The discipline to shrink: Buffett's 2004 portrait of NICO

Published: February 28, 2005 — Berkshire Hathaway 2004 Chairman's Letter

Warren Buffett ended calendar year 2004 in an awkward position: Berkshire Hathaway held $43 billion in cash equivalents, a figure he described as "not a happy position." He had found neither compelling acquisition targets nor attractive securities. The pile simply sat there, earning short-term rates, while the businesses Buffett wished he could buy remained priced beyond what he was willing to pay. 1

Yet in the same letter — published the same day, covering the same year — Berkshire's insurance operations reported a $1.55 billion underwriting profit while floating an average of roughly $45.2 billion in policyholder funds. 1

These two facts exist in tension. Berkshire was cash-poor in the sense that Buffett couldn't deploy capital as fast as he wanted to. It was also being paid over $1.5 billion per year simply to hold other people's money. The reason for both conditions is the same story: an insurance culture so disciplined that it could do something almost no institution can — deliberately shrink.

www.berkshirehathaway.com

Berkshire Hathaway 2004 Chairman's Letter

Warren Buffett's annual letter to shareholders, published February 28, 2005. Covers Berkshire's insurance operations, the float engine, NICO's underwriting discipline, and several of Buffett's most-quoted aphorisms on investor psychology and decision-making.

正在加载链接预览…

The institutional imperative that most companies can't escape

The story of National Indemnity Company (NICO) — Berkshire's original insurance subsidiary, acquired in 1967 — is told in full in the 2004 letter through what Buffett calls a "Portrait of a Disciplined Underwriter": a 25-year table of premiums written, employee headcount, expense ratios, and underwriting margins from 1980 through 2004.

The headline number from that table is almost alarming. NICO's premiums written fell from $366.2 million in 1986 to $54.5 million in 1999 — an 85% contraction over thirteen years. 1

This was not an accident. It was a decision, renewed every year, not to write underpriced business. When insurance markets softened — when competitors cut rates to gain volume, when Wall Street applauded growth in premiums written — NICO simply stopped writing. Volume dwindled. Headcount stayed flat because NICO's management had also made a separate, unusual promise: no one would be laid off because of declining business volume.

Buffett is explicit about why this discipline is nearly impossible for most organizations to maintain:

"But living day after day with dwindling volume — while competitors are boasting of growth and reaping Wall Street's applause — is an experience few managers can tolerate." 1

The institutional machinery of almost every large organization works against contraction. Managers are paid on revenue. Analysts reward growth. Employees fear for their jobs when premiums shrink. Each of these pressures creates an incentive — often a subliminal one — to write business that looks acceptable on paper but will cost money once claims arrive, sometimes years later.

Insurance has one feature that makes this psychology particularly dangerous: the lag between pricing a policy and learning whether it was priced correctly can stretch years, sometimes decades. An underwriter who writes foolish business in year one may not see the consequences until year five or seven. That gap is long enough for a culture to convince itself, repeatedly, that this cycle will be different.

NICO's solution to this structural problem was not a formula. Don Wurster (NICO's CEO since 1989, by the time of the 2004 letter) simply refused. The no-layoff guarantee removed one of the most corrosive pressures — the fear that discipline today would cost jobs tomorrow. When employees know the firm won't shrink the workforce to compensate for a soft market, they have less reason to push volume at the expense of rate adequacy.

Float: getting paid to hold other people's money

The consequence of NICO's discipline — and of the same culture operating across Berkshire's broader insurance group — is what Buffett calls the "propellant" of Berkshire's growth.

Float is the money insurance companies hold temporarily. Policyholders pay premiums upfront. Claims are paid later — sometimes years later, as with asbestos litigation or complex casualty settlements. In the gap between receipt and payment, the insurer holds and invests the funds. The cost of float is determined by underwriting results: if the insurer earns an underwriting profit (collects more in premiums than it pays in losses and expenses), the float was free — or better than free.

By year-end 2004, Berkshire's total float was $46.094 billion, distributed as follows: 1

Segment	Float (year-end 2004)
General Re	$23.12 billion
Berkshire Hathaway Reinsurance	$15.28 billion
GEICO	$5.96 billion
Other Primary (incl. NICO)	$1.74 billion
Total	$46.09 billion

That $46 billion had grown from $20 million when Berkshire acquired NICO in 1967. Buffett's description of this growth is measured, not triumphant:

"Float is wonderful — if it doesn't come at a high price. ... When an underwriting profit is achieved — as has been the case at Berkshire in about half of the 38 years we have been in the insurance business — float is better than free. In such years, we are actually paid for holding other people's money." 1

The "about half" qualifier is honest and important. Berkshire's insurance operations have not been profitable every year. The standard in the industry, Buffett notes, is considerably worse — most insurers consistently operate at an underwriting loss, which means their float carries a positive cost. They are, in effect, borrowing money to invest, at a price. Berkshire's discipline has produced float that has been, across the full history, closer to free than to expensive.

www.berkshirehathaway.com

Berkshire Hathaway 2004 Chairman's Letter — insurance float table

Pages 9–11 of the letter contain Buffett's detailed breakdown of Berkshire's $46 billion float across four insurance segments, including a discussion of underwriting profit and the cost of float over 38 years.

正在加载链接预览…

The NICO float itself — $1.74 billion under "Other Primary" — is described by Buffett as "almost certain to be proved the negative-cost kind." This is the reward for thirteen years of shrinking. Wurster spent a decade watching competitors grow faster and report better short-term results, while NICO's premium volume fell. The float that accumulated over those years carries almost no cost, because it was built by refusing to price risk below its true value.

What this means for investors reading capital allocation decisions

The float engine is one of the most-cited mechanisms in Berkshire's history, and it often attracts the wrong lesson. Investors sometimes conclude that the float itself — the large pool of investable funds — is the structural advantage. It is not, or not primarily.

The structural advantage is the underwriting culture that produced low-cost float. Float at a positive cost is worse than leverage, because it is less visible and harder to unwind. The insurance industry generates large floats routinely; the question is always the price.

What Buffett is actually describing in the NICO portrait is an organizational immune system: a culture, reinforced by specific structural choices (no-layoff guarantee, management whose compensation is not revenue-based, owner-operators rather than professional managers seeking short tenures), that resists the institutional imperative to chase volume. The float is the symptom. The culture is the thing.

For investors evaluating any company that controls a large pool of other people's money — insurers, banks, asset managers, pension funds — the diagnostic question is not "how large is the float (or AUM, or deposit base)?" It is: what does management do when writing less business is the correct thing to do?

That question almost never has a clean answer in a quarterly earnings call. Buffett's 25-year NICO table is one of the few times in public financial history when the answer has been laid out in full, one row at a time, over a generation.

www.berkshirehathaway.com

Berkshire Hathaway 2004 Chairman's Letter — NICO portrait of a disciplined underwriter

Pages 7–8 of the 2004 letter include the full 25-year table of NICO's underwriting history, including premiums written, employee count, expense ratios, and underwriting profit margins from 1980 through 2004.

正在加载链接预览…

The deeper implication reaches past insurance. Any business where customer acquisition cost is front-loaded and cost discovery is delayed — subscription software with high churn, consumer lending, real estate development — has a structural incentive to over-sell, because the seller often collects revenue before the problem appears on the balance sheet. The NICO discipline applies in every such context: the willingness to lose market share in a pricing-soft environment is the test.

Buffett closes the section with a note about scale and commitment that applies equally to the $1.5 billion in underwriting profit and to NICO's disciplined thirteen-year contraction:

"Investors should remember that excitement and expenses are their enemies." 1

Written in 2005, the sentence was aimed at the individual investor's habit of trading too much and paying too much. But the organizational version of that principle is exactly what the NICO portrait illustrates: the years when competitors were most excited about insurance growth — the years of boasting and Wall Street applause — were the years NICO was most disciplined about writing less.

Cover image: AI-generated illustration

参考来源

1Berkshire Hathaway 2004 Chairman's Letter