Gold ETF (GLD) vs S&P 500

GLD closed at $418.50 on April 30, 2026, reflecting gold spot at approximately $4,680 per ounce after pulling back from a January 2026 ATH of $5,600. SPY traded at approximately $712, near its all-time high after recovering from the March 2026 Iran-war drawdown that briefly took the index to $650.

The GLD/SPY ratio sits at 0.59, well above the 2010 to 2022 average of 0.30 to 0.40 and reflecting the cumulative dominance of gold over equities since 2022. Over a 24-month window (April 2024 to April 2026), GLD has returned approximately +95 percent versus SPY at approximately +28 percent, an unusual divergence not seen since the 2009 to 2011 post-GFC recovery period when GLD outpaced SPY by a similar margin. The current configuration places GLD ahead of SPY on every meaningful trailing window from one month through 36 months.

GLD is a physically-backed gold ETF managed by State Street. Each share represents approximately 0.0922 ounces of gold held in HSBC's London vault. The fund holds approximately 880 tons of physical gold (April 2026), making it the largest gold ETF globally with $156.98 billion in assets under management.

GLD's tracking mechanics: the fund's expense ratio is 0.40 percent annually, which compounds to roughly 8 percent of cumulative tracking deficit over 22 years since the 2004 launch. At gold $4,680, the unaccented per-share value would be approximately $431.50 (gold/10), but GLD trades at $418.50, a $13 per-share gap that reflects accumulated expense-ratio drag plus tiny daily creation/redemption frictions. The 0.40 percent annual fee is meaningfully higher than IAU (0.25 percent) and AAAU (0.18 percent), which is why some allocators prefer those competitors for buy-and-hold positions. Daily tracking error against gold spot has historically averaged 0.02 to 0.04 percent, very tight.

Gold began 2024 trading near $2,070 per ounce. By December 2025 spot had reached $4,200; by January 2026 it broke $5,600 in a parabolic move driven by the combination of (1) record central bank buying near 1,000+ tons annually since 2022, (2) US fiscal deficits exceeding $2 trillion in fiscal 2025, (3) BRICS dedollarization announcements, and (4) Trump tariff policy adding inflation premium on top of pre-existing services-inflation stickiness. After the January 2026 peak, gold pulled back to $4,680 by late April as Iran ceasefire optimism and a brief commodity rotation reduced the safe-haven premium.

The 24-month return is the largest 24-month gold rally since the 1979 to 1980 Volcker crisis episode (which saw gold rise from $200 to $850 in 18 months). Importantly, the 2024 to 2026 run has occurred without a corresponding equity bear market: SPY has gained 28 percent over the same window, breaking the historical "gold rises while stocks fall" template. This is the first sustained period since 2010 to 2011 where gold and stocks have both rallied substantially while gold has dramatically outperformed.

Gold has outperformed SPY in only three sustained multi-year windows since GLD's 2004 launch: 2008 to 2011, 2018 to 2020 briefly, and 2022 to present. The current run is now in its third year with no clear top.

The causal driver this cycle is unusual: prior gold outperformance episodes coincided with falling real yields. The 2008 to 2011 episode took 10Y TIPS yield from +2 percent to -1 percent. The 2020 episode took TIPS yield from +0.5 percent to -1 percent during COVID Fed easing. The 2022 to 2026 episode is different: 10Y TIPS yield has averaged +1.5 to +2 percent throughout, with no decline in real yields. Gold rising despite stable real yields indicates that the marginal price-setter for gold has shifted from US-rate-sensitive Western institutional investors to less-rate-sensitive central banks (China, Russia, Turkey, India, Saudi Arabia) and dedollarization-motivated retail investors in emerging markets. The 2025-2026 leg of the rally has been disproportionately driven by central-bank buying and ETF inflows from non-US investors, which is why the traditional "real yields up means gold down" relationship has broken.

Central bank net gold purchases averaged 470 tons per year between 2010 and 2021. Beginning in 2022, the rate jumped to roughly 1,080 tons annually and has held there through 2025. World Gold Council data shows 2024 central bank net buying at 1,037 tons, with 2025 tracking similar.

The largest accumulators have been China (officially adding 5+ tons monthly through 2024 to 2025, possibly higher unofficially), Turkey (300+ tons), India (200+ tons), Poland (130 tons), and Saudi Arabia (160 tons). The structural shift was triggered by the February 2022 freezing of Russian central bank reserves, which demonstrated to non-aligned central banks that USD reserves carry political-asset confiscation risk. Central bank gold demand at 1,000+ tons annually represents 25 to 30 percent of annual mine supply (~3,500 tons), creating a persistent bid that did not exist in the 2010s. ETF outflows in 2022 to 2023 (Western institutions selling ~600 tons) were absorbed without price decline because central banks were buying the same size on the other side.

2008 to 2011 GFC and aftermath: GLD +60 percent versus SPY +5 percent (October 2008 to August 2011). The driver was Fed quantitative easing taking real yields negative for the first time in modern history, plus a $700 billion TARP fiscal expansion that raised debt-monetization fears.

2018 to 2020 brief: GLD +35 percent versus SPY +20 percent over a 24-month window ending August 2020. Driver was the COVID shock initially (March 2020 Fed emergency cuts) followed by sustained Fed easing and the largest fiscal expansion in modern history.

2022 to 2026 (current): GLD +95 percent versus SPY +28 percent (April 2024 to April 2026). Driver is structural central bank buying, fiscal-deficit dollar-debasement narrative, and dedollarization. Importantly, this is the first time gold has outperformed SPY without falling real yields, suggesting the structural bid has decoupled from traditional US-rates analysis. Probable forward path: gold continues to outperform SPY until either central bank buying decelerates materially (no current sign) or the US fiscal trajectory improves enough to restore confidence in USD as a store of value.

GLD has underperformed SPY in three meaningful regimes. The 2011 to 2015 disinflation cycle: gold fell from $1,920 (peak August 2011) to $1,050 (December 2015), a 45 percent drawdown, while SPY gained 70 percent on the back of post-GFC recovery. Driver: rising real yields plus restored dollar credibility plus equity earnings recovery.

2015 to 2018: gold flat at $1,200, SPY +50 percent on tax cuts and synchronous global growth. Driver: low and stable inflation, Fed normalization, no flight-to-safety demand.

Most relevantly for the current setup: the 2013 "taper tantrum" episode took gold from $1,650 to $1,200 (-27 percent) in 6 months when the Fed signaled QE wind-down. SPY was flat to up 5 percent over the same window. The lesson: gold underperforms sharply when real yields rise rapidly from negative territory toward positive, especially if the rise is policy-driven rather than growth-driven. The current cycle differs because real yields are already positive and the central bank bid does not appear to be policy-sensitive in the same way.

GLD is structured as a grantor trust holding physical gold, which means under IRS rules it is taxed as a "collectible." Long-term capital gains on GLD held over 12 months are taxed at the collectibles rate: 28 percent for top earners, plus 3.8 percent net investment income tax, for an effective 31.8 percent. SPY long-term gains are taxed at 20 percent plus 3.8 percent NIIT, for an effective 23.8 percent.

The 8 percentage point after-tax differential is meaningful. On the current 24-month outperformance window, GLD's pretax 95 percent return becomes roughly 65 percent after collectibles tax. SPY's pretax 28 percent return becomes 21 percent after long-term capital gains tax. The after-tax outperformance gap shrinks from 67 percentage points pretax to 44 percentage points after-tax, a 23 percentage point compression. This is why some allocators access gold through GDX (gold-miner ETFs, taxed at standard 20 percent) or through gold-mining stocks directly, accepting higher equity beta in exchange for better tax treatment. For tax-deferred accounts (IRA, 401k), the differential disappears.

Modern portfolio theory suggests gold allocations of 5 to 10 percent of total portfolio for the diversification benefit, based on gold's low to negative correlation with equities and bonds. The current cycle has reinforced that case: gold has provided meaningful diversification against the 2022 simultaneous decline in stocks and bonds (the "60/40 worst year since 1937" episode).

A practical framework: in a normal 60/40 portfolio, replace 5 to 10 percent of the equity allocation with GLD. So a 60 percent equity / 40 percent bond allocation becomes 50 to 55 percent equity / 5 to 10 percent gold / 40 percent bond. Backtested against the 2000 to 2026 period, this allocation produces roughly 0.5 to 1.0 percentage point lower annualized return than pure 60/40, but with materially lower drawdowns during stress episodes (2000-2002, 2008-2009, 2020, 2022). The allocation has been accretive to risk-adjusted returns post-2022 because gold has run while bonds have been disappointing diversifiers. Forward-looking question: does the central-bank-buying-driven structural bid persist? If yes, GLD allocation closer to 10 percent makes sense. If central banks slow buying materially, the historical 5 percent allocation case reasserts itself.