How expensive are credit spreads, really? A 100-year analysis
A century of market history challenges assumptions about corporate bond spreads, revealing why today’s levels may persist for longer than investors expect.
Authors
By mid-2026, the additional yield (or spread) that US-dollar corporate bonds offer over government bonds had fallen close to its lowest level for nearly 25 years. Other global corporate bond markets offer similarly paltry spreads.
At first glance, this makes corporate bonds look highly unattractive compared with government bonds, although overall yields are at more appealing levels. But that conclusion is partly a function of the period of history over which we compare them. Credit spread data for major corporate bond indices only stretches back to the late 1990s (for example, the ICE Bank of America and IHS Markit iBoxx indices).
But this is not the whole story. In this article we take a much longer-term look at how credit spreads have evolved, going back to 1919. This captures the Roaring 20s, the Great Depression, WW2, the post-war recovery and everything since.
On this basis, the question of whether credit spreads really are expensive compared with history does not have a straightforward answer. Spreads spent several decades tighter than they are today, from the late 1940s until the early 1970s. Investors wondering whether to buy, sell or hold should ask themselves: could this happen again?
We examine the conditions that supported such an outcome and contrast them with today, finding a surprising number of parallels. We focus on spreads over government bonds throughout.
Credit spreads spent much of the 1940s, 50s and 60s at tighter levels than today
Baa corporate bond spread over 20yr Treasury
Past performance is not a guide to the future and may not be repeated
Spread is calculated as Baa corporate bond yield minus 10 and 20-year constant maturity Treasury yield. Baa corporate bond yield is based on bonds with maturities 20 years and above but is often presented as a spread over the 10-year Treasury, which is why both have been shown here. Source: Moody’s, Board of Governors of the Federal Reserve System (US), Robert Shiller, National Bureau of Economic Research, FRED, Federal Reserve Bank of St. Louis, and Schroders
1. Government debt levels were high, lowering spreads via a “convenience yield” mechanism
There is evidence that corporate bond spreads are lower when government debt/GDP is higher, as was the case after WW2, and higher when debt levels are lower. The reason is that many investors buy Treasuries for the “convenience” benefits they provide: liquidity (they are extremely liquid), neutrality (no need to choose between one corporate borrower and another) and surety (appeal to risk-averse investors). These benefits are especially relevant to large official holders such as foreign central banks and state and local governments.
This demand lowers Treasury yields relative to other bonds. It also means demand does not fall as much as might be expected when yields are low and prices high. In economic terms, demand is less price-elastic.
It is perhaps easier to see how this impacts credit spreads in the situation opposite to today, where government debt levels are lower and Treasuries relatively scarce. In this environment, the “convenience yield” that many investors are prepared to pay increases, pushing Treasury yields lower relative to corporate bond yields, resulting in higher credit spreads.
In contrast, when Treasuries are abundant, as they were after WW2 and are again today, that convenience yield falls. Treasury yields rise relative to corporate bond yields. The result is a narrower credit spread.
Krishnamurthy and Vissing-Jorgensen demonstrated this relationship statistically in their 2007 paper, The Demand for Treasury Debt. Using data from 1925 to 2005, they found that the debt/GDP ratio was negatively correlated with credit spreads, after controlling for corporate default risk.
I have extended their analysis to 2025 and shown it in a simpler form below. Unlike the academic analysis, this does not control for other influences on spreads, most importantly default risk. Even so, the relationship was relatively strong until the financial crisis, close to the period covered by the original paper. It has been weaker since then, although some of the more elevated spreads in this period reflected higher default risk and do not necessarily invalidate the argument.
Credit spreads have tended to be lower when debt/GDP has been higher
Source: Henning Bohn, 2008. "The Sustainability of Fiscal Policy in the United States" (in: R. Neck and J. Sturm, "Sustainability of Public Debt", MIT Press, pp.15-49, Moody’s, US Office of Management and Budget, FRED, Federal Reserve Bank of St. Louis, and Schroders
Could this repeat today:
The parallels are strong. Debt levels are high and forecast to keep rising, partly because ageing populations will increase health and pension costs. For example, the US Congressional Budget Office forecasts that US debt/GDP will rise from around 100% in 2025 to 156% in 2055. Many other economies face similar challenges. If the “convenience yield” argument holds, this environment of abundant government bonds could usher in an extended period of lower credit spreads than investors have become used to.
This does not mean spreads would fail to rise during recessions or when default risk increases. They still would. But it suggests that the “natural” level to which spreads revert could be lower than investors expect. This is one of the strongest arguments for spreads staying lower for longer.
2. Historically strong economic growth and zero/near-zero defaults.
The post-war economy provided an unusually supportive backdrop for corporate borrowers. Real GDP growth was often above 4% a year on a rolling ten-year basis. Excluding the deep recession of 1946, it averaged almost 4% a year between 1947 and 1969. From then until the global financial crisis, the average was around 3%; since then it has been closer to 2%, or 2.5% if measured from 2010.
Post-war growth was strong
Past performance is not a guide to the future and may not be repeated
Source: LSEG Datastream, Schroders
Strong growth was accompanied by extraordinarily low defaults. Astonishingly, not one investment grade issuer defaulted between 1946 and 1969. Even within the riskier sub-investment grade market, defaults averaged only 0.4% a year, compared with a long-run average of 2.9%. In half of those 24 years, no sub-investment grade issuer defaulted. Outside this period, there has never been a single year with none at all.
Investment grade default rate
Sub-investment grade default rate
Past performance is not a guide to the future and may not be repeated
Source: Moody’s
The relationship became self-reinforcing. Strong growth supported companies and kept defaults low. Investors then became more relaxed about credit risk and demanded a smaller risk premium. Lower borrowing costs further reduced default risk. As with all things of this nature, it worked fine until it didn’t.
Persistent and rising inflation from the mid-1960s increased uncertainty and led investors to demand higher risk premiums, including wider credit spreads. The return of defaults was also a wake-up call.
Could this repeat today:
On growth, while not our central scenario, it is possible if AI productivity gains really take off. The effects would not be evenly felt. Some sectors would benefit, including the hyperscalers, others in the AI tech stack and businesses that become more productive through the use of AI. Others, particularly more consumer-oriented sectors, could face greater credit risk. We consider the sector make-up of the corporate bond market later.
On defaults, with corporate borrowing costs well above their ten-year average and comfortably ahead of inflation, another extended period with no investment grade defaults is highly unlikely unless financial repression returns.
Defaults are also only one component of the loss investors suffer. Recovery rates matter too: bondholders do not usually lose everything when an issuer fails. The assets of a company may be sold and the proceeds distributed to creditors according to seniority. Average recovery rates on unsecured bonds are around 40%, while secured loans have historically recovered more than 60%.
Moody’s notes that recovery rates have been trending down, in part because covenant-lite loans “have allowed asset-stripping and priming that tends to diminish recovery prospects”. Lower recoveries mean higher losses for lenders. In theory, that should result in wider spreads, not narrower ones.
There is also considerable uncertainty over recoveries if a large technology borrower were to default, for example because its technology had been overtaken. Hyperscalers are a growing part of the market, but there is little history on which to base such an assessment.
3. Policy and regulation channelled investors into credit
Policy played an important part. The Fed capped long-term Treasury yields at 2.5% until 1951, holding the absolute cost of borrowing low in the early post-war years, often well below inflation. This reduced default risk and the premium investors demanded for taking it. As with quantitative easing more recently, it may also have encouraged Treasury investors to move further out along the risk spectrum in search of a satisfactory yield, increasing demand for corporate bonds.
Treasury yields were capped until 1951
Real Baa yield was often deeply negative
Source: LSEG Datastream, FRED, Federal Reserve Bank of St. Louis, Schroders
Regulation also shaped demand. Regulation Q prevented banks from paying interest on deposits and capped some other rates. This encouraged investors to move further out along the risk spectrum in search of yield. With fewer alternatives than investors have today — there was not really a high-yield bond market and equity ownership was less widespread — corporate bonds benefited. These rules had been introduced after the Great Depression, amid concern that competition for deposits had encouraged banks to take excessive risks.
Could this repeat today:
As things stand, there is little economic justification for yield curve control. Growth is reasonable and inflation moderate. But the post-WW2 economy was also growing strongly. The policy was not introduced because the economy needed support; it reflected the Treasury’s desire to keep borrowing costs low and manage the debts accumulated during the war.
There are parallels today. Government borrowing is forecast to rise in the US and other major economies. Political pressure on the Federal Reserve has also grown. President Trump has called openly for lower rates and singled out mortgage costs as a particular concern. Yield curve control may not be necessary, but that does not mean it is off the table.
Regulatory change could also increase the appeal of credit, either directly or indirectly through policies that boost demand for government bonds and encourage investors to reach for additional yield elsewhere. Possible examples include:
- Stronger duration-matching incentives for insurers or pension funds to hold long-dated government and/or corporate bonds.
- Lower capital charges for insurance companies or banks holding corporate bonds.
- Changes to central bank collateral eligibility or haircuts for corporate bonds.
- In some countries, public sector pensions are unfunded. There is no pool of assets set aside to be invested to meet current and future liabilities. Instead they are paid out of current taxation in a “pay-as-you-go” structure. The present value of future liabilities is also not captured as part of government debt/GDP statistics. If liabilities were recognized in debt/GDP, fiscal optics could push “prefunding” and issuance to seed the fund. This would create a large, structural, asset pool, spurring significant demand for many asset classes, including government and corporate bonds. The impact on debt/GDP makes this politically unpalatable for all but the bravest of politicians.
- Tax-incentivised regulatory frameworks and/or introduction of tax‑advantaged government‑bond funds for households. A net-of-tax yield uplift could spur retail demand. E.g. there is strong evidence that the UK capital gains tax exemption on UK government bonds boosts demand from UK retail investors for short-dated UK government bonds. Recent Bank of England research found that one particular gilt is so popular with retail investors that they own almost 100% of the free float
These examples are illustrative rather than exhaustive and should not be interpreted as an endorsement. The mechanism is uncertain, but encouraging demand for government debt is a plausible policy goal when the government itself is the direct beneficiary. Corporate bonds could benefit as investors seek higher yields.
4. Institutional demand exploded
Alongside regulatory/policy drivers, the investor base was being transformed. Insurance companies shifted heavily from government to corporate bonds. In the five years from 1945, they bought more corporate bonds in dollar terms than they had held at the start of the period. Corporate bonds rose from roughly one-third of their bond portfolios in 1945 to around two-thirds by the early 1950s. By the late 1960s, the share was above 90%.
US insurance company bond portfolio split between Treasuries and corporate bonds
Source: Board of Governors of the Federal Reserve System (US) via FRED®, and Schroders. Data to 2025. This analysis is based on the market value of foreign and corporate bonds combined, not solely corporate bonds. However, analysis of the book value of domestic corporate and Treasury bonds leads to a similar pattern and does not impact our conclusions.
Pension funds added another powerful source of demand. In the late 1940s, their net purchases of corporate bonds averaged only around 10% of those made by insurers. By the late 1950s they were matching them. In the 1960s they exceeded them in every year bar one.
Together, insurers and pension funds grew to own more than 80% of the market at their peak in the 1960s. Retail investors, who had accounted for more than half of the market in earlier decades, were displaced. This surge in incremental institutional demand helped to drive spreads to exceptionally low levels.
Insurance companies and pension funds quickly grew to dominate the corporate bond market
Source: Board of Governors of the Federal Reserve System (US), Corporate and Foreign Bonds; Asset, Market Value Levels, retrieved from FRED, Federal Reserve Bank of St. Louis, Schroders. Holdings data covers corporate and foreign bonds, the longest consistent historical series available by owner sector. The figures therefore include some non-US corporate debt holdings and should be interpreted as indicative of broad ownership trends rather than exact ownership of US corporate bonds. For earlier data at least this category is likely to have been dominated by domestic corporate bonds. Latest data is as of 2025.
Could this repeat today:
Regulatory incentives could create incremental demand, but corporate bonds already make up a much larger part of institutional portfolios than they did in 1945. A shift on anything like the same scale is therefore unlikely.
5. Relatively defensive utilities made up a large and growing % of the market
There is one important reason not to draw too close a parallel with today. Utilities made up a significant and growing proportion of the investment grade corporate bond market, particularly electric and telephone companies. By 1975, they were around 60% of the market. As regulated businesses, they benefited from relatively stable cashflows and government oversight, contributing to lower perceptions of risk and tighter spreads.
Sector allocation of US corporate bond market
Source: 1900-1950: Trends and Cycles in Corporate Bond Financing, Hickman, 1952; 1975-2025 Bloomberg Barclay US corporate index. 1900-1950 covers the entire bond market, not investment grade specifically. Railroads included within Industrials from 1975 onwards. For reference, railroads make up only 1.2% of the market value of corporate bonds outstanding as at 31 December 2025.
Could this repeat today:
Today’s corporate bond market contains more cyclical and less-regulated businesses, while utilities account for a much smaller share. On this measure, the “natural” spread should be higher than it was during the post-war period.
Summary
Then | Now |
1. Government debt levels were high, lowering the “convenience yield” on Treasuries, and causing their yields to rise relative to corporate bonds. | Government debt levels are high and forecast to rise. |
2. Historically strong economic growth and zero/near-zero defaults. | Potential growth upside from AI but some sectors would be negatively affected. There would be collateral damage, impacting credit risk. Near-zero defaults unlikely. |
3. Policy and regulation channelled investors into credit. | Politics is becoming increasingly entangled with monetary policy. Potential in Treasury market, with knock-on consequences for corporate bonds. |
4. Institutional demand exploded. | Not possible on the same scale but incremental shifts are possible, especially if incentivised by regulation. |
5. Relatively defensive utilities made up a large and growing % of the market. | Much more cyclical exposure, utilities a small %. |
Source: Schroders
There is historical precedent for credit spreads to stay tighter than today’s levels for several decades at a time. A short-term focus, driven primarily by data availability, means many investors are blind to this. Those waiting for the perfect moment to enter the market should bear it in mind.
Today’s environment is clearly different to the post-war era but some of the forces that held spreads down then are also present now. One that few have focussed on is the potential transmission mechanism from higher government debt levels to a lower convenience yield, to tighter credit spreads. Others, such as government interference in monetary policy, are more speculative but cannot be discounted.
To repeat from earlier, as it is important: this does not mean spreads would fail to rise during recessions or when default risk increases. They still would. But it means that the “natural” level to which spreads revert could be lower than investors expect.
For investors taking a long-term view, today’s credit spreads still offer a pickup over long-term default losses – a smaller pickup than investors have become used to, but a positive one nonetheless. For those intending to hold to maturity, there is therefore less reason to hold off buying corporate bonds. Active approaches have an advantage because current spreads are doing little to distinguish between good and bad borrowers. AI, geopolitics and other pressures will affect companies and sectors unevenly. The time to be more discriminating is now.
For more tactical investors, very tight spreads increase the risk of a reversal. From today’s starting point, it would take only a relatively small rise in spreads for corporate bonds to underperform government bonds – very relevant if you plan to buy now and sell later. Margins of safety are wafer thin.
Authors
Теми