Are Chinese Bonds Still Worth Adding? From the Stock–Bond Relationship to a First Bond Allocation
Stocks and bonds do not always move in opposite directions. Even with China's ten-year government bond yield near historical lows, starting a bond allocation from zero can still make sense—but short-duration bonds, five-year government bonds, and thirty-year bonds serve very different purposes.
As of September 17, 2026. My view on bond allocation is this: if a portfolio has no bonds yet, it can begin building a defensive layer centered on short- and intermediate-duration interest-rate bonds. But at today's low yields, buying a full allocation to long-duration bonds at once offers an unattractive balance of risk and reward. These views are consistent. The first addresses portfolio structure; the second addresses entry price and duration risk.
Stocks and bonds do not always move in opposite directions
Bond prices and yields move in opposite directions, but that relationship explains only the price of a bond. It does not explain why stocks and bonds sometimes rise or fall together. Growth, inflation, and funding conditions can simultaneously change expectations for corporate earnings, equity discount rates, and government bond yields. When yields move, the key question is: what is driving them?
The chart groups scenarios by two underlying drivers; it does not infer the macroeconomic state from stock and bond returns. In the lower-left quadrant, if property activity, credit, and earnings weaken while price pressures ease, equity earnings may come under pressure and expectations of rate cuts may support long bonds. In the upper-right quadrant, stronger credit demand and earnings alongside rising core prices may help stocks, while higher yields weigh on long bonds. The upper-left quadrant deserves particular caution: growth remains weak, but inflation, bond supply, or funding conditions push rates higher, so both stocks and bonds may struggle. In the lower-right quadrant, improving earnings, moderate inflation, and ample liquidity may help stocks while bonds remain stable or strengthen. These are qualitative paths; actual outcomes still depend on the size of the moves, policy responses, and market expectations.
In practice, I would track three groups of evidence together: core prices and the producer price index (PPI); property sales and medium- to long-term corporate credit; and corporate earnings and money-market rates. A rise in yields looks more like a growth-driven shift only when demand and earnings improve together. A rise in rates alone does not imply that stocks will benefit.
Is China experiencing broad-based deflation?
Slowing inflation, low inflation, and sustained deflation are different conditions. In August 2026, China's consumer price index (CPI) rose 0.8% year over year, core CPI excluding food and energy rose 1.0%, and PPI rose 3.8%. On these broad price measures, it would be inaccurate to say China had entered “broad-based deflation.” National Bureau of Statistics CPI data · PPI data
Positive price-index readings do not erase pressure on domestic demand and balance sheets. From January through August 2026, real estate development investment fell 19.9% year over year and the floor area of newly sold commercial housing fell 12.1%; in August, total retail sales of consumer goods grew only 0.4% year over year. Taken together with CPI and PPI, these figures point more toward “low consumer inflation, a rebound in industrial-goods prices, and continued property-sector contraction.” National Bureau of Statistics real estate data · Retail sales data
Weaker property prices and household wealth expectations may lead people to save more, consume less, or repay debt early, in turn weighing on corporate revenue and credit demand. This is a debt-deflation risk chain worth monitoring, not a claim that a broad CPI deflation spiral has already taken hold. It also helps explain why the bond market may still price in weak demand even as price data recover.
Pressure on property activity and prices → Weaker consumption and credit demand → Lower expectations for nominal growth → Downward pressure on long-term yields
This is a transmission path to watch, not an assertion that every link has occurred at the same time. A PPI rebound can coexist with it.
How far has the bond rally gone? Low yields and historical percentiles
The government bond yield curve published by the Ministry of Finance for September 14, 2026 showed approximately 1.23% for one year, 1.42% for five years, 1.69% for ten years, and 2.15% for thirty years. Moving from five to ten years added only about 27 basis points of yield while taking on more duration risk. Ministry of Finance—China government bond yield curve
The chart shows yields at four maturities, not future fund returns. The extra yield between five and ten years should be weighed against the additional duration risk.
A third-party historical ranking, using a ten-year yield of 1.68% on September 3, put its “price-direction percentile” at roughly 96.2% over the past decade and 98.5% over what it calls its full historical sample. Government bond rate thermometer These rankings reverse-map low yields into relatively expensive bond valuations. They are not the actual price percentile of any ETF. Differences in sample start dates, yield-curve definitions, and observation dates mean that figures after the decimal point should not be treated as precise trading signals. The more robust conclusion is that ten-year yields remain near historical lows, leaving a thin coupon cushion for long bonds.
Low yields do not guarantee that the bond rally is over. If growth weakens further and rates fall again, long bonds could still rise. If nominal growth recovers or funding conditions change, long bonds could suffer a material drawdown first. I would assess the long-term need for bonds separately from whether this is a good time to extend duration.
Duration: why thirty-year government bonds can fall quickly
An approximation for a bond's price sensitivity to yield changes is: percentage price change ≈ −modified duration × change in yield. The following assumptions illustrate scale; they are not forecasts of fund net asset values:
- Assume a ten-year bond has a modified duration of 8 and its yield rises from 1.69% to 2.00%. A 31-basis-point increase implies an approximate price change of −2.5%.
- Assume a thirty-year bond has a modified duration of 18 and its yield rises from 2.15% to 2.50%. A 35-basis-point increase implies an approximate price change of −6.3%.
Actual returns also depend on coupons, convexity, portfolio holdings, fees, and an ETF's premium or discount to net asset value. The point is that low credit risk in government bonds does not make a long-duration government bond ETF's value stable.
A first bond allocation: give each layer a job
If bonds are meant to reduce stock-driven portfolio volatility, provide liquidity, and fund rebalancing, I would define four layers before choosing instruments. Position sizes depend on investment horizon, liquidity needs, and tolerable drawdown; a universal allocation percentage cannot replace those decisions.
- Liquidity layer: short-term financing instruments and ultra-short-duration bonds. These can be useful for rebalancing but carry credit and liquidity risk; they are not cash equivalents.
- Core layer: government bonds with roughly three to five years to maturity. They take on relatively moderate interest-rate risk, but their value can still fall when rates rise.
- Intermediate- to long-duration layer: five- to ten-year government bonds or policy bank bonds. They add duration sensitivity during an economic slowdown; policy bank bonds have different credit characteristics from government bonds.
- Optional directional layer: thirty-year government bonds. They are highly sensitive to rate moves and can be omitted without a clear macroeconomic view.
Starting from zero, I would rather build the short- and intermediate-short-duration base first, then decide whether to extend duration as yields and fundamentals change. In particular, if the ten-year yield returns to 1.8%–2.0%, I would first determine whether the move reflects bond supply, funding conditions, or a genuine improvement in demand and corporate earnings. That range is a scenario to watch, not an automatic buy signal.
Which instruments are worth studying?
When researching an instrument, look at its underlying bonds before its fund structure. Short-term financing ETFs and ultra-short-duration bond funds mainly hold short-term credit bonds; they illustrate why “low duration” does not mean “no credit risk.” An ETF holding roughly five-year government bonds is one way to study intermediate duration. ETFs holding five- to ten-year government bonds or policy bank bonds help compare longer duration and different credit characteristics. A thirty-year government bond ETF is better studied as an interest-rate-sensitive instrument than treated as a low-volatility core holding. Shanghai Stock Exchange introduction to bond ETFs
Corporate bonds and local-government-financing-vehicle bonds add credit-spread and issuer risk; convertible bonds have a significant equity component. If the bond allocation's primary purpose is defense and rebalancing, these instruments should be assessed separately from pure interest-rate bonds. When comparing bond funds, I would examine duration, yield to maturity, credit quality, leverage, liquidity, and fees, in that order, rather than starting with a ranking of last year's returns.
Conclusion: watch why yields move
The central tension in China's bond market is that weak property activity and domestic demand still support low interest rates, but long-end yields are already low, leaving little room for error when chasing more duration. For someone starting a bond allocation from zero, the first priority is a resilient bond holding that can be used when stocks are volatile. Thirty-year government bonds should be studied as a separate directional instrument.
Going forward, I would watch the ten-year government bond yield, core CPI, PPI, property sales, medium- to long-term corporate credit, and earnings expectations together. If yields rise alongside stronger stocks, demand, and earnings, that looks more like a shift toward reflation. If yields rise without evidence of better growth, a bond pullback should not automatically be read as good news for stocks.
This article records a personal research framework and conditional judgments. It is not investment advice, a return guarantee, or a basis for buying or selling securities. The historical percentiles, duration sensitivities, and scenario analyses are research tools, not objective ratings. Markets and products can change; readers should independently verify information and bear their own risks.