Forecasting Finnish Payment Preferences

The core finding of this section is the significant shift away from cash and towards electronic payment methods in Finland between 1984 and 2002. In 1984, a substantial 80% of purchases (by value) were made using cash. However, by 2002, this figure had plummeted to below 50%. This dramatic decrease in cash transactions is directly contrasted by a sharp rise in the use of cards, particularly debit cards, demonstrating a clear trend towards the electronification of the Finnish payment system. The study highlights the rapid adoption of card payments, with their total value increasing fifteenfold between 1984 and 2002. This period also witnessed a marked decrease in the use of cheques, which are now rarely used for purchases. The overall data unequivocally reveals a major transformation in how Finns conduct everyday transactions, favoring electronic means over traditional cash-based payments. The data underscores a compelling transition from cash-dominant transactions to an increasingly electronic payment landscape.

The analysis relies on data covering the period from 1984 to 2002. While data on card payments were readily available, the analysis of cash payments presented unique challenges. Due to the untraceable nature of cash transactions, the value of cash payments at the point of sale (POS) was estimated residually by subtracting the value of non-cash purchases from the total value of POS purchases. Data on other payment methods, such as lunch coupons (from Luottokunta) and e-money (from the European Central Bank's Blue Book), were also incorporated to create a comprehensive picture of the Finnish payments landscape. Data limitations were acknowledged, primarily in the absence of information on newer payment methods like mobile phone and internet-based systems; however, the authors justify this omission by noting the relatively minimal usage of these options during the study period. The data aggregation and methodology highlight the importance of utilizing a combination of datasets to compensate for data scarcity related to cash payments and newer digital payment options to understand the larger trends.

Figure 2 and Figure 3 (referenced in the document, but not included in this response) visually illustrate the changes in Finnish payment methods over time. Figure 2 shows the absolute values of different payment methods over the study period, clearly highlighting the significant growth in card payments and the continued dominance of cash, albeit declining. Figure 3, a percentage breakdown, demonstrates the relative decrease in cash payments as a proportion of total POS payments. This visual representation reinforces the quantitative data, showcasing the increasing prevalence of card payments alongside the sustained – yet diminishing – importance of cash throughout the 19-year period. The visual aids provide compelling evidence of the gradual but consistent transition from a cash-based to a card-based system, further illustrating the shift towards the electronification of the Finnish payment system. The visual representations directly show the transition of Finnish payment preferences.

This section investigates the applicability of learning curve models to forecast the decline of cash payments in Finland. The study utilizes a 19-year dataset (1984-2002), extending the timeframe of previous research by Snellman and Vesala (1999), who used data from 1988-1996. Unlike the earlier study, which found an S-shaped curve suitable for predicting cash displacement, the current analysis reveals that this model does not adequately fit the extended dataset. The learning curve model, in this instance, predicts the eventual complete disappearance of cash. However, this prediction lacks statistical significance due to the insignificance of key coefficients (alpha, beta, kappa) and the presence of high autocorrelation in the residuals. Despite a surprisingly high R-squared value, the model's limitations suggest that alternative approaches are needed for a more robust analysis of the electronification of payment methods in Finland, highlighting the inadequacy of a previously successful model in explaining the longer trends observed in the Finnish payment system.

The analysis reveals significant limitations in using the learning curve model for predicting the long-term trend of cash displacement in Finland. The model, while successful in previous shorter-term studies, fails to capture the nuances of the 19-year data set. The lack of statistical significance in the estimated parameters and high levels of autocorrelation indicate substantial issues with the model's fit. The apparent reason for the model's failure is the absence of a visible S-shape in the longer time series, a characteristic feature assumed in the learning curve approach. The researchers suggest that the previously observed S-shape might have been partly an artifact of a recession in the early 1990s which temporarily halted the electronification trend. This lack of fit necessitates the exploration of alternative models, like error correction models, which can better account for the complex factors influencing payment method preferences and the transition towards digital payment systems in Finland, particularly considering the longer historical context.

This section introduces error correction and partial adjustment models as superior alternatives to learning curve models for analyzing the electronification of payment methods in Finland. Recognizing that the shift towards electronic payments is a gradual and long-lasting process, the researchers argue that these dynamic regression models are better suited than the previously tested learning curve models. The models incorporate independent variables to explain the changes in the cash share of POS payments, accounting for factors influencing consumer payment preferences and the broader payment infrastructure. This approach provides a more comprehensive understanding of the underlying dynamics driving the decline of cash payments and the rise of electronic payments in the Finnish context. The choice of these models acknowledges the limitations of the previous approach and proposes a refined methodology to capture the complexities of the observed trends.

The core of this subsection focuses on the selection of variables used in both the error correction and partial adjustment models. These models include the number of EFTPOS terminals and cash dispensers, as these directly impact the ease and convenience of electronic versus cash transactions. The nominal interest rate is also included, reflecting the opportunity cost of holding cash. Finally, the unemployment rate is used as a proxy for overall economic conditions, considering its potential effect on consumer spending habits. The researchers acknowledge the influence of payment method pricing but exclude this variable due to data unavailability. They justify this decision by indicating that previous research had established the significant relationship between EFTPOS terminals, cash dispensers and overall cash demand; these relationships were the primary basis for the selection of variables for the models used in this study. The careful selection of these variables is vital for explaining the observed trends in the data.

The results from both the error correction and partial adjustment models show consistent findings: increased numbers of EFTPOS terminals and higher interest rates lead to a decrease in cash usage at POS. Conversely, higher unemployment rates and increased numbers of cash dispensers correlate with increased cash usage. The effect of cash dispensers is noteworthy, as it contradicts some prior research which indicated that fewer dispensers lead to higher cash demand. The study attributes this difference to the focus on cash usage at POS rather than overall cash demand, suggesting the overall decline in cash dispensers since the early 1990s has decreased cash usage at POS while possibly increasing overall cash demand. The study highlights the importance of the specific focus of the dependent variable. The results of the models indicate a continued decline in cash usage, and the projections indicate that even by 2010, the saturation level of cash usage at POS might not yet have been reached. The discussion also touches on the challenges in interpreting the adjustment parameters, particularly when they fall outside the standard range (0-1).

This subsection focuses on using the error correction and partial adjustment models to forecast future cash usage. The forecasts are built upon projections of independent variables such as the number of EFTPOS terminals, cash dispensers, interest rates, and unemployment rates, drawing upon forecasts from the Bank of Finland (2004) and the Ministry of Finance (2003). The models predict a continued decline in cash use, with the cash share of POS payments projected to fall to only 25-30% by 2010. These forecasts assume that current trends will continue. However, the researchers acknowledge that unexpected changes in these variables could alter the accuracy of the forecasts. The findings highlight the potential for accelerating the electronification of POS payments by increasing EFTPOS terminals and reducing cash dispensers. The researchers note that the causality between cash dispensers, EFTPOS terminals and payment habits is not entirely clear, suggesting that future research needs to investigate these relationships more closely. This section emphasizes the importance of considering the underlying economic factors impacting payment preference alongside the technological infrastructure development.

This section presents forecasts generated by the error correction and partial adjustment models, projecting the future of cash usage in Finland. Both models, while differing slightly in their specifics, converge on a similar prediction: a continued decline in the cash share of point-of-sale (POS) payments. The forecasts indicate that by 2010, cash will account for only 25-30% of the total value of POS purchases. These predictions rely on several assumptions regarding future trends in key variables. These assumptions include a continued increase in the number of EFTPOS terminals, a decrease in the number of cash dispensers, and specific projected values for interest rates and unemployment rates, based on forecasts from the Bank of Finland (2004) and the Ministry of Finance (2003). The models suggest that accelerating the transition to electronic payments could be achieved by strategically increasing the number of EFTPOS terminals and simultaneously reducing the number of cash dispensers. The forecasts emphasize that unexpected changes in these underlying factors could significantly affect the accuracy of the projections; these projections provide a significant insight into the future of the Finnish payment systems.

The study compares the results and forecasts from the error correction and partial adjustment models. The forecasts from both models show remarkable similarity, reinforcing the prediction of a continued decline in the cash share of POS payments. Despite the similarities, the study notes that the error correction model, with its many parameters, requires cautious interpretation, while the partial adjustment model, although having fewer parameters, still has a high parameter-to-observation ratio. The limited number of observations in the dataset is identified as a key limitation to the analysis, necessitating prudence in interpreting the statistical significance of the findings and the precise accuracy of the forecasts. The study acknowledges the limitations of both models due to the relatively short time series data used in the analysis; the models, nevertheless, provide considerable insights into the changing payment preferences and the electronification of the Finnish payment system.

The forecasts presented suggest that the electronification of POS payments in Finland is set to continue at a substantial pace. The models provide evidence that increasing the density of EFTPOS terminals while simultaneously decreasing the availability of cash dispensers would significantly accelerate this shift. However, the study also notes complexities in establishing clear causality between the number of EFTPOS terminals and cash dispensers and changing payment habits. It is possible that changes in payment habits could be impacting the development and availability of EFTPOS terminals and cash dispensers rather than the other way around. Therefore, the study suggests that future research needs to carefully examine these causal relationships. The study concludes that the findings will allow for more informed policy decisions and infrastructural investments to support the ongoing transformation of the Finnish payment landscape. These implications suggest that the transition towards a cashless society in Finland is likely to continue and might be further accelerated by targeted policy interventions.

The study acknowledges a significant limitation: the absence of time-series data on the pricing of different payment methods. This omission prevented the researchers from incorporating a crucial factor influencing consumer payment preferences into their models. The authors explicitly recommend future research to address this gap. They propose conducting cost-benefit analyses of various payment methods, considering the perspectives of banks, stores, and customers. Understanding the cost structure associated with each payment method (cash, debit cards, credit cards, etc.) is critical for a more complete analysis of payment choices. The lack of pricing data represents a significant limitation in understanding the intricacies of consumer choices and market dynamics within the Finnish payment system. Incorporating pricing would lead to more accurate and nuanced modelling of payment method adoption and displacement.

The researchers suggest several improvements to the existing models. They recommend revisiting the error correction and partial adjustment models, grounding them more firmly in established money demand theory. This theoretical grounding would provide a stronger foundation for interpreting the results and enhancing the predictive power of the models. The current analysis, while insightful, could be strengthened by explicitly integrating the theoretical framework of money demand, offering a more robust explanation for observed trends. By refining the models and linking them to well-established economic principles, researchers can generate more precise and reliable forecasts of future payment method trends in Finland. Further development of the models is encouraged for improved accuracy and predictive capabilities.