Of Stocks & SCOTUS: The Perils Of Memory

Authored by Nicholas Colas via DataTrekResearch.com,

We have been thinking a lot about human memory over the last few days. Part of that is some lingering thought about the 10-year Lehman bankruptcy anniversary earlier this month. Another is the fact that Monday will be the one-year mark for DataTrek publishing these notes. And, of course, the Supreme Court nomination hearings today put human memory in a starring role.

The science of memory shows this subject is much more complex than most of us realize. One classic example:

  • Daniel Kahneman, who won a Nobel Prize for Prospect Theory, once did an experiment with subjects who were undergoing a colonoscopy.
  • Some patients received the standard exam, which featured an especially painful part near the end of the procedure. This is one reason colonoscopies had a bad rap for so long.
  • Others had their procedure altered so the last part was not so painful.
  • The latter remembered the whole event as being much less painful than those where the last bits were especially difficult. Same procedure, but entirely different memories based on whether the last slice of time was easy or hard.
  • As an aside, we’ve watched Kahneman’s TED talk more times than we can remember. It is worth a look if you haven’t seen it.

The researcher we follow most closely on the topic of memory is American academic Elizabeth Loftus of UC Irvine. One of her specialties is how eyewitness testimony in criminal cases can be incorrect, leading to false convictions. The crossover to the stresses of investing and how those affect our memories is what drew us first to her work. Here is a sample:

  • Loftus’ basic mental framework on how memory works is that it resembles a Wikipedia page more than an immutable record of the past. Over time, we can edit a given memory “page” and (importantly) others can also edit that page as well.
  • Language has a profound impact on memory. Some of her early work involved showing subjects video of simulated car accidents. When she posed the question “How fast were the cars going when they smashed into each other?” subjects reported witnessing much higher speeds than when the query was “…when the cars hit each other”. “Smashed” versus “hit” made a big difference.
  • Stress doesn’t make memories less subject to manipulation. In a study where US military personnel were subject to harsh interrogation techniques for training purposes, researchers could feed these individuals suggestive information that led them to misidentify their interrogators later on.
  • It is possible to implant powerful false memories in adults that go back to early childhood. Loftus’ studies show about 25% of subjects can be convinced of prior traumatic events that never actually occurred.
  • Her TED talk is also worth a look or a read (transcript included in this link).

As far as investing goes, all this is important because memory is the foundation upon which we build expectations. But how we remember events – even ones fraught with emotion and stress – are not always faithful representations of what actually occurred. Moreover, since thousands of market participants collectively decide daily prices, there are just as many memories at work.

Summing up, we take these lessons to heart in a few different ways that might also be useful to you:

  • Since language matters to memory formulation and subsequent recalibration, we are always mindful that exaggerated descriptive words – written or spoken – can alter memory.
  • Stress doesn’t imprint memory any more reliably than when things are calm. Editing down the line can still happen.
  • We try to check our memories, especially when we are “sure” of prior events. It is one reason we always have a data section in these notes. It forces us to go back and reassess what we think we remember.
  • In the end, we think of memory as a mix of fact and all the things that have happened since a given event. Our memories may be mostly accurate, mostly not, or somewhere in between. Fact-checking them when necessary is the only constant.