Sunday, August 25, 2019

The Dunning-Kruger Effect



A friend is undergoing social media detoxification and is no longer following up on interesting developments in the personal finance space. When I asked him why, he replied with one of the most hilarious quips I heard last week :

Some of these spaces have discussions that are so pointless, these forums should be re-purposed to become a research platform for social psychologists to study the Dunning-Kruger effect.

So delightfully vicious !

But I need to reflect on this exchange so I thought I should put up some information on this effect because it is vitally important that, in my course, we address this issue to the best of my ability. Students pay good money to attend an investing workshop to pick up Early Retirement skills, it is reasonable to say that some will become over-confident after a few rounds of dividends pay-outs. After that, some of them will confuse their luck with their skill.

The Dunning-Kruger effect affects everyone. As we pick up some investment skills, we experience a stage at first where we, at first, gain a lot of confidence so we start to behave with a lot of hubris and bluster. I was pretty much like that when I started studying finance years ago. After reading a textbook by Aswath Damodaran, I thought I was a whiz kid who can apply discounted cash flow to any asset I encounter and figure out it's intrinsic value. It took a lot of digging, losing money, and reading academic papers to realize that professionals have ditched DCF after under-performing for many years. With social media, we are seeing a lot of such folks out there at the peak of Mt. Stupid.

Perhaps one of the surest signs of meta-ignorance may be an inability to shut the fuck up about Warren Buffett.

The Dunning-Kruger effect diminishes once you study the subject matter deeper because you gain some insight to your ignorance. You need to have knowledge to know that you are wrong. Otherwise, we're wired to think that we are right. My pet theory is that this affects men a lot more because we need to over-estimate our attractiveness so that we can reproduce and we are wired to gloss over our inadequacies. 

Ok, back to addressing the Dunning-Kruger in an investment program.

This is what how my program avoids it :

a) The course is based on experimentation and results.

I never conduct the same course twice. Every lesson is based on a set of experimental results about the markets and which strategies still work as compared to those that don't. At least for the strategies taught, there is no dogma that says that any particular strategy is right. For the past two sessions, REITs have run up so much, I changed my strategy twice in a row.

Another useful aspect of my program is that I like to collect investment mistakes. I record the rationale behind all decisions made by myself and my students so that future batches can see how great decisions end up with huge losses.

I think this gives them a better perspective on hubris.

b) Not even the back-testing approach is spared 

Right now my challenge is to take note of weaknesses in the course meta-strategy. My approach suffers from non-stationarity. Moving forwards in time, there is a risk that return and risk characteristics change materially.

So far this has not affected the money made in the portfolios but I am very curious as to whether how long can a strategy is known to persist over time given that a lot of folks have a Bloomberg terminal.

c) Students are encouraged to start small and 

Students need time to get through the "valley of despair" where they really learn how to invest. The trick is to make them apply a tiny part of their portfolio (around $10,000) and leverage it up to experience the reality and stress of having a leveraged account. I try to get them not to scale up until they experience 3-6 months worth of dividend payouts.

Hopefully they can learn a thing or two about their risk appetite in the process.

I don't think my program can address the entirety of this effect. The investment journey is a long one and you cannot expect a 2 day Masterclass to make you an investing genius and resolve issues of over confidence. In many cases, sometimes course design aims to achieve the opposite, I still need my students to be motivated enough to act.

Achieving this balance is the hard part about investment course design.









2 comments:

Lim JH said...

Interesting, makes me wonder where am I on this dunning kruger scale.

Christopher Ng Wai Chung said...

The beauty is that without a solid measure on where we are on the curve, we will never really know whether we're still on Mount Stupid.

The only defense is to always look for counter-factuals and be honest when an investment does not pan out.