The National Statistics Office said that in the near future and in response to proposals made by the GRTU, it will be giving information on the deflators relating to sub-components of the main GDP aggregates. This new information may be included in the release covering the last quarter of 2007.

The NSO was replying to a paper by Prof. Joseph Falzon, commissioned by the GRTU, Chamber of Small and Medium Enterprises, and entitled The Real GDP Growth in Third Quarter 2007: a Reconsideration.

The NSO said the paper sets out to compare: (i) the price inflation rate as recorded by the Retail Price Index (RPI) between the third quarter of 2006 and the third quarter of 2007 with

(ii) the price inflation rate as implied by the household consumption price deflator (CPD) within National Accounts data for the same period.

The paper identifies a difference in price inflation rates obtained through these two approaches and subsequently proceeds to question the validity of the household consumption expenditure data at constant prices for the third quarter of 2007. In particular, it argues that the latter variable should be at a lower level if the RPI were to be used for deflation.

The compilation of National Accounts Data

The NSO said that Gross Domestic Product (GDP) is the most frequently used indicator in the national accounts. GDP combines all the output (or production) carried out by all the firms, nonprofit institutions, government bodies and households in a given country during a given period, regardless of the type of goods and services produced, provided that the production takes place within the country’s economic territory. In most EU countries it is calculated quarterly.

The index of variation in GDP (or any other variable) at current prices breaks down precisely into the product of the index’s variation in volume and the index’s variation in prices, the latter being known as the “deflator” or the “implicit price index”. This deflator can be used as a measure of economy-wide inflation and as such, differs from the consumer price index.

Generally speaking, the detailed statistics available to national accountants are of three types:

(i) statistics expressed in quantities, such as the number of number of microchips produced;

(ii) statistics expressed in current prices, such as figures taken from company accounts;

(iii) price indices, such as the numerous components of the household consumer price index (HICP).

If statistics expressed in quantities are not available, national accountants combine statistics of type (ii) and (iii) to obtain an indicator of volume. If one divides the change in magnitude expressed in current prices (in value) by the change in the price of the corresponding products, one obtains a measure of the change in volume. This is what national accountants call “deflation”.

The process of deflating national accounts data

At the outset one has to observe:

a. In Malta, the NSO uses the Expenditure approach of GDP estimation to arrive at the GDP in real terms;

b. Several countries use the production approach to derive real GDP estimates;

c. Deflation of the GDP is a long and complicated exercise - some deflators are standard ones, primarily in the case of certain aggregates; in other cases deflators have to be built up gradually by the statistician as and when they are considered necessary and on a quarterly basis;

d. The methodology has not changed since the adoption of ESA 95 in 2003.

It is not possible to derive all the deflators from what is published even though the quarterly GDP release runs into 20 pages and gives detailed information on a sectoral basis. It is, therefore, to be conceded that the author of the paper did not have at his disposal all the information needed to make his analysis a thorough and conclusive one, the NSO said.

For this important reason, one may infer that the conclusions drawn in the paper are the result of an over-simplification of a highly complex exercise. For example, the deflation of household final consumption expenditure has to be carried out in a much more detailed manner by using the internationally harmonised Classification of Individual Consumption by Purpose (COICOP). In deriving the value of household consumption expenditure at constant prices, the NSO uses specific deflators applied to data at current market prices to all sub-groups within COICOP.

In line with standard practice across the European Statistical System, these detailed deflators are not normally published, although they are available upon request. The conclusion is therefore evident. The deflation process attempted by anyone who is not in possession of the whole framework of information cannot be considered correct, complete or conclusive.

The same observation obtains in respect of other aggregates such as government consumption expenditure. This may be broadly divided into (a) wages and salaries and

(b) ordinary consumption of goods and services. In this case, it does not emerge from the paper that the author had at his disposal a wages and salary index or two separate indices in respect of market goods and services, to apply to the two different categories of expenditure. Therefore, no distinction could be made between the amount expended on goods and that paid for services by Government unless one has obtained such information from the Treasury.

Comparison of the RPI

with the HICP

Another shortcoming incurred by the author concerns the general deflation index that is used. Throughout the paper the RPI is referred to. One infers that it is used extensively to prove that the deflated estimate of household final consumption expenditure published by NSO is incorrect. It is conceded that the result arrived at by the author is different from the one arrived at by NSO. This is hardly surprising - they cannot tally for the simple reason that, in accordance with international good practice, the RPI is not employed in deflating household consumption expenditure. The correlation between the consumer price index and the RPI is in fact lower than the correlation between the consumer price index and the HICP.

The general methodology employed in all EU countries takes into account the HICP and not the RPI. Prior to the introduction of the HICP, NSO made use of the RPI but since

2000, the HICP has been extensively used for deflation purposes. The reason is a very valid one: consumption expenditure takes into account all expenditure within the domestic market namely, that incurred by residents and by tourists.

There are some differences in the basket of goods and services pertaining to the two indices. Furthermore, there are also considerable differences in the weighting frames of the two indices, as can be deduced from the following abridged table. Of particular significance (given the general thrust of the paper), is the difference in weights, especially in respect of the overall food category.

Conclusion

• In the near future and in response to proposals made by the GRTU, NSO will be giving information on the deflators relating to sub-components of the main GDP aggregates. This new information may be included in the release covering the last quarter of 2007.

• Similarly, a proposal will be put for discussion at the next RPI Advisory Committee meeting of 20 February 2008 so that, if approved, quarterly deflators will be discussed in this committee before they are published.

• These two proposals will enhance the transparency clause of the European Statistical Code of Practice.

• The correctness of the deflators derived from the HICP depends on the way this consumer index is computed. In this regard, one has to record that this index was subject to a Eurostat audit in October 2006 and November 2007. A copy of the last report is being attached for information.

• NSO would have cooperated with the author of this academic exercise had he consulted the Office before reaching his conclusions as he has done on several previous occasions. As reiterated on several instances, NSO is always ready to offer its assistance to interested analysts.