What are Administrative and Secondary Sources?

Store cards are a typical example of a new type of private sector data source. In return for benefits such as discounts and exclusive special offers, users of store cards give the stores a lot of data every time they use them. If you have a store card, the store knows or can derive the following data about you:

• Name, address, sex, age
• Family circumstances (e.g. if you regularly buy baby products, toys, pet food, or products such as meat in a certain quantity or size, it is easy to estimate the composition of your household)
• Indicators of work status and income (e.g. the time at which you shop can indicate whether or not you work, and the type of goods purchased can indicate disposable income)
• Other household indicators, such as car ownership (purchases of petrol and car care products), religion (purchase of goods linked to a particular religion, e.g. halal or kosher meat), etc.

Image Modified

This may seem a rather extreme example of a potential source, and one that is unlikely to be considered for the purposes of official statistics in the near future. However, several countries have considered the use of till roll data from major retailers as a source of data on retail sales and prices, and Statistics New Zealand has produced an experimental data series using electronic card transaction data[5].

The use of store card data could be seen as the next logical step, particularly if coverage can be improved by linking data from different store card schemes, as well as data from other commercial sources. If this sort of administrative data source is ignored by official statisticians, how long will it be before private sector businesses with access to these data, start to offer plausible, and more cost effective alternatives to key official statistical outputs such as population census data?