No Stop-Loss on a Forecast

In April 2014, a survey asked sixty-seven economists which way ten-year Treasury yields would move over the next six months. All sixty-seven said up. Not a majority — all of them, 100%.

Yields fell.

MarketWatch ran the follow-up under a headline you could frame: “Yes, 100% of Economists Were Dead Wrong About Yields.” And this wasn’t a one-off wobble; for the better part of a year the same crowd had been near-unanimous that rates were about to jump, while the ten-year quietly slid from over 3% to under 2.3%.

Nice miss.

Here’s the thing, though. I didn’t open with that to sneer at economists. Becoming any good at this comes in two stages, and the joke only makes sense once you see them both.

Stage one, you learn. You read a great deal, you listen to the experts, you take in every framework you can, and out of all of it you build a system — one that fits your own temperament, your appetite for risk, your patience, your laziness. That stage ends the day you’ve got a mental machine that will always, in any weather, tell you what to do next. Stage two is the strange one: from then on you forget the experts. You read them for fun, the way you’d read a gossip column. Because the moment an expert’s opinion starts moving your actual decisions, that isn’t information — it’s a hole in your system, and the market will find it and charge you for it.

A long-term investor has no business caring what the monthly jobs report said, or what George Soros thinks about the euro this week. A short-term trader, only very rarely — and only if reacting to headlines is literally the game being traded.

Now, the tempting way to write this post is to line up the famous bad calls and laugh. Roubini calling one crash and then burying that single hit under a decade of doom that never arrived. Soros predicting the euro’s collapse so many times that anyone who traded on it felt like a fool a few years later. The Jim Rogers fans who took his word and built what amounted to a commodities warehouse in the back garden.

But that’s the cheap version, and it misses the real point.

Because Soros and Rogers understand markets far better than you or I ever will. That isn’t in question. So why is their opinion still useless to you?

Because of what’s missing from it. When one of these people tells you where the market is going, do they hand you the probability they attach to it? The size you should put on? The level at which you’re wrong and should get out? No. And those three things — probability, position size, the stop — are what actually decide how you end up. The direction is the least important part of the trade. Rogers once shrugged that oil at $40 didn’t bother him in the slightest; better for long-term prices, he said. Fine for him. You, sitting on a 50%-plus drawdown, might feel differently. His opinion was calibrated to his balance sheet and his nerves, not yours. It cannot be transplanted into your account, because the thing that makes it work — the risk tolerance behind it — stays in his body.

Someone did try to score all of this properly. A research outfit called CXO Advisory spent years translating the market calls of US media pundits into concrete long/short positions and checking them against what the market actually did. Between 2005 and 2012 they logged 6,582 forecasts from sixty-eight named experts — bulls, bears, chartists, value people, the lot. The final tally, averaged across the gurus: 47.4% correct. Below a coin flip. Plot the accuracies and you get a bell curve, the exact shape you’d expect if every one of them had been guessing.

Believe me, you can hit 47% too. You don’t need the newsletter.

So far, so easy — bad forecasters are a soft target. Let me make the harder case, because the harder case is the one that actually costs people money: even good advice, concrete and specific and handed to you by someone genuinely excellent, will not save you.

Picture the setup. A row of respected managers agree, roughly, that the bull has another year, maybe eighteen months, left in it. Curious detail: they agree on the timeline but their actual recommendations scatter to the winds — one says raise cash, another says buy these names, a third says those. Meanwhile Buffett is out calling valuations reasonable, and another well-known strategist figures the run has six to eight years to go. (I’ve always had a sly suspicion about that “year to eighteen months” number, by the way. It’s just long enough that we’ll have forgotten the call by the time it’s due.)

Now one of them — the sharpest of the bunch, a man running one of the best funds in the country — actually gives you something usable. Two clean rules. Sell now: unemployment is at a multi-decade low, there’s not much road left upward. And buy back when some developed economy officially declares a recession — because by the time that announcement lands, the market has already fallen hard, the bad news is in the price, and you’re buying the wreckage cheap.

It sounds airtight. So why would following it be a mistake?

Start with the easy objection. If this man is that good — good enough that you’d stake your plan on his read — then why are you fiddling with his advice at all? Give him the money. He will interpret, time, and execute his own rules incomparably better than you ever will; he can sit there reading his own interviews and acting on them. Trusting someone halfway is the worst of both worlds.

But set that aside and look at the rules themselves.

Take the buy signal — buy when the recession is official. Here are the dates the US actually announced its recessions, and how much further the market fell after each one:

• December 2008 — roughly another 25% to go.
• November 2001 — roughly another 30%.
• January 1982 — about 15% more.
• April 1991 — barely a scratch, around 4%.

Did that show up in the confident version of the advice? It did not. The signal isn’t a floor; it’s a starting gun with fifteen to thirty percent of downside still ahead, depending on how bad the recession turns out to be. You don’t dodge the bear. You half-dodge it, on a good day.

Now the sell signal — unemployment at its lows. Pull the US jobless rate back to 1948 and you’ll see cycle after cycle bottoming at different levels, some lower, some higher; there’s no clean line that says “here, now, sell.” And it wasn’t even a forty-year low — the late nineties ran lower. Call it a twenty-year low. When was the last one of those before it? April 1997. From that point the S&P 500 went on to climb something close to 100% over the next three years before it finally topped out. Three years and a double. A bit of uncertainty in the signal, wouldn’t you say?

So let your imagination off the leash. You sell, or you just stop buying — and the market keeps ripping, month after month. At some point you crack and pile back in. Maybe the same brilliant manager publishes a fresh note three years later explaining the bull has, oh, another year to eighteen months left. And the buy side is no kinder: you step in on the recession headline and the thing drops another 20-30% under you. In March 2009, with the index near 700, you could find serious people forecasting a fall to 300 — another halving from the bottom. Nobody was going to feel that for you. You’d leaned on someone else’s advice, and the advice had wandered off somewhere around the halfway mark, leaving you to carry the rest alone.

There is exactly one way out of this, and it’s the same door every time. Never act on someone else’s opinion. If you can’t decide, on your own two feet, that it’s time to sell — then don’t sell. If you’ve got a system that makes the call, follow the system, by all means. Though I’d love to see the back-tested rule that’s screaming sell now, today, right this minute.

And the question everyone actually wants answered — one year, two years, eight years left in the bull? Let it go. Nobody knows. Not you, not me, and least of all the talking heads who get paid to sound like they do. The point was never that the experts are idiots and I’m clever. The point is that a forecast reaches you with all its working stripped out — the odds, the size, the exit, the stomach behind it — and without those you can’t build a real decision on it, however smart the person who said it.

So drop it, and walk your own path. We’ll find out how long the bull had left the same way everyone always does. Afterwards.