The substitution of cash by cards as a means of payment during the pandemic

The pandemic has led to cash being replaced by card payments, as shown by an analysis using anonymised internal CaixaBank data. This substitution effect is seen both at the aggregate level and at the sector level, particularly in purchases of food and durable goods.

Content available in
Tarjeta de crédito y mascarilla quirúrgica

The outbreak of the pandemic has brought about changes in consumption habits. One of them has been the boost which e-commerce has experienced, especially in the durable goods segment, which has resulted in online shopping closing the gap with face-to-face purchases.1 However, shopping in physical retail establishments is still the most widely used option among most consumers, and this is what we will focus on in this article.

  • 1. See the Focus «The awakening of e-commerce in the retail sector» in the MR12/2020.

In physical shops, the pandemic has also brought about changes. We will explore them using fully anonymised internal CaixaBank payment data. More specifically, our data allow us to analyse the extent to which cash has been replaced by cards as a payment method (in the rest of the article we will refer to this change of habits as the «substitution effect»).2

In the first chart, we show how face-to-face card expenditure registered on CaixaBank Point of Sale (POS) terminals has evolved relative to cash withdrawals carried out at CaixaBank ATMs (withdrawals serve as an approximation of the use of cash to make payments). As can be seen, the two series follow a similar pattern in the pre-pandemic period and during the months of widespread lockdown in Q2 2020. However, since June 2020, when many of the restrictions were lifted, a gap has opened up between the evolution of face-to-face card expenditure and that of cash withdrawals. This gap was particularly pronounced between June and October 2020 and has persisted since November 2020, albeit with a certain declining trend.3

  • 2. The ECB recently published an article in which they also document this substitution effect. See A. Zamora-Pérez (2021). «The paradox of banknotes: understanding the demand for cash beyond transactional use». ECB Economic Bulletin, Issue 2/2021.
  • 3. The gap in the average year-on-year growth between the two series was 21 pps between June and October 2020, and 13 pps from November 2020 to February 2021.
Spain: face-to-face card expenditure registe red on CaixaBank POS terminals versus cash withdrawa ls at CaixaBank ATMs

After following a similar pre-pandemic trend, the opening up of the gap between the two series is indicative of the substitution effect.4 In addition, the substitution effect could be even stronger than that seen in the chart: our data are based on cash withdrawals, but there is evidence that during the pandemic households have made greater use of cash as a store of value, rather than as a means of payment (as demonstrated in the article cited in footnote 2).

Beyond the aggregate analysis, the shift in consumer behaviour can also be sensed in specific sectors. Unfortunately, due to the anonymous nature of cash payments we have no way of knowing how much cash Spaniards spend in each sector. However, it is possible to address this issue by comparing the evolution of face-to-face card expenditure according to our internal data with the evolution of turnover in department stores offered by the National Statistics Institute at the sector level.5 Given that turnover is largely made up of payments made by cash and card, comparing the two series before and during the pandemic can capture the substitution effect at a more disaggregated level (see second and third charts).6

  • 4. To check that our internal data are representative, we have confirmed that the proportion of consumption observed in CaixaBank’s Consumption Tracker, both in aggregate and by different age groups, is very similar to that identified in the National Statistics Institute’s household budget survey in the categories of goods that are comparable between the two databases (food and beverages, clothing and footwear, and restaurants and hotels).
  • 5. To carry out this comparison, we focus on the food and durable goods sectors (electronics, textiles, furniture, etc.). According to data from CaixaBank’s Consumption Tracker, these two sectors accounted for 50.4% of the total face-to-face card expenditure in 2019.
  • 6. More specifically, turnover consists of income received in cash, by card payments, and by banks transfers and direct debits. In the sectors explored, direct debits should account for only a minor portion of total turnover, so they would not significantly alter the conclusions of our analysis.
Spain: face-to-face card expenditure registered on CaixaBank POS terminals versus NS I turnover data (food) and Spain: face-to-face card expenditure registe red on CaixaBank POS terminals versus NS I turnover data (durable goods )

As can be seen, the gap is positive and significant for expenditure on food and durable goods when comparing pre-pandemic behaviour with data for the period from June to October 2020 (when there were fewer restrictions on activity). However, when compared to the period between June 2020 and February 2021, the difference is significant only in the food category.7 This result is consistent with the information from the first chart which, as we have seen, shows that the gap between face-to-face card expenditure and cash withdrawals has gradually moderated since November 2020.

In short, our analysis indicates a change of habits among consumers as a result of the pandemic, with cards taking on a more prominent role as a payment method to the detriment of cash. The million-dollar question is to what extent this substitution effect will be structural, something for which we will need to see how the data evolve over time once normality (or the «new normal») is established.

  • 7. To determine whether the difference between the series based on our internal data and that of the National Statistics Institute (NSI) is significant, we used the difference in differences (DD) statistic: the difference between the gap in year-on-year growth per the internal data and per those of the NSI between two different periods (pre-pandemic and during the pandemic).