Statistics speak, they don't !

Dr. Saadat Husain

Photo: Tanvir Ahmed / Driknews

Political masters and executives have a love-hate relation with the statistical bureaus of their countries. They are ecstatic when the published figures support their case. They accept the figures without any reservation and use them to press their point. The same people are very upset and virulently critical against the statistical agencies if the released numbers do not endorse their claim or point of view. News papers and magazines are replete with disparaging comments from ministers and high apparatchiks when they did not find the released statistics palatable. They either trash away the statistics as if these are trivial or offer alternative numbers for public consumption. The divergence and incongruity of statistics released by different government and semi government organizations give them a handle to undermine any vector of numbers not preferred by them.

In Bangladesh no unequivocal number exists on population, food production, literacy, unemployment, growth, poverty or inflation. Different agencies release divergent figures on each of these items. The same agency or ministry also does not steadily stick to one scalar or vector; nor does it follow any readable functional form. People are baffled to find motley numbers crowding crucial areas of national life. The inconsistencies and pitfalls provide the politicians, executives, media people, NGO operatives, development partners and even researchers a formidable handle to establish their claim by quoting statistics at random. They oftener than not get away by quoting counter intuitive figures because most of the readers or audience are not very rigorous in catechizing the statements made in different fora. Controversies persist unabated.

Photo: Art Box Images / getty

Most of the macro variables mentioned above do not yield to accurate counting. Production of industrial goods is recorded at various stages and is subject to both mechanical and accounting checks; its quantity is by and large accurately known. Such is not the case for agricultural production which is spread over thousands of acres of land in rural areas and is not subject to mechanical checks. Quantity of agricultural produce has perforce to be determined through eye estimation; that also through sampling over hundreds of plots. Some agencies may follow different methodology. The results may naturally differ. We cannot therefore come across a single figure for agricultural production unless the raw numbers are reconciled through thread bare analysis by experts in the field. The literacy figures are worked out through survey, large or small. The methodology and results are at times affected by the predisposition of the organizers. That holds for high profile research projects as well. It is said that monetarists always find the empirical evidence in their favour; so do the Keynesians. Population census is carried out once a decade, that also not with lapidary perfection. Some person may be left out of counting ,some one may be double counted. Standard adjustment mechanisms are followed and the results are brought to an acceptable level. Mid term population figures are based on pure estimation with lot of room for controversies.

Macro variables like growth, poverty or inflation, to start with, are theoretical constructs. These are based first on conceptual paradigm and then on some functional form. Neither the concept nor the functional form is unequivocally agreed. The divergence may emanate right from this stage and pervade through the results. This will naturally spill over to empirical stage.Operational definitions are seldom unique. Literacy, unemployment, inflation, poverty have multiple definitions. Each definition will yield a particular set of data, palpably different from those yielded by other definitions. Before dismissing the numbers one needs to understand the definition used to generate them. Operational definitions are sometimes formulated having regard to the ground reality. When precise definitions do not work, rough and ready definitions have to be crafted in order to tide over the difficult situation. The underlying philosophy is that some thing is better than nothing. Numbers may be wanting, they are nonetheless usable in many cases. Some definitions are blatantly incorrect and the numbers generated by them have to be summarily discarded. For example even highly educated people misconceive gross domestic product(GDP) as the value of total production. They are not conversant with the concept of value added. They are baffled to know that despite high volume of production value added may be negative because the price of the product may be lower than the input cost. So if some body tries to pass the value of total production as GDP his numbers should be readily disregarded. Similarly if a list of non profit organizations is prepared by excluding all organizations which have earned profit the list will generate wrong numbers and will be utterly misleading.

Photo: ALAM

However, numbers which are not accurate but are based on sound conceptual framework and robust algorithm will serve a practical purpose and should not be trashed away just because they are not digitally precise. In some areas precision is highly warranted, fortunately accommodation and adjustment will do in most other areas. Numbers may be imprecise, still they are often enlightening no doubt.

Take the case of population and food production. There is intense debate in both the areas. BBS, ministries, agencies, development partners and international organizations release their own numbers without any coordination amongst them. Readers are confused and irritated to see multiple sets of numbers and the vitriolic polemics following them. It is not clear why government does not authorize one particular agency, in all fairness the BBS, to release national statistics where they have the capability and strengthen the advisory committee with legal fortification. Population figures vary by a range of 15million. Food production and requirement figures are not reconciled either. As a result controversies on food gap rage the ground. There is no single figure on growth rate as well. Government agencies, research organizations and development partners come up with varied figures leaving people to make their own guess. It varies between 5.3 to 6.7. Readers rake their heads to decide as to which rate they should use in their plan, papers and deliberations. It would seem that numbers do not speak, they only create noise. How should one take the numbers in this context.

The situation is not so disappointing as it seems to be in the first reading. The figures vary within a range and that range is not so yawning as to render the figures completely irrelevant or useless. With a careful cogitation they can still be made usable. For example growth figures vary but still they beckon. One can make a fairly decent assessment about the economic situation of the countries from the divergent rates released by the agencies. Growth rate within a range of 5.0 to 6.9 implies that the economy is doing moderately well at less than coruscating level.

If it is below three percent there is reason to be worried. If it is 7% or more the economy has reached a distinctly higher level of performance and a new horizon of hope is prominently discernible. The main task will be to arrest the dirty appurtenances of high growth. It is therefore not a big deal whether the growth rate was 5.7 or 6.5, they do not really mean much of difference. The worrying thing will be if the rate is secularly declining over a few years. An economy with a sizable service sector will normally post a growth rate around 5%. It is not very sensible to be sensitive about a miniscule difference of less than 1% in the growth rate within this range. For understanding the growth performance of the economy range is more important than the point. From the range we clearly understand that Bangladesh has a moderately performing economy, not utterly disappointing but nothing to be euphoric about it.

Coming to population we should start with the 2001 census figure which was by and large acceptable to citizens in general. This was 130 million. We have also accepted that the population growth rate is 1.4 % per annum. If that be the case population at the beginning of 2011 would be around 150 million, much less than 165 million as announced by a development partner. This has been justifiably trashed by some minister but freely quoted by some others giving people the freedom to choose any figure in between to their convenience. One point, however, stands out: population of Bangladesh is between 145million to165 million. Closing in from both sides it may be around a narrow range of 150 to 155 million. Even if we settle at the lower end it is a daunting figure that makes our food security extremely vulnerable and our living environment perniciously deplorable. We will not be far away from reality if we work with 155 million as our present population. Working with the production figures, not egregiously disparate, one would arrive at the compelling conclusion that production of cereal crop is not less than 34 million tons. But for highly skewed distribution of income that should suffice for existing population of the country. The author firmly believes that there is no shortage of rice in the markets which are still replete with the produce of last season or last year. Our statistics, however crude and disparate, have not landed us in an impossible world. We can meaningfully use them in a real world situation. It bears mention here that food price or for that matter price of any product in the market does not depend on the quantity of its production only, it is a function of the complex interaction of many other factors like money supply, cost of production, border price and most importantly the market power of producers, traders, consumers and the government.

Statistics is not a precision science. It never boasts about crunching out the most correct and infallible numbers. Its forte is couching qualified statements within a confidence level made well known to readers. The statements are stochastic, probability based. They appear sound to enlightened readers because deterministic statements promising certitude are prone to be proved wrong in reality. Stochastic statements with minor imprecision may be helpful in making predictions because they present practically workable numbers and at the same time cautions the planners and others not to apply such numbers blindly without a careful review.

We are exasperated to see disparate numbers on the same subject. A single number will not, however, guarantee high quality and credibility of the numbers released by different agencies. Most of the developing countries are committed to mediocrity. Their reporting system, the collection and compilation of data and the articulation in presenting them endemically lack rigour and application. Things are done in a cavalier fashion. No serious effort is made to improve the quality of data through elutriating by expert consultations. A high profile expert committee supported by competent functionaries and professionals can improve the quality of data and minimize the problem of discrepancy before the numbers are released for public consumption. Numbers will speak out if we learn the art of speaking to them.

The writer is Chairman, Bangladesh Public Service Commission.