Tax compliance remains a thorny issue for governments worldwide—especially in developing countries where informal economic activity, mistrust in government, and limited enforcement capacity create persistent gaps between potential and actual tax revenues. In this post, Ivo Vlaev asks: can a few carefully worded text messages move the needle on tax compliance in Pakistan?
The Setting: Service Providers in Khyber Pakhtunkhwa
Pakistan’s tax system, like many in low- and middle-income countries, struggles with low compliance. The province of Khyber Pakhtunkhwa (KP) introduced a digitised tax registration system in 2020 to bring more small service providers (e.g. consultants, tailors, electricians) into the fold. But getting them to file returns regularly remained a challenge.
The researchers partnered with the KP Revenue Authority (KPRA) to conduct a randomized controlled trial involving over 18,000 service providers who were registered but not yet consistent filers. This offered a unique opportunity to test the potential for small, inexpensive behavioural science interventions – nudges – to help people do what they already know they should. Text messages are one such nudge – with the advantage of being cheap, scalable, and have measurable outcomes. In a new study, Abdelnabi, Afraz, Ahmed, Ayub, Makki, Said, Schiektekat and Vlaev (2024) investigate how different combinations of text messages might affect filing behaviour.
The Interventions: Nudges with a Twist
The team didn’t just send reminders. They tested a matrix of behavioural strategies, drawing from well-established behavioural principles like loss aversion, social norms, and active choice. Two waves of messages were sent—one before the deadline (reminder), and another after the deadline (warning) if filing hadn’t occurred, along these lines:.
Reminder messages (before deadline):
- Basic Reminder – Plain and factual: “Reminder: Tax return is due.”
- Reciprocity Message – Social value framing: “Taxes help improve public services.”
- Loss Aversion Message – Loss-focused: “Failure to file may result in account deactivation.”
Warning messages (after deadline):
- Basic Warning – General warning.
- Financial Penalty – Specific mention of fines.
- Active Choice – Framing in terms of personal responsibility: “Your non-compliance is a choice.”
This experimental design allowed the researchers to see which combinations worked best, and for whom.
Key Findings: Some Nudges Work, Some Don’t
The study found that nudges work, but only for some people. The combination of Loss Aversion + Active Choice performed best among longer-registered taxpayers, increasing filing rates by 1.4 percentage points.
But among recent registrants, most messages had no effect, or even a slight negative one. This suggests that newer taxpayers may be less responsive to threat-based framing or more confused about their obligations.
Overall, the average effects were modest – just a few percentage points – but significant given the scale and cost (they were, after all, mere text messages!).
Behavioural Takeaways: What This Tells Us
1. Losses Loom Larger Than Gains
Loss aversion again proves powerful. People are more motivated by what they might lose (financial penalties, account deactivation) than by the benefits they might gain.
2. Active Framing Works
Framing non-compliance as an active decision seemed to shift responsibility onto the taxpayer’s shoulders, which may have helped nudge them out of inertia. The phrase “you are choosing not to file” appears to cut through more effectively than passive reminders.
3. One Size Does Not Fit All
The same message has different effects depending on user type. Newer registrants may feel overwhelmed or unsure; aggressive messaging might backfire. For experienced taxpayers, however, tough love may resonate more.
Let’s consider an example to make this more tangible. Suppose you’re a tailor in Peshawar who recently registered for tax online. You receive a text message saying:
“File your tax return by the 15th to avoid penalties and account suspension.”
If you’re new to the system, this might cause anxiety or confusion (“What penalties? What form? Where do I log in?”). But if you’ve filed before, the nudge may prompt action (“Oh yes, I had better do this before I lose access”).
Now imagine a message that reads:
“You are choosing not to file your tax return. Continued non-compliance will lead to penalties.”
This message reframes inaction as a decision. For someone used to ignoring vague reminders, this direct framing might break the pattern. And again, it worked – but only in specific contexts.
4. Reminders Are Not Magic
Text nudges alone can’t overcome all barriers. Filing taxes still requires navigating digital systems, understanding rules, and having the capacity to pay. Behavioural interventions must be part of a larger policy toolkit to increase compliance.
Broader Implications
Pakistan is not alone. Countries like India, Kenya, and Peru have experimented with similar low-cost interventions to improve compliance. In the UK, behavioural insights have been used to encourage timely tax payments by adding lines like: “Most people in your town have already paid their taxes.”
This social norm nudge has had a measurable impact in trials conducted by the UK’s Behavioural Insights Team (BIT). But Pakistan’s context is different: low trust, low literacy in tax systems, and often limited digital familiarity. That’s why this study matters – it tests what works in context, with real users and real consequences.
This study opens several avenues for future research and policy development:
- Adaptive Messaging: Could we use machine learning to personalise nudges over time?
- Multimodal Nudges: What if we paired texts with automated calls or visual explainers?
- Trust-Building Nudges: Would messages about how tax revenue is spent build legitimacy and improve compliance?
- Simplification: How much of the non-compliance is due to complexity? Should we test simplified forms and pre-filling?
And importantly, could such messages be used outside of taxation – in encouraging vaccination, registering for health insurance, or renewing vehicle licenses?
Final Thoughts
The results of this study are modest but powerful in implication. It shows that nudges—especially well-framed ones—can make a dent in complex problems like tax compliance. But it also cautions us: not all nudges work equally well for everyone.
Behavioural science is not a silver bullet. But in a world where attention is scarce and administrative capacity is limited, a well-timed text message might just be enough to turn intention into action.
If policymakers want to harness the power of behavioural science, they need to move beyond “one-size-fits-all” and embrace the nuance. Just like taxes, good policy is all about the details.
Ivo Vlaev is Professor of Behavioural Science at the University of Warwick.