Failure modes · Horizon
Julian Robertson Closed Tiger at the Nasdaq Peak. Being Right Was Not Enough
Tiger had identified the technology excess years earlier and refused to participate. Client withdrawals and relative-performance pressure forced the closure in March 2000, the month the index topped. The subsequent 80 percent decline proved the analysis and arrived too late for the fund.
Tiger Management reached roughly $22 billion in 1998. Robertson’s view that technology valuations were irrational kept the fund out of the hottest names. From 1998 to 2000 the funds lost about 19 percent while the Nasdaq rose sharply. Investors withdrew $7.7 billion. In his March 2000 letter Robertson wrote that there was no real indication a quick end was in sight. He closed the fund that month. The Nasdaq then fell nearly 80 percent.
The analysis was directionally correct over the subsequent two years. The capital was no longer there to benefit. De Long and colleagues had already described the mechanism: noise traders can push prices further from value, and the agents who would bet against them face finite horizons and the risk of redemption before convergence. Robertson’s experience is the case study.
What personal capital changes
Buffett closed his partnership in 1969 for related reasons—he could not find bargains and disliked the speculative environment—but he controlled the capital. He could wait. External money, especially money that arrives after a period of strong performance, often leaves when relative results disappoint, regardless of whether the absolute view is about to be vindicated.
For an individual investor the structural protection is the absence of outside redemptions. That protection only works if the investor does not create an internal redemption by abandoning the position when peers are still making money. The Robertson episode shows the external version. The Druckenmiller episode shows the internal version. Both end with the correct analysis producing no profit.
What to do with this
If you hold a position or an underweight that you believe is correct but that can continue to hurt relative performance for an unknown period, write the maximum time or maximum relative drawdown you will tolerate before you must reduce. If the capital is not yours alone, the number needs to be agreed with the people who can force the exit. Being early remains indistinguishable from being wrong until the capital survives long enough to be early.