Questioning official statistics is a standard practice, especially for research and analysis. It doesn't mean that questions are intended to undermine national attainments. Unfortunately, it has become increasingly difficult in Bangladesh to raise such questions. Policymakers in general are usually reluctant to entertain any valid criticism of official statistics particularly when these are not pleasant for them. But the quality of official statistics is under questions for several reasons.
One of the most critical things is estimation of national income as well as growth of Gross Domestic Product (GDP). The very little difference that is found between projection and final estimation of growth indicates that the country's statistical body is quite efficient in making estimation on national economy. For example, in FY15, growth was initially estimated at 6.51 per cent while final estimation put it at 6.55 per cent. In the current fiscal year, preliminary estimation showed 7.05 per cent growth against the target of 7.0 per cent. Â
But industrial growth estimation is, to some extent, confusing. According to preliminary estimation done by the Bangladesh Bureau of Statistics (BBS), industrial sector posted 10.1 per cent growth in FY16. But the Quantum Index of Industrial Production (QIIP) - released so far in the official website has updated data for eight months (July-February) - shows that growth in medium and large-scale manufacturing sector registered 12.50 per cent growth during the eight months on average, despite decline in growth in January and February.
While the QIIP numbers measure volumes, GDP numbers show value addition. So, synchronising values and volumes are important. The question is whether the statistical bureau is rightly doing this or not.
The Bangladesh Bureau of Statistics (BBS), the national statistical body, is not very strong due to negligence by the policymakers over the years. It is still struggling to improve its quality. However, BBS has advanced a lot in last one decade, especially after reintroducing the statistics division under the planning ministry in 2010. The division was abolished in 2002.
Annual budgetary allocation for the BBS and statistical offices across the country is moderate. In the proposed national budget for 2016-17, total allocation for the statistics and informatics division stands at Tk. 5.0 billion of which Tk. 3.45 billion is for the BBS. Allocation for divisional statistics offices is Tk. 25.7 million while allocations for district and upazilla offices are Tk. 16.2 million and Tk. 66.14 million respectively.
In the new fiscal year (FY17), a good number of projects are scheduled to be implemented. These include: implementations of the national household database project, national household income and expenditure survey project and digitalisation of data and statistics. Â
Improving the quality of statistics is a challenging but essential task. It requires good investment, comprehensive plans and meticulous efforts. Responding and clarifying the questions about official statistics is also a part of ensuring quality statistics. Â
Questioning official statistics is a regular practice in many countries including India. Two months back, two Indian economists, CP Chandrasekhar and Jayati Ghosh, in an article in the Business Line pointed out several flaws of the Indian official statistics. They pointed out that wide differences in official data lead to questions about their relevance for decision making. They termed the official figures 'ostensibly meticulously constructed.'
The left-oriented Economic and Political Weekly (EPW) is harsher in criticising the official statistics. In the editorial comment of its latest issue (June 11, 2016), it said that the latest data on India's GDP has raised more questions than providing answers. The editorial titled 'Lies, Dammen Lies and Statistics', said, "Window dressing may yield short-term political benefits. But the damage that fudged figures can cause to the credibility and image of the government is potentially huge."
Again, early this month, Indian business daily Mint ran a critical report on the latest quarterly release of GDP growth. When Indian policymakers are exited about 7.9 per cent growth in the first quarter of 2016 (ended on March 31, 2016), the report pointed out that the growth would be actually 3.9 per cent if the 'discrepancies' are left out of the GDP data. The newspaper report mentioned that the growth in 'discrepancies' was responsible for a large part of the growth in GDP, measured at constant prices during the quarter under review, and actually contributed 51 per cent of the achieved growth.
Now, what is the response from the government? The chief statistician of India's Central Statistics Office (CSO) told Mint that they would eliminate the 'discrepancies' and publish 'supply and use tables (SUT) which will try to map the sources of production of items and their end use.' Thus, there is an accommodative approach by the Indian officials to deal with the questions about or criticisms of official statistics.
It is to be noted that 'statistical discrepancy' in the GDP calculation of Bangladesh declined significantly in the current fiscal year. According to BBS data, gross domestic expenditure stood at Tk. 8827.52 billion in FY16 while gross domestic production stood at Tk. 8830.54 billion. Thus, there is a discrepancy of Tk. 3.02 billion which was Tk. 136.15 billion in FY15.  Discrepancy is generally the difference between the output side of GDP and the expenditure side. Compared to the output side figures, expenditure side figures are mostly indirect estimation, except for international trade where direct data is available from the central bank.
It is not clear how discrepancies have been reduced within a year. In fact, this is a question one may ask and statistical authority needs to clarify. The policymakers should be knowing that economic data and statistics are public goods. So, official statisticians need to do a better job of explaining the data. It is not necessarily because their numbers are wrong but because credibility of the numbers has to be established. Policymakers should ensure fully professional work in the statistical bureau.
asjadulk@gmail.com
Â