With the publication of the IMF paper ‘Measurement and Use of Cash by Half the World’s Population’, and the Cash Essentials rebuttal (see News in Brief), a new paper about demand forecasting by the Banca d’Italia (BdI) is timely and adds to the discussion 1.
Its context is that the traditional forecasting models used by the National Central Banks (NCBs) within the Eurosystem have not coped well with the series of economic shocks over the last few years.
The Eurosystem Research Network on Cash (EURECA) was set up in response to changing cash usage in Europe and the need to research and understand better cash related data, whether payment diaries, circulation data or something else. The first project of EURECA looked at better cash demand forecasting. The report, written by BdI, is the joint work of the NCBs of France, Germany, Italy and Spain, who account for 80% of banknotes issued in the Eurosystem.
The benchmark model, as it is known, for the Eurosystem has been forecasts based on ARIMAX models of banknotes issued nationally by denomination. These NCBs have developed Structural Time Series Models (STSMs) as an additional forecasting tool and used it on their own data. They found that, for most denominations, the results were more accurate than the benchmark model, albeit with the caveat that the projection period used was only 12 months long.
Banknote production requirements are calculated using two different approaches so as to gain robust results: (i) bottom-up based on national forecasts provided by the NCBs and (ii) top-down using a centralised euro area forecast made by the European Central Bank (ECB). This approach combines national expertise with a euro-area-wide perspective.
However, the national forecasts of the bottom-up approach are not harmonised: each NCB of the Eurosystem can decide for itself which forecasting models to choose and how to evaluate them.
At the launch of the euro, what was known as the ABCD model was used for forecasting based on the difference between withdrawals from and lodgements to NCBs and the quantity of fit banknote retrievable through sorting activities. The NCBs looked at the co-integration relationships with relevant macroeconomic variables. In the 2009 economic crisis the ABCD model could not fully reflect developments that affected cash demand.
In 2019 the ABCD-2 model was adopted by the ECB. This uses a small basket of models to arrive at a forecast. In France, Germany and Spain the NCBs use ARIMA models. BdI derives its forecast of banknotes in circulation from combining the predictions of withdrawals and lodgements generated by a basket of models which also includes ARIMA models and exponential smoothing.
All of the models fall within traditional time-series econometric techniques such as exponential smoothing, ARIMA, VAR and SUR.
The four NCBs have found that the traditional framework and benchmark models have not coped well with shocks such as the global financial crisis of 2008 or the European sovereign debt crisis 2009-10, regulatory interventions such as the decision to stop issuing the €500 banknote in May 2016 or the COVID pandemic of 2020-21. Within each country there have also been national shocks which, again, the benchmark model has struggled to cope with.
Another driver of the need for better forecasting is the ‘cash paradox’, ie. fewer cash transactions but ever more notes in circulation.
In addition, NCBs face a particular forecasting problem caused by the nature of the euro and the Eurosystem. The euro is slightly different from most other currencies both because of foreign demand for the euro, which leads to the export of euros outside the eurozone, and the flow of notes between countries within the monetary union. For countries with large numbers of tourists, this is particularly relevant. See the box below for examples of the impact of these factors on the four NCBs in this report.
The STSM approach is a classical decomposition of the time series into trends and seasonal, cycle and irregular components and augmented with regression variables. It was used to forecast net issuance for each country and denomination.
Each component is separately modelled by an appropriate dynamic stochastic process which usually depends on normally distributed disturbances. Long-term developments in the economy are characterised by the trend component. Mid-term dynamics can be modelled directly by the cycle component.
All dynamics in the time series data are analysed simultaneously. Missing data and time-varying regression coefficients are easily handled in state-space frameworks.
The state-space form provides the key to the statistical treatment of STSMs. It enables maximum likelihood estimators (MLE) of the unknown parameters in a Gaussian model to be computed via the Kalman filter and the prediction error decomposition. Once estimates of these parameters have been obtained, it provides algorithms for estimating the unobserved components and predicting future observations. An advantage of STSMs and Kalman filtering techniques is that a variety of explanatory variables, dummies, and missing observations can be included in the model without difficulty.
An important benefit of the STSM forecasts is that they originate from realistic model representations of the macroeconomic time series rather than black-box methods.
According to the forecast accuracy measures employed, the STSMs outperform the benchmark models for each denomination in Spain. In France and Italy, STSMs do a better job at forecasting banknotes in circulation for all but the €50 denomination and in Germany for four out of six denominations (€10, €20, €50 and €200).
Although the statistical informative value of this comparison is limited by the 12 month projection period, on balance STSMs seem to be a promising extension to time series models currently employed, at least for France, Germany, Italy and Spain. Inevitably, this report concludes that, in order to assess the robustness of this finding, further exercises of this kind are needed.
There is a degree of heterogeneity between the largest countries. However, while developments of cash in circulation in Italy, France and Germany broadly followed the pattern traced by the euro area, there are deviations caused by local circumstances.
Italy: during the sovereign debt crisis, a legislative intervention fixing the limit for cash payments at €1,000, in combination with new anti-money laundering controls, took circulation into negative growth rates in the period 2012-2016. At the same time, the cash-to-GDP ratio stopped increasing and remained broadly stable between 8% and 9% from 2012 to 2019.
France: cash deposits made at Banque de France counters by banks and other cash handlers decreased by 250 million banknotes per year in the years 2014-2016 and by 500 million banknotes per year in the years 2017-2019, reducing the level of inflows from more than 7 billion banknotes in the early 2010s (peak) to about 4.6 billion in 2019.
The preference of French citizens for cashless means of payment is relatively high and the digitalisation of the economy has continued, accentuated by the COVID-19 crisis. However, the decline in flows of banknotes did not prevent net issuance from growing at an average sustained pace of 7% per year (in value) over the period 2010-2019. In terms of nominal GDP, cumulated net issuance (CNI) increased at a steady pace to around 6% of GDP in 2019 compared with 2% in 2002.
Spain: after the euro changeover, banknotes in circulation accounted for 7% of nominal GDP. CNI reached its peak at the end of 2006 and afterwards showed a continuous descent until the outbreak of the COVID-19 pandemic. It even turned negative at the end of the first quarter of 2020 (0.2% of nominal GDP). At that time, the central bank was issuing fewer banknotes than were being returned.
This can be explained by the dynamics of foreign net cash inflows (eg. cash flowing into Spain from abroad mainly due to tourism) which strongly increased in the last decade, providing a substantial supplementary source of banknotes.
Germany: the development of German-issued euro banknotes in circulation was similar to that of total euro banknotes in circulation, albeit with significantly higher annual growth rates. In 2019, a historical high of 83 million tourists visited the country, tourism reached a weight of 12.4% of GDP and the typical tourism-related branches generated 2.7 million jobs (12.9% of total employment).
In the period from January 2013 to December 2019, the average growth rate of German-issued banknotes in circulation was about 3.2 percentage points above the average annual growth rate of total euro banknotes in circulation. This growth rate differential can be attributed to the traditionally important role of foreign demand for German issued banknotes.
According to an update of the estimate of foreign demand based on net shipments and travel data in Bartzsch et al. (2011), about 60% of German-issued euro banknotes on a value basis were held outside of Germany at the end of 2019 (40% outside the euro area and 20% in the rest of the euro area).
However, while the share of foreign demand in German-issued banknotes had been rising continuously from about 40% at the end of 2004 to 70% at the end of 2013, since then it has been declining. Correspondingly, the share of German domestic demand had been rising steadily from about 30% at the end of 2013 to about 40% at the end of 2019. This increase in the share of domestic demand is due to a significant increase in banknotes held as a domestic store of value. Their share in German-issued banknotes in circulation increased from 17% to 33% in the period under review whereas the corresponding share of domestic transaction balances fell from 11% to about 8%.
