# Finding the right mix of cash and precious metals

I'm trying to decide how to split my portfolio between cash and precious metals. For that purpose, the big question in my mind is, what's the chance that gold will go significantly down from here (before it goes higher), say down to \$1000/oz. For example, in a crash, if deflation gets out of hand, it's possible (although not certain), that the dollar price of gold could crash. So I'm trying out scenarios: gold goes higher, gold goes lower... And I discovered an interesting idea regarding how you account for potential losses.

Suppose you had \$100,000 in investable assets in 2012 and bought gold at \$1600/oz with 50% of that. Assume you didn't make or lose any money since then and you had the rest in cash. Now your portfolio is obviously worth less, around \$92k (assume a valuation of gold at \$1350/oz for simplicity). I always thought that in this hypothetical case, you've lost \$8k. But have you really? Bear with me.

I recently re-read Fooled by Randomness by Nassim Taleb. It's one of those books you want to re-read every year lest you forget the great lessons contained and go back to being a fool of randomness. He presents the following thought experiment: Suppose you someone offers you this weird game of Russian roulette. You load a bullet into a gun, spin the cylinder and fire into your head. If you come out intact, you get paid \$10 million. You accept the "challenge," fire and make it all in one piece. You get \$10 million.

How are you to account for what happened? If you only look at results, you think "I made \$10 million. I'm a genious, yay!" But if you go back to the moment before you pulled the trigger, you can see there were 6 possible ways for the experiment to unfold, and you were dead in one of them. So a better way to think of how you'll be after the experiment is: 1/6th chance dead, 5/6th chance with an extra \$10 million. Or think of it as 6 possible lives, 1 of them with you in a coffin. He goes on to explain that many extremely successful people you hear about are successful because they took massive risks and were lucky not to fire the bullet in the chamber, but we only get to see their "\$10 million," not the massive risks and the alternate futures they could have ended up in. But that's beside my point here.

So how does this apply to your hypothetical portfolio losses? You thought you lost \$8k, but that's because you are ignoring other potential outcomes that existed back in 2012 that didn't come to pass. For example, suppose that in 2012 you had the foresight to stay all in cash, no precious metals. Amazing, right? You'd still have \$100k. But maybe, in an alternate future, you woke up on the morning of February 5th, 2015 and gold had been re-monetized at \$10.000/oz because the Chinese got fed-up with bankrolling the profligate US government and tried to dump their treasuries and force a new monetary system backed by gold. Sorry, too late to get some gold at \$1600. Your portfolio is still \$100k nominally but it buys a lot, lot less real stuff now. Was this scenario within the realm of possibility back in 2012? Sure.

So in trying to assess how much you lost I think it's more reasonable to ask this question: knowing what you knew at the time of the purchases in 2012, how much should you have bought? You couldn't know that the actual course of events would take gold down to \$1050, so you probably should have bought at least some gold. 10%, 50%, 90%? That's up to you to decide, but you probably realize that with the information available at the time, 0% was the wrong answer.

So suppose that you were to reassess what you knew back then and conclude that a better decision, without the benefit of hindsight, would have been 40% (maybe you underestimated the chance that gold could go down because you ignored information that you had available at the time). How much did you actually lose up to today?

If you allocated 50%, today you'd have \$92,188 = \$50,000 + \$50,000 * \$1350/\$1600.

If you allocated 40%, today you'd have \$93,750 = \$60,000 + \$40,000 * \$1350/\$1600.

Your loss, in this all-possible-worlds way of viewing things is \$1,562. You would have allocated that \$40k to gold come what may, so the only loss that counts is the one suffered on the \$10k that you shouldn't have allocated to gold.

Why does all this hypothesizing matter at all? I think it matters because we don't know where the price of gold is going from here. I think higher is the most likely path, particularly now that Democrats and Republicans have finally shaken hands and agreed to throw all traces of fiscal responsibility out the window. But I could be wrong. I've certainly been very wrong in the past in forecasting the direction of the gold price. If instead of going up, gold first goes down to \$1000/oz and I have to sell some, expecting further losses before I see gains (say in a scenario where debt deflation is getting out of hand), how much have I lost, really? I think that the above reasoning illuminates a better way to conceptualize the loss than the more intuitive traditional way which ignores alternate histories that didn't come to pass.

The same thing can be said for the scenario in which gold goes much higher from here, never to come back to current levels. What happened to the purchasing power of the cash you kept? Did you lose all of that purchasing power, only a fraction, or none?