New and better statistics

Most objective observers would agree that there have been significant improvements in the national statistical services over the last several years. New methods have brought in vastly improved statistics, ranging from national accounts, various price indices, purchasing power parities, government finance, labour market data, social statistics and various others, with a novel degree of interconnectedness between the multitude of new or improved statistics.

The National Statistics Office has brought up our standards to those of the developed world. Nothing less is required of a prospective member of the European Union. I am proud to have been the minister responsible for the passage of the relevant legislation and the setting up of the Statistics Authority and the NSO, which have ushered in radical advances in comprehensiveness, timeliness, precision and impartiality. My links with the statistics office go much further back than during my ministerial responsibility over the past ten years or so.

I remember a time in the not too distant past when no single PC existed in the Statistics Office. To overcome the resistance to their introduction, I had to find funds, purchase and take six PCs myself to the Statistics Office, set them up and encourage the staff to transfer their work from hand-filled worksheets, which were the staple tool, to electronic spreadsheets. This was a time when all sorts of difficulties and resistance stood in the way to introducing modern and efficient systems.

At that time, when I was still a consultant in the Prime Minister's office at Castille, I was probably regarded almost as an alien. Many used to look at me as an outsider who was not satisfied with traditional tools and who insisted that even typing be done on a computer rather than using a typewriter which necessitated a physical cut and paste system to incorporate any changes in a document. This was not ages ago, but the late 1980s!

Times have changed and it is not without some satisfaction that I see today's set-up, the offices, the trained staff and the professional level of the NSO. Well done to all those who contributed to this change and who helped make such a revolution in our statistical services.

It is time also for the political debate to rise above the regular swipes at and aspersions on the accuracy of the statistical announcements. It is time for all debate on statistical matters to be at an entirely technical level. Such a situation will come about once all of us acknowledge the impartiality and professionalism that characterises the statistical system.

It is a pity that statistical findings are sometimes turned into a big dartboard, something to aim propaganda shots at. We deserve better, especially because the critics who would have jettisoned Malta's bid for EU membership would have found it hard to summon the interest and the resources to upgrade statistics services.

A substantial and reliable statistical system is essential to policy-makers. You need to know where things stand before you can evaluate policy and identify corrective measures. In addition, national statistics are the foundations of the financial relation between the member states of the European Union. Precise statistics are essential for us to have the correct budgetary relationship with the rest of the EU.

Still, certain aspects of the new statistics are complex enough to require careful interpretation. Take the Labour Force Survey, whose introduction was a milestone in Malta's recent statistical history.

Before, we had only the ETC administrative headcount of the employed and the unemployed. Now we have a system that looks at the employment picture from two angles. The ETC's record of employment continues to monitor places of work and the ETC's own unemployment register. On the other hand, since 2000 the Labour Force Survey gathers its information entirely from households. Since the rules of unemployment registration differ between countries, the findings of such a household survey are essential for international comparability of unemployment figures.

In addition, headcounts like the ETC's suffer from lags in notification by employers of their employment terminations and hiring. But it is not only the ETC's method that is fraught with problems. One can equally find faults with the LFS.

Household surveys, like the LFS, are conducted through the questioning of a sample of households. Statisticians infer conclusions about the entire population from statistics gleaned from a far smaller sample of 2,500 households. Such inference is subject to error.

To complicate matters further, there are differences between the two methods in the definition of what constitutes employment and unemployment. In particular, the ETC recognises the unemployed as only those on its registers. The results of the household survey include respondents who define themselves as unemployed even if they are not registered as unemployed.

The difference between the ETC's unemployed and the LFS's numbered 4,421 in December 2003. This discrepancy was up from 3,359 a year earlier. Since the primary differences in the two systems' method of counting the unemployed is that one depends on self-declaration, the other one on registration, these are self-declared unemployed who are not registering as unemployed. An in-depth investigation of unemployed non-registrant would lead to a better understanding of the dimension of the unemployment problem in Malta.

Additional clues on this subject came in the just-released NSO's disclosures on the narrowly defined "jobless households." Of the unemployed in these households, 57 per cent register with the ETC.

More generally, it would be useful to see where the ETC's results stand in relation to the boundaries of statistical confidence intervals around the LFS's findings on employment. For example, September's LFS reported public sector employment at 48,143, down from 51,658 a year earlier. The ETC reported 47,000, down from 47,692 a year earlier. Since the statistician should be able to depend on the accuracy of the government's tally on its own employment, it would be useful to split the causes of the discrepancy into definitional differences and sampling error.

Regardless of these complexities, it is useful to use a statistical system that tackles labour market measurement from two angles, since the two can serve as a check on the other. Furthermore, the LFS provides a measure of the slack there is in the Labour market. A kind of potential resource that is available, if only it can match appropriate demand for this resource.

A survey of the skills of those who declare themselves as unemployed would also highlight better the type of human resources available, and any training that needs to be provided to match better available skills with demand. Such an exercise can today also be supported with EU funds under a number of diverse programmes.

European Parliament

This week is a crucial week for Malta at the European Parliament meeting in Strasbourg. The amendment that I passed though the Regional Committee in Brussels a week ago has now been integrated into the whole report as a separate clause 58 and now reads:

The European Parliament...

Considers also that islands such as Malta and Cyprus are treated in a similar way to other eligible islands in the region that are at a similar level of development, so as to ensure a level playing field.

This will now be voted on by the whole assembly, after debate that is scheduled to take place on Tuesday afternoon.

Also this week there will be the final hearing for the Court of Auditors, to which I have been nominated. Such hearings, as was seen by many in the case of the Commissioners, are always quite searching and sometimes rather unpredictable, depending on which members are present for that particular session, and what their specific interest happens to be. God willing, Malta will also have another successful hearing this week!

www.josefbonnici.com