Where is the Economic Risk of COVID shock highest: developing countries

- Summarised from Ilan Noy, Nguyen Doan, Benno Ferrarini and Donghyun Park's article in COVID-19 in developing countries

The economic effect and the health effects of the COVID are not as correlated as we might think. For example countries like Seychelles and Fiji had less than 20 reported cases but being highly tourism-dependent would be badly hit. Other countries with more cases also differ in their capabilities to access resources required to pull their economy out of the recession. This study studies mainly the distribution of risk amongst developing countries because it is this category which is most vulnerable to economic risk overall.

To study the economic impact of different methods have been used:

• Coupling epidemiological with macroeconomic models, which are similar to the one’s used to study climate change, that is, the integrated assessment models (IAM’s). In the climate IAM’s the climate and economy are connected through temperature and productivity. Similarly in the pandemic, IAM’s the connection of COVID to the economy could be on the supply (productivity losses) or demand (lockdowns) side
• Using disequilibrium models (e.g. Mandel and Veetil, 2020) since the global COVID shock might impact and change structural parameters. Thus equilibrium models using the same structural parameters might be erroneous. Disequilibrium models, however, require constant input-output data
• Extrapolate impact of similar past events like SARS-2003 or global flu pandemic of 1918-19. The limitation of this procedure is that SARS was limited within Asia and the 1918-19 pandemic was much different from today in terms of the tools we have today to deal with

Therefore, this study used disaster risk modelling framework. The framework was taken United Nations Disaster Risk Reduction Office (UNDDR) which assess disaster risk using four concepts – hazard, exposure, vulnerability and resilience. The interaction between these have economic consequences.

These four concepts in the UNDDR are defined as –

• Hazard is the number of cases of infections due to COVID. The study hypothesises that the economic risk is decoupled from hazard (infection) risk. Thus, it's determined by exposure, vulnerability and resilience. I think this hypothesis was taken because – it’s a pandemic so will affect economies of most nations for sure, thus hazard to a large extent is homogenous. The argument the authors put up is that at the time of writing developed countries was worse hit than developing countries but the ripples originated by the former would be felt by the latter anyways. Also, the number of cases depends on the aggressiveness of testing, which differs due to the resource or political factors, thus might be a biased metric
• Exposure are the economic activities located in areas that are exposed to COVID or that are indirectly exposed by behavioural changes of people exposed to COVID
• Vulnerability is the adversity with which the COVID can affect the economy
• Resilience is the ability of the economy to bounce back after experiencing a shock. The degree of resilience is determined by the speed of recovery and ability to have minimum post-shock income losses

The study analyzes at the grid level ‘g’ depending on data availability or else at the country level. The risk was then measured as,

measured as, Each of the exposure, vulnerability and resilience is composed of multiple variables, based on literature measuring disaster risk and experience from the pandemic. Principal Component Analysis (PCA) was then conducted to compute a standardised index for each of the above-mentioned components. These are then used to measure risk as shown earlier. In simple specification weighting was kept equal, that is, (beta_i = beta_j). In other beta_i was based on a regression algorithm using the number of disability-adjusted life years lost to communicable diseases, in each country.

The study finds that the economic risk of a pandemic is not highest in the regions which have the greatest exposure (Peru, Russia, Turkey). It is highest in most of Africa, South Asia (esp. Pakistan, Nepal and parts of India) and Laos. This is because of their high vulnerability with low income and the poor healthcare system. East Asian regions (China and Vietnam) have lower risk due to their high resilience. This comes from less fractioned socio-cultural characteristics, high capacity for policy mobilisation associated with a high ratio of domestic credit to the private sector and high levels of government expenditure. Brazil, Mercosur countries, Turkey, China and Russia are also relatively safer economically as their economy is relatively more export-oriented than the vulnerable tourism sector. 

To find the highest risk the focus was then shifted to low and middle-income countries. It was found that amongst bigger countries, the Democratic Republic of Congo had the highest economic risk. Much of the Central African regions were concentrated in high risk as well. 

The risk was then estimated using the ad-hoc weighting scheme. DALY (Disability Adjusted Lost Years) measures the overall disease burden of a country and basically is the sum of years lost due to ill-health, disability or premature death. As previous DALY’s is obtained from past events, it could be useful to understand the role of hazard, exposure, vulnerability and resilience. The regression was run with country-level DALY as the dependent variable and exposure, vulnerability and resilience (along with a constant) as independent variables. The regression coefficients (weights and constants) are put back to calculate risk as, 

The spatial map with DALY weighting was found to be similar to the equal weighting case in terms of high-risk areas. However, much of Central Asia and South East Asia were considered less risky now as they are relatively poor but not densely populated too. With DALY as can be seen exposure weights higher and so grids with dense population have more economic risk. 

Although with the advancement of healthcare system relative to 1918-19 pandemic, COVID is unlikely to have such catastrophically death numbers, however, it may have bigger economic consequence. This is because of more globalized trade-investment flows, greater labour movements and tourism, and social media-induced behavioural amplifications, transmits the shock quickly and globally.

As an economic consequence, direct losses includes lost income and output due to health and illness as well as increased medical expenditures. Cost of life when calculated using statistics (value of statistical life) lead to lower direct losses as compared indirect losses. This is especially true in countries where COVID has not spread indiscriminately but the economy is exposed to the shock (e.g. tourism-dependent Fiji). 

Apart from the measurable metric like public health expenditure, factors like social capital and economic informality may play an important role in determining economic risk. Economies with larger informality make it difficult for the government to intervene productively. Also, societies with a higher degree of bonding social capital can reduce the distributional economic consequences etc.

Click here to read the ebook

Images were taken from the paper itself

Good videos to understand PCA: [1], [2]. [3]