A.N.T.E – An Intervention to Enhance Private Pension Savings of the Self-Employed

This project addresses the problem of young self employed people, here "Lotte", who often fail to save for their retirement due to biases such as hyperbolic time discounting, present bias and lack of financial literacy. I conducted a behavioural audit of t he young self employed's failure to invest in retirement and propose an intervention to nudge the target group towards such investment, using well established behavioural economics techniques. The intervention aims to change the behaviour of the young self employed by strategically targeting their behavioural biases at each step of the process. The intervention consists of a four step approach that I call A.N.T.E:

1. Allude/Anchor: Lotte receives a gift , creating a sense of indebtedness. This step
   levera ges the principle of social reciprocity to make her more receptive to the
   subsequent stages.
2. Nudge: She receives a letter from the future with a fictitious bank statement,
   emphasising the importance of saving for retirement. This step addresses hyperbo lic
   time discounting by presenting a hypothetical substantial future endowment,
   countering Lotte's tendency towards immediate gratification. The letter also uses loss
   aversion, framing the decision as a way to avoid losing future financial security.
3. Talk through: A follow up letter explains her financial situation, the benefits of early
   retirement planning, and the role of self employed individuals in shaping their pension
   plans. This step empowers Lotte to make informed decisions and serves as a gain
   fram e to set a positive tone for the upcoming intervention.
4. Empower: Lotte meets with a financial consultant who provides personalized
   guidance on retirement savings options. This step capitalizes on social reciprocity
   (step 1), inertia, and indebtedness, as it encoura ges her to follow through with the
   intervention.

The success of the intervention can be assessed by comparing the participation rates of the experimental group (those who received the intervention) and a control group (those who received information materials only). In addition, anonymised financial data could be used to measure the actual increase in participants' retirement savings over a period of 2, 5 and 10 years.

Ignoring Pillar 3 Investment – Blunders and Biases

In terms of our analogy, let’s assume that Lotte is not a coffee afficionado; she has always viewed coffee as a means to an end, she consumes it without giving it much thought. To her, black coffee in the morning is a given and she never cared much for adding cocoa powder to the mix. As trivial as the example may seem, the real-world parallel to this analogy may have dire consequences for Lotte, since having a spartan black morning coffee may make her run on fumes in the evening; or, put differently, it may well preclude her from being able to afford a nice cup of Cappuccino out in town once she retires. Why, like many other of the growing group of self-employed in the Netherlands, does Lotte fail to save for her retirement?

Like many others within her age group, for Lotte, the utility of her income translates more or less directly to its hedonic capacity. She likes to go out frequently with friends, likes to shop for clothing, and enjoys going on city-trips. Afterall, spending on experience is spending on happiness, right? Lotte is acting under a self-control problem that is otherwise known as present bias (Chakraborty, 2019). She values the short-term gratification of consumption over the delayed gratification of long-term saving. Lotte’s utility function with regards to the financial means of which she disposes thus shows a pattern of hyperbolic time discounting (Laibson,1997), such that, when assessing her financial household, Lotte places higher value on imminent hedonic pleasure and the money that is at immediate disposal for purchase, than on the financial security that is immanent in accumulating assets for future use. The flipside of this bias in intertemporal decision making implies that economization and saving become more attractive to her the further they are delayed to the future (Thaler & Benartzi, 2004), and hence Lotte keeps on procrastinating the matter of planning for her retirement. She is naïve in the behavioural economic sense of the word, as she is unaware of the fact that the decisions made by her future self will be affected by the same time discounting tendencies.

Furthermore, besides prioritizing disposability of financial means in the now, her present bias also extends to non-financial costs. In a CentERdata survey amongst a sample of employees, students and job-seeking individuals, a large majority expressed their endorsement of mandatory pension schemes, 60% of which justify this endorsement by pointing to the associated savings in terms of the time and the effort which, in the absence of compulsory schemes, have to be invested into financial education and individual planning (Van Rooij et al., 2007). Thus, a lack of expertise in financial matters along with inertia and the reluctance to allocate time and invest effort into attaining the relevant information and education seem to be further factors in Lotte’s ignorance of pillar 3 investment. This explanatory factor is further backed up by self-report data that demonstrates that the percentage of those feeling inadequately informed about the pension system is higher in the group of self-employed than in the group of traditionally employed, with easier access to information about pension funds within conventional employment infrastructure hypothesized as the main differential factor (Karpowicz, 2019). Another psychological aspect that influences Lotte’s decision making and gets in the way of adding cocoa powder to the coffee is that she is loss averse. Her current net income sustains a lifestyle which she would be reluctant to give up. Her monetary means, and the fact that she pushes the tedious and time-consuming task of commiting her free-time to pension planning to the indeterminate future, endow her with a lifestyle of which she is very fond. Besides her individual living standard, the resulting household expenses, and her basic living cost, she is part of a social system of friends and family, with fixed habits and rituals that are of high importance to her. Thus, there is further endowment nested within her social network and giving up parts of both her temporal and her financial freedom would not only affect her living standard, but could potentially exclude her from social activities to some extent. Stated in terms of prospect theory, the proximate losses in living standard loom much larger than the distal gains in pension returns.

Theoretical Framework - A Basis in Biases

In the following, we will look at the specific design features and formulations in each stage of the intervention steps in order to assess both the mechanisms employed to drive behavioural change, as well as their specific reference to and acting upon the behavioural biases that we have established previously. Why does Lotte receive a gift of luxury cocoa powder in the first step? This gift serves two purposes. First of all, if Lotte has heard of the Cappuccino analogy, the cocoa powder, along with the greeting card, may serve as an initial prompt to get her thinking about her own retirement plans. Secondly, regardless of whether Lotte knows the analogy or has never heard of it, the gift serves as a strategy to install a notion of indebtedness in Lotte: She has received a gift of high quality, maybe she even googles the brand and finds out that it is a luxury product, yet, upon wondering why someone would make her this gift, it is difficult for her to come up with a valid reason. In the absence of a valid explanation, once contacted by the source of the present, she will be more likely to make a concession. The principle of social reciprocity (Cialdini, 2001) may thus prompt her to be more receptive to the advice she will encounter in step 3 of the intervention and may render her more willing to accept the reservation that has been made.

The second step of the intervention serves to mitigate the effects of hyperbolic time-discounting under which Lotte is acting with regard to her present-day spending behaviour. While, as delineated above, her current net income and the goods and experiences she invests in, endow her toward a present bias, designing the letter such as to appear maximally realistic while presenting her with a hypothetical yet substantial endowment in the future may partly reverse this bias. The idea here is thus to use the endowment effect in terms of a future projection. Related studies have shown that letting participants interact with an age-rendered avatar of themselves in a virtual reality environment leads to empathizing with this future self (Hershfield et al., 2011). Participants take the perspective of their future selves and adjust their saving behaviour toward delayed gratification in order to benefit the future self. The idea of the future endowment as suggested here, is that a similar effect may be evoked not by confronting Lotte with an avatar of herself, but by creating a frame for her to imagine herself 40 years later. The idea of herself disposing of 552.000 € is pleasant, she might come up with a lot of purposes on which to spend her money: Travelling, going to Restaurants, meeting her old friend Marteen for a coffee, and taking part in social life in general. She could maintain the lifestyle she is loving today and is endowed with an idea of a prosperous future self.

While the proposed investments that Lotte will encounter in the final step of the intervention are quite low in risk - she will be familiarized with global index funds and ETFs that have a history of stable positive dividends - many naïve subjects still perceive any type of investment into stock as a risky endeavour. Pointing her to the possibility of old-age poverty and a potential decrease in general living standard thus serves to create a loss frame which, as postulated by Tversky and Kahnemann (1984), better prepares decision makers for accepting risk than does a framing in terms of potential gains. While this loss frame has no in situ effect, since Lotte is not making any investment decision within this acute frame, it may well prepare her to better accept risk once the arguments that constitute the frame are repeated by an expert in the following counselling meeting (step 3).

The third letter presents Lotte with what she can actually do to achieve the pension goals as delineated in the first letter. The opportunities that come with her status as a self-employed are now presented in a gain frame, in order to set a positive tone for pitching the ultimate and central element of the intervention. Lotte is now superficially informed about the steps that she may take in order to materialize the scenario as presented in step 2. This is by no means an exhaustive account of her options with regard to her pension planning, yet only serves as a basic guideline on which she may follow up. She is further informed that this follow-up has already been planned for her – reserving a fixed appointment with a counsellor for her relies both on her inertia to cancel the meeting as well as on her indebtedness and resulting concession that have been primed by the earlier donation of a gift. The function of the last step of our intervention is self-contained and does not need further elaboration. And even if Lotte does not turn up, steps one to three will most likely have served the purpose of sensitizing her for the subject matter. From now on, Lotte will never have a coffee without at least briefly thinking about her retirement plans.

To evaluate the success of the intervention in convincing self-employed citizens like Lotte to start investing time and resources into their private retirement plans, several steps should be undertaken. First of all, to determine the size of the sample of self-employed that are subject to the intervention, the tax office provides the necessary basic data. Next, since the main goal of the intervention was to motivate the self-employed to start caring for and privately investing into their pension, an operationalisation of this dependent measure is
needed. To this end, the financial advising company provides data on how many of those who were subjected to the intervention actually turn up to the counselling meeting. During a piloting period, wherein the above-described behavioural insights intervention is compared to a mere informational campaign (control group receiving information material and invitation to meeting), attendance between the experimental group and the control group is compared. Once the piloting phase terminates, the percentage of those turning up to the counselling session still remains a valid measure of the interventions success with regards to elicit motivation. Last but not least, the tax office provides anonymized financial data of the participants 2, 5 and 10 years into the intervention. This data serves to assess in how far the intervention was a success not only in terms of igniting motivation but also in terms of the overarching goal, the actual increase of pillar 3 investment of the self-employed. This analysis should differentiate those who took part in the counselling meeting and those who did not, in order to identify those individuals who took action subsequent to receiving council and those who became active themselves.