When the note ban was announced a year ago, many feared that it would hit agriculture and informal sectors the hardest given the widespread use of cash for transacting in these sectors. Based on analysis of data from 2,953 mandis across India for 35 major agricultural commodities for the period 2011-2017, this column finds that there are lingering impacts of demonetisation on farmers and adverse distributional consequences overall.

One year has passed since India’s experiment with demonetisation rendered 86% of the currency invalid. Even at the time the note ban was announced, many feared that a monetary contraction of that scale would hit the informal sectors and agriculture the hardest given the widespread use of cash for transacting in these sectors. Despite the modest contribution of agriculture and allied sector to GDP (gross domestic product) (16.1% in 2014-15 at 2011-12 constant prices), almost 54.6% find employment in this sector¹. A disruption to sale of produce would therefore potentially have huge implications for rural incomes and lead to broader impacts for the economy as a whole.

Early evidence based on field reports seemed to confirm these fears (see here, here and here, for example). A fortnight after the event, we wrote that official data too showed an implosion of trade in the regulated markets (mandis), as indicated by a steep fall in arrivals. Yet, some observers (see here and here) argued that these fears were perhaps exaggerated and the impacts were small and insignificant. The Economic Survey 2016-17 too noted with cautious optimism, based on rabi ²sowing, that the impacts of demonetisation on agriculture might have been overstated³. Unfortunately, there is not much greater clarity today than there was at that time on the causal impact of demonetisation on farmer welfare. This is in part because of the challenges in securing appropriate data and the methodological issues in isolating demonetisation from other competing explanations.

In a recent paper, we use data comprising 8.5 million unique observations on daily arrivals and prices from 2,953 mandis across India for 35 major agricultural commodities⁴ for the period 2011-2017 to analyse the impact of demonetisation on the total value of domestic agricultural trade⁵ in the mandis and its recovery over the three months period following demonetisation (Aggarwal and Narayanan 2017). In the absence of data on farmer incomes or their consumption expenditure, we believe that the value of domestic agricultural trade passing through regulated mandis offers the best, if coarse, proxy for farm incomes in a context where few farmers store for extended periods and most sell their marketed surplus soon after harvest. We then decompose the effects to examine the impacts on arrivals and prices.

The challenges in measuring the impacts are manifold. Simple before-after 8 November comparisons for the year 2016-17 are likely to be contaminated by seasonal trends. At the same time, a comparison of post-8 November figures for different years would not make sense, since no two years are like one other. Several confounding factors such as bumper harvests (which was the case in 2016-17), and cross-border trade policy changes such as exports ban or imports are also likely to have an impact on the value of agricultural trade domestically. To account for these we use the changes in value of trade post-8 November and pre-8 November over the crop year July-June for the years 2011-12 to 2015-16 to represent the counterfactual, that is, the change in the value of trade that ought to have been in 2016-17 in the absence of demonetisation and measure the extent to which the value of trade in 2016-17 (the year of demonetisation) departs from this number⁶.

The hypotheses

Conceptually, an acute shortage of cash in marketplaces such as mandis, where cash is the dominant form of transaction, is expected to result in a reduction in demand from traders, and thus a reduction in traded quantity as well as prices. Farmers too may be unable to pay to bring the produce to markets, on account of cash constraints, reinforcing the decline in arrivals. Further, if farmers anticipate a fall in prices and thus hold back the produce from sale, supply would contract; provided this counterbalances their increased need for cash, this would push them to consider selling produce sooner rater than later. Such supply contraction could however put upward pressure on the prices. Thus, while the impact on arrivals is likely to be negative, the net impact on prices depends on the relative strength of the contraction of demand and supply. The impacts on the value of trade in turn depend on the strength of impacts on arrivals and prices.

There are several reasons however why one may not see the expected impacts of demonetisation in the data. Many creative ways to circumvent the ban surfaced in the weeks after demonetisation. For example, our field visits revealed that in many mandis , old currency continued to be accepted for payment, at a discount, despite the ban. This would get reflected via lower prices if they were recorded as such. There were also reports that those who had stashes of old currency, possibly black money, were buying up agricultural produce rather than deposit these in banks. Most often however, we found that goods were passing through but not money, so that farmers, agents and traders were transacting on credit. Sometimes multilateral arrangements evolved where farmer bought inputs for the impending agricultural sowing from family members of traders who they had just sold to on credit, thus settling the transaction in kind. In each of these cases, one would not expect to see a sharp impact on agricultural trade – values, prices or arrivals. Likewise if successive notifications easing restrictions on access to new currency released constraints on cash, we would see a muted impact. Further, not all agricultural trade passes through mandis and there is significant undocumented bilateral trade through village traders. If itinerant small traders who pick up produce at the farm gate were themselves cash-starved, it is possible that trade that would have otherwise occurred locally within the village, made its way to the mandis. Similarly, if farmers are cash-constrained and are unable to pay transporters to travel to the mandis, this too would show up as contraction of supply but trade may have merely been diverted to local traders. This last possibility however suggests that while we expect mandi-based trade to reflect the trend in overall trade, this may not necessarily be the case. In effect, the actual impact of demonetisation on domestic agricultural trade is an empirical question.

Impacts on value of mandi trade, arrivals and prices

We find robust and significant negative impacts on the value of agricultural trade in the period immediately following demonetisation. The worst impacts occurred within a fortnight − demonetisation displaced agricultural trade to the extent of 15-17%. Thereafter, we observe a steady revival, which plateaus after a couple of months, virtually stalling after the 75-day mark. By the end of 90 days after demonetisation, mandis were still seeing a loss of value of trade to the extent of 7-7.2%.

In general, perishable commodities fared worse than non-perishables – the value of mandi-based trade fell by 23% in the week following demonetisation and by the end of 90 days, but was still 18% lower than the usual. Some commodities fared disastrously. For example, the value of trade declined for tomato by 36% in the week following demonetisation (compared to 13-27% for other perishables). Even by the end of 90 days, the value of tomato trade was down 29%. In contrast, the maximum decline for non-perishables was around 11% during the week following demonetisation. It recovered after that, and by the end of the 90th day, the decline was close to 1.3% per mandi per day over the three-month period. Amongst non-perishables, we see a significant impact on soyabean , where value of trade declined by 69% in the week immediately following demonetisation.

Most of this decline in the value of trade for perishables is on account of the significant decline in prices rather than of arrivals. Tomato prices sank by 48% in the week immediately following demonetisation. Others such as cabbage, guava, saw a decline between 11-22%. There was little recovery in the prices of perishables even at the end of 90 days. In contrast, except for perishables such as apple, onion, which have relatively longer shelf life, we do not see significant impacts on arrivals of perishables. This is justified by the nature of these commodities; farmers might be willing to sell at very low prices but bring it to the markets nevertheless, as long as the costs of doing so do not exceed what they might expect to get. On the other, we find that most of the decline in value of trade for non-perishables could be attributed to decline in arrivals, indicating that farmers chose to hold the commodity instead of taking a price hit. As for non-perishables, arrivals declined substantially – soyabean arrivals, for example, declined by 66%. Similarly, jowar, bajra , rice, groundnut, turmeric saw a decline in arrivals in the range of 11-43% in the week following demonetisation. Most of these commodities however recovered significantly by the end of 90 days.

**From mandi trade to welfare losses**

It is not straightforward to move from these estimates of loss of value in mandi -based trade to anticipating farmer’s income losses. The estimates presented above could overestimate the negative impact if we believe that commodities destined for mandis were instead diverted to local markets on account of the cash crunch. While this is possible, the more plausible story based on field visits, is that farmers and traders resorted to credit-based sales, with farmers often relying on borrowings at usurious interest rates to tide over this exigency. This implies that one would only detect impacts on income and consumption after a lag. The persistence of negative price impacts is consistent with the idea that demand for several commodities has not yet recovered. This seems plausible given that both farm workers and informal workers in non-farm sectors too have faced prolonged income losses and may have suppressed consumption or reallocated expenditures to the basic food items.

Both these phenomena imply that the long-term impacts on farmer incomes are likely greater than implied by our estimates. Interestingly, if one were to go by the metric used by the Economic Survey – rabi sowing – as an indicator of stress on the agricultural sector, it turns out there are marked declines in sowing for the kharif⁷ season 2017-18 as of September 2017. These shortfalls in sowing figures, despite adequate and timely rainfall, are suggestive then, perhaps, of a lagged response to the monetary shock of demonetisation. There exists ample evidence on poverty traps that suggests that in the absence of markets for credit and insurance, shocks can push people into poverty and entrap them. While we will probably never know the true impact on farmer incomes due to demonetisation, the results from our analysis point to lingering impacts of demonetization on farmers and adverse distributional consequences overall. They also point to demonetisation as perhaps one source of farmer discontent in recent months across the country.

The authors thank, without implicating, Ajay Shah, Ashok Kotwal, Bharat Ramaswami, Jean Drèze, Milind Murugkar for discussions on the draft on which this piece is based.

Notes

Data are for the year 2014-15 at 2011-12 constant prices (Government of India, 2016).
Rabi crops (or Rabi harvest) are agricultural crops sown in winter and harvested in the spring in South Asia.
“Contrary to early fears, as of 15 January, 2017 aggregate sowing of the two major rabi crops − wheat and pulses (gram) − exceeded last year's planting by 7.1% and 10.7%. Favourable weather and moisture conditions presage an increase in production. To what extent these favourable factors will be attenuated will depend on whether farmers’s access to inputs, fertilizer, credit, and labour was affected by the cash shortage.” (Government of India, 2017)
These 35 commodities represent most of the 12 commodity groups - cereals, pulses, oilseeds, fibres, sugar and beet, plantation crops, spices, fruits, vegetables, flowers, aromatic crops, and honey. The commodities are bajra (pearl millet), jowar (sorghum), maize, paddy, ragi (finger millet), rice, wheat, Bengal gram (chickpea), arhar (pigeon pea), soyabean, mustard, groundnut, cumin, coriander seed, dry chillies, turmeric, turmeric raw, arecanut, cashewnuts, copra, cotton, sugarcane, brinjal, cabbage, cauliflower, okra, onion, potato, tomato, apple, banana, guava, lemon, lime, sweet lime, orange and other citrus. Of these, apple, banana, okra, brinjal, cabbage, cauliflower, guava, lemon, lime, sweet lime, orange and tomato are categorised as perishables. All others are deemed non-perishables, although this is only in a relative sense.
The value of trade is computed as the total arrivals multiplied by the minimum price for the day as reported in the database.
We use lagged rainfall shocks (positive and negative), day of the week effects, month effects and control for the dates around Diwali, and use state and time interaction effects to control for both production patterns and policy trends that might be confounders. This difference-in-differences (DiD) approach is elaborated in the paper. We check if the assumptions of DiD hold, conduct robustness and falsification checks and validate our key findings using a synthetic control method.
Kharif crops or monsoon crops are domesticated plants that are cultivated and harvested in South Asia during the rainy season, which lasts from April to October depending on the area.