Mar 1 / Natália Gandolphi

Coffee Weekly Report - 2024 03 01

  Back to main blog page
  • This report explores the correlation between the NY-LN arbitrage and the spread between arabica and conilon prices in Brazil.

  • Analyzing daily data from April 2008 to February 2024, a strong positive correlation is observed between the NY-LN arb and the Ara-Con USD/bag spread, influenced by global arabica and robusta supply and demand dynamics. The correlation is weaker with the Ara-Con BRL/bag spread due to currency fluctuations.

  • Examining developments from 2022 to 2024, the USD/bag spread is 30% below the historical average, favoring arabica for international buyers, while the BRL/bag spread is 12% above the average, favoring conilon sales domestically.

  • Historical correlations faced disruptions in 2010 and more significantly between 2016 and 2018, attributed to Brazil's unique fluid market dynamics and recovery from the El Niño phenomenon.

  • We observe that future trends may depend on conilon availability in the 24/25 cycle, influenced by blend decisions, production, and price dynamics. The potential for lower correlation between Brazil's internal and external coffee markets is outlined but remains uncertain.

Domestic price dynamic in Brazil

In this report, we will dive into the relationship between the NY-LN arbitrage and the local spread between arabica and conilon prices in Brazil. First, let's understand the series that will be used for this analysis: i) the daily arbitrage (or difference) between NY arabica prices and LN robusta prices, in c/lb, and ii) the spread between the local reference prices for arabica (type 6, SP/MG) and conilon (13Up, Vit-ES), in both BRL/bag and USD/bag, released by Cepea.

First, looking at the daily distribution using data from April 2008 to February 2024, we can observe that the NY-LN arb and the Ara-Con USD/bag spread are strongly and positively correlated. Naturally, the global arabica and robusta S&Ds will impact the NY-LN arb, and that dynamic will trickle down into the world's largest coffee producer and exporter, Brazil.

Therefore, it stands to reason that the two series will correlate well. The correlation against the Ara-Con spread BRL/bag is lower though, with the Brazilian currency impact affecting the relationship (the historical correlation between NY-LN Arb and Ara-Con Spread USD/bag is 0.95, whereas for BRL/bag it is 0.67).

Now, we need to look at the developments for the Ara-Con spread. Considering data from 2022 to 2024, the USD/bag spread peaked in 2011, whereas the BRL/bag spread peaked in 2022, despite both showing very similar patterns. Currently, the USD/bag spread is 30% below the historical average (32 USD/bag vs. 46 USD/bag), whereas the BRL/bag spread is 12% above the historical average (160 BRL/bag vs. 143 BRL/bag).

This suggests, therefore, that there is currently more incentive for international buyers to choose arabica over conilon (since the difference between the two in USD/bag is lower than average). In the domestic market, however, the difference between prices favors conilons, although the difference is more narrow and therefore more subject to change.

Now that the series used in this analysis have been understood, as well as their own patterns, let's look at their correlation results vs. the NY-LN arbitrage.

Figure 1: NY-LN Arbitrage vs. Spread Arabica-Conilon in Brazil

Source: Cepea, ICE, Refinitiv, hEDGEpoint

Figure 2: Future Curve – NY Arabica (c/lb)

Source: Cepea

Historically, the USD/bag spread has a stronger correlation (as explained earlier in this text). Still, there are two points in time when both heavily lost correlation to the NY-LN arb. The first point in time was in 2010, whereas the second one, more substantial and that lasted longer, took place between 2016 and 2018.

This happened because the NY-LN arb received some support, whereas the local spreads did not follow suit. This happened due to the implicit characteristic seen in Brazil, being the largest global producer/exporter, and also one of the top countries in consumption figures as well: the decision to either export or consume a certain type of coffee locally is more fluid than in other origins, that usually focus either on arabica or robusta, and also do not rank as high as Brazil in terms of consumption.

Figure 3: Correlation between NY-LN Arb and Spread in BRL/bag and USD/bag

Source: Cepea, ICE, Refinitiv, hEDGEpoint

Back in 2016, Brazil and Vietnam were both affected by the El Niño phenomenon - although the two countries did eventually recover, especially in the 18/19 cycle, and especially Brazil. It took a while for the correlation between the local market in Brazil and international prices to return to normal levels. Currently, with a much more fluid dynamic between what gets exported and what is consumed internally in Brazil, and the same expectation of a faster rate of recovery for Brazil compared to Vietnam, the question arises: to what extent can correlation reduce between NY-LN arb and local spreads in Brazil?

In Summary

Ultimately, it will heavily depend on final conilon availability in the 24/25 cycle. Blend decisions - which reflect on exports - are bound by total availability (production + stocks) first, and by price dynamic second.

At this point, arabica seems to have been less affected by El Niño when compared to conilon, similar to what happened between 14/15 and 16/17 (20% cumulative drop in arabica production, 49% cumulative drop in conilon production), although the impact of the crop failure itself is not comparable. Therefore, the ingredients for lower correlation between the internal and external market in Brazil are set, but it’s yet uncertain if the market will bring enough heat.

Weekly Report — Coffee

Written by Natália Gandolphi
Reviewed by Lívea Coda


This document has been prepared by hEDGEpoint Global Markets LLC and its affiliates ("HPGM") exclusively for informational and instructional purposes, without the purpose of creating obligations or commitments with third parties, and is not intended to promote an offer, or solicitation of an offer, to sell or buy any securities or investment products. HPGM and its associates expressly disclaim any use of the information contained herein that may result in direct or indirect damage of any kind. If you have any questions that are not resolved in the first instance of contact with the client (, please contact our internal ombudsman channel ( or 0800-878-8408 (for clients in Brazil only).

To access this report, you need to be a subscriber.