A falling investment can recover and still fail its owner. Someone who needs the money for surgery in three months may have to sell before the recovery. An owner with dependable long-term capital can have more time. The investment is the same; the consequences of its decline are different.
Howard Marks offers no universal percentage of capital that everyone can safely risk. His answer begins with the loss you could bear, the cash you will need and the reliability of your financing. A return can also fall short of what your money needs to accomplish. Accept risks you understand only when the prospective reward justifies them, while leaving enough room to keep investing if events disappoint. Eliminating every uncertainty would also eliminate worthwhile opportunities.
Marks puts permanent loss of capital first. Volatility measures how much prices move, which cannot fully describe whether a business will destroy your money or whether you paid too much. Yet a temporary markdown becomes consequential if panic or a payment deadline forces you to sell. A smooth price history proves little, either. Favorable conditions can leave a dangerous portfolio untested for years.
When the averages cannot help
Riskier investments, when fairly priced, offer higher expected returns because buyers require compensation for uncertainty. The word expected matters. It describes an average of possible outcomes weighted by their probabilities. Greater exposure widens the possibilities, including worse losses; it does not promise the higher return. Even an accurate estimate cannot make the favorable outcomes arrive when your finances need them.
There are two distinct problems here. First, the model may misunderstand what is possible. Human behavior changes markets, and losses can exceed anything in the historical sample. Second, even correctly understood probabilities permit unusual outcomes to arrive in clusters. A strategy can run out of capital during that bad sequence, before its long-run average has any practical value. Better calculations alone cannot remove either danger.
Confident economic forecasts offer limited comfort. Predictions that recent conditions will continue may often be right, but markets usually already expect continuity. The major changes that would make foresight valuable are harder to predict. Marks therefore favors researching businesses and arranging a portfolio for several possible futures. Survival should not depend on switching perfectly from aggression to caution just before trouble arrives.
Life insurance provides his useful analogy for accepting risk. Insurers know their customers will die; the uncertain part is when. They assess applicants, spread policies among different people and charge premiums to compensate for expected claims. Investors can follow a similar discipline with a borrower: investigate the exposure, avoid concentrating it and demand adequate compensation. The analogy has a limit. Investment losses are harder to estimate than mortality. A convincing analysis still leaves uncertainty about both their size and likelihood.
Room for an imperfect judgment
A low purchase price can absorb some disappointment. Marks illustrates this with an asset estimated to be worth one hundred dollars. Paying ninety leaves ten dollars between price and estimated value. Paying seventy leaves thirty. The larger discount gives more room for the estimate to be optimistic without necessarily causing a loss. That is a margin of safety. But the value must be dependable. Continuing business losses or obsolescence can consume the apparent cushion. A cheap quotation cannot replace research.
The portfolio needs room for error too. Owning many names helps little if they all depend on the same favorable event. Marks asks how the holdings respond together. Different industries can still share a need for easy credit, and margin calls or frozen markets can force apparently unrelated assets down at once. Diversification requires differences in exposure, not merely a longer list of investments. Even sensible diversification cannot promise that losses will never coincide.
Borrowing and position size determine how much those losses can hurt. A commitment large enough to improve the result materially can also damage it materially. Limiting concentration and borrowing reduces dependence on one outcome. Claims paid before other investors' claims can provide additional protection, but dividing an asset's payments into layers does not erase the underlying risk. These choices aim for tolerable outcomes across circumstances, accepting some gains forgone in the most favorable one.
Protection itself has a price. Buying protection against every imaginable disaster can consume too much return or prevent investing altogether. Even preparation for a very rare event can meet something worse. Marks gives priority to limiting large losses because they leave less capital available to recover and compound. He does not claim that every loss can be prevented or that the most cautious portfolio always wins.
The balance also depends on the asset. An ordinary bond's promised upside is limited, so avoiding borrowers who fail to pay is central. Shares have greater potential upside, making participation in gains important too. Commentator Christopher Davis warns that an equity manager can take defense too far by hiding in supposedly safe holdings to avoid controversy. Competence, temperament, patient backers and dependable capital determine how much aggression an investor can sustain.
The remaining decision is how much potential gain to surrender for protection you can afford. Good years can make that protection look unnecessary because its benefit is a loss that never happened. The aim is to retain enough capital and composure to make the next decision, including after a future your model failed to describe.
Three things to keep
Judge risk by the losses, cash needs and financing demands you can survive.
A meaningful discount and different portfolio exposures leave room for assumptions to fail.
Protection costs potential gains, and even careful preparation has limits.