One of the greatest problems that any government has to face - especially when it joins the eurozone - is perceived inflation.

Perceived inflation is just as insidious as real inflation, as it prompts demands for wage increases - which lead to more inflation.

Consumer Price Indices, like the local Retail Prices Index and its counterpart for EU purposes, the Harmonised Index of Consumer Prices, are tools that are used to monitor price changes over time and any resulting inflation.

But will the public believe their perceptions or the official figures? The National Statistics Office is confident that the RPI and the inflation rate are good indicators of price increases so long as consumers realise what their inherent flaws are.

"One of the problems with the RPI is that it is based on the prices of a fixed basket of goods and services and not on actual consumer purchases. Consumers adjust their purchases according to market prices and their disposable income. For example, if oranges go up to Lm3 a kilo, my wife might substitute oranges for apples instead as the latter might be more affordable. The index would not capture such substitution but the price recorders would log oranges at Lm3," NSO director Alfred Camilleri told The Times.

"Another problem is the progress in technology. Say I bought a computer for Lm800 some years back. It is now half the price but, in fact, for the lower price I am getting much better quality in terms of technology. So the price is not really half but even less. Efforts have been made to improve our methodologies on quality adjustment so that cases like this can be reflected more appropriately in our indices."

The NSO is also planning to update the Household Budgetary Survey on a regular basis rather than every five years as at present. This will ensure that any changes in public consumption patterns are captured as soon as possible.

Of course, this will also have a positive impact on the other statistics in which the HBS plays a role, such as the calculation of consumption expenditure in the Gross Domestic Product.

The spending patterns determine the weighting given to expenditure groups - for example, how much people spend on food as a percentage of their total expenditure.

"There have been significant changes even in the past few years. For example, food represented 29.92 per cent of spending in 1995 but just 23.82 per cent in 2002.

"The thing is that people do not eat more when their level of income increases: rather they consume more services like entertainment, or buy a better car, or travel more, or eat out.

"As a result, the impact of food prices on the RPI has gone down."

Apart from the weighting, the NSO has to ensure that the price recorders monitor the thousands of products in over 250 index expenditure aggregates as accurately as reasonably possible.

"We don't monitor the prices of each product or service on the market as that would not be possible but a representative sample made up of market leading goods and services and brands, working on the theory that they set the prices and that other less popular goods, services and brands follow them, whether upwards or downwards. Following this argument, it doesn't matter whether the market leader is a local or foreign brand.

"For each expenditure item we monitor three or four brands from a large number of retail outlets and from many localities in Malta and Gozo and average their prices."

Alfred Mizzi, manager for consumer prices, said the NSO also tried to avoid items that only appeared briefly on the shelves as it was crucial when trying to establish trends to ensure that you are comparing like with like in a consistent manner.

"One of the biggest headaches is clothing prices: it is so complex. How do you compare like with like?"

The NSO gets its prices from a selection of about 700 retail outlets - small corner shops, supermarkets and even street hawkers - in different localities in Malta and Gozo. All the major shopping areas are represented, such as Valletta, Victoria, Sliema, Hamrun, Mosta, Zabbar and Paola, as well as tourist areas like Qawra where prices might be a bit higher. In addition, it carries out surveys on education-related services like schools, private tuition, school transport and the services provided by medical practitioners. It is estimated that the NSO collects around 180,000 price quotations annually in order to compile its prices index.

Another complexity for people to follow is the link between the RPI and inflation.

"The impact that changes in the RPI have on inflation depends on the weighting of that sector in the RPI, so even a substantial drop in prices - say of pasta - would have a much smaller impact on the inflation rate.

"However, you cannot overlook the psychological factor: although the RPI reflects different consumption groups, most of them are not consumed on a regular basis," Mr Camilleri explained.

"You go to the supermarket every week and instantly realise that something, say a detergent, has gone up in price. But you would not even notice that the price of fridges has gone down unless you happen to need one, and even then, that is hardly likely to be every year, let alone every week.

"So our perception on prices depends to a large extent on what is most visible to us on a daily basis: food."

In fact, in the 15-month period to March, the food index went up by four per cent and beverages and tobacco by 12 per cent. Personal goods (which include medicines) went up by 7.4 per cent and "other goods" (which include detergents) went up by 10 per cent. In that same period, the inflation rate changed from 1.79 per cent in January 2003 to 1.75 per cent in March 2004.

"People not familiar with the link between RPI and inflation would assume that one or the other is wrong. But both are perfectly correct."

The situation will be further complicated for the man in the street by the fact that the NSO will on May 18 also start to issue a Harmonised Index of Consumer Prices, which includes tourist accommodation and other expenditure and, therefore, will not be relevant to consumers for local inflation purposes.

Of course, the importance of the inflation rate goes beyond fuelling wage increase demands: it is also one of the Maastricht criteria for joining the eurozone...

The criterion is that inflation should be within 1.5 per cent of the average rate of the three lowest inflation rates in countries in the eurozone.

"The problem is that as the economy picks up, inflation might rise. So this could become a problem," Mr Camilleri said.