1969, and the slogan Garibi Hatao was addressed to the voter as a contract with the leader personally. D.K. Barooah, her party’s president, is reported to have compressed the resulting doctrine into a single sentence in 1975: India is Indira and Indira is India. The Bharatiya Janata Party’s 2014 campaign was designed around the same compression, with Modi named as the party’s prime ministerial candidate in September 2013, a campaign committee under his direction from July 2013, and constituency level organisation subordinated to the national leader’s tour and image. Political scientists describe the form as presidentialisation, the migration of authority within a parliamentary system toward the leader’s person, and the comparative literature, Poguntke and Webb’s The Presidentialization of Politics (2005) among it, treats the form as a general tendency of media era democracies. The Indian evidence carries the tendency further than the comparative norm in both the eras compared here, because in both the leader’s person became the party’s principal asset and the government’s principal warrant. What the comparative frame cannot supply is the institutional sequel: presidentialised leaders in other systems inherit offices designed for presidents. Indian Prime Ministers inherit an office designed for a first among equals, and must rebuild it around themselves. The rebuilding is the subject of the mechanism sections.

Feature Two: The Cabinet as a Ratifying Chamber

The second shared feature concerns the cabinet meeting itself. In both eras the meeting ratifies decisions whose substance has been settled at the level of the Prime Minister’s Office, and the documentary record of each era contains a canonical instance. For the Indira era the instance is bank nationalisation. On 16 July 1969 Indira Gandhi took the finance portfolio from her deputy Morarji Desai, and on 19 July 1969 an ordinance nationalised fourteen commercial banks. The cabinet had not deliberated the proposal as a collective body; the ordinance route and the portfolio manoeuvre together ensured that the decision arrived as a fact the government then defended. Haksar’s memoranda, preserved in his papers, record the preparation inside the secretariat.

The Emergency supplied the harder instance. On the night of 25 June 1975, President Fakhruddin Ali Ahmed signed the proclamation of Emergency on the Prime Minister’s advice, transmitted without a prior cabinet meeting. The cabinet met on the morning of 26 June 1975 and was informed of a proclamation already in force; the ministers endorsed what they had not decided. The Shah Commission, appointed in May 1977 under Justice J.C. Shah, documented the sequence from the government’s own records and treated it as the defining procedural fact of the period: the collective organ of the executive had been presented with a concluded act. The Emergency’s institutional architecture is the setting in which that ratification habit reached its extreme form, and the habit outlived the Emergency itself, because the secretariat that had learned to decide first and inform later remained in place after 1980.

The post-2014 instance is demonetisation. On 8 November 2016 the central board of the Reserve Bank of India met and recommended the withdrawal of the 500 and 1000 rupee notes, and the Union Cabinet met the same evening, shortly before the Prime Minister’s televised address at 8 pm announced the decision to the country. Ministers were informed of a decision whose operational design, the secrecy, the printing schedules, the banking instructions, had been prepared inside a small circle around the PMO and the Reserve Bank over preceding weeks. Press reports of the meeting recorded that ministers left their mobile telephones outside the room. The cabinet’s approval was constitutionally sufficient and deliberatively empty: no minister in that room could have altered the design, and none was presented with an alternative. The pattern repeats in quieter cases. The decisions of 5 August 2019 on Jammu and Kashmir moved from a cabinet meeting that morning to introduction in the Rajya Sabha within hours, a sequence possible only because the drafting had been completed inside the executive’s inner circle before either the cabinet or the Parliament saw the instruments.

Defenders of both eras answer that speed and secrecy are sometimes the substance of a decision, and the answer deserves its weight. A currency withdrawal announced in advance defeats itself, and a nationalisation debated for months invites capital flight. The adjudication does not turn on whether ratification is ever justified. It turns on frequency and default. Where the ratifying meeting is the exception, the cabinet system survives its suspensions. Where the meeting becomes the standard terminus of decisions made elsewhere, the cabinet system has changed its function, and the change, in both eras, is a fact of the record rather than an interpretation of it.

The contrast case clarifies the charge. Between 2007 and 2008 the Manmohan Singh government processed the civil nuclear agreement with the United States through months of genuine intra-governmental contest: a UPA and Left coordination committee heard the negotiators, the cabinet’s security committee weighed the text, and the Prime Minister staked his office on the outcome in a confidence vote on 22 July 2008. Whatever that government’s weaknesses, the cabinet system deliberated, because the Prime Minister lacked the authority to pre-empt it. The post-2014 pattern differs not in the existence of committees, which multiplied, but in the direction of travel: texts arrive at committees settled in substance, and the committee’s work is clearance. The security decisions of February 2019, after the Pulwama attack of 14 February, display the form at its most defensible and most characteristic at once: the Cabinet Committee on Security authorised the air strike of 26 February 2019, a real authorisation by the competent body, taken at a single sitting on the Prime Minister’s recommendation, in a process no wider deliberation touched. Emergency powers in a crisis are what such committees exist for. The ratification charge rests on the migration of the crisis form into ordinary business, taxation, currency and state reorganisation, where no comparable urgency argued for the narrow room.

Feature Three: The Parallel Centre Beside the Formal One

The third shared feature is a second centre of power that operates beside the formal architecture and answers to the leader personally. The Indira-era instance is Sanjay Gandhi. Between 1975 and 1977 the Prime Minister’s younger son held no government office and no party office of consequence until his appointment to the Congress general secretaryship came later, yet officials took instructions from him, his Maruti project enjoyed the sponsorship of the state, the sterilisation programme of 1976 to 1977 ran to quotas his camp enforced, recording more than eight million sterilisations in a single year, and the demolitions at Turkman Gate in Delhi in April 1976 proceeded under the authority his name supplied. The Shah Commission’s record shows administrators explaining their conduct by reference to a man the Constitution did not mention. Authority flowed around the formal chart because the person at the top allowed a channel to run beside it.

The post-2014 instance is Amit Shah, and the comparison must be drawn exactly, because the difference inside the shared feature is instructive. Shah became president of the Bharatiya Janata Party in July 2014 and ran its organisational machinery, its membership drives, its candidate selection and its state campaigns as a parallel structure of authority beside the government. From 31 May 2019 he has held the Home Ministry, the ministry that controls the Indian Police Service cadre jointly with the states, the central armed police forces, and the Union’s relations with state governments. The Cabinet Secretariat’s June 2019 committee notifications placed Shah on every standing committee of the cabinet, the only minister besides the Prime Minister so placed. PMO and Home Ministry coordination, on appointments, on state politics and on security, therefore runs through a single political principal who also commands the ruling party’s machine.

The shared feature is the channel beside the chart: in both eras, a figure whose authority derives from personal proximity to the Prime Minister exercises power that the formal architecture assigns elsewhere. The difference is one the mechanism sections develop. Sanjay Gandhi’s channel was informal in the strict sense, carrying no office, no file and no accountability to any record, and it collapsed the day its protector lost power in March 1977. Shah’s channel runs through offices the Constitution recognises, a party presidency won in the party’s internal process and a cabinet ministry held at the Prime Minister’s pleasure. A parallel centre built inside formal offices behaves differently from one built outside them: it leaves records, it survives reshuffles, and it can be audited against the chart it shadows. The feature is shared; its institutional form is not, and the form governs the costs.

The aftermath of the parallel centre marks the difference in form. When the Janata government appointed the Shah Commission in May 1977, the informal channel became an object of formal inquiry, and officials testified that instructions had arrived from a person no file recorded; the evidentiary embarrassment of that testimony, administrators unable to produce the orders they had obeyed, is the permanent weakness of the informal design. Sanjay Gandhi’s five point programme, covering family planning, afforestation, dowry abolition, literacy and the eradication of caste, was announced beside the government’s own twenty point programme, and state administrations reported against both, a doubling of the chain of command that no organisation chart admitted. The post-2014 parallel centre leaves the opposite evidentiary trail. Shah’s party presidency generated records of the ordinary political kind, his ministerial decisions issue over his own signature, and his presence on every cabinet committee stands in a published notification of June 2019. A critic auditing the channel finds documents. A critic auditing Sanjay Gandhi’s channel found memories, and the difference is not a detail of bookkeeping: it determines whether the parallel centre can be examined, priced and, in principle, dismantled by ordinary institutional means.

Feature Four: Pressure on Regulatory and Constitutional Autonomy

The fourth shared feature is the systematic narrowing of the autonomy of bodies designed to stand at a distance from the executive: the Reserve Bank of India, the Election Commission, the Comptroller and Auditor General, and the process of judicial appointment. The Indira era supplies the stark instances. In April 1973, the day after the Kesavananda Bharati judgment, the government appointed Justice A.N. Ray as Chief Justice of India over three senior colleagues, and the supersession announced that judicial advancement would price in acceptability to the executive. A second supersession followed in January 1977, when Justice M.H. Beg was elevated over Justice H.R. Khanna, the lone dissenter in the habeas corpus case of April 1976. The Reserve Bank through the 1970s operated as an adjunct of the finance ministry, its governors serving in an atmosphere in which directed credit and deficit financing were settled politically and executed monetarily.

The post-2014 record is documented with a precision the earlier era cannot match, because the information environment, treated as a constraint below, forces episodes onto the record. In the autumn of 2018 the government and the Reserve Bank entered a public contest. On 26 October 2018 the Deputy Governor Viral Acharya, speaking in Mumbai, warned that governments which undermine a central bank’s independence incur the wrath of financial markets. Press reports through October and November 2018 described consultations initiated under Section 7 of the Reserve Bank of India Act 1934, the provision that empowers the government to direct the Bank in the public interest, a provision no government had formally invoked. On 10 December 2018 Governor Urjit Patel resigned, citing personal reasons, eighteen months after the demonetisation his institution had been required to execute. In May 2019, Election Commissioner Ashok Lavasa recorded dissents in the Commission’s orders on campaign complaints against senior ruling party figures and objected when the dissents were omitted from the published orders; his objections became public through the press. The Comptroller and Auditor General’s report on the Rafale aircraft purchase, tabled in February 2019, was criticised by opposition parties and former officials for the redactions in its pricing chapter, and the criticism itself entered the public record.

The adjudication of this feature requires care in both directions. The analogue reading treats the 2018 Reserve Bank contest as the Emergency pattern restored, and the record does not carry that weight: no direction under Section 7 issued, the Governor’s resignation was his own act publicly recorded, and the Election Commissioner’s dissent survived in the newspapers if not in the orders. The rupture reading treats the same episodes as routine friction, and that reading fails too, because friction that ends in a Governor’s resignation and a Commissioner’s public dissent is not routine. The feature is shared at the level of direction: both executives pressed against autonomous bodies and both treated autonomy as an obstacle to be managed. The level of form differs, and the form is where the comparison becomes exact, in the mechanisms and constraints that follow.

Two longer arcs sharpen the feature. The Reserve Bank’s subordination in the Indira decades had a fiscal mechanism: the government’s deficit was financed automatically through ad hoc treasury bills, a device that converted the Bank into the exchequer’s cashier and survived until the agreements of September 1994 between the Bank and the government, under Governor C. Rangarajan, phased the device out by 1997. Autonomy at the Bank was rebuilt as machinery before it was defended as doctrine, which is why the 2018 contest was fought over a consultation clause rather than over the printing of money. The Comptroller and Auditor General traces the opposite arc across the same decades: an office the executive could usually ignore became, under Vinod Rai from 2008 to 2013, an auditor whose reports on spectrum and coal allocations set the political agenda against the government of the day, and the post-2014 controversy over the Rafale audit’s redactions is best read as an executive contesting an office whose demonstrated capacity to damage it was recent memory. The judiciary’s arc completes the set: sixteen high court judges were transferred during the Emergency in 1976 in circumstances the Shah Commission later examined as punishment postings, an instrument of discipline the post-2014 executive, facing the collegium and the basic structure doctrine, does not possess in the same form.

The Election Commission’s arc supplies the benchmark against which both eras read. Under T.N. Seshan, Chief Election Commissioner from December 1990 to December 1996, the Commission converted its constitutional text into an assertive enforcement of the model code that governments of both national parties experienced as a constraint with teeth, and the multi-member structure created in October 1993, partly in reaction to Seshan himself, institutionalised the assertiveness beyond one personality. The dissent episode of 2019 is measured against that benchmark by the Commission’s own alumni: a body that had enforced the code against ruling parties for three decades published orders from which a member’s dissent had been removed. The appointment method frames the whole feature. Governors of the Reserve Bank, Election Commissioners and the Comptroller and Auditor General are appointed by the executive they are meant to check, under processes Parliament has repeatedly examined and repeatedly left unchanged, the appointment reforms proposed in the Dinesh Goswami Committee report of 1990 among the unadopted. Both eras exploited the method. The post-2014 difference lies in the density of the surrounding architecture that records the exploitation: a dissent can be omitted from an order, and the omission itself becomes a document.

The Four Features Side by Side: Table 1

The argument so far can be stated in a single artifact. Table 1, the Indira-Modi PMO four-feature comparison table, sets each shared feature against the mechanism that produced it in the Indira era, the mechanism that produces it after 2014, and the constraint that bound each era’s version. The table is the article’s citable core: a reader who accepts the table’s rows has accepted the namable claim, because the rows show common features carried by uncommon machinery.

Feature	Shared structural pattern	Indira-era mechanism	Modi-era mechanism	Constraint difference
Personality as institution	The office extends the Prime Minister’s personal authority and answers upward to the person	A secretariat rebuilt around P.N. Haksar’s access after 1967, loyalty screened through the Prime Minister’s household	A campaign centred on the leader, a PMO staffed by confidants, direct leader to citizen channels	Indira faced a party of regional bosses; Modi faces a party his own machine organises
Cabinet as ratifying chamber	Decisions settle before the cabinet meets; the meeting endorses	Ordinance of 19 July 1969; Emergency proclaimed on 25 June 1975 before the cabinet met	Demonetisation cleared on 8 November 2016 the evening of its announcement; Kashmir instruments moved on 5 August 2019 within hours	The Transaction of Business Rules text is common; the post-2014 committee machinery routes more business around the full cabinet
Parallel centre	Authority flows through a figure beside the formal chart, answerable to the leader	Sanjay Gandhi, holding no office, directing officials and programmes from 1975 to 1977	Amit Shah, party president from 2014 and Home Minister from 2019, present on every cabinet committee	The informal channel left no record and collapsed in 1977; the formalised channel leaves records and survives reshuffles
Pressure on autonomous bodies	The executive treats regulatory and constitutional autonomy as an obstacle	Judicial supersessions of 1973 and 1977; a Reserve Bank run as a finance ministry adjunct	Section 7 consultations reported in 2018; a Governor’s resignation; a Commissioner’s recorded dissent	Supersession cost the Indira executive little; the post-2014 episodes generated litigation, publicity and delay within a denser legal architecture

Read vertically, the rows also show why the features persist across regimes that agree on little else. Each feature answers a standing temptation of the Indian executive: a vast administration, a federation of strong states and a Parliament of many parties reward the Prime Minister who shortens the chain between decision and execution. The temptation is structural, and governments of opposed ideologies have yielded to it in the same four places. What a period’s politics decides is not whether the temptation operates but what machinery it builds and what resistance it meets, which is why the table’s last three columns carry the argument.

The table also locates the disagreement with the two popular readings precisely. The analogue reading describes the second column well and stops there. The rupture reading describes the fourth column and stops there. The third and fifth columns, the Indira mechanisms and the constraint differences, are the evidence both readings omit, and the next three sections take the mechanisms in turn: the planning body, the committee architecture and the personnel system.

Mechanism One: A Think Tank Where an Allocating Commission Stood

The first mechanism difference is the planning body, and it is the clearest documentary contrast between the eras. The Planning Commission, created by a cabinet resolution of 15 March 1950 and chaired by the Prime Minister, held allocation authority: plan assistance flowed to states through formulas the Commission administered, the Gadgil formula of 1969 among them, and central schemes carried conditions the Commission’s approval enforced. Indira Gandhi’s PMO reached into ministries and states through that allocating body, because a secretariat whose head chaired the Commission could price its preferences into the plan itself. The Commission was an instrument of centralisation with money attached.

NITI Aayog, created by a cabinet resolution of 1 January 2015 in the first year of the post-2014 governance architecture, holds no allocation authority. Its founding resolution describes a policy think tank: a Governing Council of chief ministers and lieutenant governors, a vice chairman and chief executive appointed by the Prime Minister, and a mandate to design frameworks, foster cooperative federalism and monitor outcomes. The money that once moved through plan transfers had already been rerouted by the Fourteenth Finance Commission, whose award raised the states’ share of the divisible tax pool to 42 percent from 2015, shrinking the discretionary stream the old Commission had administered. NITI Aayog ranks states on indices, runs the Aspirational Districts Programme launched in January 2018 across 115 districts, and publishes evaluations. It persuades and measures; it does not allocate.

The mechanism difference changes how centralisation works. The Indira-era PMO centralised by holding a resource the states and ministries needed, and compliance was purchased through the plan. The post-2014 PMO centralises by holding the measuring rod: rankings, reviews and outcome dashboards define the terms on which ministries and states are judged, while the resource itself flows by constitutional formula through the finance ministry’s budget. A ministry can defy a ranking at a political price; it could not defy an allocation without a fiscal price. Which lever binds harder is a genuine question, and scholars divide on it. The point for the comparison is narrower and firmer. The instrument on the desk is a different instrument, built in a founding resolution whose text survives, and any account of the Modi PMO that describes it as the Indira secretariat restored has to explain away the document of 1 January 2015.

The planning body’s federal forum tells the same story in miniature. The National Development Council, created in 1952 to associate the states with plan formulation, met through the decades as the federation’s summit on planning; Indira Gandhi used its sessions to discipline state plans and to carry programmes, the twenty point programme of 1975 among the era’s instruments, that states were expected to execute. The Council has not been convened since December 2012, and the post-2014 government let it lapse without abolition, replacing its function with NITI Aayog’s Governing Council, a body that meets at the Prime Minister’s summons, deliberates without approving allocations, and issues no resolutions a state must answer. The 2015 exercise in which a sub-group of chief ministers, convened under NITI Aayog, recommended the rationalisation of centrally sponsored schemes shows the new form at work: states consulted collectively on the design of transfers whose size the Finance Commission and the budget had already fixed. Consultation without allocation is the mechanism’s essence, voice without the purse, and it inverts the Indira-era arrangement in which the plan’s purse made the centre’s voice heavy in every state capital.

Mechanism Two: Committees and Groups That Report to the Centre

The second mechanism difference is the architecture of committees and groups through which the post-2014 PMO coordinates the government. The Indira-era cabinet worked through a small set of committees, the Cabinet Committee on Political Affairs dominant among them, and through informal circles whose membership shifted with the Prime Minister’s confidence. The formal committee system was narrow, and the real coordination happened in the secretariat and the household. The post-2014 system inverted the design: the formal committees were widened and multiplied, and a second layer of official groups was built beneath them, reporting upward to the Prime Minister personally.

The committee layer was reset after the 2019 election. On 6 June 2019, as the Cabinet Secretariat confirmed in answer to a Lok Sabha question of 10 July 2019, the government constituted two new standing committees, the Cabinet Committee on Investment and Growth and the Cabinet Committee on Employment and Skill Development, alongside the six reconstituted committees, for a total of eight. The Prime Minister chairs six of the eight; the Home Minister chairs the remaining two and sits on all of them. The design places the two subjects the government treated as its economic core, investment and employment, inside small committees chaired from the top, rather than inside the full cabinet where twenty-five ministers deliberate.

What is the Cabinet Committee on Investment and Growth?

The Cabinet Committee on Investment and Growth is a standing committee of the Union Cabinet constituted on 6 June 2019, chaired by the Prime Minister, with the Home Minister, the Finance Minister, the Transport Minister and the Commerce Minister as members. It clears stalled investment proposals outside the full Cabinet’s calendar.

When was the National Security Council established?

The National Security Council was established by the Vajpayee government on 19 November 1998, with Brajesh Mishra as its first National Security Advisor and the head of its secretariat. The Council gave the Prime Minister a standing apex forum for security and strategic decisions, serviced by a National Security Council Secretariat.

The group layer is the more original construction. In January 2016 the Prime Minister constituted eight Groups of Secretaries, each charged with a theme of governance, and received their presentations in person, two groups presenting on 21 January 2016 and the exercise closing with the Prime Minister’s response to all eight. In October 2016 ten new sectoral groups replaced the thematic eight, reporting by the end of November, and the pattern of groups presenting directly to the Prime Minister, with rapporteurs and joint secretary teams beneath them, became a standing method. A monthly review platform, PRAGATI, operating from 25 March 2015, put the Prime Minister in video conference with union secretaries and state chief secretaries over stalled projects, compressing the distance between the PMO and the district-facing end of the administrative chain. The procedural record of this machinery reaches the citizen in fragments: committee notifications, presentation files and answers to parliamentary questions, a documentary trail that researchers assemble from office files and slide decks, the kind of material a reader can open in a browser through a presentation viewer when the original office software is unavailable.

The contrast with the Indira mechanism is structural. Indira’s coordination ran through persons: Haksar’s secretariat, the household, the kitchen cabinet of trusted ministers. Its records are memoirs and commission testimony, because the machinery was deliberately undocumented. The post-2014 coordination runs through notified committees and constituted groups whose compositions are published and whose presentations follow a calendar. Personal authority is the shared feature; in one era it travelled through confidants, and in the other it travels through an architecture of committees and groups that the confidants administer. Architecture can be examined, dated and compared. A household cannot, and the difference governs what a later government could dismantle and what it would inherit.

The group layer replaced something, and the replacement is the point. The Manmohan Singh governments had coordinated through Groups of Ministers and Empowered Groups of Ministers, ministerial bodies chaired by senior cabinet figures, Pranab Mukherjee pre-eminent among them, which deliberated across ministries and reported to the cabinet. On 31 May 2014 the incoming government abolished the standing Groups of Ministers and their empowered variants in a single decision, announcing that ministries would decide and be accountable. Coordination did not disappear; it migrated downward and upward at once, downward to the Groups of Secretaries who draft and upward to the Prime Minister who receives their drafts, bypassing the ministerial middle where the earlier coordination had lived. A Group of Ministers is a committee of principals who can bargain. A Group of Secretaries is a committee of officials who can only recommend, and whose recommendations travel to a single reader. The migration is the mechanism in one sentence: the post-2014 architecture removed the layer at which ministers collectively shaped policy before it reached the top, and replaced it with a layer that prepares policy for the top directly.

The architecture has a quiet channel that completes it. The Transaction of Business Rules permit decisions by circulation, in which a case travels on file to the ministers concerned and is approved without a meeting, a device every era has used for urgent business. Commentators on the post-2014 record note the widening use of circulation and committee clearance for business the earlier system would have tabled in full cabinet, and the committee of secretaries, chaired by the Cabinet Secretary, which in the formal design composes inter-ministry disputes, works inside a reporting line that runs toward the PMO’s review calendar, so that the secretaries who sit in it take their substantive direction from the groups and platforms described in this section. The formal chart still shows the Cabinet Secretariat at the centre of coordination. The meeting record shows the centre of coordination moving into the building the chart labels as a staff office, and the move, effected without amending the chart, is the mechanism’s signature.

Mechanism Three: Personnel as the Transmission Belt

The third mechanism difference is the personnel system, the belt that transmits the centre’s preferences into the ministries. The belt exists because of a placement decision older than either era: the Department of Personnel and Training sits inside the Ministry of Personnel, Public Grievances and Pensions, a ministry the Prime Minister holds, assisted after 2014 by a Minister of State in the PMO, Jitendra Singh. Whoever controls that department frames the empanelment of officers for senior posts, the tenure of secretaries and the approval pipeline that ends at the Appointments Committee of the Cabinet, where the Prime Minister and the Home Minister decide.

The Indira-era belt ran on direct intervention. Haksar’s secretariat screened postings for political reliability, the language of commitment supplied the doctrine, and the transfer of inconvenient officers was a recognised instrument of discipline across the states as well as at the centre. The mechanism was personal and therefore fast, and it was also visible, which made it a standing charge against the government and a documented theme of the Shah Commission’s inquiry into administrative demoralisation.

The post-2014 belt runs on procedure, and the procedures are notified. Empanelment for joint secretary and secretary ranks passes through expert panels and, from 2016, a 360 degree appraisal that gathers feedback from an officer’s superiors, peers and subordinates into the file the committee reads. Secretary-level tenures show the design’s other face: Cabinet Secretary P.K. Sinha, appointed in June 2015 for a two year term, received successive extensions that carried him to August 2019, and extensions for favoured officers across the services lengthened the effective tenure of officials the centre trusted, while officers outside the trust faced the standard rotation. In 2018 the government advertised ten joint secretary posts for lateral entry from outside the services, a breach in the cadre wall that the Indira era, for all its interventions, never attempted through open advertisement. Christophe Jaffrelot’s Modi’s India (2021) reads the personnel record as the construction of a parallel administration loyal to the executive’s project, and the documents support the direction of his reading: appraisal design, extension patterns and lateral recruitment together give the centre finer control over who staffs the state than the Indira secretariat possessed. The control travels through files and criteria rather than through telephone calls from the household, which makes it slower to build, harder to reverse and far better documented. A mechanism that leaves a paper trail also leaves the material for its own audit, and the personnel notifications of 2014 to 2024 are the trail on which this section’s claims rest.

The personnel mechanism operates against a judicially drawn counter-design, and the friction between the two is itself evidence for the constraints argument. On 31 October 2013 the Supreme Court, in T.S.R. Subramanian v. Union of India, directed fixed tenures for civil servants, civil services boards to govern transfers, and the recording of oral instructions, a judgment aimed precisely at insulating postings from political discretion. The post-2014 personnel system was built inside that judgment’s perimeter: empanelment criteria and appraisal formats are the lawful instruments through which preference travels when the blunt instruments of transfer and supersession carry legal risk. The perimeter was tested in the Central Bureau of Investigation crisis of October 2018, when the government divested Director Alok Verma of his powers in a midnight order after his public feud with his deputy, and the Supreme Court in January 2019 set the divestment aside before the statutory selection committee, chaired by the Prime Minister, removed Verma days later by majority. Every actor in that sequence used a lawful instrument. The sequence nonetheless displayed the personnel system as the government’s most contested transmission belt, fought over by the court, the selection committee and the PMO within a single quarter.

Constraint One: A Federal Architecture That Answers Back

Mechanisms describe what the centre built. Constraints describe what the centre met. The first constraint is federal, and it is an inheritance the Indira era did not face in the same form. Indira Gandhi governed a federation whose states her party mostly ruled, and where it did not, Article 356 supplied the remedy: her returning government in 1980 dismissed nine opposition-ruled state governments in a single stroke, repeating the Janata government’s own mass dismissal of 1977. The coalition decades then changed the federation’s operating terms. Governments at the centre depended on regional parties from 1989 onward, and the interregnum governments of 1996 to 1998 governed at the sufferance of the states’ parties. The Supreme Court’s judgment in S.R. Bommai v. Union of India, delivered on 11 March 1994, subjected Article 356 proclamations to judicial review and imposed the floor test, and the mass dismissal instrument passed out of usable politics.

Into that inheritance the post-2014 government introduced the Goods and Services Tax, and the tax built the constraint into a standing institution. The Constitution (One Hundred and First Amendment) Act 2016, assented to on 8 September 2016, inserted Article 279A and created the GST Council, which held its first meeting on 22 and 23 September 2016. Inside the Council the Union holds one third of the voting weight and the states together hold two thirds, and a decision requires three fourths of the weighted votes of members present, a formula under which the Union cannot carry a rate or a rule against a united bloc of states, and the states cannot carry one against the Union. In practice the Council decided by consensus through its first years, which is itself the constraint operating: consensus is the price a centre pays when it lacks the votes to command. The GST Council is the cooperative-federalism architecture as constraint, a body the Modi government created for its own tax reform and which then bounded the government’s freedom in indirect taxation for every year that followed. Indira Gandhi faced no equivalent: the taxes that mattered were Union taxes, the Planning Commission’s transfers were discretionary, and no standing body existed in which states held a structural veto over the centre’s fiscal design.

The Sarkaria Commission, whose 1988 report remains the federation’s reference text, recommended consultation through an Inter-State Council that was duly created in 1990 and then convened so rarely that its dormancy became the measure of the centre’s preference for bilateral dealing. The post-2014 government inherited that preference and the institutional facts against it: a Rajya Sabha in which it lacked a majority through its first term and beyond, state governments of opposition parties across the east and south, and a GST Council in which those governments sit as voting members. The reorganisation of Jammu and Kashmir in August 2019 shows the constraint’s exact shape. Parliament could bifurcate a state and demote it to Union Territory status, against the stated position of the state’s own mainstream parties, only because the state stood under President’s Rule and its legislature’s powers had devolved to Parliament itself. The route was lawful and the condition was exceptional: the centre required the suspension of the federal normal to act against a state’s will, and the exception, litigated for four years afterward, priced the act. Indira Gandhi’s governments dismissed state ministries as routine administration. The post-2014 centre must find or await an exceptional condition, and the difference between routine and exception is the constraint.

Constraint Two: Judicial Review Armed With a Basic Structure

The second constraint is judicial. The Indira era began before the basic structure doctrine existed and fought the doctrine to a draw: the Supreme Court announced it in Kesavananda Bharati v. State of Kerala on 24 April 1973, the government answered with the supersession of judges the following day, and during the Emergency the Court itself capitulated in the habeas corpus judgment of 28 April 1976. The doctrine nevertheless survived, was reaffirmed in Minerva Mills in 1980, and hardened across the following decades into the ground rule of Indian constitutionalism, a development the series examines at length in its later study of the doctrine’s career. The post-2014 PMO therefore works downstream of a settled doctrine and a public interest litigation architecture built since the 1980s, in which any affected party can move a constitutional court quickly and the proceedings are reported the same day.

The constraint is not theoretical, and its clearest post-2014 operation struck the Modi government’s own institutional programme. The government carried the Constitution (Ninety-ninth Amendment) Act and the National Judicial Appointments Commission Act in 2014, with near-unanimous parliamentary support, to replace the collegium system of judicial appointments with a commission in which the executive held a defined place.

What did the Supreme Court decide in the NJAC case of 2015?

On 16 October 2015 a five-judge Constitution Bench struck down the Ninety-ninth Amendment and the National Judicial Appointments Commission Act by four votes to one, holding that judicial primacy in appointments formed part of the basic structure. The collegium system revived, and the government’s appointment reform ended in court.

A constitutional amendment, passed by both Houses and ratified by the states, was annulled; the government’s subsequent negotiation over the memorandum of procedure for appointments proceeded on the Court’s terms. Friction continued, notably over the elevation of Justice K.M. Joseph, whose recommendation the government returned in April 2018 and the collegium reiterated, after which the appointment issued in August 2018 with a seniority the government had altered. The episode shows contestation inside the architecture, and the architecture holding. Indira Gandhi, facing a Court whose chief justices she could select by supersession and whose habeas corpus bench surrendered in 1976, operated against a judiciary whose constraining capacity depended on the courage of individual judges. The post-2014 executive faces a judiciary whose constraint is doctrinal, institutionalised and self-defending. The difference is not that one executive was checked and the other free. It is that the check in the later period is built into the structure the executive must plan around, and planning around it is a cost the earlier executive rarely had to pay.

The doctrine’s operation is not confined to appointments. In July 2016 a five judge bench in Nabam Rebia v. Deputy Speaker restored a dismissed Arunachal Pradesh government and read down the Governor’s discretion, annulling in substance a political operation the Union government had backed; the Uttarakhand floor test of May 2016, ordered by the courts during President’s Rule, returned another dismissed government to office for the vote that decided its fate. On 12 January 2018 the four senior-most judges of the Supreme Court after the Chief Justice held a press conference in New Delhi to state, in public, their charge that cases were being assigned in ways that threatened the institution, an event without precedent in the court’s history and one that converted an internal grievance into a public constraint on both the court and the government. None of these episodes reversed a settled policy of the Modi government. Together they mark the environment in which its legal advisers draft: every instrument is written for a court that may read it within the week, on a petition moved by any citizen, under a doctrine the government cannot amend away. Indira Gandhi’s legal advisers drafted for a Parliament that could amend the Constitution by special majority and a court that had been taught the price of resistance. The drafting difference is the constraint in daily form.

Constraint Three: An Open Economy Narrows the Room

The third constraint is economic. The 1991 reform’s institutional consummation dismantled the closed economy inside which the Indira-era PMO had exercised discretion. In the licence era, the state’s permissions were the economy’s commanding heights, and a Prime Minister’s Office that controlled industrial licensing, import allocation and directed credit controlled the material life of the country without reference to any market’s opinion. India’s membership of the World Trade Organization from 1 January 1995 bound its trade policy in treaty, the fiscal responsibility framework of 2003 bound its borrowing in statute, and the integration of Indian firms and banks into global capital markets meant that executive decisions acquired an immediate priced reaction.

The constraint operates as a cost register rather than a veto. Demonetisation demonstrated the register clearly: the decision of 8 November 2016 was taken, executed and sustained, and the costs arrived as a growth slowdown measured by the Central Statistics Office, a cash shortage borne by the informal economy, and a lasting charge against the Reserve Bank’s credibility that the Bank’s own annual report of August 2017, recording that 99 percent of the withdrawn notes had returned, kept in public view. The Indira-era equivalent decisions, bank nationalisation in 1969 or the wheat trade takeover of 1973, registered their costs inside a closed system where the state could suppress the price signals that would have announced them. A post-1991 executive retains the power to act as abruptly as any predecessor. What it cannot retain is silence about the consequences, because the consequences are quoted daily in markets the government does not control and in statistics its own commissions publish.

The constraint taught its most quoted lesson to the previous government. Between May and August 2013, when the United States Federal Reserve signalled an end to its bond purchases, the rupee fell from the mid fifties to nearly sixty nine against the dollar, and the Manmohan Singh government, whose PMO the analogue reading treats as the era of weakness, was driven to emergency measures: import compression on gold, a special dollar window for oil companies, and the appointment of Raghuram Rajan at the Reserve Bank on a stabilisation mandate the markets were invited to inspect. A currency had disciplined a government without a vote being taken anywhere. The post-2014 executive internalised the lesson in advance: macroeconomic management after 2014, the formal inflation targeting framework adopted by agreement in February 2015 and by statute in 2016, the fiscal deficit path defended through two budgets of political temptation, reflects an executive that treats market judgment as a standing member of its audience. The closed economy of the Indira decades had no member of that description. Its costs were paid by citizens in queues and shortages, not by governments in yields, and the difference in who pays first is a difference in how power behaves.

The inflation targeting framework deserves a place inside this constraint, because it is a constraint the post-2014 executive built against itself. The Monetary Policy Framework Agreement of 20 February 2015 committed the Reserve Bank to a consumer price target of four percent, and the Finance Act 2016 wrote the target and a six member Monetary Policy Committee into the Reserve Bank of India Act, three members from the Bank and three appointed by the government, the Governor holding a casting vote. A government seeking monetary room for electoral purposes after 2016 faces a committee whose votes are recorded and published, a mechanism the Indira-era executive, financing its deficit automatically through ad hoc treasury bills, would have found unintelligible. The framework did not prevent the 2018 contest over the Bank’s surplus and lending restrictions. It determined the terrain on which the contest was fought, in published votes and statutory clauses rather than in private pressure, and terrain determines outcomes often enough to count as a form of power in itself.

Constraint Four: An Information Environment That Watches Back

The fourth constraint is informational, and it separates the eras as sharply as any. Indira Gandhi’s government could impose pre-censorship on the press in June 1975 because the information architecture of the time ran through a small number of printing presses, a state broadcaster and licensed newsprint. The instruments of silence were physical, and the executive held them. The post-2014 executive governs inside an environment built by the Right to Information Act 2005, in force from 12 October 2005, which subjects the PMO itself to information requests outside its exempted categories; by round the clock television news, which converts every appointment notification into a public event within the hour; and by social media platforms on which the government’s own communication, from the Prime Minister’s radio addresses begun on 3 October 2014 to ministry announcements, is answered, archived and contested in the same medium.

The constraint cuts in one direction for secrecy and in another for narrative. Secrecy has become expensive: the Section 7 consultations of 2018, intended as an internal correspondence between the government and the Reserve Bank, were in the newspapers within weeks, and the Governor’s resignation letter was a public document the day it was written. Narrative, by contrast, has become cheaper for a leader with a direct channel, and the post-2014 PMO exploits that asymmetry skilfully, carrying its case to the citizen without the mediation of editors. The Indira-era PMO faced the opposite asymmetry: secrecy was cheap and narrative was dear, purchasable only through a managed press and a state broadcaster. Neither environment is simply freer. They distribute advantage differently between the executive and its watchers, and the distribution is a structural fact about the period, not a choice the period’s Prime Minister made. A PMO planning a contested decision after 2014 must price in disclosure. A PMO planning one in 1975 could price in silence, and the price difference shapes which decisions get taken and how they are defended.

The physical basis of the older silence is worth recording exactly. On the night of 25 and 26 June 1975 the electricity supply to the newspaper presses on Delhi’s Bahadur Shah Zafar Marg was cut, editions were seized, and pre-censorship was imposed under rules that made the censor’s office a department of every newsroom; the Shah Commission later reconstructed the machinery from the government’s own files. No equivalent instrument exists in the later environment, not because a statute forbids it but because publication no longer passes through a small number of cuttable wires. The same environment disciplines through the Right to Information pipeline in quieter ways: applications to the PMO on appointments, expenditure and correspondence are answered, refused on stated exemptions or fought before the Central Information Commission, and each outcome enters the record that journalists and scholars mine. An office that knows its files may be read by strangers writes different files. Officials of the post-2014 PMO conduct business with that knowledge as a background condition, and the condition, unlike a censor, cannot be switched off for a difficult week.

The Complication: Architecture on Paper, Operation in Practice

The strongest counter to the integrated comparison deserves a full statement, because the article’s claim fails if the counter stands. The counter holds that the constraint differences are formal rather than real. The basic structure doctrine did not save the Election Commission’s published orders from omitting a dissent. The open economy did not prevent a currency withdrawal prepared in secret. The information environment publicised the Reserve Bank contest and the contest ended with the Governor gone. On this view the architecture exists the way a speed limit exists on an empty road, and the honest comparison is between two executives equally unconstrained, one of which merely governs in a period that documents its own unconstrained conduct more thoroughly.

The answer turns on a distinction the article has used throughout: constraint architecture is the formal structure that exists before any episode, and constraint operation is what happens inside an episode. The counter is correct about several operations and wrong about what the operations cost. Urjit Patel resigned, and the resignation was the story of the season: the government’s preferred narrative of a routine change at the Bank never recovered, the Section 7 correspondence entered the public record, and the successor’s appointment was scrutinised against the vacancy’s circumstances in a way no Indira-era appointment at the Bank had ever been. Ashok Lavasa’s dissent was omitted from the Commission’s orders, and the omission became a documented controversy that followed the Commission into litigation and memoir. Justice Joseph’s elevation was delayed and his seniority altered, and the alteration was reported, debated and recorded against the government’s stated position. In each episode the executive prevailed on the immediate point and paid in delay, publicity and precedent. Indira Gandhi’s supersessions, by contrast, cost her government almost nothing at the time they occurred: no doctrine stood against them, no information architecture carried them to the public as a constitutional event, and the price was deferred to a reckoning that arrived only in 1977, on other grounds.

The distinction also explains the pattern the counter cannot. An executive facing expensive constraints tests them where the prize justifies the price and routes around them elsewhere, which is the post-2014 pattern exactly: aggressive where the architecture is thin, as in personnel extensions and committee design, and cautious where the architecture is thick, as in the GST Council’s consensus practice and the abandoned appointments commission. An executive facing no constraints has no reason to discriminate. The discrimination visible in the record is the architecture operating, not as a wall but as a tariff, and a tariff is a real constraint on any government that must husband its political capital.

One concession belongs inside the argument rather than against it. Architecture made of statute can be amended by statute, and the post-2014 Parliament amended where it could assemble the votes: the Right to Information (Amendment) Act, passed in July 2019, replaced the fixed tenure and salary protections of information commissioners with terms the central government prescribes by rule, a change the opposition and former commissioners criticised as converting an independent referee into a contract employee. The episode refines rather than refutes the comparison. The government paid for the amendment with a parliamentary fight, a public campaign by the Act’s own former custodians, and a permanent entry in the record of its intentions, costs of a kind the Indira-era executive, amending the Constitution itself by special majority during the Emergency, incurred only in the deferred currency of 1977. Architecture that can be amended at a price is still architecture. The analytical error is to treat the price list as a wall, or the absence of a wall as the absence of a price.

Adjudicating the Three Readings

The three readings can then be judged against the evidence assembled. The simple-analogue reading, in its scholarly form, rests on Pratap Bhanu Mehta’s essays of 2014 to 2024, which argue that the post-2014 government hollows institutions from within, capturing them through appointments, investigative agencies and the corrosion of norms, and which treat this hollowing as the continuation, by new means, of a decay that earlier governments began. Mehta’s evidence on appointments and agencies is strong, and his insistence that the executive’s relation to institutions is the central question of the period is the premise this article shares. Where the analogue reading overreaches is in its treatment of continuity: a decay narrative cannot explain why the decaying executive built a GST Council that binds its own fiscal hands, or why its appointments commission died in court, or why its planning body surrendered the allocation power that made the old Commission an instrument of the centre. Continuity explains the features. It cannot explain the mechanisms or the constraints, and a reading that explains one layer of three is a partial reading.

The rupture reading fails symmetrically. Christophe Jaffrelot’s Modi’s India (2021) is the most formidable version of it: a de facto ethnic democracy, a plebiscitary leader, institutions staffed into compliance, a vigilante fringe licensed from above. Jaffrelot’s documentation of the personnel mechanism, the ideological staffing of cultural and educational bodies and the majoritarian reframing of citizenship is the strongest empirical spine any of the three readings possesses, and this article’s mechanism sections corroborate its direction on appointments. But the rupture reading cannot accommodate the continuities of feature: the ratifying cabinet of 8 November 2016 is unintelligible without the ratifying cabinet of 26 June 1975, and a political science that treats 2014 as year zero must discard the Shah Commission’s record as prehistory. Inder Malhotra’s biographical work on Indira Gandhi, written from inside the reporting of that era, established the personalisation of the secretariat as a documented fact decades before 2014. A reading that cannot see the older office cannot measure the newer one.

The integrated comparison survives because it is the only reading that prices all three layers. The features are shared, and the analogue reading is right to insist on them. The mechanisms differ, in the planning body, the committee architecture and the personnel system, and the rupture reading is right to insist on the novelty of the instruments. The constraints differ, in the federal architecture, the basic structure doctrine, the open economy and the information environment, and neither popular reading has much to say about them. The namable claim follows: Modi-PMO and Indira-PMO share structural features but operate through different institutional mechanisms and within different structural constraints, and the differences are not minor. They determine what each office could do, what each action cost, and what of each centralisation would survive its author.

The comparative literature supplies a final test. Presidentialisation theory, developed for the established parliamentary democracies, predicts that leader centred executives grow as parties weaken and media personalise politics, and it expects the growth to be uniform across the system. The Indian record fits the prediction at the level of feature and defeats it at the level of mechanism, because the theory has no account of a federal tax council carrying a states’ veto, a basic structure doctrine, or an information statute reaching the executive’s own files. India is not a case of the general tendency with local colour. It is the case that shows the tendency’s institutional mediation: the same personalising pressure, run through different machinery against different resistance, producing a different distribution of power. That is the analytical yield of the integrated comparison, and it is available only to a reading that refuses both the comfort of analogy and the drama of rupture.

What the Integrated Comparison Establishes

Three consequences follow for any reader who has to judge the next claim about the Prime Minister’s Office. The first concerns durability. Indira Gandhi’s centralisation was carried by persons and collapsed with them: the secretariat she built was renamed and cut within months of her 1977 defeat, and the parallel centre vanished with the household that sustained it. The post-2014 centralisation is carried by notified architecture, committee compositions, appraisal criteria, a think tank’s founding resolution, and architecture survives its builders in a way households do not. A future government of any party will inherit the Groups of Secretaries method, the eight committee structure and the personnel pipeline, and will have to repeal them deliberately rather than watch them evaporate. Centralisation by design is therefore the more consequential kind, even where its day to day intensity matches an older personalisation. The point can be put as a test for the next change of government. A successor who wished to restore the distributed executive would have to do visible work: reconstitute the committees, restore a planning or convening body with a charter of its own, rewrite the appraisal criteria, and return the personnel pipeline to a referee outside the PMO. Each act of restoration would be a public decision with a named author, debated as such. Indira Gandhi’s successor in 1977 had to do almost none of that work, because the centralisation he inherited lived in persons who had left the building. The contrast is the durability claim in operational form, and it converts the article’s comparison into a prediction the record of the next transition will settle one way or the other.

The second consequence concerns where reform pressure belongs. Arguments aimed at the features, at the personalisation of campaigns or the theatre of leadership, address the layer least amenable to institutional remedy. Arguments aimed at mechanisms and constraints have handles: the voting design of the GST Council, the tenure protections of regulatory heads, the statutory footing the PMO has never been given, the appraisal criteria inside the personnel department. The comparison identifies which handles exist by showing which layers actually differ between the eras.

The third consequence is a discipline for the argument itself. Every future episode, a resignation at a regulatory body, a committee that clears what a cabinet should have debated, a dissent that goes unpublished, will be claimed by the analogue camp as repetition and by the rupture camp as novelty. The integrated comparison supplies the questions that settle such claims: which function moved, through which mechanism, against which constraint, and at what recorded cost. An episode that answers those four questions has been analysed. An episode described only by its resemblance to 1975 or its distance from it has merely been labelled, and labelling, in either direction, is what this article set out to replace.

Frequently Asked Questions

Q: What is PMO centralisation?

PMO centralisation is the concentration of four executive functions inside the Prime Minister’s Office: the initiation of policy, the clearance of files, the approval of senior appointments and the review of implementation. In a cabinet system those functions are distributed among ministries, the cabinet and its committees and the Cabinet Secretariat. Centralisation is measured by traces rather than impressions: where a proposal was first drafted, whether the cabinet discussed it before or after the decision, whose names the appointment notifications carry, and who summons secretaries to review progress. By those measures an office can dominate a government without any change in the formal rules, because the Prime Minister’s Office has no statutory charter whose amendment would announce the change. The term is therefore a description of where functions sit rather than of how forcefully a Prime Minister behaves in a meeting room, and two governments working under identical formal rules can sit at opposite ends of the scale, which is why the debate about the office is settled by examining records and not by collecting impressions of personality.

Q: How is Modi’s PMO different from Indira’s PMO?

The two offices share four structural features: both extend the Prime Minister’s personal authority, both reduce the cabinet to ratification, both sustain a parallel centre beside the formal chart, and both press on the autonomy of regulatory bodies. The differences lie in mechanisms and constraints. Indira Gandhi’s secretariat worked through an allocating Planning Commission, personal confidants such as P.N. Haksar and an informal household; the post-2014 office works through a think tank without allocation power, notified cabinet committees and Groups of Secretaries, and a personnel system of empanelment, appraisal and lateral entry. It also works against constraints Indira never faced in settled form: the basic structure doctrine, the GST Council’s voting design, an open economy and a right to information regime. The comparison is therefore asymmetrical in a way single labels hide: continuity at the top of the institution, novelty in its middle layers, and resistance at its edges, where courts, councils, markets and an information statute bound what the office can attempt and what each attempt will cost it.

Q: Who runs the Prime Minister’s Office under Modi?

The office is headed by the Principal Secretary to the Prime Minister, the senior-most civil servant in the building and the channel through which files, appointments and reviews move. Nripendra Misra, a retired officer of the Uttar Pradesh cadre, held the post from May 2014 and announced his wish to step down in August 2019; P.K. Mishra, who had served as Additional Principal Secretary from 2014, succeeded him in September 2019. Beside the Principal Secretary stands the National Security Advisor, Ajit Doval, in place from 30 May 2014 and holding the rank of a cabinet minister from June 2019. A small team of secretaries and additional secretaries handles ministries in clusters, and the Minister of State in the PMO assists the Prime Minister, who holds the portfolio personally. Below them the office divides its work by ministry clusters. The Minister of State through both terms from 2014, Jitendra Singh, carries the portfolio’s parliamentary business, including the Department of Personnel and Training whose appointment pipeline the office directs.

Q: Did the Cabinet system weaken under Modi?

The formal system stands intact: the Transaction of Business Rules still reserve classes of business for the cabinet and its committees, the committee system expanded to eight standing committees in June 2019, and collective decisions carry the same legal force as in any earlier period. What changed is the location of deliberation. Major decisions of the period, demonetisation on 8 November 2016 and the Jammu and Kashmir instruments on 5 August 2019, reached the cabinet with their design complete, and the meeting’s function was endorsement rather than examination. Defenders answer that the committee layer deliberates in the cabinet’s place and that the Prime Minister’s review meetings with secretaries supply the scrutiny a large cabinet cannot. The charter survives; the habit of deciding inside it has thinned. The honest verdict separates the charter from the habit: the system’s forms were maintained and in the committee layer enlarged, while its deliberative centre of gravity moved to smaller rooms whose proceedings leave thinner records, so later historians will know what was decided more easily than what was argued.

Q: What is the role of the National Security Advisor?

The National Security Advisor heads the secretariat of the National Security Council, advises the Prime Minister on security, strategic and foreign policy questions, and coordinates the intelligence and security agencies for the Prime Minister’s Office. The post was created on 19 November 1998, when Brajesh Mishra combined it with the principal secretaryship under Atal Bihari Vajpayee; later holders separated the two offices. Ajit Doval, appointed on 30 May 2014, chairs the Strategic Policy Group after its reconstitution in October 2018 moved the chair from the Cabinet Secretary to the Advisor, and in June 2019 he was reappointed with the rank of a cabinet minister. The role carries no statutory charter: its weight in any period reflects the Prime Minister’s trust and the Advisor’s access. Successive holders have defined the post differently, Mishra as a coordinator of the whole security establishment, M.K. Narayanan and Shivshankar Menon as advisors with diplomatic weight, and Doval, a former intelligence chief, as an operational principal whose remit runs from internal security to the management of India’s neighbourhood crises.

Q: How does the PMO control policy decisions?

Control runs through procedure rather than proclamation. The Transaction of Business Rules let the Prime Minister call in any case and permit departures from the reserved lists, so a proposal the PMO favours travels fast and one it questions waits. Cabinet committees chaired by the Prime Minister settle whole sectors between cabinet meetings. Groups of Secretaries, constituted from January 2016, draft policy in themes and present directly to the Prime Minister, framing the options before ministries see them. The appointments pipeline, ending at the Appointments Committee of the Cabinet, decides which officers will implement each policy. Monthly review meetings under platforms such as PRAGATI then track execution with state chief secretaries present. Each channel is formally regular; their combination is the control. None of these channels requires the office to issue an instruction a ministry can quote against it, which is the design’s elegance and its accountability problem in one: control that travels through agenda setting, drafting and posting leaves few orders to audit, and the audit must therefore reconstruct decisions from calendars, notifications and outcomes.

Q: Is Modi’s PMO architecture novel or continuous?

Both, at different layers, and the layers must be kept apart. The structural features are continuous with the Indira era: a personality-centred office, a ratifying cabinet, a parallel centre of authority and pressure on autonomous bodies all predate 2014, and the documentary record from the Shah Commission onward establishes the pedigree. The architecture is novel: a think tank without allocation power in place of the Planning Commission, eight standing cabinet committees with two created in June 2019, Groups of Secretaries reporting to the Prime Minister, and a personnel system of appraisal panels, extensions and lateral entry. The constraints are also novel in their settled form. Continuity of feature carried by novelty of mechanism is the precise description, and it is why single-word answers, repetition or rupture, both mislead. Asked for one word, the record answers with a sentence, and the sentence matters practically: a reformer who believes the office is a repetition will prescribe remedies for 1975, and a reformer who believes it is a rupture will prescribe remedies for a regime that exists mainly in commentary.

Q: What powers does the Prime Minister’s Office hold in law?

Almost none in its own name. No statute creates the Prime Minister’s Office, fixes its staff or grants it authority over a ministry, and no constitutional article mentions it. Its power is borrowed and procedural: the Prime Minister heads the Council of Ministers under Article 74, chairs the cabinet and its committees, holds the Ministry of Personnel, and works through the Transaction of Business Rules framed under Article 77, which let the Prime Minister call in cases and permit departures from the rules. Every PMO instruction reaches a ministry as the Prime Minister’s instruction, transmitted through officers the Appointments Committee posted. The office’s legal invisibility is the source of both its flexibility and the difficulty of reforming it: there is no charter to amend, only conventions, rules and habits to change. The position has one hard legal edge: the Prime Minister’s advice binds the President under Article 74 after reconsideration, so the office that drafts the advice commands the one channel through which the executive formally acts, and every instrument of the Union issues, in law, from that channel.

Q: How did the PMO function under Manmohan Singh between 2004 and 2014?

It functioned as an administrative secretariat under a Prime Minister who did not command his own party. T.K.A. Nair served as Principal Secretary from 2004 to 2011 and Pulok Chatterjee from 2011 to 2014, and the office ran files, committees and reviews with professional competence. Political authority, however, sat partly outside the government, in the Congress presidency and the National Advisory Council that Sonia Gandhi chaired, which originated legislation on rights and welfare that the cabinet then adopted. Appointments and policy initiatives bent toward that second centre, a pattern Sanjaya Baru recorded from inside as media advisor. The decade taught the designers of the post-2014 office their founding lesson: a vacant centre does not stay vacant; it is occupied from outside, and the remedy is to occupy it fully from within the office itself. The period also normalised a defensive innovation later inherited and reversed: major files moved slowly and committees multiplied, because a Prime Minister who could not command outcomes multiplied the forums in which outcomes could be postponed, an architecture of diffusion the post-2014 designers studied as the negative image of their own.

Q: How did coalition governments limit the PMO between 1989 and 2014?

Coalition arithmetic limited the office through vetoes the Prime Minister had to price into every decision. Between 1989 and 2014 no single party held a Lok Sabha majority, so allies controlled portfolios, demanded consultation through coordination committees, and could withdraw support, as the Left parties did over the nuclear agreement in July 2008 and as the AIADMK did in April 1999, bringing the Vajpayee government down by one vote. Prime Ministers Gowda and Gujral, resting on outside support, ran small secretariats with narrow reach. Even Vajpayee, whose principal secretary doubled as National Security Advisor, bargained for each major decision. The office’s contraction in those decades was not a change in the rules; it was the price of survival, payable monthly, and it ended the moment a single party majority returned in 2014. The limit was structural rather than personal, and it can be dated in the office’s own staffing: coalition Prime Ministers appointed principal secretaries of quality, Deve Gowda’s Satish Chandran and Vajpayee’s Brajesh Mishra among them, and the office still contracted, because no staffing decision can repeal the arithmetic of a hung Lok Sabha.

Q: Who was Brajesh Mishra, and why was his dual role unusual?

Brajesh Mishra, a retired Indian Foreign Service officer, served as Principal Secretary to Prime Minister Atal Bihari Vajpayee from 1998 to 2004 and, from 19 November 1998, as India’s first National Security Advisor. The combination was unusual because it fused the two channels through which a Prime Minister governs: the administrative channel, which moves files, appointments and cabinet business, and the strategic channel, which coordinates intelligence, defence and foreign policy advice. In one pair of hands the PMO’s domestic machinery and the security state’s apex secretariat answered to a single official, giving the Vajpayee office a reach its coalition arithmetic seemed to forbid. No successor has held both posts together; the offices separated after 2004, and the brief fusion remains the standing example of how personnel choices, not rules, set an office’s real weight. Mishra’s standing also rested on the Prime Minister’s trust in a way the formal chart never recorded: a retired diplomat without a party base, he was heard in every room the security state assembled, and his departure in 2004 closed the experiment, which no later government has repeated.

Q: What changed inside the Prime Minister’s office under the Janata government after 1977?

The Janata government treated the office itself as an exhibit of the Emergency and set out to shrink it. Morarji Desai renamed the Prime Minister’s Secretariat as the Prime Minister’s Office in 1977, cut its staff, and restored cabinet deliberation as the working method of his government, taking decisions in meetings where ministers could and did argue. The government also appointed the Shah Commission in May 1977, whose inquiry into the Emergency documented how the secretariat had operated between 1975 and 1977 and fixed that record in public. What the Janata government did not change was the underlying architecture: the Transaction of Business Rules, the committee schedules and the personnel ministry in the Prime Minister’s charge all survived intact, so the contraction lasted only as long as the Prime Minister who willed it, and the office regrew after 1980. The episode fixed the office’s reputation for a generation: civil servants who had watched the secretariat of 1975 to 1977 described the Janata contraction with relief in their memoirs, and the relief itself measures the distance from the founding baseline of a servicing secretariat.

Q: Why was the Planning Commission abolished rather than reformed?

Three arguments carried the decision, announced in the Prime Minister’s Independence Day address of 15 August 2014 and executed by the resolution of 1 January 2015. The first was federal: chief ministers across parties, recorded in consultations through 2014, resented a body in New Delhi that approved state plans and attached conditions to transfers, in a federation the Constitution designs as a union of states. The second was fiscal: the Fourteenth Finance Commission’s award, raising tax devolution to 42 percent, had already shrunk the discretionary plan transfers that justified the Commission’s allocation role. The third was functional: the government held that a commission designed in 1950 for a command economy could not be reformed into the think tank a market economy needed, and that a new institution, unburdened by the allocation habit, would design policy frameworks instead. Critics answered that abolition discarded expertise the new body had to rebuild. Defenders add that the Commission had become an allocator without a legislature’s accountability, distributing transfers through formulas no Parliament voted on, while the Finance Commission, a constitutional body, stood ready as the lawful home for the fiscal function.

Q: What is empanelment in the senior civil service?

Empanelment is the screening process by which officers of the All India Services and the Central Services are approved as fit to hold posts at a given rank in the Government of India, joint secretary, additional secretary and secretary. Panels of experts review each officer’s record, and from 2016 a 360 degree appraisal has added structured feedback from superiors, peers and subordinates to the file. Only empanelled officers can be posted at that rank, and the Appointments Committee of the Cabinet approves the actual posting from the panel. The process matters to the centralisation question because it is the gate through which every senior official passes: whoever designs the criteria and reads the appraisals shapes the administrative leadership of every ministry, years before any individual posting is announced. The gate has widened at its edges without changing its design: lateral entry, advertised from 2018 for joint secretary posts, routes a small number of outside specialists around the empanelment ladder, and the debate over that breach, between cadre control and imported expertise, is a debate about who the gate is for.

Q: What is the Principal Secretary to the Prime Minister, and how is the holder chosen?

The Principal Secretary is the senior-most official in the Prime Minister’s Office, the head of its staff, the channel for files moving to the Prime Minister and the officer through whom the office’s decisions return to the ministries. The post carries no statutory tenure or prescribed selection procedure: the Prime Minister chooses, the Appointments Committee of the Cabinet records the approval, and the holder serves at the Prime Minister’s pleasure, frequently on contract after retirement from the services. The occupants mark the eras: P.N. Haksar under Indira Gandhi from 1971 to 1973, P.N. Dhar after him, A.N. Verma under P.V. Narasimha Rao, Brajesh Mishra under Atal Bihari Vajpayee, T.K.A. Nair and Pulok Chatterjee under Manmohan Singh, and Nripendra Misra under Narendra Modi from 2014. Because the appointment is personal, the office’s reach in any period can be read from the holder’s standing with the Prime Minister. No selection record is published for the choice, and that absence is itself part of the centralisation debate: the government’s most consequential staffing decision is its least documented one.

Q: What is the difference between constraint architecture and constraint operation?

Constraint architecture is the formal structure that exists before any political episode: the basic structure doctrine, the GST Council’s voting formula, the Right to Information Act, the fiscal rules, the treaty commitments. Constraint operation is what happens when an executive tests that structure in a live case: whether the court holds, whether the council’s consensus price is paid, whether the correspondence leaks. The distinction matters because critics and defenders cite different layers. A resignation forced by publicity shows architecture operating through cost; a dissent suppressed in an official order shows operation failing on that point while the architecture survives, documented and available for the next contest. Judging a period requires both layers: architecture tells a government what an action will cost before it acts, and operation tells the citizen what the action cost in fact. The distinction also disciplines prediction: a government deterred by architecture behaves differently before the episode, trimming proposals, pre-negotiating with states, drafting for judicial review, and that anticipatory trimming, invisible in any single case, is where constitutional constraint does most of its work in a working executive.

Q: How does the information environment limit what a Prime Minister’s Office can do?

It limits secrecy first. The Right to Information Act 2005 reaches the PMO itself outside exempted categories, 24 hour news converts notifications into public events within the hour, and social media archives every announcement against its later outcome. The 2018 contest between the government and the Reserve Bank, designed as an internal correspondence, was conducted in the newspapers within weeks, and the Governor’s resignation was a public document on the day it was written. An office that cannot rely on silence must draft every contested decision for eventual disclosure, which changes the decision: options that cannot be defended in public are quietly dropped, and timelines are set by the news cycle rather than by administrative convenience. Indira Gandhi’s government, able to impose pre-censorship in June 1975, faced no equivalent discipline, which is why the information environment ranks as a structural constraint and not a backdrop. The limit has a second face: the same environment punishes hesitation and rewards the office that announces first, so the constraint selects for speed and message discipline rather than for deliberation, a selection effect entered on the other side of the ledger.

Q: What evidence would count against the PMO centralisation thesis?

The thesis makes testable claims, and specific records would weaken it. Cabinet minutes showing major proposals amended or rejected in full cabinet against the PMO’s preference would count. Appointment notifications elevating officers known to have opposed the government’s line, in numbers comparable to loyalists, would count. A GST Council decision carried against the Union’s stated position, or a regulatory head completing a full tenure after public disagreement with the government, would count. Memoir or archival evidence that Groups of Secretaries’ recommendations were routinely set aside by ministries would count against the mechanism claim specifically. The thesis stands on the traces the record does contain: ratified decisions, extension patterns, a resigned Governor and a suppressed dissent. A rival account that produced contrary traces of equal documentary weight would deserve to displace it, and the invitation to produce them is part of the claim. The question is included in this article’s list deliberately: a thesis about an office that cannot state what would refute it is a mood about the office, and the centralisation argument has been conducted as a mood, by both camps, for longer than the record justifies.

Q: Why does the Department of Personnel and Training sit inside the Prime Minister’s charge?

The placement descends from the design of the Ministry of Personnel, created in its modern form in 1985 under Rajiv Gandhi and held by the Prime Minister personally, with a Minister of State for the daily work. The department administers the service rules of the Indian Administrative Service and the central services, runs empanelment, processes postings for the Appointments Committee of the Cabinet and supervises the Central Vigilance Commission’s administrative side and the training academies. Placing the government’s personnel ministry under the Prime Minister reflects the theory that the head of government must command the staffing of the government. The same placement is the hinge of the centralisation question, because it gives the Prime Minister’s Office lawful reach into every ministry’s leadership: the officers who run a department owe their posting, their tenure and their next rank to a pipeline whose head the Prime Minister holds. Reform proposals return to the placement in every generation, a statutory civil service board, a personnel ministry separated from the PMO, fixed tenures by law, and their repeated failure to pass measures how completely the arrangement serves whoever holds the office that the arrangement empowers.

Q: What do insiders’ memoirs reveal about how the PMO works?

Memoirs reveal the informal layer that notifications conceal. P.N. Dhar, Indira Gandhi’s principal secretary from 1973 to 1977, described in his 2000 memoir how decisions moved through the household and how the Emergency was prepared inside a circle the cabinet never saw. Sanjaya Baru’s The Accidental Prime Minister (2014), written from the media advisor’s chair, documented how the Manmohan Singh PMO bent toward a centre of authority outside the government. Inder Malhotra’s biographical reporting assembled the Indira-era secretariat from interviews across decades. Read together, the memoirs agree on one structural point across eras and parties: the office’s real constitution is the pattern of access to the Prime Minister, who is heard, whose draft survives, whose call is returned. The formal rules describe the building. The memoirs describe the court, and the centralisation debate is an argument about which description governs. The memoirs must be read against the procedural record, because memory arranges itself around the author’s defence, and the strongest use of the genre is triangulation: where a memoir, a notification and a commission report describe the same decision, the description can be treated as the office’s real history.