The Fed Hiked Into an Oil Shock Last Month. What History Says Happens to Gold Next

The Fed's September 2026 hike was the first in over three years, driven by an oil-shock inflation. Across every Fed tightening cycle since the 1970s, gold's fate hinged on real rates and the official-sector bid, not on hikes, which frames four analog-based paths to end-2027. Late September's core PCE report release (3.0% YoY) and a weak September jobs report complicated the narrative for the Fed and future rate hike possibilities. And gold is still struggling around $4,130 per troy ounce.

Ruddy Setiadi Gunawan

Main findings

On September 16, 2026, the Federal Reserve raised its policy rate for the first time in more than three years, a quarter-point move to a 3.75–4.00% target range, because an oil shock had pushed inflation back up. We tested what gold did in every comparable Fed tightening cycle since the late 1970s and tracked the 2026 evidence. Four findings:

1: rate hikes alone say little about gold. Across ten completed tightening cycles since 1977, gold rose in five, fell in four and was flat in one. What separates the outcomes is the *real* policy rate, funds rate minus inflation, and the dollar, not the number of hikes.

2: what sorts the record is where real rates ended, not where they started. In cycles where the real policy rate ended at +4.5pp or more (1980–81, 1983–84, 1988–89), gold fell 16–27% every time. Where it ended at +2.6pp or less (1977–80, 2004–06, 2015–18, 2022–23), gold rose in three, +320% in 1977–80, and was flat in 2022–23. The three that ended near +3pp were small and mixed. Starting points sort nothing: 1980–81 and 2022–23 both began with negative real rates. September 2026 begins at +0.3pp, and where it ends is the open question.

3: the old "gold versus real yields" rule broke in 2022–24, mostly because of central banks, and has partly reasserted itself since. From 2006 to 2021 gold fell in about two of every three months when real yields rose (a −0.53 monthly correlation). From January 2022 to December 2024 real yields rose 2.8 points and gold still gained 46%. Since then the direction is textbook again: gold ran another 90% to its February 2026 average as real yields eased, then fell about 18% to late-September spot as they climbed to 2.93%. Central banks buying is the leading explanation for this break, although 2026's quarterly buying did not pump the price anymore.

4: an oil shock is not automatically a gold bull or bear case. On the monthly averages oil is up ~67% in 2026 until late September and gold is slowly going lower; on daily spot, oil is up ~68% (EIA, through October 6) and gold is down ~4% from New Year until the afternoon of October 8. The inflation channel helps gold, but the hawkish Fed hurts it, and in September the Fed hawkish narrative won decisively.

Spot gold slid through $4,200 to ~$4,135 by September 28, as much as 26% below the record, as the 10-year Treasury yields reached 5.27% and the dollar going stronger. On October 8, crude also jumped about 5% on a record run of tanker attacks in Hormuz, and gold again fell flat.

The important signals appeared late last month and early this month, with core PCE reported at 3.0% YoY (below consensus), payrolls +29,000 (below 89,000 expectation), and 4.2% unemployment. October Fed rate hold odds moving to roughly 75–85%, yet gold still closed near $4,130 after previous spike faded. The signal is priced in, and gold price recovery waits for the Fed to act. The scenarios through end-2027 still span roughly between −14% to +33%.

Where things stand

Gold spot, Oct 8 a.m.: 4130 USD/troy oz Kitco spot bid, Oct 8, 5:11 p.m. ET. Down about -0.21% in the past 7 days. World Bank monthly-average basis: $4,319 in Sep 2026.
Below January record: -26.1 % Oct 8 morning spot vs intraday record $5,589.38 on 2026-01-28 (CBS News).
Fed funds, upper bound: 4 % Target range 3.75–4.00% since the 2026-09-16 decision (effective 2026-09-17).
2026 core PCE, SEP median: 3.4 % FOMC Summary of Economic Projections, 2026-09-16.
WTI crude: 90.57 USD/bbl Spot Price per 2026-10-08 (latest published).
Central-bank buying, Q2 2026: 289 tonnes WGC Gold Demand Trends; Q1 2026 revised to 56.5t.

Every figure on its stated basis and date; the bases differ by design (benchmark vs spot vs futures) and are reconciled in the methodology.

Nine months, four regimes

2026 has been four different gold markets in nine months.

January was the melt-up. Gold crossed $5,000 for the first time on January 26 and printed an all-time high near $5,589–5,595 on January 28–29 (CBS News recorded $5,589.38). Silver spiked above $120. On Friday January 30 came the largest one-day fall since 1980: most-active futures settled 11.4% lower at $4,745.10 (widely reported), with spot closing near $4,895 about $700 off the record (The Conversation documents the session). The trigger was political: early news of Kevin Warsh's nomination as Fed chair. Exchanges raised margins repeatedly afterward (initial requirements 8% to 9% in early February, speculative toward 19%, as reported), leverage, not valuation, set the tone.

February to April was the war bid. The 2026 Iran war began February 28 with US–Israeli strikes; WTI jumped from the $60s to a $114.58 peak on April 7 (EIA Cushing spot). Gold's monthly average peaked at $5,020 in February, then drifted lower as oil climbed.

May to August was the inflation-and-ETF phase. CPI inflation accelerated from 2.4% year-over-year in January to a 4.2% peak in May (BLS via FRED). The July FOMC held at 3.50–3.75% on a 9–3 vote with three dissents *for* a hike (minutes), the loudest possible signal that September was live. Gold ETFs took in $18bn (+121t) in August alone, lifting global holdings to a record 4,189t (WGC).

September was the reckoning and it is still running. The Fed hiked interest rates, and gold sold off into and after it: December futures $4,293–4,324 on September 24 (Yahoo), with spot near $4,360 a day earlier before the slide. The pressure then became a break: after gold's worst week in a month on a hawkish outlook (lows ~$4,244, FXStreet), a 10-year Treasury yield at 5.27% pushed gold spot through $4,200 to ~$4,135 by September 28, down 3.5% on the day to the low of ~$4,110, roughly 26% below the record.

Higher yields and a firmer dollar kept pressuring gold price. On October 8 the oil narrative has also returned. Brent rose as much as 5% to about $102.55 and WTI to about $90.50 after Kpler counted a record ten tankers struck in the Strait of Hormuz in the week to October 4, with tanker transits down to seven on October 6, the fewest since July 23 (CNN). Gold found no haven bid in this, an early gain to about $4,144 faded to roughly $4,130 by late afternoon in New York time, and there's still no end in sight for the current price pressure on gold.

Gold, January 2025 – October 2026

Period	Gold (XAU/USD) (USD/troy oz)
Jan '25	2710
Feb '25	2895
Mar '25	2983
Apr '25	3218
May '25	3309
Jun '25	3353
Jul '25	3340
Aug '25	3368
Sep '25	3668
Oct '25	4058
Nov '25	4087
Dec '25	4309
Jan '26	4753
Feb '26	5020
Mar '26	4856
Apr '26	4721
May '26	4587
Jun '26	4228
Jul '26	4073
Aug '26	4411
Sep '26	4319
Oct 8	4130

Monthly average, USD/troy oz, World Bank Pink Sheet, Jan 2025 – early October 2026. Monthly averages smooth the extremes: the highest point is February's $5,020 average, against an intraday record near $5,589. Key dates: gold crosses $5,000 Jan 26; intraday record ~$5,589–5,595, Jan 28–29; Jan 30 crash (futures −11.4% to $4,745.10); war begins Feb 28; WTI peaks Apr 7; CPI peaks at 4.2% y/y in May; record ETF inflows Aug; Fed hikes Sep 16; spot ~$4,135 by Sep 28.

Source: World Bank Pink Sheet, monthly prices (CC BY 4.0); final point: Kitco spot

Rate hikes and gold: the record

For each Fed's hiking cycle, gold is measured on one basis throughout, the World Bank's monthly average, from the first-hike month to the last-hike month. As mentioned before, gold rose in five cycles, fell in four and was flat in one. The average full-cycle change is positive but dominated by two outliers (the period of 1977–80, and 2004–06). The median cycle, however, is roughly flat.

What the record supports is a sort by where *real* rates ended, and the dollar's strength. Where the real policy rate ended at +4.5pp or more, gold fell every time: 1980–81 (−27%), 1983–84 (−17%), 1988–89 (−16%). Where it ended at +2.6pp or less, gold rose in three cycles: 1977–80 (+320%), 2004–06 (+52%), 2015–18 (+16%), and was flat in the fourth, 2022–23, despite real rates rocketing from −8.4pp to +1.8pp.

The three cycles that ended near +3pp were small and mixed: 1987 (+13%, with real rates easing from an already-high +5pp), 1994–95 (−1%) and 1999–2000 (+5%). Neither the starting level nor the size of the rise sorts the record: 1980–81 and 2022–23 both began with negative real rates, and real rates rose about 3pp in 1977–80 and 2004–06 while gold climbed (in 2004–06 peaking near +1.2pp, April 2006, as the dollar fell). The level real rates reach, not the act of hiking, is what the record punishes, and 2026's end level is not yet known.

Gold in every hiking cycle

Period	First → last hike (%)	First 12 months (%)
1977–80	319.7	31.1
1980–81	-26.5	-34.6
1983–84	-17.1	-6.2
1987	12.7	16.9
1988–89	-16.4	-12.2
1994–95	-1.3	-1.3
1999–00	5.4	9.6
2004–06	52	9.9
2015–18	16.2	7.5
2022–23	0.2	-1.8

Gold price change studied during the Fed tightening cycles, World Bank monthly-average basis: first-hike month to last-hike month, and first 12 months. Scale: 1977–80's +320% sets the vertical axis, while the other nine full-cycle changes run from −27% to +52%.

Source: Calculated from World Bank Pink Sheet (gold) and FRED FEDFUNDS/DFEDTAR/DFEDTARU (cycles)

Replaying six Fed hiking cycles

1977–80: the Fed behind an oil shock

The 1979 oil shock hit mid-cycle, CPI inflation climbed from 5% to nearly 15%, and the Fed raised rates only about as fast as prices rose: the real funds rate hovered around zero for three years - negative in 22 of the cycle's 39 months, never below −2.2pp, and turned decisively positive (+2.6pp) only at the very end, in the March-1980 credit squeeze. Gold quadrupled: +320% across the cycle, +31% in the first year. This is what 'behind the curve' looks like: hikes happened, but real rates went nowhere while inflation nearly tripled. It is also the only cycle in the sample driven by an oil shock like 2026's, which is why it anchors our bullish scenario.

-6: 89.4 index
-5: 83.3 index
-4: 86.4 index
-3: 87.9 index
-2: 98.5 index
-1: 101.5 index
0: 100 index
+1: 103 index
+2: 112.1 index
+3: 112.9 index
+4: 111.4 index
+5: 106.8 index
+6: 108.3 index
+7: 109.8 index
+8: 113.6 index
+9: 120.5 index
+10: 122.7 index
+11: 121.2 index
+12: 131.1 index
+13: 134.8 index
+14: 139.4 index
+15: 132.6 index
+16: 133.3 index
+17: 139.4 index
+18: 143.2 index
+19: 156.1 index
+20: 160.6 index
+21: 172 index
+22: 156.1 index
+23: 157.6 index
+24: 172 index

1980–81: Volcker overwhelms the shock

The mirror image. Paul Volcker's Fed jammed the funds rate above inflation: the real policy rate swung from −3.3pp to +9.4pp in ten months, the dollar surging after the 1981 tightening. Gold, which peaked with the 1980 inflation panic, fell 27% across the cycle, 35% in the first year. The inflation was the same war-born kind; the *reaction function* was the opposite. Same shock, opposite outcome, the Fed's response, not the shock, did the work.

-6: 106.1 index
-5: 88.4 index
-4: 82.5 index
-3: 82 index
-2: 95.9 index
-1: 102.6 index
0: 100 index
+1: 107.5 index
+2: 105.4 index
+3: 99.7 index
+4: 85.8 index
+5: 88.8 index
+6: 79.7 index
+7: 79.6 index
+8: 79.1 index
+9: 76.6 index
+10: 73.5 index
+11: 65.2 index
+12: 65.4 index
+13: 70.8 index
+14: 69.9 index
+15: 65.9 index
+16: 65.4 index
+17: 61.2 index
+18: 59.6 index
+19: 52.6 index
+20: 55.8 index
+21: 53.3 index
+22: 50.2 index
+23: 54.1 index
+24: 58.1 index

1983–84: conventional disinflation

A quieter version. Real rates started high (+5.2pp) and were pushed to +7.3pp by August 1984, the dollar in a structural bull market. Gold slid 17% with a 20% peak drawdown. 1983 was no oil-shock inflation, the Fed was cementing disinflation. Takeaway for 2026: when policy is already restrictive in real terms and getting more so, gold's hedge bid loses to the opportunity-cost bill.

-6: 104 index
-5: 100.5 index
-4: 98.8 index
-3: 106 index
-2: 114.5 index
-1: 116.9 index
0: 100 index
+1: 103.1 index
+2: 104.3 index
+3: 98.3 index
+4: 96 index
+5: 99 index
+6: 98.1 index
+7: 93.8 index
+8: 91 index
+9: 92.4 index
+10: 88.3 index
+11: 91.9 index
+12: 93.8 index
+13: 90.7 index
+14: 89.8 index
+15: 90 index
+16: 82.9 index
+17: 82.9 index
+18: 81 index
+19: 81.2 index
+20: 81.4 index
+21: 76.2 index
+22: 72.1 index
+23: 71.2 index
+24: 74.8 index

2004–06: the honest exception

Seventeen quarter-point hikes, and gold rose 52% anyway, which was the strongest non-1970s result. Why it doesn't break the story: real rates were negative when the cycle began and, even at their April-2006 peak, barely reached +1.2pp (the hikes merely normalized an emergency setting), the dollar was falling through the post-2002 diversification wave, and the 2000s commodity/emerging-market boom did the rest. 2004–06 is the counterexample that proves the driver is the *level and direction of real rates and the dollar*, not the act of hiking.

-6: 103.8 index
-5: 105.6 index
-4: 103.3 index
-3: 103.8 index
-2: 102.8 index
-1: 98 index
0: 100 index
+1: 101.5 index
+2: 102.3 index
+3: 103.3 index
+4: 107.1 index
+5: 112 index
+6: 112.8 index
+7: 108.2 index
+8: 107.9 index
+9: 110.7 index
+10: 109.4 index
+11: 107.7 index
+12: 109.9 index
+13: 108.2 index
+14: 111.7 index
+15: 116.3 index
+16: 119.9 index
+17: 121.7 index
+18: 130.1 index
+19: 140.3 index
+20: 141.6 index
+21: 142.1 index
+22: 155.9 index
+23: 172.2 index
+24: 152 index

2015–18: the glacial normalization

Nine hikes over three years, the slowest tightening in the sample, from a zero floor. Real rates stayed near zero (+0.3pp by the end); the broad dollar rose just 3%. Gold ground upward, +16% across the cycle, with a 14% drawdown mid-way. The closest structural cousin to a 'hike-and-hold' path: policy normalizes, inflation never overheats, gold neither melts up nor breaks down.

-6: 109.9 index
-5: 104.8 index
-4: 103.9 index
-3: 104.6 index
-2: 107.7 index
-1: 100.9 index
0: 100 index
+1: 102 index
+2: 111.5 index
+3: 115.7 index
+4: 115.4 index
+5: 117.2 index
+6: 118.6 index
+7: 124.3 index
+8: 124.5 index
+9: 123.3 index
+10: 117.8 index
+11: 115.1 index
+12: 107.5 index
+13: 110.8 index
+14: 114.7 index
+15: 114.4 index
+16: 117.8 index
+17: 115.8 index
+18: 117.1 index
+19: 115 index
+20: 119.2 index
+21: 122.1 index
+22: 119 index
+23: 119.1 index
+24: 117.5 index

2022–23: the rule breaks

On the textbook model this should have been gold's worst cycle: real rates exploded from −8.4pp to +1.8pp (policy), TIPS from −0.7% to +1.6%, the dollar jumped, and gold finished flat (0.2%) after a 15% drawdown. The missing bear fuel was the buyer base: central banks absorbing supply at a >1,000t annual pace from 2022. That break is the single most important fact for reading 2026, the year the old direction reasserted itself, with gold falling from February as real yields climbed.

-6: 91.1 index
-5: 91.2 index
-4: 93.5 index
-3: 91.9 index
-2: 93.2 index
-1: 95.3 index
0: 100 index
+1: 99.4 index
+2: 94.9 index
+3: 94.3 index
+4: 89 index
+5: 90.6 index
+6: 86.3 index
+7: 85.4 index
+8: 88.6 index
+9: 92.3 index
+10: 97.4 index
+11: 95.2 index
+12: 98.2 index
+13: 102.7 index
+14: 102.3 index
+15: 99.7 index
+16: 100.2 index
+17: 98.5 index
+18: 98.4 index
+19: 98.4 index
+20: 101.8 index
+21: 104 index
+22: 104.4 index
+23: 103.9 index
+24: 110.8 index

2026: what's different

The highlighted line is the median of all ten completed cycles, the one that anchors the base case below: close to flat, 104.7 two years out. An illustrative menu, not a forecast. The dashed ghosts are the six cycles replayed above; their own median runs higher (114.2 at +24 months). Today's start cuts both ways: like 1977, the hike is oil-shock-driven with the real funds rate barely positive (+0.3pp as of August) and an SEP showing 3.4% core, behind the curve, but not 1970s-behind. Like 2022, real yields already sit at 2.93%, within sight of the 2008 peak — and the central-bank bid that broke the old rule slowed sharply in Q1 before rebounding. The first cycle to start with *both* the 1977 problem and the 2022 solution in play.

-6: 103.7 index
-5: 96.7 index
-4: 100.7 index
-3: 103.8 index
-2: 101.5 index
-1: 101.1 index
0: 100 index
+1: 101.7 index
+2: 101.9 index
+3: 101.5 index
+4: 104 index
+5: 103.8 index
+6: 103.9 index
+7: 105.4 index
+8: 105 index
+9: 105 index
+10: 103.2 index
+11: 102.3 index
+12: 103.1 index
+13: 105.3 index
+14: 103.7 index
+15: 102.9 index
+16: 102.5 index
+17: 101.5 index
+18: 102.2 index
+19: 100.9 index
+20: 100.8 index
+21: 100.8 index
+22: 102.1 index
+23: 104.1 index
+24: 104.7 index

Gold, World Bank monthly-average basis, indexed to 100 in each cycle's first-hike month, from six months before the first hike to 24 months after. Each step highlights one of six cycles; the others stay as ghosts. The four not replayed (1987, 1988–89, 1994–95, 1999–2000) are in the bar chart above. The final step shows the median of all ten completed cycles.

Source: Calculated from World Bank Pink Sheet (gold); FRED (FEDFUNDS, DFEDTAR, DFEDTARU)

Two eras of gold and real yields

2006–2021

The textbook era. A real inverse link: −0.53 monthly correlation; in the 136 monthly changes where the yield moved >5bp, gold averaged +3.1% when it fell and −1.5% when it rose (across all months: +2.2% / −1.2%). Same months, same basis, both charts.

Gold, monthly average (2006–2021)

Period	Gold (USD/troy oz)
Jan '06	550
Feb '06	555
Mar '06	557
Apr '06	611
May '06	675
Jun '06	596
Jul '06	634
Aug '06	633
Sep '06	598
Oct '06	586
Nov '06	628
Dec '06	630
Jan '07	631
Feb '07	665
Mar '07	655
Apr '07	679
May '07	667
Jun '07	656
Jul '07	665
Aug '07	665
Sep '07	713
Oct '07	755
Nov '07	806
Dec '07	803
Jan '08	890
Feb '08	922
Mar '08	968
Apr '08	910
May '08	889
Jun '08	889
Jul '08	940
Aug '08	839
Sep '08	830
Oct '08	807
Nov '08	761
Dec '08	816
Jan '09	859
Feb '09	943
Mar '09	924
Apr '09	890
May '09	929
Jun '09	946
Jul '09	934
Aug '09	949
Sep '09	997
Oct '09	1043
Nov '09	1127
Dec '09	1135
Jan '10	1118
Feb '10	1095
Mar '10	1113
Apr '10	1149
May '10	1205
Jun '10	1233
Jul '10	1193
Aug '10	1216
Sep '10	1271
Oct '10	1342
Nov '10	1370
Dec '10	1391
Jan '11	1360
Feb '11	1375
Mar '11	1423
Apr '11	1481
May '11	1513
Jun '11	1529
Jul '11	1573
Aug '11	1759
Sep '11	1772
Oct '11	1666
Nov '11	1739
Dec '11	1640
Jan '12	1654
Feb '12	1745
Mar '12	1676
Apr '12	1649
May '12	1589
Jun '12	1599
Jul '12	1594
Aug '12	1630
Sep '12	1745
Oct '12	1747
Nov '12	1722
Dec '12	1685
Jan '13	1672
Feb '13	1628
Mar '13	1593
Apr '13	1488
May '13	1414
Jun '13	1343
Jul '13	1286
Aug '13	1352
Sep '13	1349
Oct '13	1317
Nov '13	1276
Dec '13	1222
Jan '14	1244
Feb '14	1300
Mar '14	1336
Apr '14	1298
May '14	1289
Jun '14	1279
Jul '14	1311
Aug '14	1295
Sep '14	1237
Oct '14	1222
Nov '14	1175
Dec '14	1201
Jan '15	1251
Feb '15	1227
Mar '15	1179
Apr '15	1199
May '15	1199
Jun '15	1182
Jul '15	1128
Aug '15	1118
Sep '15	1125
Oct '15	1159
Nov '15	1086
Dec '15	1076
Jan '16	1098
Feb '16	1200
Mar '16	1245
Apr '16	1242
May '16	1261
Jun '16	1276
Jul '16	1337
Aug '16	1340
Sep '16	1327
Oct '16	1267
Nov '16	1238
Dec '16	1157
Jan '17	1192
Feb '17	1234
Mar '17	1231
Apr '17	1267
May '17	1246
Jun '17	1260
Jul '17	1237
Aug '17	1283
Sep '17	1314
Oct '17	1280
Nov '17	1282
Dec '17	1264
Jan '18	1331
Feb '18	1331
Mar '18	1325
Apr '18	1335
May '18	1303
Jun '18	1282
Jul '18	1238
Aug '18	1202
Sep '18	1198
Oct '18	1215
Nov '18	1221
Dec '18	1250
Jan '19	1292
Feb '19	1320
Mar '19	1301
Apr '19	1286
May '19	1284
Jun '19	1359
Jul '19	1413
Aug '19	1500
Sep '19	1511
Oct '19	1495
Nov '19	1471
Dec '19	1479
Jan '20	1561
Feb '20	1597
Mar '20	1592
Apr '20	1683
May '20	1716
Jun '20	1732
Jul '20	1847
Aug '20	1969
Sep '20	1922
Oct '20	1900
Nov '20	1866
Dec '20	1858
Jan '21	1867
Feb '21	1808
Mar '21	1718
Apr '21	1760
May '21	1850
Jun '21	1835
Jul '21	1808
Aug '21	1785
Sep '21	1775
Oct '21	1777
Nov '21	1822
Dec '21	1790

World Bank monthly average, 2006-01 to 2021-12.

Source: World Bank Pink Sheet (CC BY 4.0)

10-year TIPS real yield (2006–2021)

Period	Real yield (%)
Jan '06	2.01
Feb '06	2.05
Mar '06	2.2
Apr '06	2.41
May '06	2.45
Jun '06	2.53
Jul '06	2.51
Aug '06	2.29
Sep '06	2.32
Oct '06	2.41
Nov '06	2.29
Dec '06	2.25
Jan '07	2.44
Feb '07	2.36
Mar '07	2.18
Apr '07	2.26
May '07	2.37
Jun '07	2.69
Jul '07	2.64
Aug '07	2.44
Sep '07	2.26
Oct '07	2.2
Nov '07	1.77
Dec '07	1.79
Jan '08	1.47
Feb '08	1.41
Mar '08	1.09
Apr '08	1.36
May '08	1.46
Jun '08	1.63
Jul '08	1.57
Aug '08	1.68
Sep '08	1.85
Oct '08	2.75
Nov '08	2.89
Dec '08	2.17
Jan '09	1.91
Feb '09	1.75
Mar '09	1.71
Apr '09	1.57
May '09	1.72
Jun '09	1.86
Jul '09	1.82
Aug '09	1.77
Sep '09	1.64
Oct '09	1.48
Nov '09	1.28
Dec '09	1.36
Jan '10	1.37
Feb '10	1.42
Mar '10	1.51
Apr '10	1.5
May '10	1.31
Jun '10	1.26
Jul '10	1.24
Aug '10	1.02
Sep '10	0.91
Oct '10	0.53
Nov '10	0.67
Dec '10	1.04
Jan '11	1.06
Feb '11	1.24
Mar '11	0.96
Apr '11	0.86
May '11	0.78
Jun '11	0.76
Jul '11	0.62
Aug '11	0.14
Sep '11	0.08
Oct '11	0.19
Nov '11	0
Dec '11	-0.03
Jan '12	-0.11
Feb '12	-0.25
Mar '12	-0.14
Apr '12	-0.21
May '12	-0.34
Jun '12	-0.5
Jul '12	-0.6
Aug '12	-0.59
Sep '12	-0.71
Oct '12	-0.75
Nov '12	-0.77
Dec '12	-0.76
Jan '13	-0.61
Feb '13	-0.57
Mar '13	-0.58
Apr '13	-0.65
May '13	-0.36
Jun '13	0.25
Jul '13	0.46
Aug '13	0.55
Sep '13	0.66
Oct '13	0.43
Nov '13	0.55
Dec '13	0.74
Jan '14	0.63
Feb '14	0.55
Mar '14	0.56
Apr '14	0.54
May '14	0.37
Jun '14	0.37
Jul '14	0.28
Aug '14	0.22
Sep '14	0.46
Oct '14	0.38
Nov '14	0.45
Dec '14	0.51
Jan '15	0.27
Feb '15	0.26
Mar '15	0.28
Apr '15	0.08
May '15	0.33
Jun '15	0.5
Jul '15	0.5
Aug '15	0.56
Sep '15	0.65
Oct '15	0.57
Nov '15	0.69
Dec '15	0.73
Jan '16	0.67
Feb '16	0.47
Mar '16	0.34
Apr '16	0.19
May '16	0.21
Jun '16	0.17
Jul '16	0.04
Aug '16	0.09
Sep '16	0.12
Oct '16	0.1
Nov '16	0.32
Dec '16	0.56
Jan '17	0.42
Feb '17	0.4
Mar '17	0.49
Apr '17	0.39
May '17	0.47
Jun '17	0.46
Jul '17	0.55
Aug '17	0.43
Sep '17	0.37
Oct '17	0.5
Nov '17	0.5
Dec '17	0.5
Jan '18	0.54
Feb '18	0.76
Mar '18	0.75
Apr '18	0.74
May '18	0.84
Jun '18	0.79
Jul '18	0.77
Aug '18	0.79
Sep '18	0.88
Oct '18	1.04
Nov '18	1.11
Dec '18	1.02
Jan '19	0.92
Feb '19	0.8
Mar '19	0.66
Apr '19	0.6
May '19	0.57
Jun '19	0.37
Jul '19	0.31
Aug '19	0.04
Sep '19	0.11
Oct '19	0.15
Nov '19	0.17
Dec '19	0.14
Jan '20	0.04
Feb '20	-0.11
Mar '20	-0.12
Apr '20	-0.45
May '20	-0.44
Jun '20	-0.54
Jul '20	-0.83
Aug '20	-1.01
Sep '20	-0.98
Oct '20	-0.92
Nov '20	-0.84
Dec '20	-0.98
Jan '21	-1
Feb '21	-0.92
Mar '21	-0.66
Apr '21	-0.71
May '21	-0.85
Jun '21	-0.82
Jul '21	-1.01
Aug '21	-1.07
Sep '21	-0.97
Oct '21	-0.95
Nov '21	-1.06
Dec '21	-0.99

Monthly average of daily DFII10, 2006-01 to 2021-12.

Source: FRED DFII10 (H.15)

2022–2026

The broken era. Real yields march from −0.7% to 2.64% (September monthly average; 2.92% on October 7), territory unseen since 2008, and gold's September average is still 138% above January 2022 (127% at the October 8 morning spot). Through 2024 both lines climb together (the break); from 2025 they lean against each other again. The monthly inverse correlation survives (−0.45 through September); the level relationship does not.

Gold, monthly average (2022–2026)

Period	Gold (USD/troy oz)
Jan '22	1816
Feb '22	1856
Mar '22	1948
Apr '22	1937
May '22	1849
Jun '22	1837
Jul '22	1733
Aug '22	1765
Sep '22	1681
Oct '22	1664
Nov '22	1725
Dec '22	1798
Jan '23	1898
Feb '23	1855
Mar '23	1913
Apr '23	2000
May '23	1992
Jun '23	1943
Jul '23	1951
Aug '23	1919
Sep '23	1916
Oct '23	1916
Nov '23	1984
Dec '23	2026
Jan '24	2034
Feb '24	2023
Mar '24	2158
Apr '24	2331
May '24	2351
Jun '24	2326
Jul '24	2398
Aug '24	2470
Sep '24	2571
Oct '24	2690
Nov '24	2651
Dec '24	2648
Jan '25	2710
Feb '25	2895
Mar '25	2983
Apr '25	3218
May '25	3309
Jun '25	3353
Jul '25	3340
Aug '25	3368
Sep '25	3668
Oct '25	4058
Nov '25	4087
Dec '25	4309
Jan '26	4753
Feb '26	5020
Mar '26	4856
Apr '26	4721
May '26	4587
Jun '26	4228
Jul '26	4073
Aug '26	4411
Sep '26	4319
Oct 8	4130

World Bank monthly average, 2022-01 to 2026-10. Final point: spot on the late afternoon of October 8.

Source: World Bank Pink Sheet (CC BY 4.0); final point: Kitco spot

10-year TIPS real yield (2022–2026)

Period	Real yield (%)
Jan '22	-0.69
Feb '22	-0.52
Mar '22	-0.72
Apr '22	-0.14
May '22	0.21
Jun '22	0.53
Jul '22	0.53
Aug '22	0.39
Sep '22	1.14
Oct '22	1.59
Nov '22	1.52
Dec '22	1.36
Jan '23	1.29
Feb '23	1.41
Mar '23	1.36
Apr '23	1.19
May '23	1.36
Jun '23	1.55
Jul '23	1.6
Aug '23	1.83
Sep '23	2.04
Oct '23	2.41
Nov '23	2.2
Dec '23	1.84
Jan '24	1.79
Feb '24	1.93
Mar '24	1.9
Apr '24	2.15
May '24	2.15
Jun '24	2.05
Jul '24	1.97
Aug '24	1.76
Sep '24	1.62
Oct '24	1.81
Nov '24	2.03
Dec '24	2.09
Jan '25	2.23
Feb '25	2.03
Mar '25	1.95
Apr '25	2.04
May '25	2.11
Jun '25	2.09
Jul '25	2.01
Aug '25	1.88
Sep '25	1.75
Oct '25	1.76
Nov '25	1.83
Dec '25	1.9
Jan '26	1.91
Feb '26	1.82
Mar '26	1.91
Apr '26	1.94
May '26	2.04
Jun '26	2.18
Jul '26	2.35
Aug '26	2.4
Sep '26	2.64
Oct 7	2.92

Monthly average of daily DFII10, 2022-01 to 2026-09. Final point: the October 7 daily value (US Treasury), the latest published.

Source: FRED DFII10 (H.15); final point: US Treasury daily real yield curve

Within each tab the two charts share identical month ranges (monthly averages). The second tab then adds one daily point to each chart, spot gold on the afternoon of October 8 and the Treasury's 10-year real yield for October 7, the latest published because the Treasury posts each day's rate after the close. DFII10 is the 10-year Treasury inflation-indexed yield.

Source: World Bank Pink Sheet; FRED DFII10; final points: Kitco (October 8) and US Treasury (October 7)

How an oil shock reaches gold

An oil shock reaches gold through four channels that point in different directions.

Inflation expectations. Oil's climb from $57 in early January to $91–102 monthly averages in March–May helped push CPI from 2.4% to a 4.2% year-over-year peak in May (BLS via FRED). Gold is bought as an inflation hedge, and headline surprises raise that bid.

The Fed's reaction. The dominant topic this year. The September SEP projects 3.4% core PCE for 2026 and a 4.1% funds-rate median for end-2026 *and* end-2027, with a decent possibility of one more hike. Chair Kevin Warsh confirmed 54–45 on May 13, among the closest Fed confirmations on record, told Jackson Hole that "the Fed's predominant focus right now should be on prices" (speech). Higher expected policy rates raise real yields.

The dollar. The broad dollar index is roughly flat on the year (FRED), though reporting describes it firming around the hike as the 10-year yield passed its 2007 peak. A decisively stronger dollar is gold's classic headwind.

Supply-chain demand. A Hormuz running on an Iranian toll regime keep the focus on oil, and less on central-bank purchasing gold.

How many barrels an ounce buys

Period	Gold/WTI ratio (barrels/troy oz)	Median since 1982 (17.5) (barrels/troy oz)
1982-01	10.9	17.5
1982-02	10.7	17.5
1982-03	10.1	17.5
1982-04	11.4	17.5
1982-05	10.9	17.5
1982-06	10.1	17.5
1982-07	10.1	17.5
1982-08	11	17.5
1982-09	13.2	17.5
1982-10	12.8	17.5
1982-11	12.6	17.5
1982-12	13.7	17.5
1983-01	15.5	17.5
1983-02	16.8	17.5
1983-03	14.6	17.5
1983-04	14.2	17.5
1983-05	14.6	17.5
1983-06	13.3	17.5
1983-07	12.8	17.5
1983-08	13	17.5
1983-09	13.2	17.5
1983-10	13	17.5
1983-11	12.7	17.5
1983-12	13.3	17.5
1984-01	12.4	17.5
1984-02	12.8	17.5
1984-03	12.8	17.5
1984-04	12.5	17.5
1984-05	12.4	17.5
1984-06	12.6	17.5
1984-07	12.1	17.5
1984-08	11.9	17.5
1984-09	11.6	17.5
1984-10	11.9	17.5
1984-11	12.1	17.5
1984-12	12	17.5
1985-01	11.8	17.5
1985-02	10.9	17.5
1985-03	11.1	17.5
1985-04	11.3	17.5
1985-05	11.4	17.5
1985-06	11.7	17.5
1985-07	11.8	17.5
1985-08	12	17.5
1985-09	11.7	17.5
1985-10	11.3	17.5
1985-11	10.8	17.5
1985-12	11.8	17.5
1986-01	15.2	17.5
1986-02	22	17.5
1986-03	27.2	17.5
1986-04	26.4	17.5
1986-05	22.1	17.5
1986-06	25.4	17.5
1986-07	30	17.5
1986-08	24.6	17.5
1986-09	27.9	17.5
1986-10	28.5	17.5
1986-11	26	17.5
1986-12	24	17.5
1987-01	21.9	17.5
1987-02	22.5	17.5
1987-03	22.1	17.5
1987-04	23.4	17.5
1987-05	23.7	17.5
1987-06	22.4	17.5
1987-07	21.2	17.5
1987-08	22.8	17.5
1987-09	23.6	17.5
1987-10	23.4	17.5
1987-11	24.8	17.5
1987-12	28.1	17.5
1988-01	27.7	17.5
1988-02	26.3	17.5
1988-03	27.4	17.5
1988-04	25.3	17.5
1988-05	25.9	17.5
1988-06	27	17.5
1988-07	28.3	17.5
1988-08	27.6	17.5
1988-09	28.5	17.5
1988-10	29.5	17.5
1988-11	30	17.5
1988-12	25.7	17.5
1989-01	22.4	17.5
1989-02	21.8	17.5
1989-03	20	17.5
1989-04	18.3	17.5
1989-05	18.5	17.5
1989-06	18.4	17.5
1989-07	18.9	17.5
1989-08	19.6	17.5
1989-09	18.5	17.5
1989-10	18.3	17.5
1989-11	19.9	17.5
1989-12	19.4	17.5
1990-01	18.1	17.5
1990-02	18.9	17.5
1990-03	19.3	17.5
1990-04	20.1	17.5
1990-05	19.9	17.5
1990-06	20.8	17.5
1990-07	19.4	17.5
1990-08	14.5	17.5
1990-09	11.5	17.5
1990-10	10.6	17.5
1990-11	11.8	17.5
1990-12	13.9	17.5
1991-01	15.5	17.5
1991-02	17.7	17.5
1991-03	18.2	17.5
1991-04	17.2	17.5
1991-05	16.8	17.5
1991-06	18.2	17.5
1991-07	17.1	17.5
1991-08	16.4	17.5
1991-09	15.9	17.5
1991-10	15.4	17.5
1991-11	15.9	17.5
1991-12	18.5	17.5
1992-01	18.8	17.5
1992-02	18.6	17.5
1992-03	18.1	17.5
1992-04	16.7	17.5
1992-05	16	17.5
1992-06	15.2	17.5
1992-07	16.2	17.5
1992-08	16.1	17.5
1992-09	15.8	17.5
1992-10	15.9	17.5
1992-11	16.4	17.5
1992-12	17.3	17.5
1993-01	17.2	17.5
1993-02	16.4	17.5
1993-03	16.2	17.5
1993-04	16.8	17.5
1993-05	18.4	17.5
1993-06	19.4	17.5
1993-07	21.9	17.5
1993-08	21.1	17.5
1993-09	20.3	17.5
1993-10	20	17.5
1993-11	22.3	17.5
1993-12	24.6	17.5
1994-01	25.8	17.5
1994-02	25.8	17.5
1994-03	26.1	17.5
1994-04	23.1	17.5
1994-05	21.3	17.5
1994-06	20.2	17.5
1994-07	19.5	17.5
1994-08	20.7	17.5
1994-09	22.4	17.5
1994-10	22	17.5
1994-11	21.2	17.5
1994-12	22	17.5
1995-01	21.1	17.5
1995-02	20.4	17.5
1995-03	20.6	17.5
1995-04	19.6	17.5
1995-05	19.6	17.5
1995-06	21	17.5
1995-07	22.2	17.5
1995-08	21.3	17.5
1995-09	21.3	17.5
1995-10	22.1	17.5
1995-11	21.6	17.5
1995-12	20.6	17.5
1996-01	21.1	17.5
1996-02	21.2	17.5
1996-03	18.7	17.5
1996-04	16.9	17.5
1996-05	18.6	17.5
1996-06	19	17.5
1996-07	17.9	17.5
1996-08	17.6	17.5
1996-09	16	17.5
1996-10	15.3	17.5
1996-11	15.9	17.5
1996-12	14.6	17.5
1997-01	14.3	17.5
1997-02	15.9	17.5
1997-03	17	17.5
1997-04	17.7	17.5
1997-05	16.8	17.5
1997-06	18	17.5
1997-07	16.8	17.5
1997-08	16.5	17.5
1997-09	16.5	17.5
1997-10	15.3	17.5
1997-11	15.4	17.5
1997-12	16	17.5
1998-01	17.5	17.5
1998-02	18.8	17.5
1998-03	20	17.5
1998-04	20.1	17.5
1998-05	20.1	17.5
1998-06	21.3	17.5
1998-07	20.8	17.5
1998-08	21.2	17.5
1998-09	19.3	17.5
1998-10	20.6	17.5
1998-11	22.6	17.5
1998-12	25.8	17.5
1999-01	23	17.5
1999-02	23.9	17.5
1999-03	19.5	17.5
1999-04	16.4	17.5
1999-05	15.5	17.5
1999-06	14.6	17.5
1999-07	12.7	17.5
1999-08	12.1	17.5
1999-09	11.1	17.5
1999-10	13.8	17.5
1999-11	11.8	17.5
1999-12	10.8	17.5
2000-01	10.4	17.5
2000-02	10.2	17.5
2000-03	9.6	17.5
2000-04	10.9	17.5
2000-05	9.5	17.5
2000-06	9	17.5
2000-07	9.4	17.5
2000-08	8.8	17.5
2000-09	8.1	17.5
2000-10	8.2	17.5
2000-11	7.7	17.5
2000-12	9.5	17.5
2001-01	9	17.5
2001-02	8.9	17.5
2001-03	9.7	17.5
2001-04	9.5	17.5
2001-05	9.5	17.5
2001-06	9.8	17.5
2001-07	10.2	17.5
2001-08	9.9	17.5
2001-09	10.8	17.5
2001-10	12.7	17.5
2001-11	14.1	17.5
2001-12	14.3	17.5
2002-01	14.3	17.5
2002-02	14.3	17.5
2002-03	12	17.5
2002-04	11.6	17.5
2002-05	11.6	17.5
2002-06	12.6	17.5
2002-07	11.6	17.5
2002-08	10.9	17.5
2002-09	10.7	17.5
2002-10	11	17.5
2002-11	12.1	17.5
2002-12	11.3	17.5
2003-01	10.9	17.5
2003-02	10	17.5
2003-03	10.2	17.5
2003-04	11.6	17.5
2003-05	12.7	17.5
2003-06	11.6	17.5
2003-07	11.4	17.5
2003-08	11.4	17.5
2003-09	13.4	17.5
2003-10	12.5	17.5
2003-11	12.5	17.5
2003-12	12.7	17.5
2004-01	12.1	17.5
2004-02	11.7	17.5
2004-03	11.1	17.5
2004-04	11	17.5
2004-05	9.5	17.5
2004-06	10.3	17.5
2004-07	9.8	17.5
2004-08	8.9	17.5
2004-09	8.8	17.5
2004-10	7.9	17.5
2004-11	9.1	17.5
2004-12	10.2	17.5
2005-01	9.1	17.5
2005-02	8.8	17.5
2005-03	8	17.5
2005-04	8.1	17.5
2005-05	8.5	17.5
2005-06	7.6	17.5
2005-07	7.2	17.5
2005-08	6.7	17.5
2005-09	7	17.5
2005-10	7.5	17.5
2005-11	8.2	17.5
2005-12	8.6	17.5
2006-01	8.4	17.5
2006-02	9	17.5
2006-03	8.9	17.5
2006-04	8.8	17.5
2006-05	9.5	17.5
2006-06	8.4	17.5
2006-07	8.5	17.5
2006-08	8.7	17.5
2006-09	9.4	17.5
2006-10	9.9	17.5
2006-11	10.6	17.5
2006-12	10.2	17.5
2007-01	11.6	17.5
2007-02	11.2	17.5
2007-03	10.8	17.5
2007-04	10.6	17.5
2007-05	10.5	17.5
2007-06	9.7	17.5
2007-07	9	17.5
2007-08	9.2	17.5
2007-09	8.9	17.5
2007-10	8.8	17.5
2007-11	8.5	17.5
2007-12	8.8	17.5
2008-01	9.6	17.5
2008-02	9.7	17.5
2008-03	9.2	17.5
2008-04	8.1	17.5
2008-05	7.1	17.5
2008-06	6.6	17.5
2008-07	7	17.5
2008-08	7.2	17.5
2008-09	8	17.5
2008-10	10.5	17.5
2008-11	13.3	17.5
2008-12	19.7	17.5
2009-01	20.6	17.5
2009-02	24.1	17.5
2009-03	19.5	17.5
2009-04	17.9	17.5
2009-05	15.7	17.5
2009-06	13.6	17.5
2009-07	14.6	17.5
2009-08	13.3	17.5
2009-09	14.4	17.5
2009-10	13.8	17.5
2009-11	14.4	17.5
2009-12	15.2	17.5
2010-01	14.3	17.5
2010-02	14.3	17.5
2010-03	13.7	17.5
2010-04	13.6	17.5
2010-05	16.4	17.5
2010-06	16.4	17.5
2010-07	15.6	17.5
2010-08	15.9	17.5
2010-09	16.9	17.5
2010-10	16.4	17.5
2010-11	16.3	17.5
2010-12	15.6	17.5
2011-01	15.2	17.5
2011-02	15.4	17.5
2011-03	13.8	17.5
2011-04	13.5	17.5
2011-05	14.9	17.5
2011-06	15.9	17.5
2011-07	16.2	17.5
2011-08	20.4	17.5
2011-09	20.7	17.5
2011-10	19.3	17.5
2011-11	17.9	17.5
2011-12	16.6	17.5
2012-01	16.5	17.5
2012-02	17.1	17.5
2012-03	15.8	17.5
2012-04	16	17.5
2012-05	16.8	17.5
2012-06	19.4	17.5
2012-07	18.1	17.5
2012-08	17.3	17.5
2012-09	18.5	17.5
2012-10	19.5	17.5
2012-11	19.9	17.5
2012-12	19.1	17.5
2013-01	17.7	17.5
2013-02	17.1	17.5
2013-03	17.1	17.5
2013-04	16.2	17.5
2013-05	14.9	17.5
2013-06	14	17.5
2013-07	12.3	17.5
2013-08	12.7	17.5
2013-09	12.7	17.5
2013-10	13.1	17.5
2013-11	13.6	17.5
2013-12	12.5	17.5
2014-01	13.1	17.5
2014-02	12.9	17.5
2014-03	13.3	17.5
2014-04	12.7	17.5
2014-05	12.6	17.5
2014-06	12.2	17.5
2014-07	12.7	17.5
2014-08	13.4	17.5
2014-09	13.3	17.5
2014-10	14.5	17.5
2014-11	15.5	17.5
2014-12	20.3	17.5
2015-01	26.4	17.5
2015-02	24.2	17.5
2015-03	24.7	17.5
2015-04	22	17.5
2015-05	20.2	17.5
2015-06	19.8	17.5
2015-07	22.2	17.5
2015-08	26.1	17.5
2015-09	24.7	17.5
2015-10	25.1	17.5
2015-11	25.4	17.5
2015-12	28.9	17.5
2016-01	34.9	17.5
2016-02	39.5	17.5
2016-03	32.9	17.5
2016-04	30.3	17.5
2016-05	27	17.5
2016-06	26.1	17.5
2016-07	29.9	17.5
2016-08	29.9	17.5
2016-09	29.4	17.5
2016-10	25.4	17.5
2016-11	27.1	17.5
2016-12	22.2	17.5
2017-01	22.7	17.5
2017-02	23.1	17.5
2017-03	24.8	17.5
2017-04	24.8	17.5
2017-05	25.7	17.5
2017-06	27.9	17.5
2017-07	26.5	17.5
2017-08	26.7	17.5
2017-09	26.4	17.5
2017-10	24.8	17.5
2017-11	22.6	17.5
2017-12	21.8	17.5
2018-01	20.9	17.5
2018-02	21.4	17.5
2018-03	21.1	17.5
2018-04	20.1	17.5
2018-05	18.6	17.5
2018-06	19	17.5
2018-07	17.5	17.5
2018-08	17.7	17.5
2018-09	17.1	17.5
2018-10	17.2	17.5
2018-11	21.5	17.5
2018-12	25.5	17.5
2019-01	25.1	17.5
2019-02	24	17.5
2019-03	22.4	17.5
2019-04	20.1	17.5
2019-05	21.1	17.5
2019-06	24.8	17.5
2019-07	24.6	17.5
2019-08	27.4	17.5
2019-09	26.5	17.5
2019-10	27.7	17.5
2019-11	25.8	17.5
2019-12	24.7	17.5
2020-01	27.1	17.5
2020-02	31.6	17.5
2020-03	53.2	17.5
2020-04	102	17.5
2020-05	60	17.5
2020-06	45.2	17.5
2020-07	45.3	17.5
2020-08	46.4	17.5
2020-09	48.5	17.5
2020-10	48.1	17.5
2020-11	45.4	17.5
2020-12	39.4	17.5
2021-01	35.8	17.5
2021-02	30.6	17.5
2021-03	27.5	17.5
2021-04	28.5	17.5
2021-05	28.4	17.5
2021-06	25.7	17.5
2021-07	24.9	17.5
2021-08	26.4	17.5
2021-09	24.8	17.5
2021-10	21.9	17.5
2021-11	23	17.5
2021-12	25	17.5
2022-01	21.9	17.5
2022-02	20.2	17.5
2022-03	18	17.5
2022-04	19	17.5
2022-05	16.9	17.5
2022-06	16	17.5
2022-07	17.3	17.5
2022-08	19.3	17.5
2022-09	20	17.5
2022-10	19.1	17.5
2022-11	20.3	17.5
2022-12	23.5	17.5
2023-01	24.3	17.5
2023-02	24.2	17.5
2023-03	26.1	17.5
2023-04	25.2	17.5
2023-05	27.8	17.5
2023-06	27.7	17.5
2023-07	25.5	17.5
2023-08	23.6	17.5
2023-09	21.4	17.5
2023-10	22.4	17.5
2023-11	25.6	17.5
2023-12	28.1	17.5
2024-01	27.5	17.5
2024-02	26.4	17.5
2024-03	26.8	17.5
2024-04	27.6	17.5
2024-05	29.8	17.5
2024-06	29.5	17.5
2024-07	29.8	17.5
2024-08	32.7	17.5
2024-09	36.9	17.5
2024-10	37.6	17.5
2024-11	38	17.5
2024-12	37.9	17.5
2025-01	36.1	17.5
2025-02	40.6	17.5
2025-03	44	17.5
2025-04	51	17.5
2025-05	54.2	17.5
2025-06	49.7	17.5
2025-07	49.6	17.5
2025-08	52.5	17.5
2025-09	57.6	17.5
2025-10	67.4	17.5
2025-11	68.6	17.5
2025-12	74.4	17.5
2026-01	78.8	17.5
2026-02	77.7	17.5
2026-03	53.2	17.5
2026-04	47.9	17.5
2026-05	46.3	17.5
2026-06	51.6	17.5
2026-07	51.3	17.5
2026-08	53.3	17.5
2026-09	44.6	17.5
Oct 6–8	42.9	17.5

Monthly average gold price divided by monthly average WTI, January 1982–September 2026 (the common series starts in 1982), with the median since 1982 (17.5 barrels) flat for reference. September: 44.6 barrels. The final point is daily data rather than a monthly average, spot gold on the afternoon of October 8 divided by the EIA's Cushing spot for October 6, the latest published: 42.9 barrels. Gold has cheapened in oil prices this year but is still expensive by any standard of the past four decades.

Source: Calculated from World Bank Pink Sheet (CC BY 4.0); final point: Kitco (gold) and EIA (WTI)

Who's selling, who's still buying

The buyer base has rotated, and the rotation best explains the real-yield puzzle.

Western ETF money is price-sensitive and fast. After a record $13bn North-American outflow in March, ETFs added $18bn (121t) in August, the second-largest month on record, taking holdings to 4,189t, the highest ever, with $615bn under management (WGC). Yet Q2 as a whole was a −45t outflow as gold slid from April to June, months before the hike: ETFs follow the price. This is the marginal, real-yield-sensitive dollar.

Central banks are slow and price-insensitive until they aren't. Quarterly net purchases ran 237t, 178t, 226t, 208t through 2025. Then 56.5t in Q1 2026 revised down, 187t reclassified to OTC demand in the WGC's own erratum, before a 289t Q2 rebound (WGC). China's central bank has reportedly bought for twenty straight months through June. The ECB puts gold at 27% of world official reserves at end-2025, ahead of Treasuries (22%) and the euro (15%), while noting this largely reflects valuation effects.

The mix has shifted. Among the five sectors charted below (OTC excluded), jewellery was 44% in 2022 and about 30% in H1 2026, 23% of all demand once the half's 571t of OTC buying counts, with bars, coins, ETFs and central banks now about 61%. When the slow money stumbles and the fast money sells, gold has little buffer against a real-yield shock. Whether September was that configuration is not yet known: September ETF and Q3 central-bank data are unpublished, and the last readings before the fall, 289t of official buying in Q2, record ETF holdings in August, were strong. That is the hinge our scenarios turn on.

ETF flows vs central-bank buying

Period	Central banks (net) (tonnes)	ETFs (net) (tonnes)
Q1 2025	237	229.9
Q2 2025	177.9	171.1
Q3 2025	226.3	225.7
Q4 2025	208.2	175.7
Q1 2026	56.5	62.4
Q2 2026	288.9	-44.8

Quarterly net flows in tonnes, Q1 2025 – Q2 2026, latest WGC revision per quarter (Q1-2026 central-bank figure reflects the July 2026 erratum). Negative bars are outflows. Annual context: 2024 ETFs −2.9t vs CB 1,045t; 2025 ETFs +801t vs CB 863t.

Source: World Gold Council, Gold Demand Trends tables

Where gold could go from here

Scenario 1: hike-and-hold

The median of all ten completed cycles, nearly flat: 101.5 index by December, near 105 in mid-2027, about 103 by December 2027 (about the same as the price of ≈$4,260 on September 28). The story: one or two more hikes (the SEP's 4.1% end-2026 median implies roughly one), oil drifting toward the EIA's $74 Brent path, and central banks buying slower. Our sequencing view, pressure first, a grind after the Fed's stop-signal, reads September's price action; the median line shows no such dip. Dropped if: official buying slows below ~100t/quarter with ETFs selling (a *cut* signal is the bullish twist here, where gold up on the pivot).

Sep '26: 100 index
Dec '26: 101.5 index
Mar '27: 103.9 index
Jun '27: 105 index
Sep '27: 103.1 index
Dec '27: 102.9 index

Scenario 2: Behind the curve

The 1977–80 analog, the only oil-shock cycle in the sample, and the most extreme: +33% by December 2027 (~$5,500 gold price per troy ounce), with a mid-path pullback like the original. Hormuz stays disrupted, inflation re-accelerates toward 4%+ through upcoming CPI prints, and the Fed, pinned by an SEP showing 3.4% core, never gets real rates decisively positive. It requires September's and first week of October pressure to be the *last* flinch. Dropped if: the October FOMC hikes *and* guides to more with core PCE falling.

Sep '26: 100 index
Dec '26: 112.9 index
Mar '27: 108.3 index
Jun '27: 120.5 index
Sep '27: 131.1 index
Dec '27: 132.6 index

Scenario 3: 2022 replay

The 2022–23 analog, the shape the market priced through late September: a ceasefire holds, oil falls toward the EIA's $74 path, and the Fed pushes real rates clearly positive (TIPS at 2.93% and rising, nominal 10-year 5.27%) while the dollar firms. Gold draws down ~14% to a trough around $3,570, then recovers, in the actual episode only once the final hike had passed and the pause was unmistakable. Pressure until the stop-signal or visibly softer inflation, then the climb back to ~100. Dropped if: DFII10 stalls below ~2.5% or central-bank buying re-accelerates above ~250t/quarter.

Sep '26: 100 index
Dec '26: 94.3 index
Mar '27: 86.3 index
Jun '27: 92.3 index
Sep '27: 98.2 index
Dec '27: 99.7 index

Scenario 4: Liquidity squeeze

The average of October 2008 and March 2020: a risk-off shock, funding stress, forced deleveraging, margin spirals like January's, hits everything, gold included. Both episodes saw double-digit falls on daily prices; the monthly averages plotted here smooth them to about −9% (2008) and almost nothing (2020), so the line shows mainly the rescue: the Fed stops, liquidity returns, and gold ends up higher to ~$5,170. The January 30 crash (−11.4% in a day) is the local precedent, at 5.27% nominal yields, stress would find leverage fast. Dropped if: no shock by mid-2027, the path expires unused.

Sep '26: 100 index
Dec '26: 99.1 index
Mar '27: 117.8 index
Jun '27: 113.8 index
Sep '27: 113.2 index
Dec '27: 125 index

Analog-based illustrative paths, not forecasts. Each path starts at 100 index at the September 2026 cutoff and is derived from named historical episodes (see table below). Points are quarterly readings of monthly-average analogs, so drawdowns inside a quarter are smoothed out. The dashed ghosts are the other paths.

Source: Calculated from World Bank Pink Sheet history; analogs stated per path

Scenario assumptions at a glance

Scenario	Fed & real yields	Oil & dollar	Buyers	Index, Dec '27	We'd drop it if…
Base: hike-and-hold	1–2 more hikes, then hold; DFII10 ~2.5–3%; pressured until the pause signal	Brent toward $74 in 2027; dollar flat	CB ~100–200t/qtr; ETFs mixed	≈103 (≈$4,260)	Core PCE <3% with cuts signalled (bullish twist), or CB <100t/qtr with ETF selling
Behind the curve (1977–80)	Hikes lag inflation; real FF stays ≤ +1pp	Hormuz disrupted; oil re-spikes	CB ≥200t/qtr; strong buyers on gold bars and coins	≈133 (≈$5,500)	Oct FOMC hikes with hawkish guidance as core falls
2022 replay (2022–23)	Real rates push above 3%; 10-yr ~5.3%; dollar firms — priced now	Ceasefire; Brent to $74; war premium unwinds	ETFs sell; CB slows again	≈100 after a −14% trough (≈$3,570)	DFII10 stalls <2.5% or CB ≥250t/qtr
Liquidity squeeze (2008/2020)	Shock then rescue: Fed stops, liquidity returns	Risk-off first; oil falls with demand	Forced selling, then fast return of ETFs	≈125 (≈$5,170)	No funding shock by mid-2027; path expires

Analog-based illustrative paths, not forecasts; no probabilities assigned. The index column is the path's December 2027 value (Sep 2026 = 100); the USD figures are illustrative arithmetic on the September 28 spot snapshot (~$4,135; spot closed the first October week near $4,142). Only the 2022 replay shows a pressured first leg in the index itself; the 'pressure until the stop-signal' sequencing in the text is our reading of the market, layered on these paths.

Source: Calculated; analogs as stated

Banks' Predictions

Bank	Forecast	Key assumption	Bear case / prior	Gold at publication*	Published
Citi	$4,800 in 3 months; $5,000 in 6–12 months	Oil normalizes (Brent $70 Q4, $65 in 2027) → Fed easing channel opens	Raised from $4,500 (0–3m) in late August	$4,366–4,405 (Sep 10)	Sep 7 (raised late Aug)
JPMorgan	≈$4,300 Q3 average; $4,500 by Q4 2026	Hike cycle matures without a renewed shock	Near-term average sits below target	$4,366–4,405 (Sep 10)	reported Sep 10 (standing target)
Bank of America	$4,360 average for 2026	Average-basis call after the January blow-off	Cut 14% from $5,093	≈$4,073 (Jul monthly avg)	Jul 8 (reported Sep 10)
HSBC	$4,560 avg 2026; ≈$4,925 2027 (Jul 9) — cut Oct 1 to $4,490 / $4,825	Grind higher as buying broadens into 2027; "could be nearing a bottom" even after the cut	—	≈$4,073 (Jul monthly avg)	Jul 9; cut Oct 1
Goldman Sachs	$4,900 by end-2026, with "net upside risk"	Central banks keep buying ~50t/month (40t in 2027); China bought ~75% more than reported in July	$4,440 by end-2026 if the Fed keeps hiking	futures $4,293–4,324 (Sep 24)	Sep 23

Bank gold forecasts reported July–September 2026. BofA (Jul 8) and HSBC (Jul 9) dates are the banks' own publication dates; the figures were carried unchanged in a September 10 roundup, and HSBC cut again on October 1 (noted in its row). Citi raised its near-term target in late August and reaffirmed it September 7; Goldman's call is September 23 research. Targets are the institutions' own definitions, year-end levels or period averages. *Gold at publication is the level quoted in the same report (Sep 10 roundup), the month's average on the original date (July rows, per World Bank monthly data), or the nearest dated opened quotes (Goldman row). UBS excluded: its September numbers could not be verified from an opened source.

Source: Kitco News Sep 23 (Goldman Research); GoldSilver roundup Sep 10 (carrying July-dated BofA/HSBC/JPMorgan targets); TradingKey Sep 7 (Citi); Invezz Oct 1 (HSBC cut)

Bottomline

Banks made all those predictions before the most recent development, and their predictions were in line with several of our 4 predictive scenarios. Gold has always been a long-term HODL, but this year has been proven to be a somewhat difficult year to predict due to the uncertainties regarding the Fed's rate hikes, the bond yields, and the ongoing war in Iran.

While we believe the four predictive scenarios above are the likely outcomes, but don't treat them as financial advice.

Gold Silver Crude Oil Brent Crude