Another data dump for those interested.
(1) NdPr selling price (contracted to one major customer):
Y axis in AU$, data April inclusive.
(a) Raw data of average sell price (ASP, dark turquoise) and NdPr spot price (China ex VAT; orange).
(b) Same data modelled as per previous post in this thread (NdPr price a 3m moving average, shifted forward one month).
Comments/thoughts: As we all know, we have had a recent dip in the NdPr spot price, but given the suspected price model, it has/is having less impact than some may think. May data will be very telling (I'll update once I get it), I'm expecting a dip which will then be reversed in June. As of April, we are holding up well at AU$127/kg.
(2) Estimated NdPr production cost (estimations based entirely of LYC supplied data, plus using above price model):
(a) Tonnes of NdPr produced (left Y axis, deep turquoise line) vs estimated NdPr production cost ('COP', orange line, right Y axis AU$). X axis is HY units until end of 2022.
(b) Anecdotal support for model - LYC were a bit more transparent a few years back and provided a percentage based NdPr production cost model (2014-2019 production, page 38 here
https://wcsecure.weblink.com.au/pdf/LYC/02107222.pdf). To compare with that, I generated a percentage based model for my data (Y axis is % cost from a 2016 base).
Comments/thoughts: In my model there is a clear negative correlation (r= -0.88) between production volume and production cost (i.e. produce more and it'll cost less per unit, as expected). I suspect the uptick in cost the last two HY reports are due to significant inventory build (i.e. the revenue not realised), plus possible signs of inflation. Next HY will be very interesting as we have an increased production, but my guesstimate is that in the last couple of years our base cost of production has been somewhere around AU$40-45/kg NdPr.
Although there is a correlation between my model and LYC's own data as presented, if I would shift LYC's data one period (HY) forward the correlation becomes almost perfect (r=0.97). Perhaps it could be explained by the difference in presentation (my data is HY based, LYC's is Q based), or perhaps the base (100%) was calculated differently. Either way, I'm pretty confident of my data as it is.
Thoughts/comments welcome.
All IMO, although based on real data the data includes interpretations, assumptions, and models. DYOR.
(20min delay)