Introduction

This document explores the results from Argentina soy SEI-PCS v.1.1.0 and performs some quality checks to make a decision on the quality of the results.

Overall checks that need to be carried out:

  1. Check allocation of soy per department of production
  2. Check the percent of soy exported from the Chaco (dry and humid)
  3. Compare the results against OPEX data

1. Check allocation of soy per department of production

The new allocation algorithm includes “Simon’s constraint” meaning that a department in Argentina can never have more soy sourced from it than what is needed to crush beans into the exported bean and oil products.

This constraint is based on ratios that are determined based on the equivalence factor in each of the trade years:

  • 2015: oil = 0.200529, cake = 0.79947
  • 2016: oil = 0.202539, cake = 0.79746
  • 2017: oil = 0.201494, cake = 0.79855
  • 2018: oil = 0.207079, cake = 0.79292
  • 2019: oil = 0.201509, cake = 0.79849

The first step is to check whether the soy bean sourced from all departments does not exceed this ratio with production.

The number of “failed” entries is: 0 (NA means that there are no fails).

2. Check the percent soy produced and exported from the different ecoregions

Now we calculate the percent of soy that is produced from the Argentinean ecoregions.

From the above table we note that 11-17% of soy produced in Argentina comes from the Chaco.

We then check the relative amount of total exports that come from the different ecoregions. The table below shows that exports from the Chaco are about 6-13% of total exports considering all of the years (removing STOCK).

3. Comparison of results against OPEX

We then compare aggregated results with the data from OPEX to see how close our results are from the official data from INDEC.

Comparison with soy bean

We first look at official report from OPEX on the exports of soybean to the world (product 107A), and then specifically to China (as largest importer).

From the above comparisons of soybean exports to the rest of the world and their origin within Argentina, we note the following:

  • We estimate 3-4 times more soybean being sourced from the Province of Buenos Aires. OPEX only estimates 3 Mtonnes, while we estimate more than 10 in some cases.
  • Estimates for Cordoba are slightly better with OPEX estimates ~1-3 Mtonnes and we estimate ~2-6 Mtonnes. It is still a large difference though.
  • We estimate more sourcing from Santiago del Estero (4-9 times more 1-2 Mtonnes), as well as in Chaco (4-10 Mtonnes). This might be due to some mix of Paraguayan bean that we cannot differentiate at the moment (and so is sourced from the Chaco).
  • Our estimates are also off in Northern Argentina (Salta, Jujuy, Formosa, etc.)
  • There are large differences with Misiones as well (we source more from there)

We then repeat the same exercise with exports exclusively to China (as soy bean)

Our results typically overestimate the exports to China from The Province of Buenos Aires compared to OPEX, and underestimate the exports from Cordoba. We also underestimate exports from the Northern Provinces.

Comparison with soy cake

We then look at official report from OPEX on the exports of soy cake to the world (product 212AE), and then specifically to EU (as largest importer).

The above comparison is actually incorrect to make because the SEI-PCS results provide the source of the soy bean that is used for crushing, while the OPEX results provide the origin of crushing (supposedly) for the export. The proper comparison would be the Province of Origin in the trade data, which we know is widely “Unknown”.

For purposes of proper comparison, we assume that the port of export in Rosario area means that the Province of Origin is Santa Fe.

Clearly the “Unknown” Province of Origin is an issue for exports from Santa Fe, Cordoba and Buenos Aires when compared to OPEX. The differences are as follows:

  • Exports from Sante Fe show a 12-17.5 Mtonnes discrepancy between OPEX and customs declarations
  • Exports from Cordoba show up to 6.4 Mtonnes discrepancy
  • Exports from Buenos Aires show up to 4 Mtonnes discrepancy
  • Unknown Province of Origin in the customs declarations can have volumes ~30 Mtonnes

If we add up to 4 Mtonnes of imports from Paraguay then we have these discrepancies mapped out and the biggest errors in sourcing is due to the lack of a Province of Origin in the trade data.