← Atlas
Idea

Structure, hierarchy, and scale

Sears's 2008 reorganization tried to replace a centrally planned retailer with autonomous business units and internal market discipline. Unit leaders received profit responsibility while the holding company retained appointments, boards, and the 'Sears Way.' The experiment shows that structure never removes hierarchy or dependency: it chooses which conflicts become local, which rise to the center, and which shared obligations to customers, workers, suppliers, and pensioners risk becoming nobody's account.

Governing questionWhat happens when units receive profit responsibility but the shared customer and worker system becomes someone else's problem?

Working · Claim Cited

Sears divided an integrated retailer into accountable businesses

Edward Lampert's February 2008 shareholder letter records the formal design. It said Sears Holdings would become a holding company over operating, support, online, real-estate, and brand businesses. A single leader would be accountable for each unit's results; the real-estate business would act as an internal landlord; and a dedicated team would manage Kenmore, Craftsman, and DieHard. The stated aim was to move decisions downward and replace central planning with market forces.1

The letter also identifies what would remain at the center. Sears Holdings would define the “Sears Way,” appoint business-unit leaders, place senior executives on unit boards, and build the processes and information systems needed to support the arrangement.1 The primary record therefore supports I05-N02, the five-part redesign, and I05-N03, holding-company control. It does not support a claim that hierarchy disappeared.

Sears still presented stores, brands, services, online channels, real estate, associates, and customer relationships as assets of one retailer.1 That shared system is I05-N01. The structural problem was not whether the components depended on one another, but where responsibility for those dependencies would sit once their results were separated.

Autonomy was real, bounded, and measured from above

The same design can decentralize one decision and centralize another. Unit leaders could act over their businesses, yet the holding company constituted the units, chose their leaders, supplied their boards, and defined common doctrine. Those are different kinds of authority, not contradictory descriptions of one flat hierarchy.1

That distinction supports I05-E01 and I05-E02. I05-E01 is a typed organizational-redesign relation from the shared retail system to the five-part portfolio. I05-E02 is a power analysis: operating accountability moved downward while appointment and supervisory power remained above. Both relations are grounded in institutional intent. Neither establishes how much discretion a particular leader exercised in practice.

Hierarchy is therefore better inspected as several powers: who defines a unit, appoints its leader, allocates capital, specifies measures, controls shared infrastructure, and hears an appeal. A chart showing separate boxes can coexist with concentrated authority over every box's existence and terms. Conversely, a central standard can sometimes protect a shared obligation that no unit has an incentive to carry alone. “Centralized” and “decentralized” are incomplete until the decision and affected party are named.

Separate accounts made performance legible and cooperation contestable

Mina Kimes's 2013 Bloomberg Businessweek investigation describes the implemented arrangement as more than thirty autonomous businesses with their own leaders, boards, and profit-and-loss statements. Former employees told her that units negotiated formal agreements for internal services, bonuses followed unit performance, board service consumed executive time, and divisions could protect local results at another division's expense.2

The investigation gives concrete examples. It reports that a Sears appliance unit could earn more by promoting outside products while Kenmore was itself a separate unit, and that units competed for resources and Lampert's capital approval.2 Those accounts support I05-N04, locally legible results inside one customer experience, and I05-N05, reported cross-unit conflict. They also support the bounded operationalization relation I05-E03 and the organizational-consequence relation I05-E04.

The evidence does not justify a mechanical verdict. Kimes interviewed multiple former employees and published independently of Sears, but the investigation is one journalistic reconstruction. It does not provide a representative worker or customer sample, a counterfactual retailer, or a causal estimate of the model's net effect.2 The defensible inference is narrower: separate accounts and unit incentives created a recurring site where shared brands, services, space, and investment could become objects of bargaining.

That mechanism illustrates a residual-account problem. A unit account can clarify what the unit earns and spends while leaving the customer's complete journey outside any one account. Structure decides whether that residual is owned by an integrator, protected by a common rule, resolved through bargaining, or left unattended until a failure reaches the center.

Rival structural theories prevent a one-form verdict

Alfred Chandler's Strategy and Structure describes the multidivisional form as autonomous operating divisions joined to a general office that plans, coordinates, appraises, and allocates at the enterprise level. His historical cases include DuPont, General Motors, Standard Oil of New Jersey, and an earlier Sears.3 That model offers a different answer from making every shared relationship an internal transaction: preserve divisional operating discretion while assigning integration to a general office.

Robert Freeland's archival history of General Motors disputes a tidy efficiency story. He reconstructs shifting coalitions among owners, general executives, and division managers, and argues that informal participation and consent sometimes supported cooperation while a textbook separation of headquarters from divisions could undermine it.4 Chandler and Freeland therefore agree that formal divisions coexist with enterprise control, but disagree about how efficient form, participation, consent, and actual authority fit together.

Russell Ackoff supplies a third comparison. His 1999 chapter “The Internal Market Economy” advocates replacing central corporate allocation with internal markets and begins from transfer-pricing conflicts between profit centers.5 It is a primary theory text, not evidence that Sears copied Ackoff or that internal markets have one effect across settings. Paired with the Sears record, it clarifies the design question: which dependencies can be priced without destroying cooperation, and what governing authority remains when a price or interface fails?

These sources support I05-N08 as a field of rival structural answers and bound I05-E07 as an editorial comparison. No historical descent from General Motors or Ackoff to Sears is asserted. The comparison concerns where integration, capital allocation, consent, and appeal are placed—not a ranking of old and new forms.

Decline had multiple histories and cannot identify one cause

The structural experiment did not begin with a healthy, otherwise unchanged retailer. A peer-reviewed financial history of Sears from 1981 through 1994 finds that hoped-for cross-selling synergies from its financial-services diversification did not materialize, retail performance deteriorated, and shareholder pressure accompanied divestiture and restructuring.6 That evidence concerns an earlier Sears and emphasizes shareholder wealth. It is useful precisely because it prevents every inherited weakness from being dated to 2008.

Sears's 2017 Form 10-K records several simultaneous pressures: years of losses and negative operating cash flow, intense competition from multiple retail formats and online sellers, property and brand transactions, declining capital expenditure, debt repayment and issuance, store closures, and pension funding obligations.7 As a company filing, it is strong evidence for recorded transactions, counts, and management's stated risks. It cannot independently allocate causal responsibility among the internal structure, competitive change, operating choices, financing, or earlier weakness.

That boundary defines I05-N06 and I05-E05. The internal-market design is a plausible contributor to documented cooperation problems, not an isolated treatment whose effect can be read from bankruptcy. I05-E05 is therefore a causal qualification, not a causal estimate. A universal claim that internal markets caused Sears's collapse would exceed the evidence.

Enterprise obligations outlived the unit accounts

The 2017 filing reports that Sears closed 303 Kmart stores and 123 Sears Domestic stores during the fiscal year, had announced another 106 first-quarter closures, and recognized store-closing, severance, and impairment charges. It also reports a $295 million contribution to domestic pension plans during 2017 and expected minimum funding obligations for 2018 and 2019.7 These records support I05-N07 and the burden-transfer relation I05-E06. They do not connect each closure, job loss, or pension obligation to the 2008 structure.

The Pension Benefit Guaranty Corporation later estimated that Sears's two plans were underfunded by $1.4 billion and 64 percent funded when the federal agency sought to terminate them in January 2019. PBGC expected its guarantees to cover the vast majority of earned benefits.8 That official record shows how a legacy obligation persisted through bankruptcy and shifted to a public insurance institution. It does not measure every retiree's loss, uncertainty, or household effect.

Workers, retirees, suppliers, customers, landlords, and communities do not enter the cited record symmetrically. The filing counts closures and company costs; PBGC records plan funding and guarantees; Kimes interviewed former executives. None provides representative testimony from store workers, laid-off employees, retirees, suppliers under pressure, customers losing access, or communities around closed stores. The evidence can locate burdens without claiming to speak for the people who bore them.

A structure therefore has moral as well as informational boundaries. A person can be essential to the enterprise while appearing only as a unit cost. A community can depend on a store without occupying a box. A pension beneficiary can retain a claim after managers and units have changed. Profit responsibility is incomplete when no forum owns the dependencies and remedies that remain.

The practical question is not whether markets or hierarchies win. It is which dependencies can safely be priced, which require shared authority, who can appeal a structural decision, and who remains responsible when the chosen boundary fails.

Structured research record

The eight nodes distinguish design, practice, consequence, and comparison:

  • I05-N01 is the integrated retail system described in the 2008 letter; I05-N02 is the five-part internal-market design; and I05-N03 is the holding company's retained appointment, board, and doctrine power.1
  • I05-N04 is the tension between local profit and one customer experience; I05-N05 is reported conflict among separately measured units.2
  • I05-N06 keeps decline multi-causal rather than treating the organization as a controlled intervention.67
  • I05-N07 follows store workers and pension beneficiaries through documented closures and plan obligations while recording the absence of representative affected-party testimony.78
  • I05-N08 holds rival answers—multidivisional control and consent, internal markets, networks, and platforms—as comparisons to test rather than stages in one evolutionary sequence.345

The seven typed relations carry corresponding limits:

  • I05-E01 records the organizational redesign; I05-E02 records retained power above the units; and I05-E03 records unit accountability and separate results. The first two are directly supported by the 2008 design record; I05-E03 joins that record to later implementation reporting.12
  • I05-E04 records reported cross-unit conflict. It is evidence of a mechanism, not a net-effect estimate.2
  • I05-E05 is a causal qualification: earlier weakness and later competitive, operating, financing, and asset pressures prevent a one-cause collapse account.67
  • I05-E06 records burdens that reached workers and pension beneficiaries. It does not assign each burden to the 2008 redesign.78
  • I05-E07 is an editorial comparison of structural answers. Its internal attachments identify the General Motors and Chandler routes; the relation asserts neither historical influence nor universal effect.345

The four Workloop records remain proposals, not findings from Sears. I05-P01 proposes linked graphs for intent, authority, work, dependency, customer, resource, and knowledge. I05-P02 proposes mixed human, AI, and software teams with durable responsibility. I05-P03 proposes interface quality and dependency burden as structural measures. I05-P04 proposes recording expected effects, migration risks, and review dates for reorganizations. Each record contains an evidence requirement and an abuse question; none is validated by the cited history.

There is no structured impact record and no reading-dependency record. Impact is unclassified rather than absent, and no prerequisite sequence is asserted. There are also no typed idea_ids or work_ids; the idea, person, work, and institution routes below remain comparison coordinates unless a relation above states otherwise.

Linked analytical coordinates

The related set supplies eight internal routes. The central case is Sears under Eddie Lampert. Its principal historical comparison is General Motors under Alfred Sloan, with Alfred P. Sloan Jr. and Strategy and Structure providing the person and work routes. Delegation, decentralization, and responsibility and Cooperation, incentives, and organizational equilibrium continue into authority and joint action. Strategy, competition, and adaptation and Organizational ignorance continue into reallocation and unrecognized dependency. These are navigational coordinates; they do not create influence or equivalence claims.

The institution coordinates are prompts for comparing scale, integration, appeal, burden, and structural change. Their inclusion does not assert that the cases share causes, legitimacy, outcomes, or historical descent. General Motors and Sears are linked above, so the lists below surface the remaining seventy coordinates in their editorial order.

Long-duration, military, and early corporate coordinates:

Industrial and program coordinates:

Contemporary corporate coordinates from Amazon through Netscape:

Contemporary corporate coordinates from NVIDIA through Vanguard:

Comparative coordinates from Alaska Native corporations through BRAC:

Comparative coordinates from Cameroon through ECOWAS:

Comparative coordinates from Eskom through Haiti:

Comparative coordinates from imperial examinations through Kudumbashree:

Comparative coordinates from Kyocera through Nigeria's Ebola response:

Comparative coordinates from Odebrecht through the Senegal River basin:

Comparative coordinates from Shining Path through termination policy:

Each linked institution has its own evidentiary record. A coordinate here does not establish whether its structure was effective, democratic, voluntary, coercive, extractive, ecologically durable, or beneficial to affected people.

Evidence still needed

  • Worker-controlled or worker-authored evidence about the 2008 reorganization, unit incentives, store closures, severance, job quality, retaliation, household effects, and who could appeal a unit or holding-company decision.
  • Retiree and beneficiary evidence about pension uncertainty, benefit changes, navigation burdens, and the consequences that plan-funding percentages do not capture.
  • Customer, supplier, landlord, and community evidence capable of tracing how internal bargaining, closures, payment pressure, lost access, and local fiscal effects were distributed.
  • Board minutes, unit contracts, transfer-price schedules, capital-allocation records, performance measures, and longitudinal operating data capable of testing how often the reported conflicts occurred and what alternatives were available.
  • Comparative research that separates divisional accountability, central appointment power, internal pricing, shared-service design, capital control, interface ownership, consent, and appeal instead of treating “hierarchy” or “market” as one intervention.
  • Tests of all four Workloop proposals across mixed human, AI, and software work, including false accountability, surveillance, metric gaming, hidden coordination labor, repeated reorganization, and credible rollback.
  • Ecological and intergenerational evidence following stores, real estate, logistics, energy, materials, waste, land use, nonhuman life, and community transition after closures.

Sears makes a bounded mechanism visible: a structure can clarify local results while leaving shared dependencies and residual obligations harder to own. It does not prove that one organizational form always fails or identify a universal replacement.

Additional reciprocal institution comparisons

Newly developed institutional records add these reciprocal comparison paths:

Each path identifies a sourced case where this idea is a defining emphasis. The relation is editorial comparison, not evidence of direct influence, shared terminology, or equivalent outcomes.

Source notes

  1. Participant primary record: Edward S. Lampert, “Letter from the Chairman,” February 28, 2008, Exhibit 99.2, pp. 10–12, especially “New Structure and Personnel Changes,” on the five business types, unit-leader accountability, the “Sears Way,” appointments, unit boards, brands, the internal-landlord role, and required systems, SEC filing. The letter establishes Sears leadership's design, stated aims, and account of shared assets. It is a shareholder communication containing forward-looking claims, not independent evidence of implementation or results.

    ↩ ↩ ↩ ↩ ↩ ↩

  2. Independent investigation: Mina Kimes, “At Sears, Eddie Lampert's Warring Divisions Model Adds to the Troubles,” Bloomberg Businessweek, July 11, 2013, paragraphs beginning “Every year the presidents” and “Under the new scheme,” and the reported Kenmore, internal-service, board, bonus, and capital-allocation examples, Bloomberg. Kimes draws on interviews with multiple former employees and company records. The report provides independently published evidence of contested practice, but no representative worker or customer sample, counterfactual, or causal estimate.

    ↩ ↩ ↩ ↩ ↩ ↩

  3. Independent historical interpretation: Alfred D. Chandler Jr., Strategy and Structure: Chapters in the History of the American Industrial Enterprise (MIT Press, 1962), especially the DuPont, General Motors, and Sears chapters and pp. 309–310 on autonomous divisions joined to a coordinating general office, MIT Press publisher record. Chandler supplies a foundational M-form account and comparative history. His efficiency-oriented interpretation is not a neutral law of structure and does not document influence on Sears's 2008 design.

    ↩ ↩ ↩

  4. Independent archival scholarship: Robert F. Freeland, The Struggle for Control of the Modern Corporation: Organizational Change at General Motors, 1924–1970 (Cambridge University Press, 2000), chapters 2, 3, and 6 and conclusion, pp. 295–323, on competing versions of decentralization, changing central control, consent, and cooperation, Cambridge University Press. Freeland tests efficiency accounts against GM archival evidence and supplies counterinterpretation. The case concerns General Motors, not Sears, and does not show one universally superior balance of participation and control.

    ↩ ↩ ↩

  5. Primary design theory: Russell L. Ackoff, “The Internal Market Economy,” chapter 10 in Re-Creating the Corporation: A Design of Organizations for the 21st Century (Oxford University Press, 1999), pp. 207–223, especially the abstract's profit-center and transfer-price problem, Oxford Academic. Ackoff advocates an internal-market answer to recurring corporate-finance problems. The chapter is a theory source, not an evaluation of Sears or evidence that Sears adopted Ackoff's design.

    ↩ ↩ ↩

  6. Independent peer-reviewed financial research: Stuart L. Gillan, John W. Kensinger, and John D. Martin, “Value Creation and Corporate Diversification: The Case of Sears, Roebuck & Co.,” Journal of Financial Economics 55, no. 1 (2000), pp. 103–137, abstract and sections “The decision to diversify,” “Post-acquisition performance,” “The divestiture,” and “Conclusions,” ScienceDirect record. The longitudinal case establishes earlier deterioration, unrealized cross-selling, restructuring, and shareholder activism. Its 1981–1994 period predates Sears Holdings, and its shareholder-wealth focus does not measure worker, customer, community, or ecological outcomes.

    ↩ ↩ ↩

  7. Participant primary record: Sears Holdings Corporation, Form 10-K for fiscal 2017, filed March 23, 2018; Item 1A on competition and operational risks; Item 7, pp. 42–43, on capital expenditures, asset proceeds, debt, losses, and negative cash flow; “Domestic Pension Plans Funding”; and note 13, pp. 110–112, on store-closing, severance, and impairment charges, SEC filing. The filing is strong for recorded counts, transactions, obligations, and management's risk disclosures. It is not an independent causal analysis, and company costs do not substitute for affected-party outcomes.

    ↩ ↩ ↩ ↩ ↩ ↩

  8. Official pension-administrator record: Pension Benefit Guaranty Corporation, “PBGC to Pay Pension Benefits for Employees and Retirees at Sears and Kmart,” January 18, 2019, paragraphs beginning “PBGC has worked” through the plan-guarantee explanation, PBGC press release. The record supports the agency's $1.4 billion underfunding estimate, 64-percent funded estimate, proposed termination date, and guarantee expectation. It does not represent beneficiaries' experiences or measure every loss and transition burden.

    ↩ ↩ ↩
Research record

Evidence basis

Claim Cited. Material claims carry source locators; comparative interpretation may still evolve.

Open questions and affected lives

Benefit-to-life status: Seed

  • Who can shape or revise the structure, and who can appeal a structural decision?
  • Where might informal status, platform control, or access to information undermine formal decentralization?
  • Who bears the migration, coordination, and livelihood costs of a reorganization?
  • How can a structure represent communities, ecosystems, nonhuman life, and future generations that do not occupy roles on its chart?

These questions remain open; absence from the record does not imply absence of benefit or harm.

Ledger
I05
Research
Working
Evidence
Claim Cited
Ethical review
Seed
Structured atlas record

Lineage nodes

  1. Sears's shared retail systemjoins stores, brands, real estate, marketing, online operations, supply, service, and customer trust through dependencies no unit can fulfill alone
  2. The 2008 internal-market designdivides the company into operating, support, brand, online, and real-estate businesses with autonomous leaders and separate result accounts
  3. Holding-company controlretains authority over unit appointments, boards, and common doctrine while describing decision making as decentralized
  4. Local profit and shared experiencemakes unit results legible while encouraging bargaining over space, brands, marketing, and services that customers encounter as one retailer
  5. Reported cross-unit conflictshows unit incentives and executive disputes relocating rather than resolving interdependence
  6. Decline with multiple causesplaces the structure beside changing retail competition, losses, declining capital expenditure, debt, asset transactions, and inherited weakness rather than treating it as a controlled experiment
  7. Residual claimantsfollows store workers and pension beneficiaries through documented closures and plan obligations while identifying evidence gaps for customers, suppliers, and communities
  8. Rival structural answerscompares the multidivisional form, internal markets, networks, and platforms by where each places integration and appeal

Typed relationships

ASears's shared retail system → The 2008 internal-market design

Organizational Redesign: Sears announced in 2008 that it would reorganize its integrated operations into five types of more autonomous business units.

The announcement and chairman's letter establish intent and formal design, not how authority and cooperation worked in practice.

AThe 2008 internal-market design → Holding-company control

Power Analysis: Unit leaders were accountable for results, while the holding company appointed them and placed senior executives on unit boards.

Formal decentralization therefore redistributed operating decisions without removing central ownership or appointment power.

BThe 2008 internal-market design → Local profit and shared experience

Operationalization: Separate leadership and result accounts made individual businesses more legible and encouraged them to negotiate internal relationships as transactions.

The formal documents support separate accountability; the strength and exact reach of internal pricing varied across relationships and periods.

BLocal profit and shared experience → Reported cross-unit conflict

Organizational Consequence: Independent reporting described business-unit leaders protecting their own results in disputes over shared brands, space, and support.

The reporting relies substantially on former insiders and examples; it demonstrates conflict but does not by itself measure the structure's net causal effect.

CReported cross-unit conflict → Decline with multiple causes

Causal Qualification: The internal-market structure plausibly weakened cooperation, but Sears's decline also involved changing competition, losses, declining capital expenditure, financing and asset choices, and problems inherited before 2008.

This is a bounded synthesis rather than a causal estimate; bankruptcy is not a controlled test of organizational form.

BDecline with multiple causes → Residual claimants

Burden Transfer: By 2017, Sears reported hundreds of store closures, severance costs, and substantial pension funding obligations as cash pressure continued.

The filing documents consequences and obligations; it does not assign each closure or cost to the 2008 structure.

CRival structural answers → The 2008 internal-market design

Interpretive Extension: Multidivisional hierarchy, internal markets, networks, and platforms can be compared by who integrates shared dependencies and who can appeal when an interface fails.

This is analytical comparison, not a claim that Sears descended directly from every form or that any form has one universal effect.

Provenance and sources

Corpus
management-research-library-2026-07-10
Source section
Structure, hierarchy, and scale
Imported
2026-07-10

Online anchors

Continue exploring

73 connected institutions