Behavioural insight: How can we get the self-employed to submit on time?

Self-employed people who submit documents late and make payments late create difficulties for both themselves and the Tax Administration. Which measures has effect and which does not work in order to get more of them to submit their statements on time?

Since 2012, a project entitled «Punctual submission and payment» has been gathering information about self-employed people who do not submit on time and the reasons why. By testing measures such as sending out letters by post and through Altinn, SMS and telephone, and measuring the effects of the measures, we have learned more about what will get the self-employed to take action. The project is responsible for the most comprehensive cause-and-effect measurements in the history of The Norwegian Tax Administration. They trialled 34 different measures on a sample of 61,000 self-employed taxpayers, and 8,700 people who were VAT-registered.

By Marianne Løsset and Marianne Tenden.
This is an excerpt from an article in Analysenytt (in Norwegian only).

årstermin

We have measured the effects of 34 measures
We have learned a lot about why self-employed people do not submit. By testing and measuring the effects of 34 different measures to get more people to submit punctually, we have also found out what can get the self-employed to take action. Following the trials, we called randomly selected self-employed people and asked them what they thought about the measures we had implemented and whether they would have liked us to have done anything differently. We have for example trialled reminders sent by letter, letters with sticky notelets, letters sent via Altinn, SMS and by telephone.

How did we measure the effects?
We measured the effects by splitting the target groups into two randomly selected groups, where one group receives a particular measure and the other group does not. We then compared the number of people who submitted by the relevant deadline in the two groups. The effect consists of the difference between the group which received the measure and the group which did not. As the groups were selected randomly and can otherwise be said to be identical, we can say that the differences between the groups were due to the measure.

Targeting the recipient directly has an effect
In collaboration with the NHH business school, the Tax Administration has conducted a field experiment which showed that influencing taxpayers by sending them a letter has an effect (Bott et al, 2014).  The eight different letters that were trialled all had an effect on the reporting of capital and incomes abroad. Personal information letters, letters which appealed to the recipient’s morals, and letters which stated that «The Tax Administration knows you have capital or income abroad» had more effect than a general information letter.
The findings are in line with surveys conducted in the UK, which showed that letters with a simple message and a statement that the authority is aware of the recipient’s circumstances, had considerably more effect than a general letter (The Behavioural Insights Team, 2014).
The results of the measures we trialled in this project are in line with those of these surveys. However, the measures we trialled show that not everything works and that the key is to ‘sharpen’ the message so that the recipients feel that it concerns them.  The effects of the measures we highlight below varied between 3 and 20 percent.

Letters before a deadline should be sent to selected target groups
Sending out a reminder before a deadline may not be relevant to all recipients since most self-employed people will submit before the deadline on their own initiative. It is therefore important to identify target groups where there is a certain level of risk that the recipient will not submit by the relevant deadline if they do not receive such a reminder.
The project has, for instance, tested reminder letters sent to those who were late submitting their tax return for self-employed persons the previous year. The letters were sent out ahead of the deadline and stated that the recipients received the letter because they submitted late last year. A similar letter has also been trialled for VAT-registered businesses which submitted their annual statement late the previous year. These letters all had an effect on submission.
We also trialled reminder letters sent out ahead of the deadline for both tax and VAT for new businesses. These letters also had an effect. The letters stated: «According to the Register of Legal Entities, your company was founded in 2013. We wish to remind you that…. » and «you registered in the value added tax register during 2013.»

Letters sent out after a deadline are naturally worded more strictly
Sending out a letter after a deadline enables us to directly target recipients who have not submitted a statement by using the wording «you do not appear to have submitted…».
We have sent out such reminder letters in both English and Polish to foreign self-employed people. The letters were sent out to self-employed people who had not submitted their tax return by the deadline. Those who received the letter in Polish were given more guidance concerning logging in to Altinn, as this information was not readily available in English or Norwegian. The results from the letters showed that the reminder letter in English had much less effect than the letters in Norwegian and Polish. The difference may partly be explained by the possibility that we may not have targeted the letter well enough (Nikolaisen et al, 2014).

Handwritten and personal communication
Based on experiences gained by the tax authorities in Ireland and others, we have also trialled a greeting on yellow sticky notelets, together with a reminder on the tax return for self-employed persons.  The note was handwritten and read: «Dear Sir/Madam, I assume you will be able to submit. Please do not hesitate to contact me if you have any questions. Regards, Lars». The letter itself was a reminder letter aimed at sole proprietorships which had not submitted their tax return shortly after the deadline, and was sent out both with and without a sticky notelet. Both letters had an effect, but the letter without the yellow sticky notelet had more effect on submissions than the letter which had the notelet.

madam

The wording of the letter is important
Our surveys show that a well-worded letter is simple, brief and personal. This is in line with the results of the abovementioned surveys in the UK, which highlighted the effect of using simple language, only including what is necessary and stating the most important points first.  We tested sending out reminders after the deadline for the tax return for self-employed persons in which we did not threaten sanctions, but used a friendly tone in the letter and the heading «Have you forgotten us?». This letter had a good effect. We also received positive feedback from the self-employed people who have been in contact with us after receiving this friendly letter: «I have never in my whole life experienced such a friendly letter as I found in my mailbox from you yesterday». That is the reason why I dared to call you now» (text message from a self-employed person who tried to call after receiving a letter).

Still more to do
It is important to emphasise that the effect measurements indicate the results of each individual measure and the situation there and then. The project has not measured long-term effects. It is possible that some of the measures should be measured repeatedly over time, so that we can be more certain about what has an effect with regard to getting statements submitted on time.
Based on the testing we have conducted concerning the reasons for late submission and payment, we can see that there are more possibilities to target specific groups than we have been able to trial during our project period. We can see possible benefits from sharper measures aimed at defined target groups, such as linked to the start-up and cessation of trading by self-employed people. By using information concerning the status of enterprises (newly registered, no turnover during a certain period of time or notified as having ceased trading), we can target the information measures aimed at these groups more accurately. An example is a letter in which we say: «We can see that you have recently registered as being self-employed. We would like to point out that you must submit the tax return for self-employed persons even if you do not have any turnover».

A knowledge-based approach is effective
By testing and measuring effects, we have learned that no two self-employed people are the same. We have seen that a knowledge-based approach provides us with a good basis for identifying measures which have an effect on the punctual submission of obligatory statements.  We can also see that this approach is a step in the direction of knowing, rather than constructing our tools around assumptions. We believe that there are no short-cuts when working with a knowledge-based approach and targeting the measures at specific target groups; we must to go out and ask, test, learn, evaluate, correct mistakes and start again. Our experience shows that it is important to familiarise yourself with the target group concerned, preferably through direct contact, before planning and implementing measures.

Can the Tax Administration provide better information?
A high proportion of self-employed people cite a lack of knowledge as the reason why they have not submitted their obligatory statements. There may be a number of reasons for this.  One of the reasons may be that the information that the Tax Administration provides is not adequate, clear or sufficiently accessible. If so, we should look at what we as a tax authority can do to provide better information concerning the obligations of self-employed people to submit tax returns and statements.