|Unit||Index 2015=100, SA|
Survey-based retail sales indexes for Italy, by product type (under NACE), store type, and enterprise size. Monthly from 1996.
The data is classified based on the NACE 2.0 system.
The source writes:
The monthly retail sales survey conducted by ISTAT refers to trading undertakings through retail outlets in the fixed authorized the sale of new products. Are excluded from the survey enterprises whose main activity is the sale of monopoly goods, vehicles and fuels. Also outside the field of observation points of sale of goods used, the street vendors and markets. From the press release on the data in January 2009 the index series were circulated in relation to the base 2005 = 100 and the new classification economic activity Ateco 2007 (Italian version of the European classification Nace Rev. 2). With the spread of data from October 2009 indices have been corrected, as announced in previous months due to the introduction of a structure weights more disaggregated, which considers all the variables of stratification of the survey.
This change aims to make more robust measure of the dynamics of sales of specific components and distribution of goods. The operation resulted in a comprehensive revision of the series (from January 2005 to September 2009) indices based on 2005 previously circulated. The correction has had a very limited effect on the dynamics and the aggregate index on the part of traditional distribution, while it has changed significantly related to changes in the mass distribution and, particular those concerning the different types of exercises belonging to this sector.
In particular, the rate of change in average annual total sales to be adjusted minimum (between 0.1 and less than 0.1 percentage points) for period between 2006 and 2008.
The dynamics of the traditional component distribution also showed limited modifications: the wider review is records with regard to mean change of 2007 decreased by 0.2 points percentage than previously measured. The indices relating to supermarkets have changed the most significant for some individual monthly values, but contained an annual average: the relative changes are been revised downwards by 0.3 percentage points on average in 2006 and upward a similar amount in 2007, then remained unchanged for the average of 2008.
The series of indices recalculated from January 2005 are available ConIstat database (accessible at http://con.istat.it). Shortly retrospective reconstruction will be completed until 2000, covering the key aggregates and aims to make available sufficient time series homogeneous.
For a more detailed description of the basic characteristics of the index in 2005, and in particular the new classification of types of business belonging to large retailers, see Information Note "The new indices sales value of the fixed retail trade by 2005 = 100" of 31 March 2009.
The theoretical sample survey consists of approximately 8,000 companies operating throughout the country, an extract derived from stratification from the intersection of two variables:
The seasonally adjusted indices are obtained through the procedure by-SEATS (February 2008 version for Linux operating system). Like other procedures seasonal adjustment, including through-SEATS assumes that each series historical interim basis is represented as a combination of components, not directly observable:
TRAMO-SEATS, in particular, uses a model based approach that relies identification of a statistical model representative of the behavior of time series by seasonal variations. In order to make available to the users seasonally adjusted index of total sales and those of food and non-food, for these three indicators was carried just a slip of indices calculated from 2000, and for the period prior to 2005 (from 2000 to 2004), through appropriate coupling coefficients.
The statistical models used for seasonal adjustment are reviewed periodically to ensure their ability to correctly represent the development of the individual series. Furthermore, since the addition of a new monthly newsletter provides a better picture of the different components series, monthly data already published for the past year are subject to review.
The general index of sales adjusted for seasonal effects is calculated as synthesis seasonally adjusted index of sales of food and what sales of non-food products (using the so-called indirect method).
The dataset is rebased every five years.