Executive summary
March 2026 was gold’s worst month in more than 17 years. The metal opened at $5,321 and closed at $4,670, a 12.2% decline that Reuters confirmed was the steepest monthly drop since October 2008.
The culprit was not a collapse in gold’s fundamental case. It was an oil shock. The US-Iran war that began in late February pushed crude through $100 per barrel, reignited inflation expectations, forced the Federal Reserve to shelve rate cuts indefinitely, and sent the dollar surging toward multi-year highs. Every one of those dynamics — higher inflation, higher rates, stronger dollar — hit gold simultaneously.
And yet. Our Gold Safe-Haven Premium Score (GSHPS v2) dropped to 21.5 on March 18 — deep in the Buy zone for the first time since early 2024. The indicator is saying what the headline price is not: gold has pulled back toward the bottom of its recent premium range, market stress is elevated, and the conditions that historically precede recoveries are present.
This report covers what happened in March, why the selloff was mechanical rather than fundamental, and why the April outlook — anchored by a fragile ceasefire and a GSHPS Buy signal — may represent the best entry point this year.
Market performance
The month in numbers
| Metric | March 2026 | vs. Prior Month | vs. Prior Year |
|---|---|---|---|
| Open (Mar 1) | $5,321 | - | - |
| Monthly High | $5,321 (Mar 1) | - | - |
| Monthly Low | $4,424 (Mar 26) | - | - |
| Close (Mar 31) | $4,670 | −12.2% | +22.4% |
| Average Price | ~$4,890 | - | - |
The decline was not evenly distributed. Gold drifted lower through the first two weeks as the Iran war drove oil steadily higher, then accelerated sharply in the week of March 23–26, falling over 8% in five sessions as rate-cut expectations collapsed entirely. A 3.2% recovery on the final trading day of the month trimmed the damage, but the monthly figure remains stark.
This was the eighth consecutive month of war-era volatility. January saw the all-time high of $5,608. February absorbed the worst single-day crash since 1983. March delivered the worst monthly decline in a generation. The range between January’s peak and March’s trough — $5,608 to $4,424 — represents a 21% drawdown. By any standard, this has been an extraordinarily violent first quarter.
Regional context
- UK: Gold in sterling fell from approximately £4,150/oz to £3,640/oz — a roughly 12% decline — though the pound’s own weakness against the dollar cushioned some of the blow. Year-to-date, gold remains up approximately 8% in GBP terms
- China: SGE premiums surged to $15–20 above spot during the late-March selloff, confirming aggressive dip-buying by Chinese retail and institutional investors. Physical demand rose as the paper market sold
- India: Rupee weakness offset some of the dollar-denominated decline. Local prices remained elevated enough to suppress jewellery demand, but investment buying held firm
GSHPS analysis: why the model says Buy
Our proprietary Gold Safe-Haven Premium Score (GSHPS v2) dropped to 21.5 on March 18 — the first time it has entered the Buy zone (≤25) since February 2024.
What the components show
| Component | March 18 Reading | Interpretation |
|---|---|---|
| Gold price | $4,890 | Well below January’s $5,608 peak |
| Raw premium over 200-day SMA | +20.7% | Still elevated in absolute terms |
| Premium relative position (60-day) | 0.058 | Near the bottom of its recent range — the key v2 insight |
| VIX normalised | 0.50 | Elevated stress (VIX near 25) |
| DXY stress | 0.54 | Dollar near multi-year highs — moderate headwind |
| RSI (14) | 42.5 | Below 50 — momentum has turned neutral to bearish |
| Stress factor | 0.50 | Elevated, which in v2 lowers the score (stress justifies the premium) |
| Premium floor | 21.5 | Active — prevents false Buy when premium is still 20%+ above SMA |
| Final GSHPS v2 | 21.5 | Buy zone |
The critical reading is the premium relative position at 0.058. This means gold’s current premium, while still 20.7% above its 200-day SMA in absolute terms, sits near the very bottom of its 60-day range. Gold was 30–40% above the SMA as recently as January. The pullback is significant relative to where it has been trading.
v2’s inversion of the stress factor is doing exactly what it was designed to do: with VIX elevated and the dollar strong, genuine market stress is present. That stress justifies a safe-haven premium rather than suggesting gold is overextended. The model recognises this and pushes the score lower — toward Buy.
Historical context for GSHPS Buy signals
The backtest data shows what happened after previous GSHPS v2 Buy signals:
| Date | Score | Gold Price | Forward 30 Days | Forward 60 Days |
|---|---|---|---|---|
| Mar 2020 | 21.6 | $1,589 | +8.4% | +6.9% |
| Aug 2021 | 23.9 | $1,723 | +3.1% | +3.8% |
| Feb 2024 | 24.0 | $1,993 | +9.9% | +17.0% |
| Mar 2026 | 21.5 | $4,890 | ? | ? |
Previous Buy signals have preceded positive returns over 30 and 60 days in the majority of cases. The February 2024 signal — the most recent before this one — preceded a 17% gain over 60 days as gold broke decisively above $2,000.
The model does not predict prices. It measures the relationship between gold’s premium and the level of stress in the market. When that relationship indicates the premium has compressed while stress remains elevated, the historical track record is constructive.
What drove the selloff
1. The oil shock
The dominant factor. The US-Iran war that began on February 28 pushed oil from approximately $70 per barrel to above $100 through March. WTI crude hit $96.32 on March 18; Brent touched $118–119 intraday. The Strait of Hormuz — carrying roughly 20% of global oil production — remained closed for the entire month.
Oil above $100 does two things that hurt gold:
- It reignites inflation expectations, which forces the Fed into a hawkish posture
- It strengthens the dollar as energy importers sell local currencies to buy dollar-denominated crude
The paradox: the geopolitical crisis that should support gold’s safe-haven role instead undermined it, because this particular crisis hit energy markets first and the inflationary consequences overwhelmed the safe-haven bid.
2. The Fed’s hawkish pivot
The FOMC met on March 17–18 and held rates at 3.5–3.75%. That decision was expected. What was not expected was the tone.
Officials raised their 2026 inflation outlook, citing the oil shock. The minutes, released in April, revealed growing concern that inflation could remain above 2% indefinitely. Several policymakers suggested further rate hikes might be necessary.
The market response was immediate. Rate-cut expectations — which had been priced for mid-2026 as recently as February — were pushed to late 2027. Some traders began pricing rate increases. The shift in the rate trajectory was the single largest driver of the late-March acceleration in gold’s decline.
3. The dollar surge
DXY moved from approximately 98–99 at the start of March to above 100 by mid-month, closing near 99.96. The dollar’s strength reflected both the safe-haven bid into US assets and the mechanical effect of oil pricing: when oil rises, dollar demand rises with it.
Gold priced in dollars faces a double headwind when the dollar strengthens — the commodity loses purchasing power at the same time as alternatives (Treasuries, dollar cash) become more attractive.
4. Institutional profit-taking and ETF liquidation
Global gold ETFs recorded a record $12.8 billion in outflows during March — the worst monthly performance in over a decade. This broke a nine-month inflow streak and reduced year-to-date net flows to approximately $11.7 billion.
The pattern was clear: Western institutional investors, who had built significant positions through the Q4 2025–Q1 2026 rally, used the oil shock as a catalyst to take profits. The selling was concentrated in North American and European funds. Asian ETFs, by contrast, saw continued modest inflows — the same dip-buying behaviour visible in physical markets.
Demand analysis
Central bank buying: the floor holds
Despite the price decline, central bank demand remained resilient:
- China: Extended its accumulation streak, with holdings reaching approximately 2,309 tonnes (9.6% of total reserves). The PBOC has now bought gold in 16 consecutive months
- Poland: Continued its buying programme, now holding 480 tonnes (20% of reserves)
- Czech Republic, Malaysia, Indonesia: All reported net purchases through Q1
- Exceptions: Turkey sold approximately 60 tonnes ($8 billion) to defend the lira amid war-related energy costs. Russia sold 14 tonnes to address budget deficits
The World Gold Council maintains its forecast for approximately 850 tonnes of central bank purchases in 2026 — consistent with the 1,000+ tonne pace of 2023–2025 when adjusted for the Turkish and Russian exceptions.
Central bank buying is the structural floor beneath gold. These buyers are not momentum traders. They are executing multi-year reserve diversification strategies that are insensitive to single-month price moves.
Physical demand
Physical retail demand split along predictable lines:
- Asia: Aggressive dip-buying, particularly in China (SGE premiums $15–20) and India
- Western markets: Mixed — some profit-taking on existing holdings, but UK dealers reported firm demand for Sovereigns and Britannias at the lower prices
- ISA season: The April 5 deadline drove buying of gold ETCs within Stocks & Shares ISAs, with dealers and platforms reporting above-average activity despite the price decline
Supply dynamics
Mine production
Global mine supply remains constrained at approximately 3,600 tonnes annually. AISC (all-in sustaining costs) continue to rise, now averaging approximately $1,350–1,450/oz. At $4,670 gold, producer margins remain healthy at 60%+ — though significantly compressed from the 75%+ margins at January’s highs.
The oil shock adds a specific cost headwind: energy accounts for 15–25% of mining operating costs. With diesel and power prices elevated, margins face additional pressure that the headline gold price alone does not capture.
Recycling
Scrap supply surged in March as consumers and dealers sold into the remaining elevated prices. Recycling is price-elastic in ways mine production is not — and the rapid price decline through the month created urgency among holders looking to exit before further falls.
Macro environment
Stagflation enters the conversation
March’s macro backdrop was dominated by a single word: stagflation. The combination of:
- Oil above $100/barrel (inflation)
- Global equities falling (the S&P 500 had its worst month since September 2022)
- Rate cuts postponed indefinitely (tightening expectations)
- Consumer confidence declining
…created the textbook conditions for a stagflationary environment. For gold, stagflation is historically mixed: the inflation component supports prices, but higher real yields and dollar strength can offset.
Interest rate expectations
| Measure | End of March 2026 | End of February 2026 |
|---|---|---|
| Fed Funds Rate | 3.5–3.75% (unchanged) | 3.5–3.75% |
| 10-Year Treasury | ~4.5% | ~4.3% |
| Next expected rate cut | Late 2027 | June–July 2026 |
| CPI (Feb data) | 2.8% | 2.4% (Jan) |
| DXY | ~100 | ~97 |
| VIX | ~25 | ~20 |
The shift in rate expectations is the most significant change. In February, markets priced 42–50 basis points of cuts through 2026. By end-March, cuts had been priced out entirely. This repricing was the primary mechanical driver of gold’s decline.
Geopolitical backdrop
The US-Iran war dominated March:
- Strait of Hormuz remained closed throughout the month, causing the largest oil supply deficit on record
- IEA coordinated a 400-million-barrel strategic petroleum reserve release — insufficient to offset the full disruption
- VIX spiked to 35 intraday on March 20 when oil breached $100/barrel
- Global stocks suffered their worst month since September 2022
- The Warsh Senate confirmation hearings for incoming Fed Chair were delayed by the conflict
Russia-Ukraine talks remained deadlocked. US-China trade tensions persisted with the 15% global tariff still in force.
Technical outlook
Gold’s technical picture has shifted from constructive to neutral following the 12.2% decline.
Key Levels:
- Support: $4,424 (March low, critical), $4,300 (approximate 200-day SMA as of late March), $4,000 (psychological)
- Resistance: $4,890 (mid-March pivot), $5,000 (psychological), $5,321 (March open / February close)
The 200-day SMA, sitting near $4,300, has not been tested. Gold closed March approximately 8% above this level. A test of the 200-day would be technically significant — it would be the first time gold has traded at its long-run baseline since mid-2025.
RSI on weekly charts has fallen to approximately 42 — the most oversold reading since mid-2024. Combined with the GSHPS Buy signal, this suggests the selling pressure has overshot in the near term. However, the monthly candle (a large red bar) is the kind of technical pattern that keeps momentum sellers active into the following month.
The recovery on the final trading day (+3.2%) is modestly encouraging — it suggests buyers emerged at lower levels and prevented a monthly close near the lows.
Mining equities
Mining equities declined alongside gold but showed resilience relative to the metal.
| Index | March Return | Gold Beta | YTD |
|---|---|---|---|
| GDX (Major Miners) | −8.4% | ~0.69 | +14.2% |
| GDXJ (Junior Miners) | −10.7% | ~0.88 | +18.6% |
| Gold Spot | −12.2% | 1.00 | - |
Both GDX and GDXJ outperformed gold on the downside — the first month in this cycle where miners demonstrated defensive characteristics. This is meaningful. At January’s highs, miners were trading at approximately 0.85x NAV against a historical average of 1.2x. The March decline compressed NAVs (because the gold price fell) but share prices fell by less, implying the market was already pricing in a correction and views the structural thesis as intact.
Mid-tier producers with AISC below $1,100/oz remain particularly compelling. These companies are generating substantial free cash flow even at $4,670 gold and are using it for share buybacks rather than the empire-building acquisitions that destroyed value in previous cycles.
Sister metals
Silver
Silver fell 15.9% in March, from $89.31 to $75.14 — amplifying gold’s decline as it always does. The gold-silver ratio widened from approximately 60:1 to 62:1. Silver’s industrial exposure made it doubly vulnerable: the oil shock both threatened industrial activity (negative for silver demand) and raised inflation (which pushed rate expectations against all precious metals).
At $75, silver sits at the lower boundary of BNP Paribas’s projected $65–75 trading range for 2026. The structural supply deficit (67 million ounces forecast for 2026) and China’s export restrictions remain in force, but thrifting and economic uncertainty continue to create legitimate headwinds.
Platinum
Platinum fell 15.3%, from $2,310 to $1,956. The World Platinum Investment Council projects a fourth consecutive annual supply deficit of 240,000 ounces in 2026. The platinum-gold discount narrowed slightly to approximately 58% but remains near historical extremes.
Risk factors
Near-term
- Ceasefire fragility: The April 7 US-Iran ceasefire is two weeks old and already under strain. Iran’s military controls Strait shipping. Israel’s Lebanon campaign is excluded. Collapse would send oil back above $100 and gold into another volatile phase
- Fed hawkishness: If March FOMC minutes (released April 8) reveal serious rate-hike discussion, gold faces further downside from the rates channel
- China economic data: A material slowdown in Chinese demand would remove one of gold’s key physical-market supports
Structural tailwinds
- Central bank buying continues: 850+ tonnes forecast for 2026. De-dollarisation is a multi-decade trend, not a trade
- GSHPS Buy signal: The indicator’s historical track record at sub-25 readings is overwhelmingly positive over 30 and 60 days
- Ceasefire resolution: If the ceasefire holds, oil normalises, inflation expectations ease, rate cuts return to the table, and the entire March headwind reverses
April outlook: forward-looking sentiment
The ceasefire changes everything — or nothing
On April 7, President Trump announced a two-week ceasefire with Iran, contingent on the immediate reopening of the Strait of Hormuz. As of publication (April 9), the ceasefire is holding but under severe strain. Iran’s military is managing Strait shipping. Tehran has accused Washington of violations. Israel’s campaign in Lebanon is excluded.
If the ceasefire holds and transitions to a longer-term arrangement:
- Oil falls toward $70–80/barrel
- Inflation expectations ease
- Rate cuts return to 2026 pricing
- The dollar weakens
- Gold recovers toward $5,000+
If the ceasefire collapses:
- Oil returns above $100
- The stagflation narrative intensifies
- Gold’s safe-haven bid re-engages but is again partially offset by dollar strength and rate expectations
- Net effect is probably modestly gold-positive but with extreme volatility
GSHPS forward view
The GSHPS v2 Buy signal at 21.5 provides a quantitative anchor for the April outlook. The model’s components as of early April:
| Component | Direction | Impact on Score |
|---|---|---|
| Premium relative position | Compressed — near 60-day lows | Pushes score lower (bullish) |
| VIX | Elevated but declining from March highs | Stress justifies premium — pushes score lower |
| DXY | Easing from 100+ toward 99 | Headwind receding — modestly supportive |
| RSI | Below 50, recovering from oversold | Momentum turning — neutral to supportive |
| Absolute premium floor | Active at 20%+ above SMA | Prevents false Buy, but score still in zone |
The model’s message is clear: gold has corrected far enough, fast enough, that the risk-reward has shifted. The premium has compressed relative to recent history, genuine stress remains in the market, and the absolute premium floor is keeping the signal honest (it would not flash Buy if gold were still 40% above its SMA).
Our base case for April is that the ceasefire provides a floor, institutional buyers re-enter at levels 12–20% below January’s highs, and gold consolidates in the $4,700–$5,100 range before the next directional move.
Sentiment: cautiously bullish
We maintain a bullish outlook, but the conviction level is lower than in January or February. The structural case (central banks, dollar diversification, supply constraints) is unchanged. The tactical case (GSHPS Buy signal, oversold technicals, ceasefire potential) is constructive. But the macro backdrop (higher oil, postponed rate cuts, stronger dollar) is genuinely adverse in ways it was not three months ago.
The honest assessment: gold at $4,670 offers a better entry point than gold at $5,500. Whether April delivers the recovery depends on events in the Strait of Hormuz that no model can predict.
Price forecast
We revise our outlook to reflect March’s correction and the elevated uncertainty heading into Q2.
| Scenario | Probability | Price Range |
|---|---|---|
| Base Case | 50% | $4,750–$5,200 |
| Bull Case | 25% | $5,200–$5,600 |
| Bear Case | 25% | $4,000–$4,750 |
Revised Target: $4,750–$5,500 by end of Q2 2026
The base case assumes the ceasefire transitions to a longer-term arrangement, oil stabilises below $90, one rate cut returns to pricing by September, and central bank buying continues at pace. The bull case incorporates a swift geopolitical resolution, ETF inflows resuming, and the January record being retested. The bear case assumes ceasefire collapse, oil above $100 through Q2, rate hikes entering discussion, and a test of the 200-day SMA near $4,300.
Major bank targets as of early April: Goldman Sachs at $5,400 (year-end), CPM Group forecasts a $4,000–$5,000 trading range, BMI maintains a $4,600 full-year average.
What to watch
- Iran ceasefire durability: The two-week window expires around April 21. Extension or escalation is the single most important variable for gold, oil, rates, and equities simultaneously
- March FOMC minutes (released April 8): The detail on rate-hike discussion will set the tone for April’s rates pricing. Early signs suggest several officials raised the possibility
- April ETF flow data: Whether the $12.8 billion March outflow reverses or continues will signal institutional conviction
- China PBOC April report: Whether the 16-month buying streak extends. Any pause would be a significant sentiment negative
- Oil price trajectory: Every $10 move in crude shifts the inflation outlook, rate expectations, and dollar, all of which flow directly into gold
- Warsh confirmation timeline: The incoming Fed Chair’s commentary on balance sheet reduction and rate path will shape expectations through H2 2026
Conclusion
March was ugly. There is no point pretending otherwise. Gold’s 12.2% decline was the worst monthly performance in 17 years, driven by an oil shock that simultaneously raised inflation, killed rate cuts, and strengthened the dollar — the worst possible combination of macro headwinds.
But ugly months have preceded the best entry points in gold’s history. Our GSHPS v2 indicator, which specifically measures whether gold’s premium has compressed while stress remains elevated, flashed Buy for the first time since February 2024. The last time it did so, gold gained 17% over the following 60 days.
The structural case has not changed. Central banks bought through the selloff. China’s dip-buying was visible in real time. The supply deficit persists. The geopolitical backdrop that supports long-term gold ownership has, if anything, intensified.
What has changed is the price. Gold at $4,670 is 17% below January’s high. For investors who believed in the thesis at $5,500, the thesis is stronger at $4,670 — you are buying the same asset, with the same fundamentals, at a meaningful discount.
The ceasefire is the wildcard. If it holds, the oil premium fades, rate cuts return, the dollar weakens, and gold recovers. If it fails, volatility returns — but so does the safe-haven bid. Either way, the GSHPS model suggests the risk-reward at current levels favours buyers.
March tested conviction. April will test patience. The model says wait — and buy.
This report reflects market conditions and data available as of April 9, 2026. The US-Iran ceasefire remains fragile and rapidly evolving. GSHPS readings are based on our proprietary methodology — past signals do not guarantee future results. This is not financial advice.