13. December 2023
It is well known that payments in Switzerland have striking weekly seasonality patterns within months. Expenditures increase towards the end of the month with a subsequent decrease at the beginning of the month. Data from Monitoring Consumption Switzerland (MCS) reveal that this pattern is most pronounced for payments with TWINT and can be attributed only partly to the different composition of the expenditure basket across payment channels.
Figure 1 shows the week-on-week changes of expenditures in ecommerce during the first half year in 2023. The figure illustrates that expenditures are much larger in the last week of the month. This pattern is strongest for payments with TWINT if compared with other card payments. The difference between peak and trough is almost 80 percentage points for TWINT payments, followed by 50 percentage points for other debit card payments, and 30 percentage points for other credit card payments.
Figure 1: Expenditure changes in ecommerce
Notes: Week-on-week percentage changes for each calendar day, based on seven-day moving averages. TWINT, Other debit, Other credit.
Figure 2 shows that the different expenditure patterns are much less pronounced for expenditures at the point of sale (PoS). To put this into context, note that the volume of point-of-sale expenditures is on average six times the volume of ecommerce expenditures, ranging between a factor of around three on Sundays to approximately eleven on Saturdays. For TWINT payments instead, the volume of ecommerce transactions is larger than the volume at the point of sale: the volume of ecommerce expenditures with TWINT is on average 1.6 times the volume of expenditures at the point of sale, ranging between factor 4-5 on Sundays to 0.6-1 on Saturdays.
Figure 2: Expenditure changes at the point of sale
Notes: Week-on-week percentage changes for each calendar day, based on seven-day moving averages. TWINT, Other debit, Other credit.
Figure 3 zooms in on the different expenditures made with TWINT, distinguishing whether the TWINT payments are associated with expenditures at the point of sale (PoS) or in ecommerce. Figure 3 shows that TWINT payments in ecommerce feature much stronger weekly patterns than PoS payments and are thus far more volatile. Week-on-week, ecommerce payments increase close to 60% at the end of the month and decrease by 20% at the beginning of a month!
Figure 3: Changes of expenditures with TWINT in ecommerce and at the point of sale
Notes: Week-on-week percentage changes for each calendar day, based on seven-day moving averages.
Which categories of the TWINT expenditures are most associated with the weekly fluctuations? Figure 4 shows that expenditures for Financial Services and Professional Services fluctuate most across weeks. Digging deeper, we find that this is associated with expenditures for insurance or telecommunication services, likely associated with monthly payments scheduled to take place after payday. Regular expenditures for rail travel (SBB) are contained in the category Transport Services. Interestingly, also expenditures for retail goods (labelled as Retail: Other goods) or entertainment and sports (labelled as Entertainment & Sports) change up to 50% across weeks and tend to be highest at the end of the month. The pattern is also present but somewhat less pronounced for expenditures for Food & Beverage Services.
Figure 4: Changes of TWINT expenditures (ecommerce and point of sale) by expenditure category
Notes: Week-on-week percentage changes for each calendar day based on seven-day moving averages.
The within-month expenditure patterns in Switzerland may be associated with a different composition of expenditures made via TWINT or other payment channels. To gauge the importance of composition effects, we scale the composition of expenditures for each payment channel so that it corresponds to the average composition of expenditures across all payment channels. For example, expenditures with TWINT are more concentrated in the categories Entertainment & Sports and Financial Services than the expenditures across all payment channels.
Figure 5 shows that some, but not all, of the higher volatility of expenditures via TWINT can be explained by composition effects. In a counterfactual, in which we make the composition of the expenditure basket for each payment channel equal to the average expenditure basket, the volatility of weekly expenditures remains highest for expenditures with TWINT. Interestingly, compositional effects matter quantitatively for TWINT expenditures but not for expenditures through the other payment channels. These findings relate to a recent literature on payment channels and consumer spending which emphasizes that the ease, with which payments can be triggered, interacts with behavioral spending motives. Through this channel, the ease with which TWINT payments can be executed may thus contribute to the volatility of expenditures.
Figure 5: Standard deviation of week-on-week expenditure changes (ecommerce and point of sale) by payment channel
Notes: Week-on-week percentage changes based on seven-day moving averages. Original: based on observed expenditure composition; Counterfactual: composition of expenditure basket scaled for each payment channel to match the composition of the average expenditure basket across all payment channels. TWINT, Other debit, Other credit.