One Metric to Govern Them All: Producer Support Estimate

OECD's Producer Support Estimate is THE ONE metric to investigate if you are obsessed with a simple, but tricky question: What is the role of Government in Agriculture?

"One ring to rule them all, One ring to find them, One ring to bring them all and in the darkness bind them." - Translation of the Inscription found in the Ring

As a recovering independent management consultant with a shady conslutting past (You read it right. It’s not a typo), I have developed over time a healthy allergy to metrics and measurement.

The reasons aren’t hard to fathom.

Metrics have ring-like qualities.

They can be addictive (Who doesn’t want to slay the beast of complexity inside a system?) and often have a strange will of their own. If you don’t pay attention, they can create a spell and transform, for worse, the very system you are trying to manage.

If you don’t believe me, ask the British economist Goodhart who coined Goodhart’s law in 1975 in response to Margaret Thatcher’s trigger-bound monetary policy.

Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.

Here is a plain-speak translation: When a measure becomes a metric, it ceases to be a good measure.

Take the case of Producer Support Estimate.

This metric has been often used to make a powerful argument in favor of the controversial farm laws that received tremendous attention over the past year.

The argument goes something like this.

1) Producer Support Estimate (PSE) for India has been negative since 2000.

Translation: Farmers in India have not been supported all these years. They have been indirectly taxed by the government’s restrictive agricultural trade and marketing policies. And when you add that with poor infrastructure and lack of proper mechanisms to aggregate produce, you get the full sorry state of Indian agricultural affairs.

Here is the link if you want to play with this data further. Source: OECD Website.

This argument is not new.

During the eighties and nineties, when WTO’s Aggregate Measure of Support was the ONE metric to govern them all, Sharad Joshi, India’s prominent free-market agricultural activist, calculated Aggregate Measure of Support using Government data and argued that it was the negative subsidy that was killing Indian farmers.

Although he did concede that positive AMS didn’t mean to imply that tide would drastically turn around.

2) If a country has a higher positive producer support estimate, we could imply that agricultural policies have supported the producers of the country and would consequently have an impact on the prices received by the producers.

When you divide the Producer Support Estimate by the gross farm receipts (total value of the output of production + payments (based on output, inputs, and others) of the country, you get a macro view of how each country has been supporting their farmers respectively.

Here is the link if you want to play with this data, change countries and see how they are supporting farmers.

As you can see from data, India is not alone in countries that are “taxing” farmers. Countries like Argentina, Ukraine, Vietnam have negative Producer Support Estimates.

You can contrast this with China, which offers the largest policy support to farmers than any country in the world (when measured in absolute terms)

3) Under conditions of liberal domestic markets and freer global trade, remunerative prices received by farmers would raise.

It is the third part of the argument that requires deeper examination. For it is here where the metric turns prescriptive. And in order to unpack this argument, we need to open up the bonnet of this metric.

OECD defines PSE as “The Annual Monetary Value of Gross Transfers from Consumers and Tax Payers to Agricultural Producers, measured at the farm gate level, arising from policies that support agriculture”.

Essentially, if you break down the Producer Support Estimate, you see two components.

In short, the budgetary support works largely as input support, while the market price support works largely as output support. In India, the PSE for Indian Agriculture in 2019 was ?1,62,470 crores. Out of which the Market Price Support (MPS) was ?-4,61,804 crore, while Budgetary Support (BOT) was ?+2,99,604 crores. (PSE= MPS+ BOT).

In essence, understanding the two components which make up the Producer Support Estimate is important as few countries have higher BOT (Input Support), while few others have higher MPS (Output Support)

Image Credits: Agricultural Economist Shweta Saini’s Presentation on Producer Support Estimate

While the positive producer support estimate in Australia could be attributed to high budgetary support (shown in the orange bar), the positive producer support estimate in China could be attributed to the market price support (shown in the blue bar).

While critics have raised several issues with the methodology underpinning Producer Support Estimate, if you want to get to the heart of the assumption underpinning Producer Support Estimate, we need to understand Law of One Prices.

In Economics 101, Law of One Price states that when prices are converted into a common currency, the same good should sell for the same price in different countries.

Even though this law has been violated more than any other economic law (Refer to this and this for more details), it is of extreme importance in the way agricultural economists think about the prices of agricultural commodities. 

If there were a free market in Indian agriculture,

(I have explored this big if when I profiled the story of Indian agriculture?? and laid down the assumptions underpinning it)

theoretically speaking, we can expect the market prices of commodities around the world to differ only by an amount equal to

1) The difference in quality

2) International freight and handling charges

3) Marketing Margins, processing costs, and the cost of transportation.

The consequences of this assumption are extremely critical.

Prices of agricultural goods vary greatly from one country to another because of government intervention in agriculture.

Now the issue with Producer Support Estimate is this: It makes a problematic distinction between “provision of support” and “impact of support”.

When you are comparing a domestic price with its global benchmark price, producer Support Estimate wants you to ignore the fact that the global benchmark price may have been impacted by the very existence of support. While OECD argues in academic speak that PSE as a measure of policy effort should be kept separate from any analysis of policy effect, in the real world, you cannot make a distinction between the two.

This argument reminds me of Nassim Nicholas Taleb's delicious quote

You can see the repercussion of this in the recommendation offered by agricultural economists, based on the producer support estimate.

Consider this recommendation based on the Negative Market Price Support data

Image Credit: “Mapping Sustainable Agriculture: How much are Indian farmers supported? Findings from an ICRIER-OECD”

“The negative gap between the domestic producer price and the international reference price ( adjusted to farm level) would likely decline and perhaps disappear if the margins in the value chain between producers and consumers (or exporters are allowed to decline.”

Now contrast this recommendation with this graph which compares global farm gate prices with retail prices for wheat in US and Canada.

Image Credits: Darrin Qualman

When the margins between producers and consumers are increasing worldwide in US and Canada, is it feasible to expect the margins between producers and consumers to decline?

Why restrict to US and Canada?

When the European Commission was asked in parliament to provide details on the average share/percentage of the final price that farmers receive for their produce, relative to the share/percentage received by other actors in the food supply chain, they found that the value-added for agriculture in the food chain dropped from 31% in 1995 to 21% in 2011, mainly in favor of other food chain actors.

If you contrast this number with 28% for the food industry and 51% for food retail and food services, it is evident that the highest value addition occurred in the supply chain far away from the farmer.

Considering all these data points, is it seriously plausible to expect that the margins in the value chain between producers and consumers (or exporters )will decline in the case of Indian Agriculture?

Conclusion:

Measures turn into metrics when we want to change the system. In the case of OECD’s Producer Support Estimate, the proposed change spelled in the PSE methodology is very clear: ”Agricultural trade should be more fully integrated within the open and multilateral trading system”

My argument so far is not to propose a simplistic stance against the free market and open trading system. We need to deeply understand the causal relationships in the role played by the government in agriculture. When 1) changing PSE doesn’t necessarily imply a change in policy settings and when 2)unchanging PSE doesn’t necessarily imply no change in policies, it clearly indicates the problem in defining the causal loops, while dealing with an aggregate metric like Producer Support Estimate.

The moral of the story is clear: When causation is not clarified, metrics are dangerous to the system.