India’s Supply Chains Unchained

The issue of supply chain disruptions in food markets took center-stage during the nation-wide lockdown in India. Matt Lowe and Ben Roth assess the market outcomes in wholesale markets in 24 major states for 256 commodities during the different phases of lockdown. A thorough quantitative assessment of the food markets was certainly needed. The authors quantify the near sudden stop type outcome that happened immediately after the lockdown as an unanticipated shock. As expectations caught up and policy uncertainty reduced, volumes in markets gradually recovered and are now normal or near normal – Devesh Roy, series co-editor and Senior Research Fellow, Agriculture for Nutrition and Health (A4NH).

With the outbreak of coronavirus pandemic, Prime Minister Narendra Modi announced the beginning of a national lockdown on March 24. During the first phase of the lockdown that lasted until April 14, we estimate that the total volume of food arrivals across India had reduced by around 64%. To reach this estimate, we use publicly available daily data for 266 commodities traded at 1804 markets in 24 states of India. Despite the fall, volumes have gradually recovered, so much so that we estimate that aggregate volumes were close to the levels of 2019 by the end of May (Figure 1). We use similar patterns across six major commodity groups (Figure 2).

Figure 1: The lockdown caused volumes across Indian markets to plummet

Figure 2: Volumes evolved similarly for each of the six major commodity groups

The initial negative shock and the subsequent recovery were driven by changes on both the extensive and intensive margins. On the extensive margin, many agricultural markets shut down and later re-opened (Figure 3 tracks a proxy of the number of functioning markets, i.e., the number that report any data to agmarknet.gov.in). On the intensive margin, those that stayed open initially operated at lower volumes, and later recovered.

Figure 3: The number of functioning mandis plummeted and recovered

We had previously argued that several links in the supply chain were vulnerable during the first phase of the lockdown. Uncertainty in the rules of inter-state travel made transport across state lines difficult. Many migrant laborers returned homes, constraining the labor supply necessary to keep the wholesale markets functioning. Also, retail vendors substantially reduced their procurement, as weekly markets had shut down. Vendors were forced to roam the streets to look for customers, resulting in far more work and much fewer sales.

Given these initial vulnerabilities, the recovery is striking. Nevertheless, the nation-wide improvement may mask large differences across states. Are the markets in all the states recovering or is it the case in only those states that have fewer cases of coronavirus? To answer this question, we estimate the size of the volume shock separately for each state, and separately for the first phase of the lockdown, and compare it with the subsequent phases of the lockdown. Notably, the first phase volume shocks are increasing in the incidence of confirmed coronavirus cases, as of June 1 (Figure 4). So, while the lockdown was nation-wide, the initial effects on supply chains have been more severe in heavily-affected states.

Figure 4: States with more coronavirus cases had bigger supply chain disruptions during the first phase

Nevertheless, the story since the start of phase two on April 15 is very different. Figure 1 already presents that volumes of food arrivals have gradually recovered across India. Figure 5 shows that the volume shock from phase two onwards was not correlated to the extent of coronavirus cases in the state, i.e., volumes may have recovered regardless of the extent of coronavirus. This surprising non-correlation might signal toward the fact that even though coronavirus continues to spread in India, supply disruptions may not always be disrupted.

Figure 5: Volume shocks were not correlated with coronavirus cases during the second phase

It is most likely that state-specific variation may still mask important spatial variation. The question that may arise is how do particular districts within a state fare if they had more cases of coronavirus? We approached this question by splitting the districts into two types, those with at least one COVID-19 death by June 1 (high-COVID-19 affected districts), and those with no COVID-19 related deaths (low-COVID-19 affected districts), with roughly 50% districts in each category. Next, we estimate the additional effects of the lockdown on aggregate volumes in high-COVID-19 infected districts in comparison to low-COVID-19 affected districts in the same state.

Against our expectations, there is no evidence that districts with more number of COVID-19 cases are affected differently from districts with less number of cases in phase one or later (Figure 6). All the ‘pre-trend’ coefficients are close to zero and insignificant, demonstrating that 2020 volumes relative to previous years were evolving similarly in high- and low-COVID-19 affected districts prior to March 24. The ‘post-trend’ coefficients are close to zero, and insignificant too. Volumes in previous years continued to evolve similarly in the two types of districts following March 24, despite the districts facing different on-ground coronavirus-related fatalities. Given these findings at the district-level, the state-level variation in the effects of phase one lockdown is more likely due to variation in each state’s implementation of the lockdown, as compared to the more local variation, due to the fear of contracting and spreading coronavirus among market laborers.

Figure 6: The lockdown effects in districts with high COVID-19 and low COVID-19 districts were similar within the same state

In India, nation-wide volumes see a similar pattern in its wholesale prices. During the first phase of the lockdown, wholesale prices, conditional on availability, have increased by almost 10% (Figure 7 shows these prices trend after controlling the food type-by-mandi combination). This price increase is exactly what basic economics would predict in the presence of a large negative supply shock, at least one that is not fully offset by a reduction in demand. Having said that, even this price increase understates the overall effect on any defensible price index. With the non-functionality of many markets, some prices are effectively infinite, since the food is not available at all. From phase two onwards, wholesale prices have also returned to a downward trend. As with volumes, the effects of the lockdown on prices appear to have been temporary.

It is difficult to say whether wholesale prices are now actually lower than what it would be otherwise. However, lower prices are also a plausible implication of Econ 101. So, while supply may have rebounded, if demand remains low, as seems likely, prices will also be low.

Figure 7: After initial increase in wholesale prices, prices have returned to trend

The pattern of rising and then falling wholesale prices holds for most of the major commodity groups (Figure 8), with the notable exception of spices, which did not see a lockdown-induced price increase at all. One possible explanation is that the demand drop in spices was large enough to offset the supply disruption. This is because non-perishability and relative non-necessity of spices meant that demand was more elastic than it was for other commodity groups.

Figure 8: Wholesale prices evolved similarly for all commodity groups, except for spices

The nation-wide recovery of agricultural markets is encouraging, but also somewhat puzzling. Why have volumes recovered nearly fully, despite the continued spread of coronavirus, and the fear that densely populated mandis will increase transmission? We offer two tentative hypotheses. The first is that India has surprised the world by having relatively few confirmed cases (2,47,000 as of writing) and a relatively low case fatality rate (2.8%). Even with under-reporting, the situation appears to be far less dire than many predicted, for reasons that are not entirely clear (though India’s relatively young population certainly helps). The low expectations and more benign reality may explain the volume shock and subsequent recovery. The shock was large because states and markets took drastic measures initially. The recovery was gradual but complete since initial fears were not realized.

A second point, and less sanguine one, relates to a different realization, which is that coronavirus is a long-term affliction that can only be delayed, not deterred, by a strict lockdown. A long-term disruption to supply is not sustainable as people have to eat. This logic provides a motivation for wholesale markets returning to full business, even though the virus is going about its business too.

As ’Unlock 1’ phase gets underway, we might wonder about the future. Will the recovery in food markets continue? Or will localized supply shocks emerge wherever new COVID-19 clusters appear? While the soothsaying of economists tends to be perilous, we lean towards optimism. In particular, the striking fact in the data is that post-lockdown, volumes have not been differentially impacted in districts in each state that faced deaths from coronavirus as compared to those with zero deaths. Coupled with our other findings, it appears that supply disruptions are more likely driven by state and central government lockdown policies, and less by localized exposure to the virus and the overall fear of the disease. If state and central policies continue to relax, we suspect that supply chains will remain unchained.

Fortune-telling aside, the ongoing recovery of food volumes is good news for the food security of urban dwellers across India. Provided that coronavirus remains at bay, and wholesale markets do not spark new virus clusters, this too should be a positive news for India.

The content in this article is based on work by Matt Lowe, Assistant Professor of Economics, University of British Columbia, GV Nadhanael, PhD candidate in Economics, University of British Columbia, and Ben Roth, Assistant Professor of Business Administration, Harvard Business School, with field assistance from Simranjeet Dhir and Toshan Majumdar, JPAL. The analysis and opinions expressed in this piece are solely those of the authors.

This blog has been published as a part of the International Food Policy Research Institute (IFPRI), South Asia, blog series on analyzing the impacts of the COVID-19’s pandemic on the sub-national, national, and regional food and nutrition security, poverty, and development. To read the complete blog series click here