Gold: ETFs vs. Miners vs. Streamers vs. Spot — Where to Put Your Money
Disclaimer: These reports include AI-generated summaries, often explicitly arguing one side as strongly as possible (positive or negative) and should not be relied on. They're designed to test theses and foster polite debate and scrutiny, so please comment if you see an error!
Date: March 31, 2026
Gold Spot Price: ~$4,696/oz | GDX: $91.77 | GDXJ: $120.04 | GLD: ~$430
The Question
If you're bullish on gold, how should you express it? Physical gold (GLD), a miner ETF (GDX/GDXJ), individual miners (NEM, AEM, Barrick), or streaming/royalty companies (WPM, FNV, RGLD)? The answer depends on your time horizon, risk tolerance, and whether you're making a structural allocation or a cyclical trade. The data strongly favours one approach over the others.
The Long-Term Track Record: Miners Destroy Value
This is the most important chart bulls don't want to see. Over the 2006-2025 period, GDX (gold miners) underperformed GLD (gold bullion) by approximately -6.5% annually — roughly -350% cumulatively. Holding a gold bar in a vault generated dramatically better returns than owning the companies that dig it out of the ground.
Why? Because mining is a business that consumes capital. Miners must constantly replace depleting reserves, fight cost inflation, manage geopolitical risk, deal with labour disputes, and reinvest billions just to maintain production. Gold in a vault doesn't deplete, doesn't need a CEO, and doesn't have a balance sheet. Over decades, the compounding drag of operational risk, dilution, and capital misallocation overwhelms the operating leverage that makes miners exciting in any given upcycle.
Performance Comparison
| Period | GLD (Spot Gold) | GDX (Miners) | GDXJ (Juniors) | Winner |
|---|---|---|---|---|
| YTD 2026 | +8.6% | +7.0% | +5.5% | GLD |
| 1-Year | +51.5% | +102.8% | +116.0% | GDXJ |
| 3-Year Ann. | +38.2% | +43.7% | +56.8% | GDXJ |
| 5-Year Ann. | +22.3% | +25.3% | +24.4% | GDX |
| 10-Year Ann. | +13.9% | +17.5% | +17.5% | GDX/GDXJ |
| 2006-2025 cumulative | Strongly positive | -350% relative to GLD | — | GLD |
The recent exception: In the current gold bull market (2024-2026), miners have outperformed gold because operating leverage works in their favour when margins are expanding. At $4,700 gold with $1,600 AISC, miners earn $3,100/oz — a 7x expansion from the $400/oz margin at $2,000 gold. This is why GDX returned +155% in 2025 when gold rose +65%.
But this leverage works in both directions. From 2011-2016, gold fell 45% and GDX fell 81%. Barrick Gold fell 85% and nearly went bankrupt. The operating leverage that makes miners exciting in an upcycle makes them devastating in a downturn.
GDX: The Diversified Mediocrity Machine
AUM: $26.6B | Holdings: 53 | Expense Ratio: 0.51% | P/E: ~22x | Yield: 0.69%
Top 10 Holdings
| # | Company | Weight | 5-Year Return |
|---|---|---|---|
| 1 | Agnico Eagle (AEM) | 12.3% | +258% |
| 2 | Newmont (NEM) | 10.9% | +80% |
| 3 | Barrick Mining (B) | 7.7% | ~-27%* |
| 4 | Franco-Nevada (FNV) | 5.3% | +99% |
| 5 | AngloGold Ashanti (AU) | 5.0% | — |
| 6 | Wheaton PM (WPM) | 4.9% | +247% |
| 7 | Kinross Gold (KGC) | 4.7% | +6% |
| 8 | Gold Fields (GFI) | 4.4% | +390% |
| 9 | Pan American Silver (PAAS) | 3.2% | +86% |
| 10 | Alamos Gold (AGI) | 2.6% | +489% |
*Barrick's ticker changed from GOLD to B; 5-year return is approximate under old ticker.
The problem is obvious. The top holdings (AEM +258%, GFI +390%, AGI +489%, WPM +247%) have massively outperformed GDX's ~211% cumulative 5-year return. Meanwhile, the long tail — Kinross (+6%), B2Gold (+7%), Harmony Gold (+4%) — destroyed the returns the top names built. GDX is a blended average where excellence is diluted by mediocrity.
Why GDX Underperforms Its Best Holdings
- Forced diversification into losers. 53 stocks means the top 5 are ~41% of the fund. The other ~59% dilutes returns with underperformers that would never make a concentrated portfolio.
- Rebalancing sells winners, buys losers. Quarterly index reconstitution forces mean-reversion trades. In a sector with extreme dispersion between the best and worst operators, this systematically destroys alpha.
- 0.51% annual expense ratio compounds over time.
- Finite mine life. Every miner must spend billions replacing depleting reserves. Some succeed (AEM, AGI). Many fail (BTG, HMY). The ETF holds all of them.
When GDX Makes Sense
- You want broad gold miner exposure with zero stock-picking effort
- You're trading the sector short-term (GDX has excellent liquidity, tight bid-ask)
- You can't hold individual non-US stocks (some accounts restrict ASX/TSX listings)
- You're using options (GDX options are highly liquid)
GDXJ: Not Actually Junior Miners
AUM: $8.5B | Holdings: 117 | Expense Ratio: 0.51% | P/E: ~23x | Yield: 2.2%
The Mislabelling Problem
GDXJ has not been a "junior" gold miner ETF since June 2017, when VanEck modified the index rules to include larger companies. The fund had grown too large for the tiny junior mining universe and was forced to buy stakes it couldn't easily trade. The fix was to expand the market cap range — effectively transforming GDXJ into a mid-cap gold miner ETF.
Current top holdings include:
- Alamos Gold (AGI): $17.5B market cap
- Coeur Mining (CDE): $19.4B market cap
- Evolution Mining (EVN): Large Australian mid-cap
- Endeavour Mining (EDV): Large African mid-cap
These are not "juniors" by any definition. For genuine small-cap/exploration-stage gold exposure, you'd need individual stock selection or the Sprott Junior Gold Miners ETF (SGDJ), which uses different methodology.
GDXJ vs. GDX Performance
| Period | GDX | GDXJ | Winner |
|---|---|---|---|
| YTD 2026 | +7.0% | +5.5% | GDX |
| 1-Year | +103% | +116% | GDXJ |
| 3-Year Ann. | +44% | +57% | GDXJ |
| 5-Year Ann. | +25.3% | +24.4% | GDX (slight) |
| 10-Year Ann. | +17.5% | +17.5% | Tie |
| Since GDXJ Inception (2009) | — | +3.0% ann. | Very poor |
GDXJ's 3.0% annualized return since inception (November 2009) is abysmal. Over 16+ years, it has barely beaten inflation. The recent outperformance (1-year and 3-year) reflects the mid-cap leverage effect in a strong gold market, not structural superiority.
My take: GDXJ is the worst option among the four approaches. It's mislabelled, has no meaningful differentiation from GDX (there's even overlap — AGI appears in both), the long-term track record is terrible, and you're paying 0.51% for a portfolio that doesn't deliver what the name promises.
Streaming/Royalty Companies: The Compounders
Streaming and royalty companies (Wheaton Precious Metals, Franco-Nevada, Royal Gold) represent a fundamentally different business model. Instead of operating mines, they provide upfront capital in exchange for the right to purchase a percentage of a mine's production at a fixed, below-market price — forever.
Why Streamers Beat Miners Over Time
| Advantage | Detail |
|---|---|
| ~90% EBITDA margins | Franco-Nevada operates at 90.9% adjusted EBITDA margin. Miners: 30-50%. |
| No operational risk | No mine blowups, cost overruns, labour strikes, or environmental liabilities |
| No capex to replace reserves | Cash gets redeployed into new streams. Miners must spend billions replacing depleting ore |
| Built-in diversification | WPM: 18+ streams. FNV: 400+ assets. One mine failure barely dents the portfolio |
| Lower volatility | FNV volatility 8.1% vs. GDX 16.9% |
| Superior Sharpe ratios | FNV Sharpe: 2.20. WPM Sharpe: 2.08. Both above GDX |
| Growing dividends | Progressive policies funded by high-margin cash flows |
Performance vs. GDX (10-Year Annualized)
| Vehicle | 10-Year Ann. Return |
|---|---|
| Wheaton PM (WPM) | +29.4% |
| Royal Gold (RGLD) | +25.6% |
| Franco-Nevada (FNV) | +19.1% |
| GDX | +17.5% |
| GLD (spot gold) | +13.9% |
All three streamers have outperformed both GDX and spot gold over 10 years, with meaningfully lower volatility. The only period where miners outperform streamers is during sharp, short-term rallies (like 2025's +155% for GDX vs. +70% for WPM) — because streamers have less operating leverage.
The Trade-Off
In a momentum-driven gold rally, you'll feel like you're leaving money on the table owning WPM (+70% in 2025) instead of GDX (+155%). That's the price of quality. Over a full cycle — including the inevitable downturn — streamers compound better because they don't give back half their gains when gold corrects.
Individual Miners: The Concentrated Approach
If you're willing to pick individual names, the data strongly supports owning 3-4 top-quartile operators instead of the ETF.
5-Year Returns: Best vs. Worst in GDX
| Stock | 5-Year Return | vs. GDX (~211%) |
|---|---|---|
| Alamos Gold (AGI) | +489% | Crushed GDX by 278pp |
| Gold Fields (GFI) | +390% | Crushed GDX by 179pp |
| Agnico Eagle (AEM) | +258% | Beat GDX by 47pp |
| Wheaton PM (WPM) | +247% | Beat GDX by 36pp |
| Coeur Mining (CDE) | +117% | Below GDX |
| Franco-Nevada (FNV) | +99% | Below GDX |
| Pan American Silver (PAAS) | +86% | Below GDX |
| Newmont (NEM) | +80% | Below GDX |
| B2Gold (BTG) | +7% | Destroyed 204pp of value vs. GDX |
| Kinross (KGC) | +6% | Destroyed 205pp of value vs. GDX |
| Harmony (HMY) | +4% | Destroyed 207pp of value vs. GDX |
The dispersion is enormous. A portfolio of {AEM, WPM, AGI, GFI} returned an average of +346% over 5 years. GDX returned +211%. A portfolio of {BTG, KGC, HMY} returned an average of +6%. The ETF forces you to own all of them.
Recommended Individual Portfolio
| Name | Role | Why |
|---|---|---|
| Agnico Eagle (AEM) | Quality/Safety | Lowest AISC ($1,339/oz), 85% Canada, 67% EBITDA margin, best management |
| Barrick Mining (B) | Value/Growth | Cheapest (9.9x fwd P/E, 1.8x P/B), Fourmile + Reko Diq growth, 2.2% yield |
| Wheaton PM (WPM) | Streamer/Compounder | 90%+ margins, no operational risk, 18+ streams, 29.4% 10-yr CAGR |
| Alamos Gold (AGI) | Mid-cap leverage | +489% 5-year, strong operator, growing production |
This gives you: quality (AEM), value (Barrick), structural compounding (WPM), and mid-cap leverage (AGI) — without the ETF drag of 40+ underperformers.
The Framework: How to Choose
| Approach | Best For | Time Horizon | Risk | Expected Return vs. Gold |
|---|---|---|---|---|
| GLD / Physical | Core allocation; structural gold thesis | 5+ years | Lowest | Gold price return minus 0.40% expense |
| Streamers (WPM, FNV) | Long-term compounding; risk-adjusted return | 3-10 years | Low-Medium | +5-15% annual alpha vs. gold (historically) |
| Individual miners (AEM, Barrick) | Concentrated leveraged bet; cycle trade | 1-5 years | Medium-High | 1.5-2.5x gold in upcycles; -2-3x in downturns |
| GDX | Broad exposure; short-term trade; options | Weeks to 2 years | Medium-High | Diluted by tail holdings |
| GDXJ | Avoid | — | High | Mislabelled; poor long-term track record |
The Bottom Line
If you can only choose one: Own GLD as your core and add WPM as your satellite. This gives you gold price exposure (GLD) plus leveraged upside through the highest-quality business model in the sector (WPM) — without the operational risk, ETF drag, or tail-end losers.
If you want to be more aggressive: Replace GDX with a 3-4 name concentrated portfolio (AEM, Barrick, WPM, and one mid-cap like AGI or GFI). You capture the operating leverage without the dilution from the bottom half of the ETF.
If you're a short-term trader: GDX is fine — the liquidity and options market are excellent. Just don't confuse a trading vehicle with a long-term hold.
Avoid GDXJ. It's not what the name says, its long-term track record is poor, and there's no compelling reason to own it over either GDX or a concentrated individual portfolio.
Key Data Sources
- VanEck GDX ETF
- VanEck GDXJ ETF
- StockAnalysis — GDX Holdings
- StockAnalysis — GDXJ Holdings
- QuantPedia — Gold's Rally and the Gold Mining Stocks Trap
- FinanceCharts — GDX vs GLD Total Return
- PortfoliosLab — GDX vs GLD
- StockAnalysis — Individual stock returns: NEM, AEM, B, WPM, FNV, GFI, KGC, AGI, BTG, HMY, CDE, PAAS