Default Nudges: Fake Behavior Change

Default Nudges: Fake Behavior Change

After my last two emails, I received a large number of messages expressing shock over the state of applied behavioral economics. However, a small group pushed back, expressing that they thought nudges, default nudges in particular, are probably still effective.

The apparent success of defaulting on 401(k) enrollment or organ donation was brought up in basically all these emails.

I’d like to share a response I gave to the apparent success of these examples of defaulting. 401(k) and organ donation defaulting are examples of what I call “fake behavior change”. That is, behavior change in name only. I describe what this means in my response below:

Impact and behavior-change are two different things. I understand that someone donating an organ is a beneficial, impactful thing. However, what I’ve been interested in academically and throughout my career has been figuring out how to get people to behave differently. I’m interested in getting people who are going to purchase one thing to instead purchase another, or for someone who doesn’t run to start running each week. In my mind, a behavior is a mental or physical action that someone performs. I do not think that 401(k) and organ donation defaulting can be considered, in any way shape or form, “behavior” change. You are not getting a person to *do* anything. In most defaulting situations, you’re merely doing stuff *to* people without their knowledge or conscious consent. Let’s talk more about the 401(k) example you gave: the big breakthrough with 401(k) participation came when employers started to automatically opt all new hires into 401(k) plans. As this review states: “Under automatic enrollment (also called negative election), employees are automatically enrolled in their company’s 401(k) plan unless the employees elect to opt out of the plan. This contrasts with the usual arrangement in which employees must actively choose to participate in their employer’s 401(k).”

Yes, this increased the % of employees participating in 401(k) plans. But how is that in any way “behavior” change? It’s legal trickery happening behind the scenes and without any actual awareness, decision-making, or effort on the part of the participant. If tomorrow I was legally allowed to automatically take 1% out of my employees’ payroll accounts and donate it to a nonprofit of my choice, I’d be increasing the nonprofit donation rate — but calling this “behavior change” is a stretch. Stating that I got my employees to behave differently would be a deceptive way of describing that course of events. I think that when most people talk about changing behavior they’re not talking about bureaucratic meddling happening behind the scenes.

This is what you see with all the defaulting interventions (including organ donation). They’re not “behavior change” in any commonly understood sense of the term.

Re: the effect size of these sorts of interventions, they’re covered in the paper I linked to in my last email (defaults are structure interventions). The bias-adjusted results from PNAS are here:

A Cohen’s d of 0.2 is considered a small effect—so .12 is quite small. So even if we did consider defaulting legitimate behavior change, the effect sizes after being corrected for publication bias are nothing to write home about. And, I don’t think there’s a free lunch here—I think the negative ripple effects on other relevant behaviors would be non-zero. That’s been my experience with these sorts of interventions in general.”

My perspective is that instead of doing these sorts of ineffective nudges, it’s important to instead spend a lot of time understanding your customer and the problem they’re trying to solve—and then doing everything possible to construct the best experience possible. Nudges distract from these sorts of hard, heads-down projects, since they make the people involved think that often fundamental issues can be solved with a little nudge or tweak here and there. Not my experience. And the literature is bearing this out. The evidence that these sorts of interventions aren’t worthwhile has been collecting for years.

To summarize: 

  • Defaults don’t cause actual behavior change.

  • Defaults are a re-branding of legal coercion and trickery as something willful (“behavior change”)

  • Even if defaults were real “behavior change”, the effect sizes don’t appear to be that large in the real world.**

* No behavior occurs in isolation. For example, if you run a sale and get someone to spend $500 this week, that doesn’t mean they’ll continue to purchase at the same rate they did previously. In fact, all my experience predicts they’ll purchase a lot less. The sale merely pulled their purchases forward. It may result in an overall slightly increased annual spend, but that’s dependent on your customer base etc. If you only measured short term spend/conversion, your sale initiative might have looked like a smashing success. But if you track and look at the long-term spend/behavior repertoire, you’ll find that it’ll look more like a wash. Yes, you increased behavior A, but you might thus get a lot less of behaviors B or C. Nothing is all that simple. 

** I’ll write more about the organ donation default sham in another email. But it’s worth pointing out that defaulting people into organ donation has a questionable track record: ”Our study findings suggest no significant gain for established opt-in countries considering a switch to optout. Although historically some countries have observed impressive increases after introduction of presumed consent, such as Belgium, others have fared badly with either no difference or an actual drop in organ donation rates, including Singapore, Brazil, Chile, Sweden, and more recently Wales.”