This article is written by Flora Finamor Pfeifer, behavioural scientist at São Paulo City Hall

Even in a city of 13 million people like São Paulo, the largest in Brazil, individual behaviour can have a tremendous impact on city life and the population’s well-being.

Immunisation, washing hands and exercising could solve major public health challenges. Paying taxes on time is key to secure public resources. Certain norms, such as the extent to which citizens respect the speed limit, can mean the difference between life and death. Nudging citizens towards choices that are good for them as well as society as a whole can improve government efficiency and legitimacy, assure citizens’ rights, and avoid negative consequences that “bad” behaviours might incur.

I work for the (011).lab, a government innovation lab within São Paulo’s City Hall.

Starting a behavioural team in government comes with  challenges

Inspired by other successful examples from around the world, we started experimenting with behavioural science in 2018 to solve São Paulo’s challenges. We identify behavioural challenges to understand what drives behaviour in different contexts, and prototype evidence-based solutions.

Starting a behavioural team in government comes with its own challenges. We did not have financial or institutional support from private organisations, we lacked methodological expertise, and we had budget and bureaucratic restraints.

Nevertheless, working in government also has its upsides. For example, it enables a higher level of enforcement and of penetration in the public sector. Further, working in the municipal and local government context, in turn, gives us a higher potential of application since we are at the government level directly in touch with citizens.

We’ve learned by doing, experimenting, and failing, and over the 1.5 years, we have conducted projects in health, finance, social assistance and transportation.

A better way to collect taxes

The first policy area we tackled was taxation.

Collecting taxes is essential to public administration. A whole body of evidence pointed to the efficiency of behavioural insights applications in this area. Incorporating seemingly irrelevant factors to the communication with taxpayers, such as making the content simpler, highlighting the deadline and important information, and adding tax morale and deterrent messages, could increase compliance. This was low-hanging fruit and we decided to start our behavioural efforts with a project in partnership with the treasury department.

The property tax is one of São Paulo’s major sources of income. Non-compliance prevents an important share of the revenue (around 12%) from being collected on time — adding up to more than 1.7 billion reais ($328 million USD), according to São Paulo’s treasury department data from 2018’s fiscal year. Late taxpayers receive a letter urging them to pay in 30 days, or they will be registered as defaulters. Of those, around 55% still do not pay. If we could increase this share, the extra money collected could be spent on public goods and services, such as better schools and hospitals.

We decided to redesign this letter to late-payers to increase compliance. We changed the bureaucratic and formal jargons to plain language and structured the content so it would have a highlighted headline with the main message. In addition, all the important information was foregrounded, with a clear call-to-action, deadline, and instructions on how to do it, removing cognitive hassle factors.

Additionally, we tested versions of this simplified letter using behavioural principles and variations in tone. One version — the “social norms” letter — highlighted the fact that the majority of people had already paid their taxes. The “deliberate choice” letter initially showed empathy and understanding to the taxpayer but framed continued non-compliance as an active choice on their part that would be irresponsible. The “consequences” letter provided an overview of every potential consequence of non-complying and framed it as a poor choice for both the city and the recipient. Finally, the “visual illustration” letter had the step-by-step to payment drawn in a flow chart.

Success opens doors

Through a randomised control trial, we measured significant results in two of the letters tested — deliberate choice and consequences — with the best one, the consequences letter, improving compliance rate by 8.4% (from 48.46% to 52.53%).

A great way to learn by doing is to pick a challenge that has already been addressed successfully with behavioural insights in similar contexts to the one you're working in

Once we had confirmation that we had received positive and significant results from the two of the letters tested, we could calculate the potential impact of scaling it. We then set up a presentation with high-level officers from the treasury department and presented the whole project, the solution, the impact measured and the potential increase in revenue.

One of the challenges we faced early on was that the project was delayed due to implementation issues. Since it was our first project, we did not have complete buy-in from the treasury department and we faced several restrictions to access the debt data, which was classified as confidential.

When we were able to show the potential increase in revenue, that caught the attention of the treasury as they could suddenly see the relevance of behavioural insights. Since then, the treasury has eased our access to the data, implemented the letter version we recommended, and opened up the doors to start new projects at the treasury.

Three tips for your first BI project

If you are a public servant and are thinking about starting to apply behavioural insights in public policies, there are three recommendations that I would like to give you:

  1. Start with what has already worked: A great way to familiarise yourself with the method, and learning by doing is to pick a challenge that has already been addressed successfully with BI in similar contexts as the one you’re working in. Redesigning fiscal communications is a great way to start. You can follow the process and solutions published here, and you will probably face similar challenges.
  2. Rigorously test your interventions: Behaviour is not an exact science. Cultural and societal aspects may influence what works and what does not. Thus, it is important to test, and the best way to do so is through an RCT. It is important to have equipped staff to conduct it, because if you fail to design it you can arrive at false conclusions. By testing you also gather hard evidence that it works, which is a great help in getting buy-in from other agencies and governmental partners.
  3. Think about how to communicate: You can have a perfect project, but unless you communicate it in the right way to the right people, it will not be scaled or implemented. Structure your presentation, show your results, and try to explain it without using too much scientific jargon (behavioural sciences, nudge, and impact evaluation are actually really hard to explain!). Try to identify what high-level officials are important to be in this meeting (who is the one who makes the decision) and when will be the most appropriate time to involve this person in the project cycle. — Flora Finamor Pfeifer

(Picture Credit: Unsplash)