Dataset details: CRDA/61/DS/3

The dataset as accessioned by NDAD comprises an analysis of trends in inflation as measured by the Retail Prices Index (RPI) since 1947. This distinguishes it from the previous datasets in this series, which are snapshots of the RPI and the Consumer Prices Index (CPI).

The RPI covers expenditure on the full range of consumer goods and services purchased by the majority of households in the United Kingdom, from the basics such as food, clothing and housing, through to leisure goods and services.

The dataset contains an analysis of increases and decreases in household expenditure over time. For example, at the start of the period, food (and catering) represented the largest single category of expenditure, accounting for 35-40 per cent of the overall total.

The RPI measures the full range of household expenditure, but it does not price everything that households spend their money on. Instead, a representative set of items is chosen, for which prices are tracked on a month by month basis. In 1947, there were around 200 of these representative items in the RPI basket of goods and services. Nowadays (ie 2006), there are over 650; the increased number reflects the growing complexity of the retail market and range of goods and services available, and improvements in the representativity of the items included in the basket.

In 1947, the basket included many items which are not present in today's basket. The changes that have occurred reflect increasing affluence, evolving technology and changing tastes. Increasing prosperity and leisure time is also apparent with the increasing number of items in the household services, leisure goods and leisure services groups. Until 1962, the content of the RPI basket was generally held constant between rebasings but since then it has been updated in most years to reflect changing tastes in expenditure and to ensure adequate representation of each expenditure category.

While the majority of the data is in the date range 1947-2004, the dataset contains one table (reference CRDA/61/DS/3/5/1) of average prices of a range of foodstuffs and selected consumer goods from 1914 onwards. Unlike the remaining data, which contains averages arranged on a monthly basis, this table only contains annual figures.

For further narrative detail on trends in expenditure, the range of items in the basket, and the changes to these items over time, see the article 'Consumer price inflation 1947-2004', catalogued as CRDA/61/DD/1/5 in the Dataset Documentation Catalogue.

The tables were downloaded from the website of the Office for National Statistics website in the form of eight Microsoft Excel spreadsheets. From these files, NDAD derived 42 tables for the dataset as rendered. Some of the Excel files held a number of worksheets (pages or sheets within the spreadsheet) which were split or combined by NDAD as appropriate to the structure of the data. In some cases one worksheet produced one table in the NDAD representation, while sometimes one worksheet resulted in five tables; in other cases, up to three worksheets were combined to produce one table at NDAD. Details of the original location and context of each table is held in the Scope and content sections of the table catalogues. All tables were produced by cutting and pasting the relevant section of data into a new spreadsheet, removing all unused columns and rows if necessary, and saving this data in a new comma separated variable format (.csv) file. This processing was performed in MS Office Excel 2003 for Windows XP. The original field descriptions were taken from the column headings contained within the original Excel spreadsheet files, and field names were created by NDAD.

NDAD holds a collection of price indicators used to compile the Retail Prices Index, for the years 1947-2004 (not every year is represented). See the Dataset Documentation Catalogue, reference CRDA/61/DD/2.

NDAD holds a copy of an article Consumer Price Inflation 1947-2004 written by Jim O'Donoghue, Carol McDonnell and Martin Placek of the Office for National Statistics. This article looks at trends in inflation since the Second World War, as measured by the Retail Prices Index (RPI). See the Dataset Documentation Catalogue, reference CRDA/61/DD/5/1.

The dataset as downloaded from the ONS website originally comprised eight named MS Excel spreadsheets. The data was spread across several named worksheets within each spreadsheet. These were treated by NDAD as 42 separate tables, and organised into the following five categories:

• CRDA/61/DS/3/1 = Yearly and monthly Group indices, with percentage changes (30 tables)
• CRDA/61/DS/3/2 = Group weights (1 table)
• CRDA/61/DS/3/3 = Group and subgroups, and weights (8 tables)
• CRDA/61/DS/3/4 = Subgroups and sections, and weights (2 tables)
• CRDA/61/DS/3/5 = Average prices (1 table)

There was no original numbering sequence in the data as downloaded. The numbering structure in the dataset tables as rendered has been devised by NDAD.

The dataset comprises the following table(s):

The historic RPI data was mostly input manually from paper records, dating back to 1914 or earlier. However, the price collection methods for the historic RPI are not known.

The weighting methods for the historic RPI are broadly the same as the Retail Prices Index snapshot 1987-2003 snapshot (CRDA/61/DS/1), in that some components are weighted to ensure that the index reflects the varying importance of the goods and services, the amounts spent in different regions of the UK, and in different shops. Households spend different amounts on different types of goods and services. In order to reflect their relative importance properly in the RPI, changes in the price of these different types of goods and services are weighted according to their overall share of households’ expenditure. Between 1947 and 1961, the weights were held constant for between four and six years at a time; since 1962 they have been updated annually.

In the original file 'RPI_Groups_1947-2004r.xls', the group indices table in each of the worksheets also held the weights for that particular time period. These haven't been made into a separate table by NDAD since the weights were already held collectively in the table 'Groups weights 1947-2004' (CRDA/61/DS/3/2/1).

The reference date for 'Groups and Sub-groups' 1952-1955 was given as 15 January 1956=100. It is likely that this should be 15 January 1952=100.

Around 200 checks were carried out by NDAD to compare the transformed data against the original data in Excel. These included checks on the numbers of fields and records for each table, and some simple statistical functions (sums, means) on a random selection of tables. No discrepancies were found.