China Market Report 2014 — Recognising the Debt You Already Owed

2014 in China is best read as the year the bills from 2009 were itemised.

The Asia-Pacific Investment Report 2009 describes the stimulus mechanism: investment delivered through directed credit, with sub-national borrowing routed through separate corporate vehicles because local governments could not borrow directly. The debts were real, large, and of genuinely ambiguous status — not legally guaranteed, universally assumed to be.

The 2014 response was to count them and reclassify them. A national audit established the scale, and a programme followed to convert qualifying vehicle debt into explicit local government bonds.

Why this matters more than it sounds:

• The obligation already existed. Recognising it changed nothing economically.
• But it changed everything institutionally. An explicit bond has a named obligor, a published price, a maturity and a market. An implicit obligation has none of these.
• The interest cost fell, since explicit government debt is priced better than ambiguous corporate debt.
• And the ambiguity that made the implicit guarantee unpriceable — per the Asia-Pacific Investment Report 2013 — was removed for the portion converted.

This is the single clearest example in the archive of a government choosing to make an implicit guarantee explicit, and it is worth studying precisely because the usual pattern is the opposite.

The year's other developments follow from the same adjustment. The property cycle turned, and property in this economy is not one exposure but four. Growth slowed, and the authorities accepted it rather than responding with the credit expansion that had been the previous instrument. And a substantial equity rally began, financed heavily by margin lending — which made it a credit event as much as an equity one.

Why recognition is not the same as creation

The distinction is simple, frequently confused, and matters for reading any debt disclosure.

A debt exists when the obligation exists, not when it is recorded. The vehicle borrowings described in the Asia-Pacific Investment Report 2009 were real obligations from the moment they were incurred.

What recognition changes:

• Someone becomes responsible. An explicit government bond has a named obligor with a legal duty. An implicit obligation has an assumed one, and assumption is not enforceable.
• A price appears. Traded debt has a yield, which aggregates opinion about repayment. Unpriced debt provides no information at all.
• The maturity becomes known. Vehicle debt was frequently short-dated bank lending against long-dated assets — the maturity mismatch this archive documents in five other settings. Converting to long-dated bonds fixed that directly.
• And the interest cost falls, because ambiguity carries a premium.

What recognition does not change:

• The total obligation.
• Whether the underlying projects generate returns — many did not, per the demand-now-supply-later analysis in the 2009 report.
• Or the ultimate loss, which is determined by the assets rather than by the accounting.

Making a hidden liability explicit costs nothing economically and is politically expensive, which is why it is rare. The economics were unchanged; what changed was that someone now had to sign for it.

The reason it is usually avoided is that acknowledgement is read as an admission — the headline debt ratio rises, and commentary treats the increase as new borrowing rather than as improved disclosure. The willingness to absorb that reaction is what made the programme notable.

The maturity extension was arguably the larger benefit. Replacing short-dated bank debt against decades-lived infrastructure with long-dated bonds removed a rollover risk that had no reason to exist, and rollover risk is what actually kills institutions.

Property is four exposures

The property slowdown's significance comes from how many places it transmits to, and the list is worth setting out because each channel is usually analysed separately.

Household wealth. Property was the dominant household savings vehicle, per the Asia-Pacific Investment Report 2010 — a rational response to capped deposit rates and limited alternatives. A property decline is therefore a decline in household net worth, which affects consumption directly.

Local government revenue. Land sales were a principal source of local government income, and land was the collateral behind the vehicle borrowing. A property slowdown reduces both revenue and collateral value simultaneously — the concentration failure the archive documents repeatedly, here in its purest form.

Construction employment and upstream demand. Construction is enormously labour-intensive and commodity-intensive. A slowdown transmits immediately to employment, steel, cement and glass — and, through those, to the commodity exporters covered in the Australia Investment Report 2010 and Brazil Investment Report 2009.

Bank collateral. Property secures a large share of lending, directly and indirectly. Falling values reduce collateral coverage across the loan book, which is the reflexive loop the Asia-Pacific Investment Report 2010 describes.

Why the fourfold transmission is the point:

A property slowdown is not a sector event. It arrives in household consumption, local government finances, industrial employment and bank capital at the same time, from the same cause. Four line items, one exposure.

This is why the policy response was measured rather than aggressive. Reflating property would have required more credit, which was the thing being restrained. The authorities accepted the slowdown as the cost of the adjustment, which is a coherent trade rather than a failure to respond.

Choosing a lower number

The acceptance of slower growth was a deliberate policy position, and its logic deserves stating because it was frequently reported as weakness.

The available options:

Maintain the previous growth rate. This was achievable — through more credit-financed investment, the instrument that had worked in 2009.

Its costs: a higher credit-to-GDP ratio, more capacity in already-oversupplied sectors, and more of the debt the audit had just quantified. Each additional unit of investment was producing less output, per the capital-output analysis in the Asia-Pacific Investment Report 2012.

Or accept a lower rate and use the space to restrain credit growth, recognise debt, and let the property market adjust.

The logic of the second is straightforward once the arithmetic is accepted: if growth is being purchased with credit that produces progressively less output, then the growth rate is not the variable to optimise. Continuing means a larger adjustment later.

Why this reads as weakness in market commentary:

• Growth rates are the headline statistic and a lower number is reported as a deterioration regardless of cause.
• The benefit — a smaller future adjustment — is counterfactual and never observed.
• And the distinction between "cannot grow faster" and "chooses not to" is invisible from outside, though it implies completely different things.

The diagnostic for distinguishing them: look at whether the constraint is being applied deliberately. An economy that is restraining credit, tightening property rules and permitting defaults is choosing; one that is easing in every direction and still slowing is constrained. These are observable in policy actions rather than in the growth rate.

A rally financed by borrowing

The equity market rally beginning in 2014 was substantially margin-financed, which changes its character entirely.

How margin lending works: an investor borrows against securities to buy more securities. The position is levered, and the lender requires collateral to remain above a threshold.

Why this makes a rally different in kind:

• Buying power is created by credit, so the rally's fuel is loan growth rather than savings reallocation.
• It is self-reinforcing upward. Rising prices increase collateral value, permitting more borrowing, funding more buying.
• And it is violently self-reinforcing downward. Falling prices reduce collateral, triggering margin calls, forcing sales, lowering prices further — and the forced seller has no discretion about timing.
• The unwind is faster than the build. Accumulating leverage takes months; a margin call resolves in days.

The observable indicator is the margin balance, published by the exchanges, and its ratio to market capitalisation. Rapid growth in that ratio is the clearest available warning, and it requires no view on valuation.

A margin-financed rally is a credit expansion that happens to be denominated in equities. Its mechanics are determined by collateral requirements, not by any opinion about the companies.

The connection to the wider adjustment is direct: credit restrained in property and lending vehicles found its way into securities, which is the leakage pattern the Asia-Pacific Investment Report 2010 identifies in macroprudential policy. Restricting credit in one channel relocates it rather than removing it, unless the aggregate is what is being controlled.

The Asia-Pacific Investment Report 2015 covers the unwind.

Opening by metering

The capital account changes of this period used an unusual design that is worth understanding as a model rather than a stage.

Conventional liberalisation removes restrictions, allowing capital to move freely. The result is the trilemma constraint the Asia-Pacific Investment Report 2016 describes — free movement forces a choice between exchange rate control and independent monetary policy.

The design used instead was a channel-and-quota model:

• Specific channels are opened, connecting defined markets through defined mechanisms.
• Aggregate quotas cap the total flow through each channel.
• Eligibility is defined, limiting which investors and which securities participate.
• And the quotas are adjustable, so the aperture can be widened or narrowed by administrative decision.

What this achieves:

• Foreign participation without full convertibility. International investors access the market; the currency does not become freely convertible.
• Flows are metered rather than free, so the trilemma is not fully triggered — capital mobility is partial by design.
• And the process is reversible, which conventional liberalisation is not. Reimposing removed controls carries an enormous credibility cost; narrowing a quota does not.

The costs are equally real:

• Complexity and friction, which deter some participants and raise transaction costs.
• Index inclusion difficulties, since global benchmarks require accessibility that quotas complicate.
• And an administrative rather than a market allocation of who gets access.

The honest assessment is that this is a coherent alternative model, not an incomplete version of liberalisation. It trades efficiency for control and optionality, which is a legitimate preference given the 1997 and 2013 experiences documented elsewhere in this archive.

Permitting a default on purpose

A development that received less attention than the debt audit was arguably more important for how the system prices risk: the first genuine corporate bond defaults were permitted rather than prevented.

The prior situation, per the Asia-Pacific Investment Report 2013: an implicit guarantee was assumed across a wide range of borrowers. Defaults had been avoided through arrangements nobody was obliged to make — a local government stepping in, a bank restructuring quietly, a state entity absorbing the obligation.

What universal prevention costs:

• Yields carry no information. If everything is repaid, spreads reflect liquidity and habit rather than credit quality.
• Capital is misallocated, because the price signal that would direct it toward better borrowers is switched off.
• The contingent liability accumulates on whoever is assumed to be standing behind it.
• And the eventual reckoning grows, since more debt is extended into the same distorted pricing.

Why permitting a default is genuinely difficult to calibrate:

• Too small and nobody updates. A minor obscure issuer failing changes no one's assumptions.
• Too large and it triggers the general run that the guarantee existed to prevent — exactly the equilibrium problem the Global Investment Outlook 2012 describes, run in reverse.
• And the demonstration must be repeated to establish a pattern, since a single instance is read as an exception.

What was observable afterwards is the useful part: spread dispersion. In a market with universal implicit guarantees, similar-rated bonds trade at similar yields regardless of the issuer's actual condition. As defaults were permitted, spreads began to differentiate — weaker issuers paid more, stronger ones less.

Spread dispersion is the measurable output of a functioning credit market. Its absence means the price is measuring the guarantee, not the borrower.

This is a practical diagnostic for any market where implicit support is suspected. Compare the yield range across issuers of similar rating and different quality. A narrow range means the market is pricing something other than credit — and the analysis in the Asia-Pacific Investment Report 2013 applies.

The process was gradual and deliberately so. Removing an implicit guarantee quickly would force a repricing of everything supported by it, which is a systemic event. Doing it slowly means living with distorted prices for years, which is the cost of the safer path.

What an allocator could act on

Distinguish debt recognition from debt creation. A rising headline ratio caused by an audit is improved disclosure, not new borrowing, and the maturity extension that usually accompanies it removes real rollover risk.

Count how many channels a property market transmits through. Household wealth, local government revenue, construction employment and bank collateral are four line items and one exposure.

Check whether slower growth is chosen or imposed. An economy restraining credit and permitting defaults is choosing; one easing everywhere and still slowing is constrained. The distinction is visible in policy actions, not in the growth rate.

Watch margin balances relative to market capitalisation. A levered rally's mechanics follow collateral requirements, the unwind is faster than the build, and the indicator is published and requires no valuation view.

Expect restrained credit to relocate. Controlling one channel moves the activity unless the aggregate is the target, which is why the equity leverage and the property restraint appeared together.

Read quota-based opening as a distinct model. Metered access preserves reversibility and partial monetary independence at the cost of efficiency, which is a trade rather than an unfinished reform.

What 2014 established

• Recognising an implicit obligation costs nothing economically and removes ambiguity, rollover risk and a pricing premium.
• Property transmits through four channels simultaneously, making it one concentrated exposure rather than a sector.
• Slower growth was a choice, since the faster rate was available only through the credit expansion being restrained.
• A margin-financed rally is a credit event, with mechanics set by collateral rules rather than fundamentals.
• Quota-based capital account opening is a distinct design trading efficiency for control and reversibility.