Cost Benefit Analysis of Restricting Cash Usage

A paper by the Einaudi Institute for Economics and Finance (EIEF), ‘Cash: A Blessing or a Curse?’ ¹, examines what happened in Mexico when changes to cash availability were made. What happened to crime? What were the costs and benefits of the changes? Although there are a number of caveats about the findings, the main conclusion is that the benefits of restricting cash are less than the costs.

The academics focused on Mexico because it has good household and firm level data on the access to and use of cash. It made two policy changes that, effectively, restricted the use of cash to some groups in society. Furthermore, there is good data available estimating the cash-credit elasticity of substituting cash, which is key if you want to analyse the consequences of eliminating cash for cash-only households, a big group in Mexican society.

Household and firm level surveys show that over 90% of transactions are paid in cash. Over 50% of households are ‘mixed’, they have access to making card payments as well as using cash. In these households, 80% of payments are still made in cash. If cash and cards are not perfect substitutes for each other, then changes reducing cash usage affect this group as well.

Benefits paid by debit card

Between 2009 and 2015, a policy change was made to how payments were made in the Prospera programme, a government conditional cash transfer programme. 1.3 million people received debit cards with the aim of increasing financial inclusion and reducing the use of cash. The effect was to increase the number of ATM transactions but also to significantly increase the number of point-of-sale terminals.

The policy was implemented across randomly selected localities, which allowed event-study design to be possible. The result of the policy change was a small, but significant reduction in thefts and robberies. However, there was no reduction in homicides, the number of people working in the informal economy or any change to local tax revenues.

ATM sharing agreements

In October 2014 the regulatory requirements to implement ATM sharing agreements between banks were reduced. As a result, from 2014-2019 there was a gradual adoption of sharing agreements between banks. Where these took place, ATM fees fell. The authors were able to map the location of shared branches by municipality compared with branches where ATM services were not shared, and so could compare the impact between regions with different levels of shared branches.

Again, there was a reduction in thefts and robberies, particularly where pedestrians were victims, but no impact on homicides or the number of people working in the informal economy.

Impact of restricting cash usage

Using the observed patterns of cash usage and the estimated elasticity of crimes to cash, it was possible to estimate the effects of restricting cash usage to households right up to a complete ban on cash.

The modelling assumed the price of goods was the same whether paid for with cash or credit cards if there was a complete ban on using cash, limits on the value of cash payments or taxes on goods paid for in cash. It looked at different income groups and different categories of purchase using the national survey of household income and expenditure.

Based on the findings of the two policy changes, the focus was on thefts and robberies. Victimisation studies were used to measure the prevalence of crime and the direct costs relative to GDP. Indirect costs, for example preventative policy costs and judiciary costs, and intangible costs, such as psychological damage, were also included.

The result of this work was that the private costs of heavily taxing the use of cash were greater than the social benefits identified. The social benefits of eradicating thefts and robberies was found to be 1.3% of GDP. The cost, the private losses, of a 40% tax on cash was about 6% of cash or higher, while a total ban was estimated at 10% of GDP.

Caveats

Clearly this work has involved making a number of assumptions – for example, the estimates for the private costs of restricting the use of cash heavily rely on the available estimates of the elasticity of substitution between cash and cards from experimental and observational data for Uber services in Mexico. More work is needed to compute an elasticity applicable to the entire economy.

The work did not consider the effect of the policy on tax evasion since it did not find significant effects of either policy change on the number of people working informally. When India went through its 2016 demonetisation exercise, analysis found no major impact on tax compliance. It is possible, of course, that there is an impact of cash on tax evasion and other crimes, especially if the size of the intervention is larger. The authors consider whether a full ban on cash, for instance, could have an impact on crimes such as extortion.

Finally, the work does not consider the general equilibrium effects of the policy. One is the restriction that the storability of cash has on the level of nominal interest rates, ie. the zero lower bound. Another is the supply side response of alternative payment methods. These calculations could change the welfare effects of the policy estimated in this paper and are identified by the authors as an important area for future research.

1 - EIEF Working Paper 21/10: Fernando Alvarez (University of Chicago), David Argente (Pennsylvania State University), Rafael Jimenez (University of Chicago), Francesco Lippi (LUISS University and EIEF)