Shocks and economic disruption: Networks, insurance, and propagation

How might the economic consequences of the global pandemic spread through an economy? This column reports evidence from a household survey of village communities in Thailand, which sheds light on how a large shock experienced by one household may be mitigated by informal insurance networks. If not insured, however, it will propagate to others through local supply chains, leading to longer-term low-productivity structural shifts. The research findings indicate the importance of a robust public response and improved market infrastructure for shocks like Covid-19.

The global pandemic is first and foremost a public health emergency. But with many businesses closed and huge numbers of people under lockdown, it will have enormous economic consequences too.

The consequences will not be evenly spread. Some workers can do their jobs from home with little disruption, while others are hard hit, because their jobs cannot be done remotely or because they work in sectors with large drops in demand – or both. Those direct effects may then propagate in complex ways.

It is important to understand how the economic consequences will spread through an economy. While this is a highly complex question, our research, which analyzes data from a household survey in Thailand, sheds light on one important aspect: how a large shock experienced by one specific household propagates to others.

Households within a village community are interconnected: what happens to one household matters for others

Households in developing countries are both consumers and, as small business owners, producers. In addition to purchasing consumption goods, households transact labor and other productive inputs with each other. They also borrow and lend, and give and receive gifts. These interactions among households constitute local financial, supply chain, and labor networks within a village community.

In an ideal world in which households had perfect access to savings, credit, and insurance, an unexpected event that increases expenses or reduces income would be fully financed and would not affect production activities. In such cases, there would be no propagation to suppliers and workers through supply chain and labor networks.

But in reality, households may lack savings or access to credit and insurance to cover unexpected shocks. A household that experiences an isolated shock (an unexpected event that does not directly affect other households in the village) may need to adjust its production – reducing labor hiring and purchases of material inputs.

This response in turn transmits the shock from one household to others connected through the supply chain and labor networks, as those other households experience a drop in sales revenue, an increase in inventory, and a decrease in labor income. Effectively, a household-specific shock propagates and becomes systemic in the village. We study the dual role of insurance and propagation provided by village networks by leveraging 14 years of monthly panel data.

Household-specific shocks affect production decisions in the absence of complete markets

We find that household-specific shocks, captured by sudden increases in a household’s health spending, are partially smoothed on the consumption side through transfers from other households in the village. This result highlights the importance of local financial networks in providing insurance against household-specific shocks.

But while local networks of gifts and loans provided insurance, this informal insurance was only partial: on average, incoming transfers covered two-thirds of the increased spending needs of the households experiencing a shock.

Moreover, these shocks do not hit everyone equally: some households are well covered by incoming gifts, as households help each other. But households outside this gift network get little. Consequently, to smooth food consumption fully, some entrepreneurs who experienced expenditure shocks substantially reduced input spending and their demand for paid external labor.

Ironically, as they scaled back their production, they also reduced the work hours of (non-sick) family workers allocated to family businesses. One implication of this result is that in periods of crisis, those households who are left out of informal social protection networks may be hit the hardest.

Household-specific shocks propagate to other households through production networks

Businesses closer to households experiencing shocks in the supply chain networks experienced reduced sales due to their (indirect) exposure to the shock. Similarly, workers closer to households experiencing shocks in the labor network experienced a fall in the probability of working for local employers and a reduction in total hours allocated to wage labor.

As a result, total household labor earnings declined. A simple calculation suggests that the total magnitude of these indirect effects can be as large as, or larger than, the direct effect of the shock.

These results imply the existence of frictions in the markets for goods and labor: suppliers may not be able to find new customers when their clients face shocks; workers may struggle to find new jobs when their employers suffer. Informal relationship-based supply chains have their limitations.

This suggests that the closure of businesses due to economic lockdowns can have long-term consequences. Likewise, the effectiveness of post-pandemic policies can be magnified through positive spillovers if existing links are preserved or replaced.

Household-specific shocks propagate and become systemic with adverse structural shifts

Households who are indirectly affected do not appear to buffer these shocks by receiving gifts or loans. Instead, households shifted away resources from local retail businesses and off-household labor activities vulnerable to shocks from other households – towards farm-related businesses – activities that provide food and also tend to sell most of their output outside the village, though arguably at lower returns.

Access to insurance not only mitigates the direct impact of household-specific shocks, but also reduces the degree of propagation to others and these shifts. Thus, expanding insurance to include more formal alternatives may lead to social welfare improvements larger than the direct gains.

Once shocks become systemic, there is a limit to local insurance, and assistance from the government or market-based insurance contracts from (national-level) providers are needed. Likewise, centralized platforms that reduce dependence on local supply chains would also be helpful. 

In sum, inter-household linkages are important for households in developing countries, but these links cut both ways. On the one hand, these networks can partly insure major shocks through gifts and loans. On the other hand, networks also propagate shocks: when households experiencing shocks pull back on production activities, linked households are hurt by the reduced demand. This causes large household-specific shocks to become de facto aggregate, leading to lower-return structural shifts. 

Immediate and longer-term responses

These findings suggest that a robust response to shocks like Covid-19 is particularly important: it can help mitigate economic contagion caused by a pandemic. Households who may not appear to be directly affected by Covid-19 are likely to be indirectly affected from the economic downturn, making broad-based safety nets and better market infrastructure particularly important.

By reducing economic propagation, public and market-based private insurance have benefits that extend beyond those who are directly eligible. For example, measures that assist small and medium-sized enterprises to cover shocks and remain operational during the Covid-19 lockdown help to preserve local supply chains and labor relationships. Not only can they mitigate shocks transmitted to other businesses and households, but they can also reduce the delay in the recovery as there are costs to finding new suppliers and workers.

Longer-term efforts such as centralized platforms that improve and broaden infrastructure can help protect us against future episodes.

 

Authors:

Cynthia Kinnan is an assistant professor of economics at Tufts University and a Research Associate of the NBER. Her research lies in the areas of economic development and social networks.

Krislert Samphantharak is Executive Director of Puey Ungphakorn Institute for Economic Research (PIER) at the Bank of Thailand and Associate Professor at the University of California San Diego.

Robert M. Townsend is a theorist, macroeconomist, and development economist. He is the Elizabeth & James Killian Professor of Economics at Massachusetts Institute of Technology.

Diego Vera-Cossio is an economist in the Research Department of the Inter-American Development Bank. His area of interest is development economics.