INTERVIEW | Wednesday, 06 February 2008

Statistics are the foundation of the State

Charlot Zahra spoke to Reno Camillieri, Chairman of the Malta Statistics Authority and Acting Director-General of the National Statistics Office about the latest controversies involving the NSO and beyond

The NSO is currently conducting a revision of the Household Budgetary Survey. Can you explain why this revision is being carried out?
In the first place, the NSO is carrying out a Household Budgetary Survey (HBS) and not a revision of it. A primary objective of this exercise is to provide the information necessary for an update of the Retail Price Index basis. The RPI Index, as it is commonly known, measures changes in prices in a pre-determined basket of goods and services and is an important indicator of the impact of inflation on family budgets. The purpose of this important and costly survey is to make changes in the basket of goods and services whose prices are reviewed regularly and to update in the weighting frame of the Index. In the past, the HBS used to be carried out every ten years. Now we are undertaking this exercise every five years.
We have still to decide whether this survey will continue on an on-going basis as is done in several countries or whether to cover only a 12-month period. In the former case, the survey will assist in determining consumption expenditure for national accounts purposes and to make annual updates, if necessary, in the weighting frame of the RPI and the Harmonised Index of Consumer Prices (HICP) on an annual basis.
As in the case of other countries the RPI or the CPI is considered as an index of inflation that affects our earnings, allowances, benefits, pensions and the rents we pay.

What differences are expected in the composition of the items that will compose the basket using for calculating inflation statistics when compared to the basket used until now?
I have no crystal ball to look into to foretell the results of the current Household Budgetary Survey. Differences in the present weighting frame will definitely emerge. However, one may not expect radical differences.
I would envisage that the weight of food items in the representative basket will again be less than during the previous five years by a very small margin. Similarly the weight of beverages and tobacco may again show very small decreases.
In contrast, the weight of communications and transport as well as the weights attached to electricity and other energy items will go up. Health items may show an increase in their overall expenditure outlay. Items related to Education and Cultural activities may show the same level of overall expenditure in the HBS.
The more definite conclusion is that food items may not show the same drastic decline in their weight as resulted from the last four HBSs where the Food category weights decreased from 41.3% in 1980 to 36.8% in 1989, and from 30.0% in 1995 to 24.0% in 2001. Will the recent general rise in food prices persist during the current year when the HBS is being carried out? The answer to this question will influence the new weight of food items.

Every now and then, a controversy arises over the credibility of NSO statistics, with the latest issue being the report submitted by the GRTU about the statistics used by the NSO for calculating GDP. Do you think that undermines the credibility of the NSO as the official source of statistics or not? Why?
Within a local context, such situations tend to put the NSO in bad light, the more so when controversies involve an indicator like the RPI that attracts the interest of most of us. I remember that in the past, the RPI was under constant examination and suspicion particularly by the Trade Unions.
For this reason, I had suggested that the RPI should be examined by the social partners every month before it is published. I remember the late Salvino Spiteri, on behalf of the UHM, and Anglu Fenech for the GWU, accepted this proposal and the RPI Advisory Committee was set up. This is a unique arrangement. This high-level Board still meets every month to examine the RPI before it is published.
During 2007, the NSO published 207 statistical releases. This output is remarkable by any standard. There was no controversy on these releases except on two of them. It is the way that criticism of the NSO is projected that is unacceptable.
If a statistic is misunderstood or considered incorrect let us discuss it by all means. We have an important forum, again my baby, the MCESD, where important discussions may take place.
Perhaps of greater significance is the undue pressure and sense of apprehension that is put on the staff. I accept that everyone should be accountable but the way the NSO image is tarnished by the manner certain comments on the work of the Office are spread or commented upon in the local media may cause a lot of harm to this institution.
Two thousand years ago, Cicero wrote that ‘numerus reipublicae fumdamentum’ (statistics are the foundation of the State). Sometimes we may pause and consider this statement for the good of our democratic State?

In its report, the GRTU said that the real GDP statistics for the third quarter of this year should have been 1.94% and not 4.14% as stated by the NSO. How can you explain the difference between the GRTU report and the NSO statistics in this respect?
I cannot see any reason why this unnecessary controversy should drag on further. Both the GRTU and the NSO have stated their positions. Whoever is interested may consult the reports.
But even from this controversy I identify two positive courses of action. In response to the call for more transparency by the GRTU, the NSO will publish the GDP deflators. I shall also propose that the social partners discuss these in the RPI Advisory Committee.
I have always favoured more involvement by the social partners in statistical work even though the NSO will remain solely responsible for official statistics. On the other hand, one should appreciate the views of users, be they organisations or researchers, as outlined in the European Code of Practice.

Don’t you believe that occupying the post of head of the Malta Statistics Authority (MSA) while at the same time holding the post of acting director-general at the NSO is a conflict of interest?
I think that the phrase ‘conflict of interest’ is too strong. I am the chairman of the MSA Board. As Chairman, I consider myself to be ‘primus inter pares’. There are eight other members in the governing Board and on several occasions the Board did not accept my proposals. This is as it should be unless the Board is not composed of ‘yes’ men.
During my taking over the administration of the NSO because there was no alternative and this was done with the unanimous approval of the Board and the Minister, I made it a point to report to the Board more frequently what was going on at the NSO and was assisted in taking certain decisions.
Is the situation an ideal one? Of course not. I have been vociferous about this on several occasions both to the Board and to the Minister that I do not wish to remain one second more at NSO than is required.
Our search for someone competent to occupy the position of DG has, so far, proved fruitless. I may add that during the last meeting of the MSA Board, I was determined to call it a day, but the Board insisted that I should not leave the NSO at this delicate moment.

What concrete action has the NSO taken to select a new director-general for the NSO? Has a call for applications been issued? Was anyone approached by the Office or by the Government?
It is not the NSO that has taken concrete action but the MSA Board that is responsible to recruit a new Director General. You may recollect that when my predecessor presented his resignation with immediate effect, the Board had been faced with a very serious situation. There was no deputy Director at NSO so something had to be done to ensure the normal flow of statistics both locally and to Eurostat.
After a period of reflection, the Board proposed to issue a call for applications but there was the possibility that a competent incumbent could have been identified. This situation repeated itself on various other occasions. The rest has already been divulged to the press. This paper carried an account of our efforts to fill the position of DG last week. I do not need to repeat what I have said.
In brief, this is the present position. The issue will be taken up again immediately after the next General Elections. But I may add that when we had issued a call for applications in 2006, we had only three applicants.
It seems that, for some reason or other, certain positions are not considered very attractive. In the UK, they had to recruit a foreigner for two successive terms of office. Within the local context, a case in point is the post of Auditor General.

Do you think that the NSO should remain under the portfolio of the Finance Ministry or not? Shouldn’t it be independent?
The Malta Statistics Authority Act provides for the establishment of a Statistical Authority to supervise the national statistical system. The EU will shortly be introducing something on these lines on a European level through the setting-up of a Statistical Governance Body.
The MSA Board is composed of nominees of the social partners, the University and the Central Bank of Malta. The Chairperson of this Board is nominated by the Minister. Since the coming into force of this law in March 2001, the ‘responsible’ Minister has not refused any nominee by the social partners. The only condition governing Board members is that these must possess expertise in economics or statistics.
The Board in consultation with the Minister appoints the Director General (DG) of the NSO. The DG is answerable to the Board. and is an ‘ex officio’ member of the Board.
The law specifically lays down that the Minister cannot give instructions to the Board except in exceptional circumstances (I don’t know what this means) and in such cases these have to be given in writing.
This is important so that if this happens, the Board may include its comments thereon in the annual report that is laid on the table of the House. The DG is solely responsible for the collection, compilation and publishing of official statistics and will report to the Board at the end of the financial year.
There is no doubt that functional independence and full autonomy from Ministerial interference is guaranteed by our law. In practice, I may say that I was never subjected in any way, either directly or indirectly to Ministerial interference.
From experience, I can vouch that our legislation is among the best in so far as functional independence is concerned. I may add that the NSO does not release any statistical information to anyone, including Ministers, before publication.
I may also refer to the conclusions of a report drawn up by a Eurostat mission that was in Malta in late November 2007 on the independence of the NSO: “In all interviews carried out, the representatives of the various user groups stressed that they have every confidence in the professional independence of the NSO.”

Do you agree with the suggestion made by veteran economist Karmenu Farrugia in comments he gave to BusinessToday that the NSO is transformed into an authority and its head is appointed by the House of Representatives, just like the Auditor General and the Ombudsman, or not? Why?
This is a very interesting proposal and I have to confess that in drafting the present law way back in 1998, that was one of the options that were considered. Personally, I was not very enthusiastic about it but I cannot, in any way, object to the spirit or the intention behind it.
Having been involved in the drafting of the law, I had consulted twenty legal instruments pertaining to different countries and nowhere did I find that the Director is appointed by Parliament. In our case the Auditor General and the Ombudsman are appointed by the House as provided for in the Constitution. One of the reasons being that they may censor even Ministers as heads of the Executive.
The Director of the NSO has to act in an independent manner but his Office is mainly to provide vital information for the proper function of a democratic State. Furthermore, our recent experience has demonstrated that even Judges and the Auditor General himself have been subject to criticism.
Nonetheless, I would not dismiss this proposal lightly and if there is a wide consensus that such an arrangement may contribute to a better image of the NSO, let’s consider it very carefully.
Our main objective in the drafting of the new law was to set up an organization, call it an Agency, Institute (my preference at that time) or an Authority, that is fully independent of Government and staffed by professionals in the field of official statistics. Did we succeed in this? The latest Eurostat review report (November 2007) had this to say: …“the NSO produces and disseminates its statistics in a professionally independent manner.”

Another source of contention is the frequent revision of statistics published after a certain period of time. Why are revisions of statistics necessary?
Anyone not familiar with economic statistics cannot comprehend why revisions of statistical data is an on-going exercise. Revisions are necessary for several reasons. Let’s take, by way of example, the estimation of GDP. The estimation of GDP, apart from being a complex exercise, is as correct as your sources of raw data are.
After a time, you may become aware that these sources are defective. So you have to look for better ones. Moreover, the exhaustive factor of GDP poses a big and continuous challenge. Notwithstanding every effort to achieve comprehensiveness of this aggregate, one should be realistic and accept that this is hardly ever achievable.
A second reason is related to the problem of non-response or delay in responding to official questionnaires. In the case of non-response, the statistician has to make a ‘best of judgment’ estimate that may prove to be incorrect and be rectified later on.
A third reason concerns the adoption of new definitions of variables. The more radical revisions are usually made on the introduction of new methodologies. With the adoption of ESA 95 by Member States of the EU, considerable revisions were made to the GDP of these countries ranging from 5 % (7.0% in our case) to 20 per cent.
Before the adoption of the ESA 95, the most noted revision I remember was that made in 1987. In that year, the Italian statistical Office – ISTAT – had surprised everyone when it revised its GDP by 18 per cent with the result that the country’s per capita income surpassed that of Great Britain by 9 per cent in nominal terms and by 2 per cent at purchasing power parities. But lately, another Member State
– Greece – has revised its GDP by 25 per cent.
So in discussing certain economic aggregates, like the GDP, one should not let himself be too much influenced by what I usually refer to as ‘illusions of accuracy’. The compilation of this aggregate involves the frequent reference to a large number of variables, some of which are by-products of administrative registers and procedures that were not primarily devised to provide statistical information. For this reason, all countries revise their economic aggregates on a regular basis in accordance with an established revision policy.
This policy provides for GDP to be revised for a period of three years after the close of the period to which it refers. Eurostat goes one step further. They give ‘flash’ estimates, then provisional figures and then a final estimate. Normal revisions need not be announced beforehand but substantial revisions should be announced at least three months before publication. This is the policy that has been adopted also by the NSO.
Substantial revisions are also made to correct demographic data following a Census of Population. In 1984, I was asked by my Minister to organise a Population Census. The last census had been held in 1967 and our population statistics were truly in a bad shape.
Based on the census results I revised the 65+ population group by adding something like 8,000. The joke then was that I was giving birth to elderly people and including them in the population.

Public perception is that the cost of living is higher than what official statistics are showing. Does this mean that the tools used to calculate inflation are not effective enough or not? Why?
First of all, I do not know of any index that truly measures the cost of living. The layman’s perception of the RPI, that is an index of the Laspeyres type, is that it is an ideal barometer of the cost of living and inflation. In reality, it is neither one nor the other. However, in the absence of a proper cost of living index, it is considered as the best measure we have for this purpose.
It is interesting to mention that the RPI has a long history. As early as 1925, the Director of Labour was busy monitoring changes in a small basket of goods in order to assess whether the standard of living had deteriorated since the First World War.
Since then, this index has been worked out and published regularly. At present, the NSO is also publishing another inflation index: the Harmonised Index of Consumer Prices and the only one that can provide comparable inflation indices in the EU.
The public perception of the RPI as a cost of living index is, in the first instance, what one may call a too subjective one. If I use a lot of fuel and the price of fuel goes up, then I expect the index to rise. When I find out that the index has actually declined, I say that the index is not measuring correctly my cost of living. This is a general perception.
Theoretically, this is relatively easy to explain. The index takes into account a ‘basket’ of some 360 goods and services. Some may register higher prices during a period; others may actually show lower prices. Again, different items have different weights. The varying item weights within the RPI formula often determine, as much as the recorded price changes themselves, the behaviour of the index.
As I said we have no general index of the cost of living because a cost of living index should also include items of a capital nature that are specifically excluded from a Consumer Price Index. Some consider changes in the GDP deflator as a good indication of movements in the cost of living.
It is important that in referring to statistical measures like the RPI, one should be aware of their value and limitations. If one is knowledgeable about the true meaning of a statistic, a lot of misconceptions is avoided.

Do you believe that an October budget speech is too early for the NSO to get a clear idea of how the country’s economy is doing during any particular year as only data for the first two quarters is available, or not? Why?
The October budget poses severe problems to the NSO for January-September data in that the collection and processing of certain data takes some time. To quote one or two examples: the Eurostat calendar provides for the transmission of provisional GDP in respect of any quarter to Eurostat within seventy days from the close of the period. The monthly Harmonised Index of Consumer Prices has to be sent within a month. The first data sets are always subject to revision for the reasons already referred to in a previous answer.
It was only lately that arrangements were made with the Economic Policy Division of the Ministry of Finance that is responsible for the compilation of the Economic Survey to be supplied with data covering the first half of the year. In the past, I remember quite well, the NSO used to send the ‘best of judgement’ information up to end September. But the Budget used to be presented in November or in December.
However, one can make forecasts of how the economy will continue to perform during the rest of the year even if the only data available relates to the first six months. This may be possible by taking into account other indicators such as imports of industrial supplies, export orders, inflation trends locally and abroad, the level of the money supply, government and future private investment, trends in the labour market and other similar indicators.

[email protected]