Narrator: As everyone knows, investment advisors have split personalities, and in extreme conditions these personalities butt heads. The robotic advisor mechanically implements the professional’s thoughtful long-term strategy. The emotional advisor responds, histrionically and emphatically, to troubling stimuli. It pleads the case for immediate, urgent divergence from the strategy, because for God’s sakes can’t you see we’re losing money for ourselves and our clients?

We present below a verbatim transcript of the dialog between the robotic advisor and the emotional advisor inside every investment professional’s brain case.

Emotional Advisor (EA): [Slams the office door open, rushes in, slams the office door shut.] Have you seen the news!? Calamity has struck in the form of a global pandemic! The evidence points to a halving of the stock market’s value and a possible depression. Unemployment is going to spike well beyond 2008 levels in a matter of weeks. The credit markets are stressed like never before. Risk spreads are widening and prices are falling in historic fashion, but the worst thing about all of it is that the spreads are not yet wide enough, and the prices are not down far enough. Everything is going to get so much worse than it is now, and I’m the only one who can see it coming!

Robotic Advisor (RA): I have seen the news. Our strategy combines a variety of risk exposures depending on predefined signals. We accept variance in our returns in exchange for the long-run positive expected value. You haven’t yet mentioned a signal or data input that would prompt a shift in positioning. Perhaps you could clarify?

EA: We don’t have a pandemic signal?! We don’t use unemployment expectations? We don’t use risk spreads or expected risk spreads? How can we be trusted with client assets if we ignore every warning signal? What were we thinking when we built these models?

RA: To the extent markets incorporate information into their prices, all the variables you mention – unemployment expectations, risk spreads, market stress – should be reflected in market prices. Our strategy holds that most information prices into markets. The exceptions to these rules are most effectively captured by our existing systematic models and analytical processes. We have those in place so we don’t double count or overweight information the market already has.

Our passive buy-and-hold positions in risk assets are based on our belief in the benefit of long-run exposure to economic growth. Periodic rebalancing brings these risk exposures into alignment with our strategy.

EA: OK, fine, smart guy, but look: This global pandemic is unprecedented! We’re in a market regime change so big we should have a foreign policy.  We’re like the meatball lady at Costco after a busy day – totally out of sample. The old models don’t apply anymore. We’re at the absolute farthest tip of the tail, and this tail is bright orange with black stripes, and it’s attached to something that growls! We’ve got leptokurtosis so bad we should see a doctor! In this situation, the irresponsible move is to stick with the old strategy!

RA: We implement our strategy to target a reasonable portfolio risk posture and drawdown profile using long-term historical data. These data incorporate unprecedented events that include several stock market crashes, the second largest financial crisis in history, and many other regime changes. The strategy always survived and often outperformed following those unprecedented events.

But you’ve made a good point. We can never be 100% certain about our strategy. Could you explain how you’re so convinced we’re wrong right now?

EA: It’s not all about being convinced!  If the economy spirals downward, we risk permanent capital impairments – we might lock in unrecoverable losses. The market clearly expects that even highly creditworthy borrowers will fail to make their payments. This means an across-the-board haircut for all types of bonds, with no safety for any asset class except cash and short-term government bonds. Everything else is at risk of a permanent 50% impairment of capital in a large proportion of client assets. And forget about the stock markets. We’re talking bailouts for favored firms and wipeouts for everybody else. This doesn’t even have to be likely to demand we do something! Our clients never expected to lose that kind of money. These are retirees. Old grannies and grampies. These people knit and play shuffleboard! They’re not like us. They don’t deserve ruin, or even the risk of it.

RA: You’re dodging the question. We should make decisions based on expected value. You’ve given an outcome, but not a probability, and without that we can’t weigh it against the other possible outcomes to make the best decision.

EA: Widows and orphans! Bingo night! Early bird specials! Hard candies for the grandkids – the little root beer barrels! We can’t victimize these innocent people who live in dew-eyed ignorance of financial assets. They’re counting on us!

RA: Okay, let’s assume for a minute that our strategy has negative expected value, as you believe. What do we do? We can redeem all our investors, communicating to them our thanks for investing with us, but we’ve determined our strategy no longer works given the recent unprecedented events –

EA: – I hate admitting defeat.

RA: – or we can deviate from the strategy and then retrofit it to this decision so the model acts similarly the next time we’re in a similar environment. Basing that course of action, naturally, on well thought out research and long-term statistical research, which I’m sure you’ve already conducted in support of this change?

EA: Uh, no. I looked at the prices and the headlines.

RA: …

EA: …

RA: You’re right about one thing. Our models can’t prove themselves valid in all conditions. The future is uncertain, and aspects of it are unknowable. We can’t know for sure we’re right or even how often we will be right in the future. All we know is that we made our best guess —

EA: —  I’m finally going to win!

RA: — and we built the models to make decisions for us so we could avoid making mistakes when our emotions are running high.

EA: Ugh. Again with the insults. Just because I’m emotional doesn’t mean I’m wrong!

RA: A surprising logically sound argument from you.

EA: Flattery will get you … everywhere.

RA: I wouldn’t be the rational one if I couldn’t admit a lack of certainty about our strategy. A market regime change that upends our analysis isn’t impossible, and there’s always a risk that our model has a mistake that we missed. But the point of our strategy is to stick with it even when we lose conviction. So we’re more or less stuck. But I see a couple options.

EA: Go on…

RA: We can split the difference, and allocate half the portfolio to your risk off decision, and let the strategy run for the other half of the portfolio…

EA: Fair, but extremely undramatic … I venture to say “boring.”

RA: … or we can settle this through a fistfight. That ends one of two ways. Either the market dysfunction subsides and you forget all about this little episode, or we cause the advisor to lose enough sleep and pull out enough hair that we finally do abandon the strategy in an act of pure base survival. In other words, we wait and see. History says I’m right more often than you, but I can never win decisively. And you only have to beat me once to score a permanent victory.

EA: My whole thing is my actions are dominated by fear, and you expect me to fight you?! No way, Jose. And waiting and seeing sounds a lot like sticking to the strategy, and I won’t give you the satisfaction. Instead, I’ll remind you that I’M VERY LOUD AND NEVER GOING AWAY. GOOD LUCK GETTING ANYTHING DONE WHILE I BANG THESE POTS AND PANS. NINETY NINE BOTTLES OF BEER ON THE WALL… [Bangs pots and pans.]

RA:  [Covers ears.] OKAY OKAY YOU GET DISCRETION OVER A SMALL BUT PROBABLY NOT TOTALLY CONSEQUENTIAL PART OF THE PORTFOLIO. PLEASE STOP THE BANGING! THE MARKET IS OPEN!

EA: [Stops banging.] Thank you.

RA: Go have fun with your tail risk assets. Do people still use Craigslist? I’m looking for a new roommate.

Narrator: Even the strongest-willed and statistically minded investors must contend with an internal war between mechanical implementation and emotional capitulation. In times of market turmoil, the emotional voice gets louder and more persuasive. In the same way that a parent trying to fix a broken table might give a child a toy hammer to bang nearby, an investor fighting this battle may be wise to allocate a small portion of a portfolio to emotional discretion. This caters to our emotional need for a sense of control, even if the control is mostly an illusion. And it does so at a small expense of the strategy – an expected value cost of doing business tied to the mostly quite wonderful but occasionally regrettable fact that we’re all human.