Asia-Pacific Investment Report 2010 — Capital You Did Not Ask For

The problem in 2010 was an unfamiliar one: too much capital arriving, too fast, for reasons unconnected to the receiving economies.

The source, per the Global Investment Outlook 2010, was policy elsewhere. Zero rates and asset purchases in the developed world pushed investors toward higher-yielding assets, and the region's economies — growing quickly, offering higher rates, with strengthening currencies — were exactly what that search produced.

What arrived was not investment the region had sought. It was displaced capital, and it created four simultaneous problems:

• Currency appreciation, damaging the export competitiveness that the Asia-Pacific Investment Report 2008 shows the growth model depended on.
• Asset price inflation, particularly in property, driven by the inflow rather than by domestic income growth.
• Credit growth, as cheap foreign funding enabled domestic lending expansion.
• And inflation, both from the credit growth and from commodity prices the region's own stimulus was driving.

The standard instrument could not address this. Raising policy rates to control inflation and credit widens the rate differential and attracts more capital, worsening the appreciation. Cutting rates to deter inflows fuels the credit growth. One instrument, two problems moving in opposite directions.

The response was to add instruments, and this is the period's lasting contribution:

• Macroprudential tools acting directly on lending terms — how much a borrower can borrow against a property, what share of income can service debt, how much capital a bank must hold against particular exposures.
• Capital flow management — taxes on inflows, minimum holding periods, limits on foreign purchases of particular assets.
• And continued reserve accumulation, resisting appreciation at a cost examined below.

These worked well enough that the toolkit was adopted globally, including by the developed economies whose policy had created the problem. The Global Investment Outlook 2016 describes them as standard practice by then.

One instrument, two problems

The bind deserves precise statement because it is the reason new instruments were necessary rather than merely convenient.

The policy rate does two things simultaneously:

• Domestically, it sets the cost of borrowing, so raising it slows credit and demand.
• Externally, it sets the return on holding the currency, so raising it attracts capital and strengthens the exchange rate.

In normal conditions these point the same way. An overheating economy needs slower credit and can tolerate a stronger currency, so raising rates addresses both.

In 2010 they pointed in opposite directions:

• Credit growth and inflation called for higher rates.
• Currency appreciation and capital inflows called for lower ones.
• And the inflows were themselves a cause of the credit growth, so the two problems were linked through the very channel the instrument would worsen.

Raising rates to cool credit brought in more foreign capital, which funded more credit. The instrument fed the problem it was aimed at.

This is the policy trilemma the Asia-Pacific Investment Report 2016 sets out formally: with free capital movement, an economy cannot simultaneously control its exchange rate and set monetary policy for domestic conditions. 2010 is when the region encountered it as a practical daily constraint rather than a textbook proposition.

The available resolutions were all imperfect:

• Accept appreciation, and lose export competitiveness in an economy structured around exports.
• Accumulate reserves to resist it, at the cost described below.
• Restrict the flows, which was heterodox at the time.
• Or add a domestic instrument that does not affect the exchange rate — which is what macroprudential policy is.

Acting on the borrower, not the price

Macroprudential policy is frequently described as a softer version of interest rate policy. It is structurally different, and the difference is why it resolved the bind.

Interest rate policy acts on the price of money, economy-wide and indiscriminately. It affects every borrower, every sector and the exchange rate simultaneously.

Macroprudential policy acts on the terms of lending, and can be targeted:

• Loan-to-value limits cap how much can be borrowed against an asset's value, requiring a larger deposit.
• Debt-service-to-income limits cap borrowing relative to the borrower's income.
• Sectoral capital requirements make it more expensive for banks to lend against particular assets.
• Differential treatment by borrower type — a higher deposit for a second property than a first, or for a non-resident buyer than a resident.

Why this escapes the trilemma: these tools do not change the return on holding the currency. A cap on how much a household can borrow against a flat has no effect on the rate differential, so it slows credit without attracting capital.

Their other structural advantages:

• They are targeted. A property credit problem can be addressed without slowing manufacturing investment.
• They act on quantity rather than price, which works even when rates are already low — precisely the condition in which rate policy is weakest.
• And they are adjustable and reversible at higher frequency than rate policy, since they are regulatory rather than monetary.

The genuine limitations are worth stating:

• Leakage. Restrict bank lending and credit migrates to non-bank lenders. The constraint-produces-workaround pattern from the Asia-Pacific Investment Report 2009 applies directly.
• They are distributional. A loan-to-value cap bears hardest on first-time buyers without accumulated wealth — the buyers least responsible for the problem.
• And calibration is difficult, since the relationship between a specific limit and the eventual credit outcome is uncertain and varies by market.

Nonetheless the evidence over the following decade was favourable enough that the tools became standard, and the region's early adoption is the reason the developed world had a tested toolkit to reach for later.

Why property is always the transmission point

Property appears in nearly every credit episode in this archive, and the reason is structural rather than cultural.

Property is the dominant form of collateral. Banks lend against it because it is durable, immovable, registrable and hard to conceal. This makes property credit the largest single category of lending in most economies.

The reflexive loop this creates:

1. Credit becomes cheaper or more available.
2. Buyers can bid more, so prices rise.
3. Higher prices mean more collateral value, so more credit can be extended against the same properties.
4. Which allows higher bids. Return to step 2.

The loop runs in reverse just as readily, which is what makes property downturns so damaging: falling prices reduce collateral, which reduces credit, which reduces bids, which reduces prices.

Three features specific to inflow-driven episodes:

• The buying is not income-constrained. Domestic property prices normally bear some relationship to domestic incomes. Foreign capital does not have to satisfy that constraint, so prices can detach from local affordability entirely.
• Supply responds slowly. New construction takes years, so an inflow-driven demand increase produces price rises rather than volume increases for a long time.
• And the political consequences are severe, because housing affordability is directly experienced by voters. This is why targeted measures on non-resident purchases appeared quickly, ahead of broader tools.

Property is where credit conditions become a price, and where a foreign capital flow becomes a domestic political problem. That is why it is the sector every macroprudential toolkit is built around.

The Asia-Pacific Investment Report 2017 and Singapore Investment Report 2020 cover the eventual results of these measures, which were mixed and are still contested.

An orthodoxy that actually changed

The shift in professional opinion on capital controls over this period is genuinely unusual and worth documenting.

The prior orthodoxy, held firmly through the 1990s: free capital movement is efficient, allocating savings to their most productive use globally. Restrictions are distortionary, invite avoidance and corruption, and signal a weak policy framework. Multilateral institutions actively discouraged them.

What changed the view:

• The 1997 Asian crisis had suggested that short-term portfolio flows could be destabilising in ways that direct investment was not — a distinction that treats "capital" as several different things rather than one.
• Economies that had used controls came through subsequent episodes better than the theory predicted, which is evidence rather than argument.
• The 2008 crisis undermined the assumption that developed financial markets allocate capital efficiently, which was the theoretical basis for the prior view.
• And the 2010 inflow problem had no orthodox solution. The instruments the theory permitted did not work, which is a strong practical argument.

The revised position, which multilateral institutions formally adopted in this period: capital flow management measures are a legitimate part of the toolkit under specified conditions — as a complement to sound macroeconomic policy rather than a substitute, targeted at flow types that create genuine risks, and preferably temporary.

Why the distinction between flow types matters so much:

• Direct investment builds a factory. It cannot leave quickly, and it brings technology and market access.
• Portfolio flows buy securities. They can leave in days, and they bring only capital.
• Bank flows, particularly short-term foreign currency borrowing, can leave fastest and carry a currency mismatch — the mechanism at the centre of the 1997 crisis.

The measures adopted mostly targeted the second and third, leaving direct investment untouched. This is a targeted intervention rather than a wall, which is what made it defensible.

The Global Investment Outlook 2013 shows what happened when the flows reversed, which is the strongest retrospective argument for having managed them on the way in.

The unstated cost of resisting appreciation

Reserve accumulation to resist currency appreciation has a running cost that is rarely stated plainly, and it makes the policy a fiscal decision as much as a monetary one.

The mechanism: to prevent the currency rising, the central bank buys foreign currency and sells domestic currency. This increases the domestic money supply, which would be inflationary — so the central bank sterilises by selling domestic bonds to absorb the money it just created.

The resulting position:

• The central bank holds foreign assets — typically developed-market government bonds — earning a low yield.
• It has issued domestic liabilities — the sterilisation bonds — paying a higher domestic yield.
• The difference is a running loss, borne every year the position is held.

The loss is real and can be large. An economy holding a substantial share of output in reserves, with a several-percentage-point yield gap, carries an annual cost measured in a meaningful fraction of a percent of GDP.

Two additional exposures:

• A currency loss if the domestic currency eventually appreciates anyway, which reduces the domestic-currency value of the foreign assets. The policy's failure and its cost arrive together.
• And sterilisation has limits. Beyond a certain scale the domestic bond market cannot absorb the issuance without distorting domestic yields, which reintroduces the problem the policy was avoiding.

Reserve accumulation is usually described as prudent and free. It is prudent and it carries an annual fiscal cost, plus a currency exposure that crystallises precisely if the policy does not work.

None of this makes it wrong. The Asia-Pacific Investment Report 2008 shows the insurance value is real and was collected. But it is a purchase, not a windfall, and the price should be in the analysis.

What an allocator could act on

Check whether an economy's rate policy faces conflicting objectives. When inflows drive credit, the rate cannot address both, and the resulting policy will look inconsistent because the constraint is real.

Read macroprudential measures as targeted quantity controls, not weak rate policy. They act on the borrower rather than the price of money, which is why they slow credit without attracting capital.

Expect leakage to non-bank lenders. Restricting bank lending moves credit rather than removing it, so the aggregate exposure needs measuring outside the regulated system.

Distinguish flow types when assessing external vulnerability. Direct investment, portfolio flows and short-term bank borrowing have completely different reversal risks, and the aggregate capital account number conceals this entirely.

Treat property as the amplifier in any credit episode. It is the dominant collateral, so credit conditions become prices there first and the loop runs both ways.

Price the carrying cost of reserve accumulation. The yield gap between foreign assets and sterilisation liabilities is a real annual fiscal cost, and the currency exposure crystallises if the intervention fails.

What 2010 established

• The policy rate cannot serve domestic and external objectives when they conflict, which forced the development of additional instruments.
• Macroprudential policy acts on lending terms rather than money's price, escaping the trilemma constraint.
• Property is the structural transmission point for credit conditions, and inflow-driven buying detaches prices from local incomes.
• Capital flow management became orthodox, with the distinction between flow types doing the analytical work.
• Sterilised intervention carries a running fiscal cost plus a currency exposure that crystallises if it fails.