Adequate data needed to capture true impact of Covid-19

The speed at which Covid-19 has spread across the globe is unforeseen and its impact in terms of health problems, lack of adequate health services and death was immediately evident.

Bangladesh, like many other countries, attempted to contain the pandemic through shutdown of normal economic activities and public life during April-May. Then the country went through the gradual reopening of economic activities.

The economic crisis following the shutdown of activities included loss of income and employment.

The government responded to the problems with announcements of transfer programmes for low-income groups and incentive packages for economic activities. Implementation exercise of these policies revealed that there is inadequate data for identifying the target groups.

The impact of the shutdown and related economic crisis has drawn attention of researchers and there have been many studies on the subject. These were based on secondary data and projections using past data.

These may be useful for short-term policy interventions when there is no scope of carrying out national sample surveys.

However, the limitations of these estimates are obvious. Many parameters have changed during the pandemic and it is difficult to take those into account.

Estimates of poverty incidence during the immediate aftermath of the pandemic have been attempted and these are based on existing data and various assumptions. The research studies and media reports have highlighted severe adverse effects of Covid-19 on employment and poverty.

Many organisations have also conducted "rapid surveys" for generating estimates of poverty and income of poor households. Rapid surveys can yield quick results and are useful for short-term policy adoption, but cannot go in-depth into the details of source of income, type of employment etc. The rapid surveys could not reach a nationally representative random sample.

Therefore immediate steps are required to generate adequate data to capture the details of the downturn of the economy and the phases of its recovery and the processes through which these cycles have taken place so that a national picture emerges. Otherwise valuable information will be lost forever. Such information may be relevant not only for present day policy making but also as historical data, which may guide future generations in their fight against economic and social downturn, especially when induced by pandemics like the present one.

The following discussion highlights the need for generation of data on employment and poverty for a proper assessment of short-and medium-term impact of Covid-19. Most of it is addressed to the national statistical organisation (BBS) but will also be relevant for other organisations who take up similar surveys.

The issues raised here can also be relevant for surveys covering impact on other socio-economic aspects of households, activities of enterprises, etc. The following discussion focuses on data on a few economic and social dimensions only and it does not form a comprehensive list.

The concerns raised here may be relevant for other developing countries, which do not have in place a regular quarterly/annual data generation system. High income countries collect monthly/quarterly and annual data on key economic indicators, including unemployment, and thus the impact of Covid-19 unfolds continuously.

Survey of labour force, employment and unemployment

The Labour Force Survey (LFS) of Bangladesh Bureau of Statistics (BBS) covers more or less a comprehensive set of issues, which are considered necessary to understand the functioning of the labour market. But the coverage is adequate only in a normal situation and cannot capture the effect of a pandemic like the present one.

Therefore there is a need for reviewing the issues covered as well as the methodology that is used in the survey. Following are some aspects of labour market, which needs attention.

Loss of employment associated with Covid-19 may take various forms like work sharing and low hours of work resulting in underemployment, low productivity per person, termination from jobs, temporary loss of job, lower wage, deferred wage, denial of other benefits as provided in the country's labour law.

Contracts in the case of paid employment may be formal and written, informal and unwritten or may take in-between forms and during and after Covid-19, the form of contract may change to worse.

Loss of employment and rising unemployment are likely to have a discouraging effect on labour force entrants, who may opt out of the labour force and join the category NEET (not in employment education and training).

The discouraged workforce can pose a challenge not only as a lost potential for economic activities but also can result in social problems.

LFS adheres to definitions provided by ILO. This is useful for international comparison. But this may often result in misleading interpretation in national contexts. To capture the impact of pandemic, data collection efforts can be meaningful if these are based on concepts which are suitable in the socio-cultural context of the country. Therefore, modification of some of the definitions should be considered.

Timing of the surveys: The latest round of LFS has been completed in June 2017. The next one, even if starts early next year will mean an interval of 3.5 years. Thus, an analysis of the impact of Covid-19 will miss out information on both ends.

There is no updated information on 2018, 2019, which could form the base year for comparison. One has to depend on the assumption of continuation of parameters of 2017 and make projection for 2019 as the base year.

The survey will commence one year after the beginning of the pandemic. This will be useful to capture the medium-term impact but cannot capture the impact and the processes associated with it in the initial peak pandemic period (hopefully there will be no second wave) and the government responses during those months. To understand the process of change of the labour market in the short run and its comparison with the medium term, the survey will have to depend on memory recall.

Appropriate questions should be included in the survey to generate data through such recall. It will be a complex matter and will require quite some details including recall of three points of time: immediately before Covid-19, during April-June 2020 and at the time of the survey in 2021.

One must balance the usefulness of detailed data generation through recall with the likely decline in quality of such data.

Poverty and household income data

During the last three decades, Bangladesh has been assessing poverty incidence on the basis of data of Household Income and Expenditure Survey (HIES). This is usually conducted at 4/5 years of intervals.

The latest round has been completed in 2016. The next round has not yet started and may be expected to commence in 2021. In the normal situation this serves the purpose of poverty assessment and analysis of the factors behind the changes.

However, with the spread of coronavirus, there has been evidence from rapid surveys that poverty has accentuated. But these surveys are not based on nationally representative samples. Therefore, these cannot be compared with the national survey results. Then how would one arrive at national estimates of poverty during the various phases of the pandemic? HIES poverty estimates use detailed information of consumption of individual food items and other consumption items during a short reference period. It will not be possible to collect such details through recall method for periods three to six months earlier. The method of obtaining idea about consumption standard or poverty during the pandemic is not thus, quite straightforward.

One may resort to qualitative indicators which may be easily recalled or self assessment of the extent of decline of consumption. In addition, there may be recall based data on consumption of major items/groups of items. Questions on the extent of decline/increase in consumption of these items may provide a basis of comparison even if it does not provide estimates of poverty based on standard methodology. Indicators used in the international multi-indicator poverty assessment method may be taken into account in this context.

Organisational issues

A comprehensive plan for data generation on "Pandemic and its impact on health, economy and society" should be adopted. The major responsibilities will of course be with the national statistical organisation, BBS. Nonetheless, large surveys may not be able to go sufficiently in-depth and therefore cannot be a substitute for small and micro surveys.

The author is the executive chairperson of the Centre for Development and Employment Research and a former member of the board of directors of Bangladesh Bank.