Europe Investment Report 2009 — When the Numbers Turned Out to Be Estimates

In late 2009 a euro-area member state substantially revised its reported deficit upward. The arithmetic change was large. The informational change was larger.

Why a revision is worse than a bad number:

A bad number that is accurate is priceable. Investors can assess a sovereign with a large deficit — it is a known quantity, and the analysis is routine.

A revision means the previous number was wrong, which raises a different and much more damaging question: what else is wrong, and how would anyone know?

• Every other figure from that source becomes an estimate rather than a fact.
• The historical series becomes unreliable, so trend analysis loses its foundation.
• And the question generalises. If one member state's figures were wrong, what is the verification process elsewhere?

The answer, in 2009, was uncomfortable. Sovereign fiscal statistics are compiled by national statistical agencies — part of the state being measured. The union's statistical office aggregated and reviewed submissions but had limited powers to audit, no right to inspect underlying records, and modest resources.

This is not a scandal so much as a design assumption. The system was built on the presumption that member states report honestly, which is reasonable for routine statistics and inadequate when the numbers determine market access.

The second finding concerns the rules. The union had fiscal rules with numerical limits and a procedure for breaches. Those rules had been breached by large member states earlier in the decade without meaningful consequence, which established that the procedure was political rather than automatic. A rule that has been ignored once is not a constraint — it is a statement of preference.

And the third is that markets reassessed a category rather than individual sovereigns. Spreads widened across a group of member states with quite different fiscal positions, because the reassessment was of a class — peripheral euro-area sovereigns — rather than of each balance sheet.

Who produces a sovereign's numbers

The institutional question deserves direct treatment, because most investors assume a verification apparatus that does not exist in the form imagined.

How corporate financial data is verified:

• Prepared by the company under a defined accounting standard.
• Audited by an independent firm with a legal duty and liability exposure.
• Regulated by a securities authority with enforcement powers.
• And subject to legal consequences for material misstatement.

How sovereign fiscal data was verified in 2009:

• Prepared by a national statistical agency, part of the state.
• Submitted to a union-level statistical office which reviewed for consistency and methodology.
• That office had limited powers to audit, no right of inspection, and could not compel access to underlying records.
• And there were no meaningful consequences for misreporting.

The structural difficulty is genuine rather than a matter of negligence:

• A sovereign cannot easily be audited by an outside party without a sovereignty question arising.
• The accounting judgments are enormous. What counts as government? Are public enterprises included? How are pension obligations, guarantees and public-private partnerships treated? Each has a defensible answer and a range of them.
• And the incentive runs one way. Every judgment that reduces the reported deficit is attractive, and each is individually defensible.

A sovereign's fiscal statistics are self-reported by an entity with a strong interest in the result and no external auditor with the power to check. Investors treated them as facts because there was no alternative, not because the verification existed.

The reforms that followed gave the union's statistical office enhanced powers — the right to inspect, to require documentation, and to visit. This is a real improvement and it took the crisis to produce it.

The durable investor lesson is to ask, of any data underpinning a position, who produced it, what incentive they had, and who verified it. For sovereign data the answers are: the government, a strong one, and until 2010 essentially nobody.

A rule that has been broken is not a rule

The union's fiscal framework failed before it was tested here, and the failure is worth understanding because it explains why the numbers mattered so much.

The framework, as designed: numerical limits on deficits and debt, with an excessive deficit procedure for breaches, escalating to financial sanctions.

What happened in practice earlier in the decade: two large member states breached the deficit limit, the procedure was initiated, and it was then suspended and the framework subsequently softened. No sanction was applied.

What this established:

• The procedure was political, decided by member states judging each other, with an obvious reluctance to sanction peers.
• Enforcement was weakest against the largest members, which is precisely backwards from a systemic risk perspective.
• And the limits became targets to be reported rather than constraints to be met, which shifts the pressure from fiscal management to fiscal accounting.

The last point is the direct connection to the statistics problem:

When a rule has a numerical threshold, weak enforcement and self-reported data, the pressure moves from meeting the rule to reporting compliance with it. The incentive is to manage the number, not the deficit.

This is a general property of any threshold-based framework with self-reported measurement, and it recurs throughout this archive — in the earnings definitions of the Private Credit Report 2013 and in the leverage measurement of the Private Equity Report 2008.

The eventual response was a strengthened framework with more automatic procedures, independent national fiscal councils, and enhanced statistical powers. Whether the enforcement is genuinely more automatic remains contested, since the political dynamic that softened it originally has not changed.

Contagion by category

The spread of the crisis across member states followed a pattern that is important to understand, because it is how markets actually behave under uncertainty.

What fundamentals-based contagion would look like: each sovereign assessed individually, with spreads reflecting its own debt, deficit, growth and financing needs.

What happened instead: spreads widened across a group of member states with substantially different fiscal positions, at similar times, in similar directions.

Why markets reassess a category:

• Verification is expensive. If one sovereign's numbers were wrong, assessing each of the others individually requires work nobody can do quickly.
• The reassessment is about the framework, not the borrower. What changed was confidence in the reporting system and in the union's willingness to support members — and those apply to the whole category.
• Positioning is categorical. Investors held these bonds through allocations to "euro-area periphery," so reducing exposure meant reducing the category.
• And the safest response to uncertainty about which member of a group is affected is to reduce exposure to all of them.

This produces a specific and exploitable pattern:

Categorical repricing moves the strong and the weak together. It is right about the category and wrong about the individuals within it, which is where the dispersion opportunity sits — for anyone willing to do the verification work the market has just decided is too expensive.

The qualification that matters: the categorical reassessment was partly correct, because the shared feature was real. Every member of the category faced the same structural problem — borrowing in a currency it did not control, with no lender of last resort, per the Global Investment Outlook 2011. That was a genuine common exposure, not a market error.

So the correct reading is that markets were right about the category-level risk and imprecise about individual positions within it — which is a much better characterisation than either "contagion is irrational" or "the market correctly priced each sovereign."

The assumption that had never been priced

A structural point that reframes the entire episode: the crisis was not a departure from correct pricing. It was the beginning of a correction to a decade of incorrect pricing.

Before 2008, spreads between euro-area sovereigns had compressed almost to nothing. Bonds issued by economies with very different debt levels, growth rates, competitiveness and institutional quality traded at nearly identical yields.

What that compression implied:

• That the credit risk of these sovereigns was essentially identical, which was plainly not true on any fundamental measure.
• Or that they were effectively mutualised — that a member in difficulty would be supported by the others.
• The second interpretation is the only coherent one, and it had no legal basis whatsoever. The founding framework explicitly excluded mutual liability for member state debts.

So the market had been pricing a guarantee that did not exist, for roughly a decade.

Why this went unexamined:

• The convergence had been continuous since before the currency launched, so it looked like an established fact rather than a hypothesis.
• Regulatory treatment reinforced it. Bank capital rules assigned euro-area sovereign debt a zero risk weight regardless of issuer, so holding it required no capital — which is a regulatory statement that these assets are riskless.
• And nothing had tested it, so no evidence accumulated against it.

A price of zero for a risk is a claim about that risk, not the absence of it. Euro-area sovereign spreads had been asserting for a decade that a mutualisation existed which the treaties explicitly prohibited, and no one had to defend the assertion because it was never questioned.

The zero risk weight point deserves emphasis because it persisted long after 2009 and shaped the doom loop directly: a capital framework that treats domestic sovereign debt as riskless encourages banks to hold it in size, which is exactly the concentration the Global Investment Outlook 2010 describes.

The Europe Investment Report 2016 develops this as a portfolio problem, and it remains only partially addressed.

Ratings as an amplifier

An underappreciated feature of the crisis's mechanics was that sovereign rating changes did not merely reflect deterioration — they triggered forced actions that caused more of it.

Why a rating is more than an opinion in practice:

Investment mandates reference ratings. A great many institutional mandates require holdings to be investment grade, so a downgrade below that threshold forces sales regardless of the manager's own view. The seller is not expressing a judgment; it is complying with a document.

Collateral frameworks reference ratings. Central bank eligibility depended on minimum ratings, per the Europe Investment Report 2011. A downgrade could remove an asset's fundability, which withdraws liquidity from every institution holding it.

Bank capital requirements reference ratings, so a downgrade raises the capital a bank must hold against an exposure — encouraging it to sell precisely when prices are falling.

And index membership references ratings, so a downgrade removes a bond from benchmarks that passive and benchmarked investors track.

The cliff effect is what makes this dangerous:

• A gradual deterioration produces no forced action until a threshold is crossed.
• Crossing it produces simultaneous forced selling from every constrained holder at once.
• The selling depresses the price, which worsens the fundamentals it was responding to.
• And the sellers are price-insensitive, because they are complying rather than transacting.

A rating threshold converts a continuous variable into a discontinuous one. The deterioration is gradual and the consequence arrives all at once, from holders who have no discretion about whether to sell or at what price.

The two-sided problem this created for the agencies: downgrading promptly triggered the cliff and accelerated the crisis; downgrading slowly meant ratings lagged reality and gave false comfort. Neither option is good, and the criticism they received for both was in a sense deserved simultaneously.

The structural response was to reduce mechanistic reliance on ratings in regulation and mandates — replacing automatic triggers with requirements for independent assessment. This was agreed and implemented only partially, and rating-referenced triggers remain widespread.

The practical point for an allocator is to know where the thresholds sit in a portfolio's own documents. The forced-seller behaviour is predictable in advance, which makes the cliff a knowable event rather than a surprise.

What an allocator could act on

Ask who produced a number and who verified it. Sovereign fiscal data is self-reported by the entity being assessed, and before 2010 the union-level office had no power to audit it.

Treat a revision as an information event, not an arithmetic one. It converts every other figure from that source into an estimate and invalidates the historical series that trend analysis depends on.

Read an unenforced rule as a preference. A threshold that has been breached without consequence shifts pressure from meeting the rule to reporting compliance with it.

Expect categorical repricing and look for dispersion within it. Markets reassess groups when individual verification is expensive, which moves strong and weak together and creates the gap worth working on.

Check whether the shared feature is genuine before dismissing contagion as irrational. Here it was — every member of the category borrowed in a currency it did not control, which was a real common exposure.

Interrogate any risk priced at zero. A zero spread and a zero risk weight are both claims, and the euro-area sovereign case shows how long an unexamined claim can persist and how much can be built on it.

What 2009 established

• Sovereign statistics are self-reported by the entity being measured, with verification powers that were inadequate before the crisis.
• A revision damages credibility beyond its arithmetic, converting facts into estimates across the whole series.
• Unenforced fiscal rules shifted pressure to fiscal accounting rather than fiscal management.
• Contagion spread by category, correctly identifying a genuine common exposure and imprecisely pricing individuals within it.
• A decade of near-zero spreads was the original mispricing, reinforced by a regulatory framework that treated the debt as riskless.