Insurance Float

Concept Analysis

Definition & Origins

Insurance float is the pool of money held by an insurance company between the time premiums are collected and the time claims are paid. In a well-run insurance operation, this float represents essentially free — or even negatively-cost — financing for investment purposes. Buffett has described it as "other people's money that we get to invest" and as the structural engine behind Berkshire's extraordinary compounding.

The float concept was not new when Buffett acquired National Indemnity in 1967 — but Buffett's insight was to recognize that float managed with disciplined underwriting could be a permanent and growing capital advantage, not merely a transient liability. Most insurance accounting treats float as a liability (future claims to be paid); Buffett treats it as a costless or near-costless source of capital if underwriting discipline is maintained.

Core Ideas

Float is only attractive at low cost. The key metric is not the size of the float but the cost of acquiring it. If underwriting generates losses equal to 5% of premiums, the float costs 5% per year — still potentially attractive versus the bond yields available as an alternative source of capital, but increasingly marginal. If underwriting discipline ensures that premiums collected exceed claims paid (combined ratio below 100%), the float is free or negative-cost — an extraordinary competitive advantage.

Scale creates a compounding float advantage. GEICO's float grew from $600 million when Berkshire acquired the remaining stake in 1995 to over $20 billion by 2020. Each new policyholder adds to the float pool; the investment returns generated by that float further compound. The structural advantage of insurance float grows with scale — making Berkshire's insurance operations increasingly difficult for competitors to replicate.

Float duration matters as much as size. Long-tail insurance (policies where claims develop over many years — workers' compensation, environmental liability, medical malpractice) generates float of long duration, allowing investment in longer-term assets with higher yields. Short-tail insurance (auto, homeowners) generates float of short duration — claims are paid quickly — limiting investment to shorter-duration instruments. Berkshire's catastrophe reinsurance business generates float of intermediate but variable duration.

Catastrophe years test the float model. A diversified insurance operation's float should be robust to individual catastrophe years — the question is whether the combined ratio over a full underwriting cycle (including catastrophe years) remains near or below 100%. Years with exceptionally large catastrophes will produce above-100% combined ratios; years without will produce unusually low ratios. Buffett accepts this volatility because the long-term average is what determines whether float is genuinely free.

Float has a stated cost that can be negative. Buffett reports "cost of float" (essentially the underwriting profit or loss as a percentage of float) in every annual letter. A negative cost of float — where underwriting generates a profit — means Berkshire is paid to hold other people's money. This extraordinary circumstance, sustained across many years by disciplined underwriting, is what makes Berkshire's insurance operations uniquely valuable.

Practical Application

National Indemnity (acquired 1967): Early model of the float concept. A specialty property/casualty insurer known for writing unusual risks — policies that other insurers refused. The discipline of charging adequate premiums for unusual risk created a culture of underwriting integrity that persisted across the float model.

GEICO (acquired fully 1995): The defining float machine. GEICO's structural cost advantage (direct distribution eliminating agent commissions) meant it could simultaneously offer lower premiums to customers and generate larger underwriting profits than competitors — creating more float at lower cost. Buffett has described GEICO as the most important single holding in Berkshire's history.

General Re (acquired 1998): Cautionary example. General Re was acquired partly for its large reinsurance float but proved to have problematic underwriting standards and a contaminated derivatives book. The integration required years to clean up and represented a period where Berkshire's float cost too much. The lesson: float quality depends entirely on the underwriting culture of the operation generating it.

Catastrophe reinsurance: Berkshire writes large catastrophe reinsurance contracts — covering other insurers against major natural disaster losses — at volumes that no competitor can match because Berkshire's financial strength allows it to commit to contracts that require paying claims immediately in a catastrophe. The float from these contracts is explicitly priced to compensate for the catastrophe risk.

Common Misconceptions

Misconception 1: Any insurance float is valuable. Float generated by inadequate underwriting — where combined ratios consistently exceed 100% — represents expensive capital, not free capital. Many insurance expansions that appeared to generate impressive premium growth actually generated float at rates above market alternatives, destroying rather than creating value. Float is only valuable when underwriting discipline is maintained.

Misconception 2: Insurance is a commodity business. The ability to charge adequate premiums across market cycles — resisting the industry's tendency to cut premiums in soft markets — is the rare discipline that separates superior insurance operations from commodity providers. GEICO's structural cost advantage allows it to price below competitors while maintaining adequate underwriting margins; General Re's cultural weakness was accepting inadequate premiums to maintain market share.

Misconception 3: The float model is easily replicable. The combination of scale, underwriting discipline, financial strength, and investment expertise required to replicate Berkshire's float advantage has no parallel in the insurance industry. Smaller insurers with less disciplined underwriting and shorter investment horizons cannot replicate the model.

Thought Evolution

Related Concepts

Case Companies