Data on poverty levels in Bangladesh will be available by the end of March this year as the Bangladesh Bureau of Statistics (BBS) completed its Household Income and Expenditure Survey (HIES)-2022 last Friday.

The BBS, the lone government agency for compiling data, drastically reduced its household sample size to 14,400 for the latest survey, which comes a year late, compared to 46,080 previously.

The HIES is a core survey of the government, not only for its ability to measure poverty, but also for having a wide range of socioeconomic data crucial for policymaking and research.

It gives detailed statistics on household income, expenditure and consumption, standard of living, and nutritional status, among other things.

Field work for the HIES was completed between January and December.

"This year, we brought many changes and a strong monitoring system was put in place. The final report will be published in December 2023," said Mohiuddin Ahmed, project director of the HIES 2020-21.

Meanwhile, the number of food items listed in the survey has increased from 149 to 265 while that of non-food products and services has gone from 216 to 441, he added.

Regarding the change in sample size, Hossain Zillur Rahman, executive chairman of the Power and Participation Research Centre, said the BBS usually uses between 10,000 and 15,000 households for the HIES.

"Only the last survey in 2016 saw a dramatic jump in analysis," he added.

The economist then slammed the delaying of BBS data, stating that the survey should be conducted with a three-year interval.

"Following the last survey in 2016, we may get the final output at the end of 2023, when it will be too late to get the real picture," he said.

The BBS generally releases the HIES every five years.

Prof Selim Raihan, executive director of the South Asian Network on Economic Modeling, said determining the right sample size is very crucial considering the different aspects of population size and geographic locations.

The 2016 survey was planned to reflect the various aspects of the huge population. So, the question arises of whether the current sample size is a proper representation or not.

"If the sample size does not provide a proper representation, we would not say the outcome is accurate," he added.

HIES Project Director Ahmed then said the BBS has conducted 16 rounds of the survey till 2016. In each round, the sample size was intended for divisional-level estimates with HIES 2016 being the only exception.

"The sample size of 2016 was designed to provide district-level estimates. So, it showed a 'big jump' in the sample size," he said.

Poverty statistics and sampling experts suggested that handling such a large sample size in a round-the-year survey like HIES is challenging and results in a higher level of non-sampling error.

To avoid this, the BBS decided to conduct the HIES-2022 with 14,400 households, Ahmed added.