In this article
← All articles
Investment allocation

Did bonds protect in the October-November 2026 dip? Stocks, bonds and gold for a euro investor


Euro-area government bonds cushioned the October-November 2026 sell-off only a little: by 19 November they were up 0.7% from 12 October, while world equities were down 5.3% in euros. The iShares funds' returns include reinvested income; gold, often bought as protection, fell.

The five-week comparison below ends on 16 November and uses the 5 November low within that window, when world equities were down 4.4% and euro government bonds up 1.5% from 12 October. The 19 November fund figures are a separate update, not a new input to the charts or correlations.

The stock-bond correlation, which measures how closely the two move together, fell from strongly positive in early October to about zero. That is far from the negative values of 2008, when bonds cushioned a large crash, and far from the joint losses of 2022. For a euro investor, the useful question is what a larger fall driven by inflation and rate rises would do to the whole portfolio, bonds included.

Key takeaways

  1. Bonds softened this fall without offsetting it. A 60/40 mix of world equities and euro government bonds lost 2.0% to the 5 November low, against 4.4% for equities alone; the bonds rose on 6 of the 8 days equities fell up to the low, and on none of the 3 after it.
  2. The cushion appeared only after the peak. In 2026 up to 9 October, euro government bonds fell in 10 of the 14 weeks when world equities fell, and the 26-week correlation stood at +0.48.
  3. This was not a crash test. In the 2022 fall, euro government bonds 7-10 years lost 14.8% while world equities lost 17.0% in euros, and the 60/40 lost 16.0%, as both fell while yields rose; the same bond fund's worst fall, from December 2020 to March 2023, was 23.1%.
  4. Lower losses came with lower long-run growth in this sample. From December 2006 to November 2026, the 60/40 compounded at 6.7% a year against 8.9% for world equities, with a deepest fall of 30.9% against 52.7%.
  5. Check duration, cash and shared inflation risk, not a correlation forecast. You can measure how much your bond funds lose if yields rise and whether near-term spending sits outside both shares and long bonds; nobody can tell you which regime comes next.

What did each holding do from the October peak?

Within the five-week study, world equities fell 4.4% in euros to the 5 November low, while every government bond fund rose and gold fell. The peak is the highest close of the iShares Core MSCI World UCITS ETF measured in euros, on 12 October 2026, which was also its record. Every figure below is a total return from fund net asset values (NAVs) with income reinvested and fund costs deducted, converted into euros where the fund is priced in dollars.

Total return in euros from the 12 October 2026 equity peak to the 5 November low and to 16 November. World equities minus 4.4% and minus 2.3%. Euro-area government bonds 1.5% and 0.8%. German government bonds 1.7% and 0.8%. US Treasuries 7 to 10 years hedged to euros 3.2% and 1.1%. US Treasuries unhedged 2.9% and 2.8%. Gold minus 1.8% and minus 5.6%. Euro cash 0.2% and 0.2%. 60/40 world equities and euro government bonds minus 2.0% and minus 1.0%.

From 12 October to 5 November / to 16 November 2026, in euros: world equities −4.4% / −2.3%; euro-area government bonds +1.5% / +0.8%; German government bonds +1.7% / +0.8%; US Treasuries hedged +3.2% / +1.1%, unhedged +2.9% / +2.8%; gold −1.8% / −5.6%; euro cash +0.2% / +0.2%; 60/40 −2.0% / −1.0%. Source: iShares fund workbooks, ECB reference rates and €STR; returns since the peak (CSV), daily path (CSV).

The fall was much milder in euros than the Nasdaq's. Our Nasdaq correction analysis found the US technology index 11.9% below its record on 5 November, before a lower close on 19 November; that figure is in dollars and from the index's own record, so it is not like for like, but a world fund spreads that exposure across many other companies and currencies.

The currency also cushioned the equity loss. By 16 November the world fund was down 3.7% in dollars since 12 October but only 2.3% in euros, because the dollar rose against the euro; that effect, about 1.5 points, was larger than the 1.2 points by which the 60/40 beat equities alone by that date.

The bond funds never closed below their 12 October level. The 60/40 here was bought at the 12 October close and left to drift. Holding the same 40% in euro cash instead of bonds would have lost 2.6% to the low, and holding it in gold 3.4%; that is one five-week episode, not a ranking of the three.

Currency mattered for US bonds. The hedged Treasury fund removes the dollar's moves, the unhedged one keeps them. The euro fell from $1.1181 to $1.1014 by 16 November, so the unhedged fund kept more of its gain. That currency effect can just as easily run the other way.

When did bonds help, week by week?

German bonds and hedged Treasuries rose in each of the first four weeks after the peak, euro-area bonds in three; all fell in the week shares rebounded. The pattern is clearer by week than by day: bonds rose on 6 of the 8 days when equities fell up to the 5 November low, and on none of the 3 down days after it. Weeks run from Friday close to Friday close, except the first, which starts at the 12 October peak, and the last, which is Monday 16 November alone.

Week (2026)World equitiesEuro govt bondsGerman govt bondsUS Treasuries, hedged60/40
12 to 16 Oct−1.73%+0.60%+0.46%+1.13%−0.80%
16 to 23 Oct−0.14%−0.31%+0.54%+0.65%−0.21%
23 to 30 Oct−0.05%+0.44%+0.18%+0.74%+0.15%
30 Oct to 6 Nov−2.44%+0.89%+0.47%+0.44%−1.09%
6 to 13 Nov+3.44%−0.77%−0.84%−1.99%+1.70%
16 Nov (one day)−1.27%−0.02%+0.03%+0.20%−0.77%

Source: iShares NAVs in euros, copio calculation; weekly returns, including gold and cash (CSV).

The week to 13 November is the one where bonds failed to help, but it was not a week of joint losses: shares rose 3.44% while bonds fell. Opposite moves in both directions are what a negative correlation looks like. Equities, euro government bonds and gold all fell on the same day only three times, on 14 October, 30 October and 10 November, and the bond losses on those days were 0.18% or less.

Is the stock-bond correlation negative again?

No, it is about zero: the 26-week correlation between world equities and euro government bonds fell from +0.48 on 9 October 2026 to +0.02 on 13 November. A correlation of +1 means two assets always move in the same direction, 0 means no linear relation and −1 means they always move in opposite directions. We use weekly returns over a rolling 26 weeks, which reduces the noise from funds priced at different times of day.

Correlation of weekly returns over 26 weeks. World equities with euro government bonds in euros, and with US Treasuries in dollars: average 2007 to 2021 minus 0.11 and minus 0.34; 2008 average minus 0.45 and minus 0.38; 2022 average 0.26 and 0.04; lowest since 2023, on 19 September 2025, minus 0.44 and minus 0.54; 9 October 2026, before the peak, 0.48 and 0.42; 13 November 2026, 0.02 and minus 0.12.

26-week correlation of weekly returns, world equities with euro government bonds in euros / with US Treasuries in dollars: 2008 average −0.45 / −0.38; 2022 average +0.26 / +0.04; 9 October 2026 +0.48 / +0.42; 13 November 2026 +0.02 / −0.12. Source: iShares MSCI World UCITS ETF, € Govt Bond 7-10yr and $ Treasury Bond 7-10yr workbooks; summary (CSV), weekly series 2007-2026 (CSV).

A euro investor could read this dip in two opposite ways: bonds are a hedge again, or everything is falling together. On these measures neither is accurate. The euro correlation dropped sharply but has not turned clearly negative; in September 2025 it reached −0.44, and the dollar version −0.54. A different window, frequency or market can give a different number, so check which one a headline uses.

On 13 November the same weekly correlation was −0.38 over 13 weeks, which would read as a hedge, and +0.28 over 52 weeks, which would not.

The move was fast because most of 2026 looked like 2022. In 2026 up to 9 October, the euro correlation averaged +0.42, and the Core € Govt Bond fund fell in 10 of the 14 weeks when world equities fell. Five weeks of the opposite pattern have pulled a 26-week average down; they do not show that the regime has changed.

Our 36-year study of asset class correlations shows a related measure, the rolling 36-month correlation of monthly returns between European equities and developed-market government bonds, changing sign over decades: positive from 1990 to 1997, negative from 2000 to 2021 and positive again since 2022, even though the shorter 26-week measure here dipped to −0.44 in September 2025.

Why do stocks and bonds sometimes fall together?

Because a rise in interest rates hurts both, and which force drives yields decides whether bonds cushion a fall. A bond's price falls when market yields rise, because its fixed payments are worth less. Shares are also hurt by higher yields, which raise the rate used to value future profits and the cost of borrowing.

When investors fear weaker growth, they tend to expect lower policy rates: yields fall, bonds rise and shares fall. When they fear inflation or more rate rises, yields can rise while shares fall, and both lose. That is the mechanism behind 2008 and 2022; the bond-protection comparison of 2008 and 2022 works through the bond arithmetic.

The 2026 episode so far fits the first pattern, then partly reversed. The Bundesbank's 10-year Bund yield fell from 3.53% on 12 October to 3.29% on 6 November, then rose to 3.45% on 16 November and stood at 3.42% on 17 November. Over the same weeks bond prices rose, then gave part of the gain back.

The US 10-year Treasury yield fell 40 basis points from 9 October to the 5 November low, to 4.84%, then rose 28 points to 5.12% by 16 November. The rebound came from the inflation-protected (real) yield, up 32 points, while the market's implied 10-year inflation rate edged down 4 points. That is consistent with markets pricing tighter policy rather than higher inflation, but yield data alone cannot identify why traders moved.

Policy rates have been rising in the euro area. The ECB raised its deposit rate to 2.25% with effect from 17 June 2026 and to 2.50% from 16 September; its next monetary policy meeting is on 16 and 17 December. In a Reuters poll published on 12 November, all 71 economists surveyed expected a rise to 2.75% at that meeting. That is a press-reported survey, not a forecast of ours.

What happened to a 60/40 in earlier falls?

In every fall of 15% or more since 2006 the 60/40 lost less than equities, but in 2022 only by about one percentage point. The comparison uses the two longest-running iShares funds available here, the MSCI World UCITS ETF in euros and the € Govt Bond 7-10yr UCITS ETF, from December 2006. Each window runs from the world-equity peak to its lowest close.

Total return in euros from each world-equity peak to its trough, for world equities, euro government bonds 7 to 10 years and a 60/40 mix rebalanced monthly. 2007-09 financial crisis: minus 52.7%, 18.9%, minus 30.7%. 2015-16: minus 22.1%, 0.1%, minus 13.5%. 2018: minus 16.2%, 1.9%, minus 9.1%. 2020 Covid: minus 33.9%, minus 2.8%, minus 22.2%. 2022: minus 17.0%, minus 14.8%, minus 16.0%. 2025 tariffs: minus 20.5%, minus 1.1%, minus 13.0%. 2026 so far, to 5 November: minus 4.4%, 1.7%, minus 2.0%.

World equities / euro government bonds 7-10y / 60/40, peak to trough in euros: 2007-09 −52.7% / +18.9% / −30.7%; 2020 −33.9% / −2.8% / −22.2%; 2022 −17.0% / −14.8% / −16.0%; 2025 −20.5% / −1.1% / −13.0%; 2026 so far −4.4% / +1.7% / −2.0%. Source: iShares MSCI World UCITS ETF and € Govt Bond 7-10yr UCITS ETF workbooks, ECB reference rates; all falls (CSV).

Bonds rose in the 2007-09 financial crisis and fell sharply in 2022. In the 2020 Covid crash they fell a little too, so the 60/40's smaller loss came from holding fewer shares rather than from a bond gain. The 2022 bond losses continued after the equity low: the 60/40's deepest fall, from November 2021 to December 2022, was 16.7%.

Over the full sample the bonds lowered both losses and growth. From 8 December 2006 to 16 November 2026, the 60/40 compounded at 6.7% a year and the equity fund at 8.9%, with deepest falls of 30.9% and 52.7%; the bond fund compounded at 2.8% and fell as much as 23.3%. Those are observed results over one period with costs included and taxes excluded, not evidence that either mix is optimal for the next twenty years.

Did gold and cash do their job?

Euro cash did not fall; gold fell to the equity low and kept falling as shares recovered. Euro cash, measured by the ECB's overnight €STR rate of about 2.44%, earned 0.2% over the five weeks and never had a down day. It is the only holding here with a fixed value in euros, which is why it suits money needed soon, not why it would raise long-run growth.

Gold lost 1.8% to the 5 November equity low and 5.6% by the study cutoff on 16 November. By 19 November it was down 9.6% from 12 October. It rose on 5 of the 11 days when equities fell in the study, including 2.25% on 5 November, the window's worst equity day, when world equities lost 2.9%. Its 26-week correlation with world equities was +0.36 on 9 October and +0.08 on 13 November.

Five weeks cannot judge gold's role. Over 2006-2026 our diversification study measured a monthly correlation with global equities of −0.06. For protection that pays off in sharp falls, our tail-risk hedge comparison covers the same sell-off.

What can a household check now?

Check how your defensive holdings would behave if rates, rather than growth, drove the next fall. These are calculations on your own holdings and spending, not a view on where rates or correlations go next.

  1. Read the effective duration of each bond fund. It is on the factsheet. Multiplied by a rise in yields, it gives the approximate loss before income: the iShares Core € Govt Bond UCITS ETF reported 6.75 on 16 November, so a one-point rise would cost about 6.75%.
  2. Separate spending needs from long-term capital. List what you must pay in the next one to three years and check whether it sits in instruments whose value barely moves with shares or yields, such as deposits or money market funds; long bond funds can fall when you need them.
  3. Test a 2022-type fall. Apply world equities' 17.0% loss and your bond fund's own worst fall to your actual weights, then compare the result with your accessible cash. For the € Govt Bond 7-10yr fund that worst fall was 23.1%, and on 13 November it was still 16.3% below its December 2020 peak with income reinvested.
  4. Check whether "defensive" holdings share the same inflation and rate risk. Long government bonds, corporate bonds and rate-sensitive shares can all fall together when yields rise; our guide to bonds when rates rise explains how starting yield and duration set the damage.
  5. Note the currency of foreign bonds. A hedged fund removes exchange-rate moves, an unhedged one adds them; in this episode the difference for US Treasuries was 1.7 points by 16 November.
  6. Judge diversification on joint losses, not on an average correlation. Our study of why investments fall together counts what fell in world equities' worst months since 2006.

Sources and method

Checked on 18 November 2026, with fund NAVs to 16 November 2026 and the Bund yield and ECB reference rate to 17 November. The 19 November snapshot (CSV) uses the iShares workbooks below and that day's ECB rate of $1.0992 per euro. The Nasdaq Composite's lower close that day, 23,879.20, is from Nasdaq's price index history. The study charts and weekly calculations keep their original cutoffs.

Fund data: iShares fund workbooks, "Growth of Hypothetical 10,000" sheet (fund NAV total return, income reinvested, after the total expense ratio), for the Core MSCI World UCITS ETF (IE00B4L5Y983, USD), MSCI World UCITS ETF (IE00B0M62Q58, USD), Core € Govt Bond (IE00B4WXJJ64) and Germany Govt Bond (IE00B5V94313).

Further workbooks: € Govt Bond 7-10yr (IE00B1FZS806), $ Treasury Bond 7-10yr USD (IE00B1FZS798) and EUR Hedged (IE00BGPP6697) share classes, and the Physical Gold ETC (IE00B4ND3602), which tracks the LBMA Gold Price. Effective durations are from the same workbooks as of 16 November 2026.

Rates: ECB euro reference rates (US dollar per euro, 14:15 CET), €STR, ECB deposit facility rate and the ECB Governing Council calendar. Bund yield: Deutsche Bundesbank, yield on listed Federal securities with 10 years' residual maturity, Svensson method. US yields: FRED DGS10, DFII10 and T10YIE. Press: Reuters poll of 12 November 2026, as syndicated by Euronext.

All returns, correlations, 60/40 portfolios, growth rates and drawdowns are copio calculations; the CSV files linked above contain every derived figure, and the weekly closes (CSV) and long-run results (CSV) give the inputs behind the correlations and growth rates. Dollar fund values are divided by the ECB reference rate of the same date; euro cash accrues €STR on an actual/360 basis.

The episode 60/40 holds the Core MSCI World and Core € Govt Bond funds at 60% and 40% on 12 October and drifts. The historical 60/40 holds the MSCI World and € Govt Bond 7-10yr funds and is reset to 60/40 at the first close of each month. Correlations use the last common close of each week; period averages are means of the rolling 26-week values ending in that period. Annual growth uses 19.94 years, calendar days divided by 365.25.

Limits: five weeks is a short, open episode, and the 26-week correlation still mostly reflects weeks before it. The bond fund's worst fall and distance from its peak use weekly closes from the linked CSV. Equity NAVs reflect US closing prices while the ECB rate is fixed at 14:15 CET and euro bond NAVs are struck earlier, so daily figures carry timing noise; weekly figures reduce it.

Fund returns include fund costs but no trading costs or taxes, and the 60/40 rebalancing is costless. The € Govt Bond 7-10yr fund has a longer duration than a broad euro government bond fund, so the historical bond losses are larger than for shorter funds.

This article is educational, not personalised investment advice.

FAQ

Did government bonds protect a euro portfolio in the October-November 2026 dip?

Modestly: from the 12 October 2026 peak to 19 November, an MSCI World fund lost 5.3% in euros while a euro-area government bond fund gained 0.7%. In the five-week study through 16 November, the 60/40 mix of those funds lost 2.0% to the 5 November low, against 4.4% for equities alone.

Is the stock-bond correlation negative again?

It is close to zero on our main measure, and the window decides the answer. Measured on weekly returns over 26 weeks, the correlation between world equities and euro government bonds in euros fell from +0.48 on 9 October 2026 to +0.02 on 13 November. It averaged −0.45 in 2008 and +0.26 in 2022. The same measure in dollars, against US Treasuries, was −0.12, less negative than its −0.54 of September 2025. Over 13 weeks the euro figure was −0.38, clearly negative; over 52 weeks it was +0.28.

Why do stocks and bonds sometimes fall together?

Both lose value when interest rates rise for reasons other than weaker growth. Higher yields lower existing bond prices directly and raise the rate at which future company profits are discounted. When fear about growth dominates, yields tend to fall and bonds rise while shares fall. Inflation surprises and expected rate rises can push both the wrong way, as in 2022: in euros, euro government bonds 7-10 years lost 14.8% and world equities 17.0% from the January peak to the June low, and 20.1% and about 13% over the year.

Did gold protect in the October-November 2026 dip?

Not in euros. The iShares Physical Gold ETC lost 1.8% in euros from the 12 October peak to the 5 November equity low and 5.6% by 16 November. By 19 November it was 9.6% below the 12 October level. In the five-week study, gold rose on 5 of the 11 days when world equities fell. Five weeks say little about gold's role over a longer crisis.

How much do bond funds lose when interest rates rise?

Roughly their effective duration multiplied by the rise in yields, before income. The iShares Core € Govt Bond UCITS ETF reported an effective duration of 6.75 on 16 November 2026, so a one percentage point rise in yields across its holdings would cost about 6.75% of its value. The fund's own factsheet gives the current figure.

Optional cookies help us measure visits and advertising. Choose each purpose, or decline both. You can change your choice using Cookie settings in the footer. Cookie policy.