There are lies, damned lies and statistics

Wednesday, February 13, 2008 Vincent Lingga, The Jakarta Post, Jakarta

Conferences on corruption or climate change and political party conventions are all headline-generating events. The National Press Day celebrations, like the event in Semarang last Saturday, are spectacular occasions for media coverage.

President Susilo Bambang Yudhoyono, who attended the press gathering, knew its importance, its ability to generate a vast amount of publicity and TV footage.

But statistics? What a boring and dry subject. Hence, most of the mass media simply ignored the national conference on statisticians, which Yudhoyono opened in Jakarta early last week.

It is glad to know that the President, amid his tight schedule, could still spare some time to open the national conference of the Central Statistics Agency (BPS). Yet more encouraging is that he fully realizes and reiterates the importance of reliable and accountable statistics for the policy and decision making processes.

A statistics office, being a government institution, is often suspected of engineering or tampering with figures to satisfy particular parties. During Soeharto's authoritarian rule the BPS was often accused, though never with any strong evidence, by government critics of fixing data or figures to massage the performance records of the government in all fields.

A fitting aphorism commonly attributed to Benjamin Disraeli states: "There are three kinds of lies: lies, damned lies and statistics."

Even now when the BPS has publicly been perceived to be strongly independent, the agency still often comes under attack from critics. More recently, for example, several analysts rejected the government's claim of a significant reduction in poverty figures as being based on flawed data provided by the BPS.

But Yudhoyono rightly reaffirmed the crucial role of the BPS in gathering reliable and accountable data which is needed for policy making, pointing out that complete, reliable data gathered with a credible methodology represented 50 percent of the whole process of policy making.

"I always believe the data collected by the BPS through surveys or censuses even though the data does not bode well for the government," the President said.

Reliable data indeed underlies our knowledge and hence our actions. The point is that the role of statistics goes well beyond the production of figures. It touches upon people's everyday lives.

When a government prepares a new budget, when businesses decide on investments, stock brokers make recommendations to clients, even when families decide which school their children should attend, all their decisions are mostly based on some sort of statistical information which is converted into knowledge and use to inform their decisions or choices.

Even one of our democratic tools, general elections, depends on statistical data on voters and the reliable counting of ballots.

The main challenge for the BPS is maintaining quality data under heavy demands: The process of defining and gathering the statistics, ensuring relevance, veracity and comparability, supplying the right metadata, such as definitions, sources and disseminating and updating, all in a fast-moving technological environment.

We live in a data-rich world in which ordinary people have become familiar with notions like inflation, imports, gross domestic product, or interest rates.

Statisticians do not take decisions but they do an important job as statistics represent a fundamental tool in developing knowledge, which in turn is vital for making evidence-based decisions.

Needless to say the government should always help safeguard the independence of the BPS and give it adequate resources to improve the quality of its surveys or censuses, which in turn determine the reliability and quality of its statistical data.

Without sound data, advice sounds rhetorical, and policy prescriptions ideological.

Many of our problems -- a sudden steep rise in the prices of certain commodities or services or an unexpected shortage of food -- are often caused by policy measures that were based on inaccurate statistical information which in turn caused an incorrect analysis.

The BPS undertakes methodological surveys and research on various aspects of our economic and social life and produces statistical data on a national, provincial, regency basis for use by the government, the people and businesses in making decisions.

Certainly, a number of critics sometimes wonder whether all of the statistical data work is nothing more than statistical overindulgence. We do not rule out the risk of statistical overload and of attaching too much importance to certain figures to create a perception or impression as desired.

While the importance of quality data cannot be overestimated, quite often it is the handling and interpretation of the data by both users and suppliers that causes problems.

One can see a half-empty glass as a half-full glass or the other way around.

But obviously what counts most is to know how to treat numbers and to develop the knowledge we need to act on them. Good decisions depend on good judgment. But there are too many things, the uncertainties of life, that we do not know.