Abstract
This thesis will examine electronic (debit and credit) card transactions in New Zealand, utilising the Electronic Card Transactions (ECT) statistics from Statistics New Zealand. The aim of this thesis is to test the Statistics New Zealand claim that the ECT variables can be used as an indicator for the change in the level of consumption expenditure and of economic activity in general. This will be done via business cycle analysis and through a forecasting exercise. The ECT series are only available from the fourth quarter of 2002, so the sample size utilised in the analysis is small. Cyclical information is extracted from the original series using univariate business cycle filters, due to the small sample size considered. These are then compared to the cyclical components of GDP, consumption expenditure and its components and measures of economic activity from the labour market (unemployment rate, weekly hours worked). Results show that the cyclical components of the ECT series are highly associated with cyclical movements in both GDP and consumption (and its components). Their association with the labour market variables is slightly weaker. Differences in the timing of the cyclical co-movements are also found between the debit and credit card series. The forecasting exercise tests the ability of the ECT series to improve short term economic forecasts. The results from the bivariate forecasting exercise indicate that the ECT series may be useful for improving forecasts, particularly for consumption spending. When an ECT variable is included in a more realistic forecasting model for consumption expenditure, the performance of the model is improved dramatically, relative to the case when the ECT variable is excluded. The two sets of results provide some evidence in support of the Statistics New Zealand claim and suggest that policymakers and economic forecasters should consider how to include the ECT series into their forecasting models.