Why we can’t take World Bank’s projection on GDP growth seriously

The GDP growth predictions by various multilateral development partners have been an issue of discontent for a long time, particularly in the case of Bangladesh.

No doubt that GDP growth is a sensitive factor for the governments, especially in developing countries, as it is to date the only economic indicator that can exhibit the outcome or success of any government's economic development policies and interventions.

However, in most cases, the sole authority for the GDP calculation remains in the hands of government agencies, mainly in the Bangladesh Bureau of Statistics (BBS) for Bangladesh, which has the access to all data relevant to GDP estimates.

The GDP estimation is a daunting task, considering the involvement of data from various sectors and sub-sectors of the economy as well as various predicted indicators based on nationwide survey results done at different periods. Therefore, without having access to that data, it is very difficult for anyone to come up with a precise GDP estimation.

However, difficulties in having access to raw data on GDP lead various organisations to predict GDP growth using various macro-econometric or macro-financial models.

A few days back, the World Bank has released its projection on GDP growth for various countries, including Bangladesh, for the current fiscal year 2021. The Bank has projected that the Bangladesh's GDP will grow at 1.6% in FY21. The same was predicted for the partly Covid-19-affected preceding year, but they had later revised it at 2%.

The predictions on GDP growth by other multilateral agencies are quite different compared to the World Bank's one. For example, for the last year, the Asian Development Bank (ADB) predicted it to be 4.5%, the International Monetary Fund (IMF) predicted it to be 3.8% and the BBS's provisional estimate was 5.24%.

For the current year, ADB forecasted it to be 6.8% and the BBS predicted at 8.2%. These varying results on GDP projection from various multilateral donor agencies attributed to the fact that the model can predict different estimates based on assumptions as depicted in different estimates of these institutions.

More so because the World Bank's estimate stays very far from reality as well as from other organisations' estimates.

In the last fiscal year, Bangladesh went on a relaxed lockdown – in the name of general holidays – for April-May to stem the spread of coronavirus. After that from June, the government has started opening everything partially.

Therefore, only one quarter of the last year – April-June, FY2020 – was affected by the pandemic. If we assume that the previous three quarters before the pandemic demonstrated normal economic activities, can the loss in the only one quarter contribute to an 80% decline in GDP growth, assuming it at 8%? The answer is obviously no.

If we look at various indicators for the last fiscal year, the volume of exports declined by about 47% in the last quarter though the yearly decline was only about 11%. Remittance started growing significantly from June, which accounted for 35% growth.

During the lockdown, rural economy, particularly the agricultural sector, has remained mostly unaffected from the Covid-19 pandemic. The country saw a bumper Boro production last year, showing almost a similar growth in crop production achieved in the preceding year.

The private sector credit grew by 10.4%, which was just 4 percentage points lower than the previous year's, indicating a reasonable investment growth. The performance of all these indicators do not qualify the growth rate that was projected by the World Bank for the last fiscal year.

Based on my own experiences of macro-econometric modeling exercises for a long time, I can also highlight some technical aspects as to why a model might produce misleading results if proper care is not taken while simulating tonnes of behavioural and structural equations.

A model basically considers four economic blocks – macroeconomic bloc, fiscal or govt block, monetary block, and external block – in order to capture the interdependence among these sectors and their contribution to GDP. 

A set of putative assumptions are usually made to capture the real economic phenomena that have consequences on the growth aspect of an economy, and therefore, warrant very careful considerations and a greater understanding of the economy.

For example, if one wants to relate tax (including VAT) revenue with production, it will be misleading considering tax avoidance and corruption that are involved in it. If private sector credit growth is considered fully as an indicator of private investment, it might mislead because a part of credit becomes non-performing from the very beginning.

Similarly, the quantum index of production is not a good predictor for various reasons. Furthermore, the main critique of this kind of model, coined as the "Lucas Critique", is that it exerts no effort to capture the underlying perception of the economic agents regarding the expectations of policy interventions.

My observation also is that even if a model gives a better out-of-sample forecast based on historical data, it is difficult to come close to the actual level because various structural breaks are involved in the data.

Moreover, in a crisis situation like the current pandemic, the assumptions on various shocks could be incorrect as the situation has been changing rapidly. For example, even though some epidemiologists had earlier predicted a huge loss of lives in Bangladesh due to the Covid-19 pandemic, which by now, has been proved wrong.

In the last three (July to September) months, exports showed a 3% growth along with substantial growth in remittance inflows. It is expected that if Covid-19 remains restrained in Bangladesh paving the way for workers to continue their work, the export sector, as the supplier of low-end products, will remain buoyant in the coming Christmas and New Year's Eve in Europe and the USA.

A higher volume of remittances provides an indication that it plays a catalytic role in offsetting the income losses of families and friends of remitters due to the Covid-19 crisis. In other words, a higher volume of remittances also indicates an acceleration of effective demand in rural areas.

Some recent surveys also showed that firms have returned to more than 80% of their production capacity by now. The service sector has been almost in its full operation.

With all these positive signs of strong recovery of the economy is there any reason to take the World Bank's projection on Bangladesh's GDP growth seriously? If the pandemic situation does not deteriorate further the resilience of the economy with hardworking and aspirant people deserves much better growth than the predicted one.

The author is a Research Director at Bangladesh Institute of Development Studies (BIDS).