The Safe Asset With a Fragile Shadow
This series examines the modern sovereign-bond market: the market for government debt that anchors borrowing costs, supplies widely used collateral and is often treated as a refuge when other assets look dangerous. The central question is not whether government bonds are always safe or always risky. It is how a market built around a high-quality asset can still amplify a shock when the institutions trading and financing that asset rely on leverage, short-term funding and tightly linked balance sheets. Understanding that distinction gives readers a way to judge whether a sharp move reflects public-finance fundamentals, stressed market plumbing or both.
The apparent contradiction matters because sovereign bonds do much more than promise future payments. Their yields help set the price of mortgages, corporate loans and other securities. They are pledged in repurchase agreements, or repos, in which one party raises cash against securities and agrees to reverse the transaction later. They are also held by banks, insurers, pension funds, money-market funds, asset managers and hedge funds with different obligations and different abilities to wait out a loss. A disturbance in the bond market can therefore change financing conditions well beyond government treasuries.
Why government bonds sit at the center
A government bond can be attractive for a straightforward reason: the issuer is expected to make the promised payments. But the market values the bond continuously, and those values respond to interest rates, inflation expectations, fiscal news and the willingness of investors to bear risk. A bond that is sound when held to maturity can still lose market value when yields rise. That price change matters immediately to an institution that must meet a margin call, satisfy a redemption request or maintain a risk limit.
The bond market is also a piece of infrastructure. A sovereign yield is a reference point against which private borrowers are priced. Government securities can be turned into cash through repo without being sold outright. Their liquidity allows investors to rebalance portfolios and allows dealers to make markets in normal conditions. These functions make the asset useful precisely because many participants expect it to remain tradable and acceptable as collateral.
That expectation creates a second layer of risk. The safety of the claim does not guarantee that every holder can finance it, that every dealer can continue quoting prices or that every fund can sell without moving the market. Those are properties of intermediation, not merely of the government’s ability to pay. The Bank for International Settlements describes this as a new fiscal-financial stability nexus: public debt is increasingly held and intermediated through non-bank financial institutions whose funding models and risk controls differ from those of banks.
The key issue is therefore not to replace one slogan—“government bonds are safe”—with its opposite. It is to ask which kind of safety is being discussed. Credit safety concerns the likelihood of receiving the promised cash flows. Market safety concerns the ability to trade, finance and hedge the position when conditions change. The two usually support one another, but they can separate under stress.
Safety of the bond is not safety of the market
Consider a deliberately simple balance-sheet picture. An investor uses a small amount of its own capital and borrows the rest of the cash needed to hold a much larger position in government bonds. If the bonds move slightly, the gain or loss is calculated on the large position, not just on the investor’s capital. The strategy may be sensible in calm conditions, especially when the expected price difference is small but the position can be financed cheaply. Its vulnerability lies in the financing terms: the lender can ask for more collateral, a derivatives counterparty can demand more margin, or the investor can lose access to short-term borrowing.
At that point, the investor may need cash before the bond has become a bad long-term asset. Selling is not necessarily a judgment that the government will default. It may simply be the fastest way to meet a contractual demand. If many investors face that demand at once, sales can push prices down and yields up. The lower prices can create more margin pressure, while the larger price moves can make dealers less willing to absorb inventory. A market that looked deep in calm conditions can become thin precisely when everyone needs liquidity.
This is why fiscal space is not determined by a debt ratio alone. The BIS argues that the financial system affects both the level of sovereign borrowing costs and how sharply those costs respond to additional debt or a shock. When intermediary balance sheets are strong, volatility is contained and funding is plentiful, the market may absorb new issuance with limited pressure on yields. When risk-bearing capacity weakens, the same fiscal position can produce a larger increase in borrowing costs. Fiscal space can contract even if the long-run relationship between interest rates and economic growth has not changed.
There is no single debt threshold that resolves this uncertainty. A country’s debt burden, maturity profile, fiscal credibility and currency structure matter, but so do the institutions that intermediate the debt at that moment. The distinction also separates balance-sheet amplification from a purely self-fulfilling run. Investors’ expectations can matter, yet the market can become unstable because risk limits, collateral requirements and funding capacity force otherwise willing holders to reduce positions.
The high-level sequence is shown below. It is a map of a possible feedback loop, not a prediction that every yield move reaches the final box.
The feedback arrow does not mean that official action always worsens conditions. It means that a response can affect the incentives and balance-sheet choices that shape the next episode. A backstop may restore trading and prevent forced sales today; if market participants come to expect rescue, however, they may carry thinner liquidity buffers or use more short-term funding tomorrow. The question is not whether intervention is good or bad in the abstract, but which failure it addresses and what vulnerability it leaves behind.
The network around a government bond
The institutions surrounding the sovereign-bond market do not all behave alike. Insurers and pension funds often have long-term liabilities and can provide durable demand for government debt or use it to manage duration. Hedge funds may use futures, spot bonds, swaps and repo to pursue relative-value strategies whose returns depend on small price differences and favorable financing. Banks and broker-dealers make markets, lend securities, provide repo and prime-brokerage services, and stand between many counterparties. Money-market funds and open-ended bond funds can hold liquid assets for investors who expect frequent access to their money, even though market liquidity can deteriorate in a downturn.
The label “non-bank financial institution” therefore describes a broad set of roles rather than one common risk. The BIS reports that NBFIs’ share of sovereign-debt holdings in advanced economies rose from 44 percent in 2021 to 53 percent in 2025. That shift can increase market diversity and absorb more government issuance. It can also make the market more sensitive to funding and redemption conditions when a larger share of activity is conducted by institutions that do not use the same capital, liquidity or supervisory framework as banks.
The bank-sovereign relationship changes within this network rather than disappearing. Banks still hold government bonds and can suffer valuation losses when yields rise. They also finance hedge funds, provide derivatives and margin services, extend credit lines and depend on non-banks for parts of their own funding. A bank may therefore be exposed to a sovereign-market shock through a client’s collateral or funding position even when its direct bond holdings are not the main source of loss.
Current conditions illustrate why a careful assessment must hold resilience and vulnerability together. The Federal Reserve’s May 2026 report says that US banks had historically high regulatory-capital ratios and substantial liquid-asset buffers, while broker-dealer leverage remained near historically low levels. Aggregate funding risks were broadly in line with historical norms, and recent redemption pressure at certain private-credit vehicles remained manageable, although some vehicles limited redemptions. At the same time, hedge-fund leverage was near all-time highs and concentrated among large funds, life-insurer leverage remained elevated, fixed-rate asset losses at banks were still sizable, and Treasury-market liquidity showed sensitivity to periods of high interest-rate volatility.
Those observations do not amount to a forecast of imminent crisis. They describe a system that can function normally while retaining channels through which a sufficiently large shock could be amplified. The difference is important. A market can be calm because its participants are well buffered, because financing terms are favorable or because no demanding shock has arrived. Those conditions have different implications for what happens next.
The first question when the market moves
When sovereign yields rise, the first task is to resist a single-cause explanation. Some of the move may reflect a genuine reassessment of inflation, growth, fiscal sustainability or the supply of government debt. Some may reflect the valuation losses and funding pressure created by the move itself. The two forces can interact: fiscal news can raise yields, higher yields can tighten margins, and forced sales can push yields farther than the original news alone would have done.
A useful reading asks four connected questions. First, what changed in the government’s expected finances or in the wider economy? Second, who is carrying the position, and how much of it is financed rather than owned outright? Third, which contracts require cash when prices or volatility move—repo, derivatives, securities lending or fund redemptions? Fourth, can dealers and other investors absorb the sales, or are their own balance-sheet and risk limits tightening at the same time?
These questions keep the sovereign-bond market at the center while making its hidden dependencies visible. They also explain why the subject matters to people who never trade a government bond. A disorderly rise in sovereign yields can raise borrowing costs for households and businesses, narrow a government’s room to respond to a downturn and complicate a central bank’s effort to control inflation. The asset that anchors the system can become the place where pressures meet.
The next step is to look more closely at the balance-sheet arithmetic behind that meeting. A small relative-value opportunity can support a position many times larger than the capital committed to it. As long as funding remains cheap and liquid, that leverage may look like efficient intermediation. When the terms change, it can become the fragile shadow cast by the safe asset itself.