Historical data for precious metals miners

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Historical data for precious metals miners

Post by siamond » Wed Jan 30, 2019 10:13 am

In a separate thread discussing Vanguard's strange changes of focus about the VGPMX fund, I asked the following question and... didn't get much feedback. So let me try again, with a dedicated thread.
siamond wrote:
Fri Sep 07, 2018 10:11 am
So far, in the Simba backtesting spreadsheet, we have a VGPMX data series from 1985 till now, as a proxy for the precious metals and mining (PME) industry sector. The main point of tracking it is that this allows analytical people to develop an opinion about the diversification benefits (or lack thereof) of such sector, and check a popular theory from W. Bernstein and others that the PME sector was one of the best diversifiers around. With VGPMX radically changing its direction in Sep-18, we need to find a better proxy to keep doing so, while staying true to the original intent.

An easy non-disrupting fix could be to keep using VGPMX for 1985-2017, and then use PSAU (Invesco Global Gold and Precious Metals ETF) starting in 2018.
- Pros: PSAU seems to be one of the most passive PME funds around, closely tracking the NASDAQ OMX Global Gold and Precious Metals Index. And its ER is relatively low compared to most PME funds (0.75%). Recent trajectory is quite close to VGPMX.
- Cons: PSAU and its index have limited history (starting in 2008/09); Invesco isn't a large player; the fund itself is very small (assets $28M compared to the $1.1B of Fidelity FSAGX or the $1.8B of VGPMX).

Another approach would be to take the opportunity to entirely ditch VGPMX and assemble a data series with a longer history, more narrowly focused on PMEs. It appears that the VGPMX managers got a little 'overly active' and erred way outside the PME scope (see this post and this article; fertilizers, chemicals and coal, really?). For example, we could use INIVX (VanEck International Investors Gold, assets $560M, ER 1.43%) from 1957 to 1985, spliced with FSAGX (Fidelity Select Gold, assets $1.1B, ER 0.84%) from 1986 till now.
- Pros: this adds nearly 30 years of history, including the oil crisis period. Should be more focused on the PME sector, avoiding the vagaries of VGPMX.
- Cons: lack of continuity with previous Simba instances; staying anchored on active funds.

Checking correlation numbers, the latter approach actually reinforces the very low correlation theory (nearly zero). Returns over the 1985-2017 period wouldn't be terribly different from VGPMX (a little lower, unsurprisingly given the ER difference). The 2009+ trajectory is actually very similar between PSAU and FSAGX, so FSAGX doesn't seem 'overly active'. Finally, a comparison between various PME funds with long history shows that VGPMX dramatic fall in 2008 was somewhat atypical (see this chart for more focus).

Overall, I would vote for the second approach (ditch VGPMX in favor of INIVX+FSAGX). Feedback?

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Wed Jan 30, 2019 10:22 am

This morning, I received a long PM pointing out that VanEck INIVX was a broad International fund to start with, and didn't focus on gold miners until 1968 (do check the video). Doubts were also expressed about the representativity of this fund later on, due to a narrow focus on smaller/riskier mining companies. All good points.

There are other options to extend the data series before Fidelity's FSAGX inception. For example, ASA Gold and Precious Metals (ASA) goes back to 1959 (and is already tracked in Simba's Data_Misc tab). Franklin Gold and Precious Metals (FKRCX) goes back to 1970, which would be good enough. Or maybe we could use an index (S&P Gold Miners is said to have a long history, but I just can't find it in Total Return form).

Note that FSAGX started at the end of 1985, and between 1970 and 1985, INIVX and ASA tracked each other fairly closely (at least until 1980), while FKRCX diverged (even when ignoring its initial hiccup).

We can find some historical perspective on the ASA Web site, confirming that this fund was indeed focused on gold miners from its inception, contrary to VanEck INIVX.

Before changing (again!) the precious metals miners data series (say using ASA spliced with FSAGX), I'd be glad to get more feedback. Thoughts?

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Wed Feb 06, 2019 6:50 pm

Well... Not that many thoughts provided on this topic... :( :( :(

I guess I'll proceed switching to ASA if I can't elicit responses, but this is rather unfortunate.

User avatar
willthrill81
Posts: 8025
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Historical data for precious metals miners

Post by willthrill81 » Wed Feb 06, 2019 8:55 pm

FWIW, after I alerted Portfolio Visualizer of Vanguard's change to VGPMX, they switched the asset class for precious metals to PSAU for 2018 and forward. That seems reasonable to me.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Wed Feb 06, 2019 10:33 pm

willthrill81 wrote:
Wed Feb 06, 2019 8:55 pm
FWIW, after I alerted Portfolio Visualizer of Vanguard's change to VGPMX, they switched the asset class for precious metals to PSAU for 2018 and forward. That seems reasonable to me.
Thank, Will. Yes, I am aware, but I think this was kind of a quick band-aid on a broken leg. I explained the reasoning here.

User avatar
willthrill81
Posts: 8025
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Historical data for precious metals miners

Post by willthrill81 » Wed Feb 06, 2019 10:40 pm

siamond wrote:
Wed Feb 06, 2019 10:33 pm
willthrill81 wrote:
Wed Feb 06, 2019 8:55 pm
FWIW, after I alerted Portfolio Visualizer of Vanguard's change to VGPMX, they switched the asset class for precious metals to PSAU for 2018 and forward. That seems reasonable to me.
Thank, Will. Yes, I am aware, but I think this was kind of a quick band-aid on a broken leg. I explained the reasoning here.
Hmm...there doesn't seem to be a very simple solution.

What about taking on a more radical move and simply ditching the "precious metals" asset class in lieu of treating gold, silver, etc. as separate? Historical spot prices are easily obtained. Yes, they would ignore transaction costs, but that might be the most parsimonious solution.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Wed Feb 06, 2019 10:47 pm

willthrill81 wrote:
Wed Feb 06, 2019 10:40 pm
siamond wrote:
Wed Feb 06, 2019 10:33 pm
willthrill81 wrote:
Wed Feb 06, 2019 8:55 pm
FWIW, after I alerted Portfolio Visualizer of Vanguard's change to VGPMX, they switched the asset class for precious metals to PSAU for 2018 and forward. That seems reasonable to me.
Thank, Will. Yes, I am aware, but I think this was kind of a quick band-aid on a broken leg. I explained the reasoning here.
Hmm...there doesn't seem to be a very simple solution.

What about taking on a more radical move and simply ditching the "precious metals" asset class in lieu of treating gold, silver, etc. as separate? Historical spot prices are easily obtained. Yes, they would ignore transaction costs, but that might be the most parsimonious solution.
As a general principle, I try hard to keep in Simba historical data allowing to opine on (and possible disprove) what was once a popular idea. Because such old ideas have a way of resurfacing. So I am pretty attached to keep both the gold (the price of the metal) and the precious metals miners data series.

As to a simple enough solution, I thought that INIVX + FSAGX was a pretty good replacement, but, well, maybe not (cf. the 2nd post of this thread). Now I am hoping that ASA + FSAGX is a pretty good proposal. But maybe I am missing something again... Hence my call for comment. I am essentially asking if there is anything wrong with this new approach, given the loaded history of such precious metals miners funds.

asif408
Posts: 1567
Joined: Sun Mar 02, 2014 8:34 am
Location: Florida

Re: Historical data for precious metals miners

Post by asif408 » Thu Feb 07, 2019 8:11 am

I don't know how feasible this is for you, siamond, but since all these funds have their issues going back historically at times, would it be possible simply to take an average of several of them (or as many as existed as possible) and use that average as an approximation? It obviously wouldn't be perfect, but maybe it would be good enough for comparison data and might smooth out any outlier performance of one fund during certain periods. Obviously with the very oldest data you might only have one fund, so this wouldn't help much, but when you have at least 2 or 3 funds this might be helpful.

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Thu Feb 07, 2019 9:19 am

asif408 wrote:
Thu Feb 07, 2019 8:11 am
I don't know how feasible this is for you, siamond, but since all these funds have their issues going back historically at times, would it be possible simply to take an average of several of them (or as many as existed as possible) and use that average as an approximation? It obviously wouldn't be perfect, but maybe it would be good enough for comparison data and might smooth out any outlier performance of one fund during certain periods. Obviously with the very oldest data you might only have one fund, so this wouldn't help much, but when you have at least 2 or 3 funds this might be helpful.
Thanks for the input. This would be akin to what the Morningstar Equity Precious Metal 'index' (I'm hesitant to use this word in such context, but Morningstar sure doesn't) does, more precisely they seem to be taking some sort of weighted average of real-life funds. What you're suggesting would be very easy to do (while avoiding a Morningstar dependency), essentially giving a weight of '1' to each existing fund!

Trouble is... this is garbage in, garbage out. Now that I learned that INIVX was an International fund with no precious metal focus whatsoever until 1968, I really don't want to use it at all, at least for this time period. I also grew VERY reluctant about Vanguard VGPMX (fertilizers, chemicals and coal, really?).

This being said, if we were to restrict ourself to 1970+, the data seems cleaner (?!), and we could indeed average VanEck's INIVX, ASA Gold and Precious Metals (ASA) and Franklin Gold and Precious Metals (FKRCX). Then add Fidelity's FSAGX to the mix in 1986. This could indeed mitigate the side-effects of each fund, while hopefully staying focused on precious metals miners. Ok, let's mull this over a bit more.

EDIT: I gave it a quick try. The simple average hugs the Morningstar PME index quite remarkably. And it is pretty close to Invesco's PSAU too. I am starting to warm up to the idea...

PS. As a side note, I also compared to a handful of formal PME indices with some (limited) history, and they are wildly inconsistent. Sigh.

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Sat Feb 09, 2019 2:43 pm

I was starting to consider again PSAU as a possible investible vehicle spliced to an averaged series of active funds as suggested by asif408 and... here is what showed up on my screen:

Effective on Wednesday, February 20, 2019, the Fund will be liquidated and it will be its final day of trading. Please see the Fund’s prospectus for further information.

https://www.invesco.com/portal/site/us/ ... uctId=PSAU

EyeDee
Posts: 1304
Joined: Tue Feb 20, 2007 12:15 am

Re: Historical data for precious metals miners

Post by EyeDee » Sat Feb 09, 2019 2:58 pm

Siamond,

Have you considered USAA Precious Metals and Minerals Fund (USAGX)?

It was started only a few months after Vanguard's Precious Metals and Mining (VGPMX).

I have not looked at it closely so I am not sure how good of a fit it is and it has too high of a expense ratio to be a good choice for actual investing, but might provide you with a reasonable choice for historical information.
siamond wrote:
Wed Feb 06, 2019 6:50 pm
Well... Not that many thoughts provided on this topic...

I guess I'll proceed switching to ASA if I can't elicit responses, but this is rather unfortunate.
Randy

User avatar
Tyler9000
Posts: 369
Joined: Fri Aug 21, 2015 11:57 am

Re: Historical data for precious metals miners

Post by Tyler9000 » Sat Feb 09, 2019 2:59 pm

siamond wrote:
Thu Feb 07, 2019 9:19 am
This being said, if we were to restrict ourself to 1970+, the data seems cleaner (?!), and we could indeed average VanEck's INIVX, ASA Gold and Precious Metals (ASA) and Franklin Gold and Precious Metals (FKRCX). Then add Fidelity's FSAGX to the mix in 1986. This could indeed mitigate the side-effects of each fund, while hopefully staying focused on precious metals miners. Ok, let's mull this over a bit more.
While I'm not particularly familiar with the inner-workings of precious metal miner funds, to me this sounds like a reasonable idea. Limit the years so that INIVX actually tracks what you want to study and average a few options to minimize any fund-specific variables. Miners to me are a lot like commodities in that funds are rarely as passive as you think, but that should get you in the ballpark for general backtesting purposes.

One other option is to perhaps see if the basic index methodology can be reproduced from other sources. For example, Fama-French has a "mines" series in the Industry Portfolios section that might be useful. Now I have no idea what goes into that or whether it actually tracks the precious metals miners index you're looking for, but it might be worth studying to see if it can be used to build something with a consistent methodology.

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Sat Feb 09, 2019 3:22 pm

EyeDee wrote:
Sat Feb 09, 2019 2:58 pm
Have you considered USAA Precious Metals and Minerals Fund (USAGX)?
Ah, thank you, I had it on my sight at some point, then somehow forgot about it. Yes, its scope matches our intent well, and its inception date is Aug-84, hence 30+ years of history. A quick test shows that, after a tough start, it largely followed the trajectory of similar funds.

I couldn't find historical information indicating a change of focus along the way. I am a little wary about what happened in its first decade though. At this point, from what I can tell, it could be a good addition to the average approach suggested by asif408.

DB2
Posts: 118
Joined: Thu Jan 17, 2019 10:07 pm

Re: Historical data for precious metals miners

Post by DB2 » Sat Feb 09, 2019 3:31 pm

I like this relatively new gold/metals mining ETF. The ticker is GOAU (U.S. GLOBAL GO GOLD AND PRECIOUS METAL MINERS ETF). Since it's been out, it's outperformed the Van Ecks (GDX, GDXJ) and others. It uses a different approach/methodology to metal mining funds and Royalty companies which gives it less volatility.

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Sat Feb 09, 2019 4:38 pm

DB2 wrote:
Sat Feb 09, 2019 3:31 pm
I like this relatively new gold/metals mining ETF. The ticker is GOAU (U.S. GLOBAL GO GOLD AND PRECIOUS METAL MINERS ETF). Since it's been out, it's outperformed the Van Ecks (GDX, GDXJ) and others. It uses a different approach/methodology to metal mining funds and Royalty companies which gives it less volatility.
Thanks, I didn't know this one. It is a little too peculiar (screening and weighing methodology) and too small ($13M) to match what we're trying to do here, but it's always good to broaden our horizons. For archive, here is the US Global fund page and the ETF.com ]description of it.

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Sat Feb 09, 2019 6:12 pm

Here is a growth chart comparing the Morningstar Precious Metals Equity 'index' (I'd rather call it a benchmark) and the simple average technique suggested by asif408, starting in 1970. Then I added a comparison with Fidelity FSAGX from its inception.

Image

Aside from the hiccup in the first couple of years, there is a strong alignment between the Morningstar benchmark and our simple average. Morningstar isn't too clear about their methodology (or at least I couldn't find the details), but I strongly suspect they do something similar to our average, probably involving a few more funds.

After a bit of a tough start, FSAGX (red line) isn't too far from both benchmarks. Which is good to see, because I'd like to keep the idea of having a data series anchored on a real-life investible fund, like we do for every other series in the Simba backtesting spreadsheet, and FSAGX seems the best candidate (stable Fidelity fund; its ER is a bit lower than the others).

I guess a proposal could then be the following:
* 1970-1999: average of known funds
* 2000+: FSAGX


Why 2000? No good reason :). I just need a cut-off point somewhere, I like round numbers (!), and I think it's more credible to keep roughly 2 decades of real-life data. I don't think it matters much, in truth.

Why not directly use the Morningstar PME benchmark numbers for 1970-1999? Good question. I actually had a bit of a hard time to extract those numbers (although I got the hang of it now). Plus Morningstar's methodology doesn't appear well documented. Plus I am not too sure how public this benchmark data is (we can see it on charts, but not as annual numbers). And then our simple average model matches it pretty well, and that is much easier to reproduce for somebody taking over the Simba spreadsheet maintenance.

stlutz
Posts: 4963
Joined: Fri Jan 02, 2009 1:08 am

Re: Historical data for precious metals miners

Post by stlutz » Sat Feb 09, 2019 6:42 pm

USERX is a fund worth considering as was "the first no load gold fund in the U.S."

http://www.usfunds.com/our-funds/our-mu ... /overview/

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Mon Feb 11, 2019 12:57 pm

Sigh, we keep discovering new funds. And some of them (incl. USERX) have a weird trajectory. This is starting to make the 'simple average' strategy overly complicated AND unstable (find out about a new fund, add it, numbers change a tad, rinse and repeat with no end in sight). And heck, I am 95% sure this is basically what Morningstar does with their benchmark. And if we splice such benchmark with FSAGX (say starting in 2000, somewhat arbitrarily), then the difficulty of extracting it is just a one-time event (which I already did). And if we have an issue with this data in the future, it would be easy to come back to the 'simple average' approach, possibly focused on just a few representative funds to limit the entropy.

I really want to issue the latest Simba update (for other reasons), so I'll take any last feedback you guys might have, then proceed tonight.

Bullionaire187
Posts: 1
Joined: Mon Feb 11, 2019 1:08 pm

Re: Historical data for precious metals miners

Post by Bullionaire187 » Mon Feb 11, 2019 1:13 pm

I have done research on finding silver miner data... probably the best source is Nick Laird over at GoldChartRUs.

One of his best charts on the matter is here: https://sdbullion.com/blog/paper-silver ... rformances

MachineGhost
Posts: 108
Joined: Sat Feb 14, 2009 2:46 am

Re: Historical data for precious metals miners

Post by MachineGhost » Tue Feb 12, 2019 12:14 pm

I looked into this situation years ago. Barron's Gold Mining Index that went back to the 1930's had only about 7 components. So I vote for ASA + FSAGX. Tracking error is always a huge problem with active management natural resource funds, so maybe you'd want to switch to the GDX ETF when it became available. Representative also has to be investable point-in-time otherwise its just a fantasy.

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Tue Feb 12, 2019 3:29 pm

MachineGhost wrote:
Tue Feb 12, 2019 12:14 pm
I looked into this situation years ago. Barron's Gold Mining Index that went back to the 1930's had only about 7 components. So I vote for ASA + FSAGX. Tracking error is always a huge problem with active management natural resource funds, so maybe you'd want to switch to the GDX ETF when it became available. Representative also has to be investable point-in-time otherwise its just a fantasy.
Thanks for the feedback. Here are a few tidbits of GDX history (according to ETF.com). Only its 2014+ returns are of interest, based on such history.

GDX is one of the most popular funds in the global gold-mining segment, a trading powerhouse with deep primary, secondary and derivatives markets. Following a 2013 overhaul of the fund’s index, GDX is no longer limited to US-listed firms, nor does it filter out firms that have hedged their gold exposure with derivatives. In addition, it includes firms mining other precious metals in addition to gold, so it’s not entirely pure-play gold miners. GDX includes all the major names, but the fund’s broader take on the gold-mining space means that the big firms get less weight than in our neutral benchmark. The fund is fairly priced, though real-world holding costs (measured by tracking difference) have sometimes been highly variable. As an industry benchmark, GDX falls a bit short, but it's still hugely popular and easy to trade.

Note that its ER is 0.53% and it was launched in 2006 as a US-only fund. Nowadays VanEck Vectors® Gold Miners ETF seeks to replicate as closely as possible, before fees and expenses, the price and yield performance of the NYSE Arca Gold Miners Index (from VanEck).

I like the fact that it's a low ER fund with a clear aim at index tracking, but this does seem a little too centered on Gold for our Precious Metals aim?

User avatar
Topic Author
siamond
Posts: 4515
Joined: Mon May 28, 2012 5:50 am

Re: Historical data for precious metals miners

Post by siamond » Wed Feb 13, 2019 9:54 am

MachineGhost wrote:
Tue Feb 12, 2019 12:14 pm
I vote for ASA + FSAGX. [...] Representative also has to be investable point-in-time otherwise its just a fantasy.
I ended up agreeing with MachineGhost... Given ASA's history though, I started the data series in 1970 for the regular Simba annual returns (Data_TR_USD). While keeping its full history in Data_Misc. ASA was South Africa only before 1970, which doesn't seem representative enough. Bottomline, keep it simple: ASA (1970+), FSAGX (1986+).

You can find the latest Simba update here.

Many thanks to all contributors on this thread.

Post Reply