How a parent company can create or destroy value across unrelated businesses, and what evidence distinguishes a capability from a corporate label.
Ownership is not the same as integration
A conglomerate is a parent that controls businesses whose products, customers, or production systems are not all closely related. The units may share a treasury, procurement platform, distribution network, brand, technology, or nothing more than ownership. That distinction matters. A common legal entity does not prove that the businesses share an operating capability.
The useful question is what the parent can do that a collection of independent owners could not do as well. It may move capital quickly, preserve a business through a downturn, provide specialized expertise, or combine demand and distribution. It may also add headquarters cost, slow decisions, cross-subsidize weak units, or make it difficult for investors to see which business earns the returns.
Three possible sources of value
| Parent capability | Operational mechanism | Evidence needed |
|---|---|---|
| Internal capital market | Cash moves from a mature unit to a project that cannot yet raise external finance | Investment decisions, incremental returns, and whether weak projects are stopped |
| Shared operating resource | Units use the same technology, customer access, purchasing, or distribution | Usage, avoided cost, service quality, and the cost of maintaining the shared system |
| Risk and time horizon | Parent absorbs temporary volatility or funds a long project across cycles | Liquidity, governance, and whether risk is actually diversified rather than hidden |
Internal capital is not automatically superior to markets. Stein's analysis of internal capital markets argues that a parent can direct funds toward divisions with better opportunities, but information and managerial incentives can also cause inefficient investment. The theory provides a mechanism, not a universal prediction.
Berkshire: decentralization as a corporate capability
Berkshire Hathaway is a useful case because its structure is not presented as one integrated operating chain. The 2024 shareholder letter describes insurance, operating businesses, and investments while emphasizing decentralized management and capital allocation from the center. The parent can retain cash, buy securities, or acquire a business without requiring each unit to build a separate treasury.
That description does not establish that Berkshire creates value merely by owning diverse businesses. The relevant tests are whether managers receive useful autonomy, whether the center allocates capital at attractive incremental returns, whether subsidiaries can avoid duplicated bureaucracy, and whether the balance sheet can absorb a unit-specific shock. The same structure could fail if the parent overpays for acquisitions or protects poor operators.
GE: when the shared system stops helping
General Electric offers a contrasting case. Its 2018 annual report records plans to separate healthcare, reduce the Baker Hughes stake, sell assets, and shrink GE Capital while reorganizing the industrial portfolio. Those choices followed different capital requirements, leverage concerns, and operating cycles. The decision is evidence that management judged the existing boundaries costly or no longer useful; it is not a controlled test proving that the conglomerate form caused all of GE's problems.
For an analyst, the comparison between Berkshire and GE is productive because both owned diverse businesses but used different levels of centralization, leverage, and operating integration. A label such as “conglomerate” predicts less than the actual authority, information, and capital flows.
How value destruction enters
- Cross-subsidy. Cash from a strong unit funds a weak unit without a clear return hurdle.
- Central complexity. Shared systems become expensive or slow when units need different decisions and controls.
- Capital rigidity. Debt, tax, pension, or regulatory obligations at the parent limit the response of healthy units.
- Information loss. Segment reporting does not fully reveal transfer prices, shared assets, or the quality of management at each unit.
- Boundary failure. A parent keeps a business because of history or executive status after the original operating or financing rationale has disappeared.
How to investigate a conglomerate
Map every material flow between parent and units: cash, guarantees, people, customers, intellectual property, procurement, and debt. Ask whether the flow creates a measurable benefit and whether an independent unit could buy it more cheaply. Examine the center's record of acquisitions, divestitures, reinvestment, and stopped projects. Finally, compare the parent's financing and governance constraints with the alternatives available to a focused company.
Conglomerate structure is best understood as an operating choice. It can preserve a long-term capability that markets would underfund, or it can conceal an inability to allocate and let go. The evidence lies in the flows and decisions, not in the number of industries listed beneath the parent's name.