Are government bonds really safe?
- Robin Powell
- 45 minutes ago
- 9 min read
Most investors think of government bonds as safe — the ballast that steadies a portfolio when markets turn rough. But three centuries of UK and US data tell a more complicated story, and with gilt yields at their highest since 2008 as a new military conflict reshapes the global outlook, the evidence has never been more urgent.
If you held a mainstream UK gilt fund five years ago, you're still sitting on a cumulative loss of about 16 per cent. Not a temporary dip. Not a blip you can explain away. A sustained, grinding decline across four separate crises — each of which was supposed to be the kind of environment where gilts earned their keep.
First came the pandemic. Then an energy shock triggered by Russia's invasion of Ukraine, which sent inflation to 11.1 per cent and gilt prices into freefall. Then the Truss mini-budget, when long-dated gilts crashed so hard the Bank of England had to launch an emergency buying programme. And now, since US-Israeli strikes on Iran began on 28 February, a fresh sell-off has pushed 10-year gilt yields to 5.065 per cent — their highest since July 2008 — while 30-year yields have reached 5.703 per cent, closing in on levels not seen since 1998. That's roughly 70 basis points of upward pressure in a single month.
Before the conflict, markets had been pricing in rate cuts. Now they're pricing in hikes. Bonds and equities are falling together — the scenario that balanced portfolios are supposed to prevent.
If you're uneasy about this, you should be. Something that was meant to protect you hasn't been doing its job. And not just recently.
A team of researchers from Stanford, Columbia and Northwestern has examined government bond returns stretching back 300 years, covering every major war and the Covid-19 pandemic across both the US and the UK. What they found should change how you think about the 'safe' part of your portfolio. Government bonds, it turns out, work a bit like an insurance policy with an exclusion clause buried in the fine print. They pay out reliably in some crises. But in the ones that burn hottest — wars, pandemics, geopolitical shocks that rewrite the fiscal landscape — the policy doesn't cover you.
Once you understand the exclusion clause, though, you can do something about it.
"Something that was meant to protect you hasn't been doing its job. And not just recently."
When bonds protect you — and when they don't
Government bonds aren't a myth as a safe haven — but the reputation only tells half the story. Bonds perform brilliantly in some crises and terribly in others. The difference depends on what kind of crisis you're facing.
That's the central finding of a landmark study by Jiang, Lustig, Van Nieuwerburgh and Xiaolan (2026), which tracks real returns on US and UK government debt across three centuries — the deepest dataset ever assembled on sovereign bond performance. During financial crises and recessions, bonds did what investors expect: they outperformed stocks by 16 per cent and GDP growth by 13 per cent in the years following the onset of trouble. That track record is real. But it's earned in peacetime downturns — banking panics, credit crunches, bursts of market volatility.
Wars and pandemics are a different matter. Across every major conflict in the dataset, bondholders suffered average real losses of 14 per cent in the first four years. Cumulative returns ran more than 20 per cent below both equities and real estate. Bonds didn't just fail to protect — they were the worst-performing major asset class during the episodes many investors fear most.
The UK evidence is especially telling, thanks to a remarkable dataset compiled by Ellison and Scott (2020), who reconstructed bond-by-bond monthly returns on every individual UK government bond since 1694. More than 330 years of gilt history — the most granular record of any sovereign's debt. The chart below shows what that record looks like. Watch the grey-shaded war periods, the sharp drops during WWI and WWII on the UK line, and the recent decline during Covid.
Napoleonic Wars, both World Wars, the pandemic — each carved a visible wound in the cumulative return line.
Why? Wars trigger fiscal demands that dwarf anything a recession produces. Government spending during wartime rose by an average of 7 per cent of GDP annually in the first four years — and tax increases alone have never covered that kind of bill. Governments needed another way to pay. Bondholders have historically been the ones who footed it — through a mechanism most investors have never encountered.
How governments quietly take your money
Those losses didn't happen by accident. Governments have a playbook for financing wars, and bondholders are central to it.
When the bill for a conflict exceeds what taxation can raise, governments borrow heavily — then erode the real value of that borrowing through two mechanisms: surprise inflation and financial repression. Jiang et al. found that together, these forces opened a cumulative 31 per cent wedge between bond returns and GDP growth over four years of war. Nominal returns on bonds stayed positive — around 5 per cent — but wartime inflation averaged 20 per cent cumulatively across US and UK conflicts. The purchasing power of bondholders was quietly destroyed.
Financial repression is the less familiar mechanism, but it may be the more important one. Reinhart and Sbrancia (2015) defined it as a toolkit of policies — yield caps, forced bond purchases by banks and pension funds, gold standard suspensions, capital controls — all designed to keep interest rates below inflation. The effect is a hidden tax on anyone holding government debt.
During the financial repression era of 1945–1980, real interest rates across advanced economies were negative roughly half the time. The chart below shows how dramatic the shift was — compare the wide, negative-skewing distribution of 1945–1980 with the positive rates that prevailed after liberalisation, and notice how the post-2008 distribution drifts back toward negative territory.
The annual cost to bondholders ran between 1 and 5 per cent of GDP. Not small change. That's how the UK halved its debt-to-GDP ratio in roughly 20 years after WWII — a feat that took more than 40 years after the Napoleonic Wars, when capital moved freely and governments couldn't rig rates against savers.
British bondholders were particularly exposed because the UK government has historically favoured long-duration debt. Ellison and Scott found that issuing shorter-dated bonds would have been cheaper for the Treasury — implying the government locked investors into longer maturities because long bonds are easier to erode through inflation. From the 1990s, the Debt Management Office pushed gilt duration even longer, setting the stage for the 2022 LDI crisis. Leveraged pension funds holding long gilts collapsed so violently that the Bank of England spent £19.3 billion on emergency purchases to prevent a systemic meltdown.
The insurer, in other words, isn't just excluding your biggest risks. It's quietly raising your premiums every year.
None of this is conspiracy. It's well-documented, and it's centuries old.
"The insurer isn't just excluding your biggest risks. It's quietly raising your premiums."
Are government bonds safe in a recession? Yes — that's the catch
Before you let panic drive your decisions and liquidate every gilt in your portfolio, a word of balance. The Jiang et al. research is unambiguous: during recessions and banking crises, government bonds do what they're supposed to do. They cushion portfolios. They justify their place in a sensible allocation. The post-war era has been dominated by these kinds of downturns, which is why so many investors trust bonds without question.
For most of the crises you'll face in a normal investing lifetime, that trust is well placed.
The trouble is the exceptions.
A separate study by Baltussen, Martens and Van der Linden (2026) examined defensive strategies across 222 years of global financial history. When they looked at the 12 largest equity drawdowns since 1800 — the real catastrophes, not ordinary corrections — treasury bonds delivered negative cumulative returns of -3.5 per cent over cash. Worse, the diversification benefit of bonds depends heavily on the correlation between stocks and bonds. And that correlation shifts at the worst possible moments. When you most need bonds to zig while equities zag, they start moving in the same direction. (A disclosure: the authors are affiliated with Robeco and Northern Trust, both of which offer investment products related to this research.)
The insurance policy still covers most standard claims — and those are the claims you'll file most often. But the exclusion clause kicks in during the deepest drawdowns. Investors who assume their bond allocation protects against everything are leaning on a guarantee that doesn't exist.
The question isn't whether to own bonds. It's whether to rely on them as your only line of defence.
Two things you can do about it
The evidence points to two practical adjustments. Neither requires you to predict the next war.
First, manage duration. The longer your bond holdings, the more exposed you are to inflation erosion and rate shocks. As we've seen, the UK government has long preferred issuing longer-dated debt — not because it's cheaper, but because it's easier to erode. The current crisis shows the cost of that exposure neatly: in March 2026, short-duration US Treasury funds were roughly flat while long-duration equivalents lost more than 4 per cent. The same dynamic applies to UK gilts. Shortening the average maturity of your bond allocation won't eliminate risk, but it cuts the damage that rising prices can inflict.
Second, diversify your crisis protection beyond bonds. If the defensive value of bonds depends on a correlation regime that shifts during the worst environments, then relying on one 'safe' asset class for all scenarios is a weakness you can fix. For a retail investor, that might mean broader global equity exposure, an allocation to inflation-linked bonds, or simply holding more cash than convention suggests when geopolitical risks are elevated.
This isn't market timing. You're not trying to guess when the next conflict starts. You're building a portfolio that doesn't assume government bonds are safe in every conceivable crisis. Design decisions, not tactical ones.
Time to read the fine print and update your coverage.
The incomplete truth about government bonds
"Three centuries of evidence won't tell you when the next war starts. But it will tell you what happens to your bonds when one does."
That 16 per cent five-year loss on UK gilts isn't bad luck. It's the exclusion clause at work — a pattern stretching from the Napoleonic Wars through both World Wars and into the pandemic and the current Iran conflict. The difference is that now you can see it for what it is.
The assumption that government bonds are safe isn't wrong. It's incomplete. Bonds protect you in recessions and banking crises — the downturns you'll face most often. They fail during wars, pandemics and geopolitical shocks, when fiscal demands overwhelm everything else and bondholders quietly absorb the cost.
Knowing that changes what you do. Not dramatically — you're not tearing up your allocation. But you're shortening duration, and you're broadening your defences so they don't depend on one asset class behaving well in every scenario.
Three centuries of evidence won't tell you when the next war starts. But it will tell you what happens to your bonds when one does — and that knowledge is worth more than any false sense of security.
In a future article, we'll examine what this evidence means specifically for retirees who depend on bond income.
Resources
Jiang, Z., Lustig, H., Van Nieuwerburgh, S., & Xiaolan, M. Z. (2026). Are government bonds safe in times of war and pandemic? NBER Working Paper 34820.
Ellison, M., & Scott, A. (2020). Managing the UK national debt 1694–2018. American Economic Journal: Macroeconomics, 12(3), 227–257.
Reinhart, C. M., & Sbrancia, M. B. (2015). The liquidation of government debt. Economic Policy, 30(82), 291–333.
Baltussen, G., Martens, M., & Van der Linden, L. (2026). The best defensive strategies: Two centuries of evidence. Financial Analysts Journal, 82(1), 6–34.
One logical next step
If this has made you think about whether your current approach to investing is actually working, our Find an adviser directory is a good place to start. Everyone listed has publicly committed to low-cost, globally diversified investing — the kind of approach that reinforces what you've read here, rather than quietly undoing it.
Stay connected: YouTube | LinkedIn | X For advisers: Connect with Robin about content