China Development Bank Additional Discussion Report: Policy Asset Risk

• Report date: 2026-06-01
• Issuer / Theme: China Development Bank / policy asset risk, sovereign linkage, funding, capital maintenance, overseas policy finance
• Report type: additional_discussion
• Discussion scope: Review of potential credit-metric deterioration paths associated with the expansion of CDB's policy mandate, as discussed in the SSC discussion
• Reference context: China Development Bank issuer summary dated 2026-05-20, issuer notes / knowledge snapshot / source registry dated 2026-05-20, and discussion generated 2026-06-01

1. Purpose and Treatment

This report is a supplementary report that organises the Q&A conducted in the discussion in the context of the existing China Development Bank issuer report. The views set out here include analytical hypotheses and future verification points raised in the discussion, and do not constitute newly verified factual findings.

The foundation already confirmed in the existing report is that CDB is a policy bank 100% owned by Chinese central-government-related shareholders, and is a core quasi-sovereign financial issuer supporting infrastructure, industrial policy, the Five Priority Areas, new policy-based financial instruments, and the Belt and Road Initiative. Total assets at end-2025, a low NPL ratio, capital adequacy, and the likelihood of government support are strong credit-supportive factors. At the same time, Stage 2 balances, low-yielding long-term policy assets, reliance on bond-market funding, and the distinction between government support and explicit guarantees have already been identified as constraints in the existing report.

This discussion examined, among those constraints, not the weakness of government support but how policy mandates may accumulate on CDB's balance sheet and which metrics may show strain first.

2. Main Analytical Takeaways from the Discussion

The central hypothesis in the discussion was that CDB's credit deterioration path is more likely to proceed through a delayed deterioration in policy asset quality, combined with low margins, a renewed rise in credit costs, capital consumption, and weaker funding conditions, rather than through a near-term disappearance of expectations for government support.

In terms of policy areas, local-government-related infrastructure and new policy-based financial instruments were treated as the first areas to monitor. Both are close to CDB's core function as a policy bank, and are prone to a combination of large balances, long tenors, repayment sources dependent on local public finances, public tariffs, and land-related revenue, and delayed recognition of NPLs due to policy coordination. Overseas policy finance may not necessarily be the main driver that immediately moves CDB's overall accounting metrics, but it was framed as an area likely to affect spreads and transparency premiums because of country risk, debt restructuring, difficulties in collateral recovery, and limited disclosure.

With respect to early-warning indicators, the NPL ratio was treated as a lagging indicator. The more important metrics are Stage 2 balances and ratios, loan modifications, repayment deferrals and tenor extensions, accrued interest and capitalised interest, and delinquencies of more than 30 days but less than 90 days. In particular, a combination in which the NPL ratio remains low and flat while Stage 2 exposures, loan modifications, and accrued interest increase and provision coverage declines was positioned as a warning line before deterioration in policy assets appears in earnings or capital.

In terms of sovereign linkage, China's sovereign rating and outlook have the most direct bearing on downgrade risk. For spreads, however, the discussion noted that a widening spread differential between CDB bonds and Chinese sovereign bonds, a prolonged local-debt resolution process, and the scale of capital support for policy banks and state-owned banks could have an impact before a rating change. CDB should not be treated as identical to the Chinese government itself; expectations of government support, fiscal capacity, and transmission channels from local fiscal risk need to be assessed separately.

On funding, the discussion framed CDB's risk as more likely to emerge not as an inability to issue, but as a deterioration in the quality of funding despite continued issuance. If a sustained widening of spreads versus government bonds, persistently elevated new-issue premiums, weaker demand for long-dated bonds, shortening of issuance maturities, and a widening pricing gap between domestic financial bonds and offshore senior bonds occur at the same time as a decline in net interest margin and an increase in Stage 2 exposures, the market may be starting to price in CDB-specific policy burdens rather than only the interest-rate environment.

3. Summary of Q&A Content

3.1 Policy Areas Most Likely to Deteriorate First as Policy Finance Expands

The intent of the question was to view the main potential weakness of CDB not as weak government support, but as the accumulation of policy assets with long tenors, low profitability, reliance on local public finances, and overseas country risk, and to identify which policy areas are most likely to lead to an increase in Stage 2 exposures, a renewed rise in credit costs, and capital consumption.

The key point of the response was to rank local-government-related infrastructure, urban renewal, and regional infrastructure as the first area to monitor; new policy-based financial instruments as the second; Belt and Road and overseas policy finance as the third; and, among the Five Priority Areas, technology finance, inclusive finance, and some green finance as the fourth. This ordering emphasises balance size and proximity to CDB's core mandate. Overseas policy finance may have a large impact on loss severity and market sentiment, but whether it is large enough to move CDB's overall credit metrics requires further verification.

The point examined in the follow-up was how to interpret the low NPL ratio. The existing report shows a low NPL ratio at end-2025, but Stage 2 balances are large, and delinquencies or problem-asset recognition in long-term infrastructure and policy loans may take time to surface. In the discussion, it was emphasised that for a policy bank, rollovers, repayment deferrals, government coordination, and project-tenor extensions may cause Stage 2 exposures, delinquencies, restructurings, provision coverage, and accrued interest to move before the NPL ratio.

The credit analytical implication is that the higher CDB's policy importance, the stronger the government-support logic becomes, but the same policy importance can also increase low-profitability, long-duration, concentrated, opaque, and capital-consuming exposures. This duality needs to be monitored explicitly.

3.2 Early-Warning Indicators to Monitor Before the NPL Ratio

The intent of the question was to rank, in practical terms, which indicators are most likely to become the first deterioration signals for local-government-related infrastructure and new policy-based financial instruments: Stage 2 exposures, delinquencies, loan modifications and repayment deferrals, accrued interest, lower provision ratios, or lower capital adequacy.

The response identified Stage 2 balances and the Stage 2 ratio as the most important leading indicators, followed by loan modifications, repayment deferrals and tenor extensions, accrued interest and capitalised interest, and delinquencies of more than 30 days but less than 90 days. The Stage 2 provision ratio, loan-loss reserve ratio, credit costs, and growth in risk-weighted assets were framed as confirming indicators, while the NPL ratio, capital adequacy ratio, and declining net profit were classified as lagging indicators.

The follow-up examined situations in which Stage 2 exposures increase while NPLs remain flat, loans are still treated as performing after modification, repayment deferrals or principal grace periods are granted, accrued interest or capitalised interest rises, and migration from Stage 2 to Stage 3 occurs. For local-government-related infrastructure, even if tariff revenue or land-related revenue falls below assumptions, the first manifestation may be a reset of the repayment schedule rather than an NPL.

The credit analytical implication is that CDB's downside scenario should not begin with an "increase in NPLs", but with a path in which internal credit risk and cash-collection quality deteriorate while the NPL ratio remains low. This can be seen as a more practical decomposition of the Stage 2 monitoring issue already identified in the existing report.

3.3 Linkage with China's Sovereign Creditworthiness and Local-Debt Resolution

The intent of the question was to determine which factors are most likely to affect CDB's downgrade risk and spread widening: the deterioration of CDB's own policy assets, the Chinese government's fiscal capacity, the local-debt resolution framework, the stance on capital injections and liquidity support for policy banks, or the sovereign rating outlook.

The response stated that China's sovereign rating and outlook are the most direct factors for downgrade risk, while the earliest drivers of spread widening are likely to be market assessments of the Chinese government's fiscal capacity and concern that CDB may absorb the burden of local-debt resolution. Capital injections and liquidity support for policy banks are normally credit-enhancing, but if the required support becomes large and repeated, they may be viewed as a fiscal burden and as evidence of policy banks acting as loss-absorbing vehicles.

The follow-up emphasised the spread differential between CDB bonds and Chinese sovereign bonds as a market signal to monitor before waiting for a rating change. This is because the pricing may show that the market is beginning to factor in CDB-specific policy burdens, even while support expectations remain intact. Background factors cited included progress in resolving hidden local-government debt, the scale of capital support for policy banks and state-owned banks, the broader fiscal deficit, and the general-government debt-to-GDP ratio.

The credit analytical implication is that CDB should not be treated simplistically as the sovereign itself. Sovereign creditworthiness, local fiscal risk, the transfer of burdens to policy banks, and the explicitness of support need to be assessed separately. If local-debt resolution takes the form of low-rate refinancing, repayment deferrals, and expanded policy funding by CDB, CDB may be re-evaluated by the market both as a safe government-related issuer and as a transmission channel for sovereign and local-fiscal risk.

3.4 Reliance on Bond-Market Funding and the Quality of Funding

The intent of the question was to assess under what market conditions liquidity and refinancing risk may start to be recognised for CDB, given that it relies heavily on bond-market funding rather than deposits, and in particular how funding costs affect margins, capital headroom, and market access.

The key point of the response was that CDB's liquidity risk is less likely to appear as a near-term inability to issue, and more likely to emerge through a collision between low-rate, long-term policy lending and higher funding costs, leading to margin compression, a reduction in earnings buffers, and slower capital accumulation. In particular, if CDB bond spreads over Chinese sovereign bonds widen on a sustained basis, it could be a sign that the market is starting to price in the increased burden of CDB's role as a policy bank, rather than simply reflecting market volatility.

The follow-up examined the definition of a state in which "market access is maintained, but the quality of funding is deteriorating". Specifically, a warning line was identified when a sustained widening of spreads versus government bonds, wider new-issue premiums, lower demand for long-dated bonds, shorter issuance tenors, lower bid-to-cover ratios, and a wider pricing gap between domestic financial bonds and offshore senior bonds occur together. If these developments coincide with lower net interest margin, higher Stage 2 exposures, and expanding policy lending, they should be treated as CDB-specific credit and policy-burden risks, rather than merely as rising interest rates.

The credit analytical implication is that CDB's funding risk needs to be assessed not by asking whether it can issue, but by examining at which tenors, at what premium, from which investor base, and how much additional yield is required relative to the sovereign.

3.5 Expansion of Policy Assets and Capital Maintenance Policy

The intent of the question was to understand how CDB intends to reconcile the expansion of its policy mandate with capital maintenance, including its policies on asset growth, capital injections, issuance restraint, loan selection, risk transfer, and loss sharing with the government.

The response stated that CDB is not an issuer that prioritises profitability or shareholder returns like a commercial bank and restrains asset growth accordingly; policy objectives are likely to take precedence. Therefore, the credit focus is not whether CDB will grow, but whether capital injections, government guarantees, risk-sharing arrangements, co-financing, and loan-selection mechanisms are sufficiently in place to accompany the expansion of policy assets.

The follow-up organised the difference between situations in which capital injections or government support should be read as "credit enhancement" and situations in which they should be read as "crystallisation of policy asset risk". If support is preventive and institutional, is provided before policy expansion or before clear losses emerge, and is sufficient relative to asset growth and growth in risk-weighted assets, it should be read as credit enhancement. In contrast, support provided after increases in Stage 2 exposures, loan modifications, accrued interest, overseas policy-finance losses, higher credit costs, and lower capital ratios is both a confirmation of the government's support capacity and a confirmation that past policy asset risk has consumed capital.

The credit analytical implication is that the occurrence of government support itself should not always be treated as a positive. The timing, purpose, scale, target, concurrent indicators, and market reaction need to be assessed separately. If multiple policy banks and state-owned banks receive support at the same time, the market may price it not as CDB-specific credit enhancement but as an increase in the support cost for China's quasi-sovereign financial sector as a whole.

3.6 Transparency Risk in Overseas Policy Finance and Belt and Road Exposures

The intent of the question was to assess how overseas policy-finance and Belt and Road-related exposures may affect spreads as a transparency concern or tail risk, even if they are not large enough to move the issuer's overall rating immediately.

The response stated that in the base case, overseas policy finance may not be the main driver of CDB's overall credit metrics, but if delinquencies, debt restructurings, or significant provisions for specific countries surface additionally, they could increase doubts about transparency across CDB's policy assets. In the discussion, the Venezuela-related delinquency and large-provision case seen in the 2025 disclosure of CDB's Hong Kong Branch was cited as an example, but this is information for the Hong Kong Branch alone and was not asserted to be representative of CDB group's overall overseas policy-finance portfolio.

The follow-up organised criteria for distinguishing a one-off case from a deterioration in overall portfolio quality. If the issue is limited to a single country or a small number of cases, provisions are sufficient, there is no spread of delinquencies or loan modifications to other countries, and Stage 2 exposures and margins in domestic policy assets are stable, it should initially be treated as a one-off loss. Conversely, if delinquencies, loan modifications, and high provisions are confirmed across multiple countries, debt restructurings are prolonged, collateral values are low, disclosure by country, transaction, and Stage 2 classification is insufficient, and Stage 2 exposures and margin decline are also progressing in domestic policy assets, it should be treated as a factor that widens the risk premium on the overall overseas policy-finance portfolio.

The credit analytical implication is that more emphasis should be placed not on the amount of overseas cases in isolation, but on whether they appear at the same time as deterioration in domestic policy assets, and whether limited disclosure worsens market sentiment.

4. Relationship with the Existing Report

The issues already confirmed in the existing issuer summary are that CDB has high support-driven credit strength as one of China's policy banks, supported by government-related shareholders, policy indispensability, its position in the domestic financial bond market, a low NPL ratio, provisions, and capital adequacy. The existing report also already notes that the likelihood of government support and explicit guarantees for individual bonds are separate matters, and that CDB parent senior bonds, domestic financial bonds, offshore bonds, subordinated bonds, Tier 2 instruments, and subsidiary bonds should be checked separately.

The argument developed in this discussion expands the existing report's "constraints created by policy mandates" into a more detailed monitoring framework. In other words, rather than taking comfort from a low NPL ratio as a given, the proposal is to treat Stage 2 exposures, loan modifications, accrued interest, provisions, funding conditions, the nature of capital support, and overseas country risk as leading indicators.

Many items remain unverified. In particular, the policy-area breakdown of Stage 2 balances, repayment-source segmentation for local-government-related infrastructure, loss sharing for new policy-based financial instruments, disclosure on loan modifications, repayment deferrals, and accrued interest, the country, currency, and Stage 2 breakdown of overseas policy finance, and the time series of relative spreads between CDB bonds and Chinese sovereign bonds cannot be treated as verified facts in this discussion.

5. Monitoring / Next Check

In the next and subsequent reviews, the first point to monitor should be the divergence between policy assets migrating to Stage 2 and the NPL ratio. Even if the NPL ratio remains low and flat, increases in Stage 2 balances and ratios, loan modifications, repayment deferrals, accrued interest, capitalised interest, and short-term delinquencies could be warning signals before policy asset risk appears in earnings.

Continued monitoring is also required of local-government debt resolution and the transfer of policy burdens to CDB. If local-debt resolution is skewed towards tenor extensions, refinancing, repayment deferrals, and policy funding provision rather than substantive debt reduction, CDB may be re-evaluated by the market as a transmission channel for China's sovereign and local-fiscal risk.

On funding, the focus should be the quality of funding rather than issuance availability. CDB bond spreads versus government bonds, new-issue premiums, bid-to-cover ratios, demand for long-dated bonds, issuance tenor, and the pricing gap between domestic financial bonds and offshore senior bonds need to be assessed together with net interest margin, Stage 2 exposures, and expansion of policy lending.

For capital maintenance, the mechanisms for capital injections, government guarantees, co-financing, risk sharing, and loan selection in response to the expansion of policy mandates should be checked. Capital support should be evaluated by distinguishing whether it is a preventive measure before policy expansion or gap-filling after deterioration in existing assets.

For overseas policy finance, more emphasis should be placed on delinquencies, loan modifications, high provisions across multiple countries, prolonged debt restructuring, declining collateral values, and insufficient country-level disclosure than on a loss in a single country. If overseas cases surface at the same time as increases in Stage 2 exposures and margin deterioration in domestic policy assets, market concern over the transparency of CDB's overall policy assets is likely to intensify.

The transfer candidates for the "Follow-up on Management Strategy, Investment Plan, and Financial Policy" section of issuer_notes.md should reasonably be limited to the following five points. All of them are follow-up candidates from this discussion, and confirmation at the next update is required before any actual transfer to permanent notes.

• CDB's policy asset risk is difficult to assess based only on its low NPL ratio, and Stage 2 increases, loan modifications, and accrued interest in local-government-related infrastructure and new policy-based financial instruments need to be monitored continuously.
• If local-government debt resolution is prolonged and CDB becomes a vehicle for low-rate refinancing, repayment deferrals, and policy funding provision, CDB may be re-evaluated as a transmission channel for China's sovereign and local-fiscal risk.
• CDB's liquidity risk should be monitored by the quality of issuance conditions rather than issuance availability, and deterioration in spreads versus government bonds, new-issue premiums, demand for long-dated bonds, and issuance tenor should be checked continuously.
• At CDB, expansion of the policy mandate is likely to take priority, and capital injections or government support need to be evaluated by distinguishing whether they are preventive capital reinforcement or gap-filling after increases in Stage 2 exposures and the crystallisation of losses.
• In overseas policy finance, delinquencies, loan modifications, high provisions across multiple countries, and insufficient country-level disclosure may raise concerns about transparency across CDB's overall policy assets more than a single-country loss.

6. Unverified / Pending Items

The following items remain unverified in this discussion and should not be treated as verified facts.

• Policy-area breakdown of Stage 2 balances. It remains unconfirmed whether Stage 2 exposures are concentrated in local-government-related infrastructure, new policy-based financial instruments, the Five Priority Areas, or overseas policy finance.
• Repayment-source breakdown of local-government-related infrastructure. The degree of dependence on tariff revenue, government subsidies, land-related revenue, and local fiscal transfers remains unconfirmed.
• The equity-like or debt-like nature of new policy-based financial instruments, the credit risk borne by CDB parent, and loss sharing with the government, local governments, and private capital remain unconfirmed.
• Balance trends in loan modifications, repayment deferrals, principal grace periods, tenor extensions, restructured loans, accrued interest, and capitalised interest remain unconfirmed.
• The time series of relative spreads among CDB bonds, Chinese sovereign bonds, domestic policy-bank bonds, state-owned commercial bank bonds, and offshore senior bonds remains unconfirmed.
• The institutional design for capital injections, government guarantees, co-financing, risk sharing, and loan selection in response to the expansion of policy mandates remains unconfirmed.
• The largest country-level exposures in overseas policy finance, Stage 2 and Stage 3 classification, delinquencies, loan modifications, repayment deferrals, provision ratios, collateral values, and debt-restructuring status remain unconfirmed.
• It remains unconfirmed whether Venezuela-type delinquency and large-provision cases are exceptional one-off cases within the CDB group as a whole or part of broader overseas policy-finance risk.

7. Reference Context

The existing context referenced consists of China Development Bank's issuer summary dated 2026-05-20, issuer notes / knowledge snapshot / source registry dated 2026-05-20, and the discussion generated on 2026-06-01.

This report has not conducted new web research or primary-source verification. The Venezuela-related disclosure by CDB's Hong Kong Branch and external references such as Reuters mentioned in the discussion are treated as reference information within the discussion, and are not adopted here as newly verified facts for the CDB group as a whole.