Europe Investment Report 2016 — The Integration Premium Reprices

European assets had been priced, for most of the preceding two decades, on an assumption that ran so deep it was rarely stated: integration proceeds in one direction. Members join, competences deepen, and the union's scope expands. Reversal was not modelled because it had not happened.

2016 ended that. A member state voted to leave, and the market had to price something it had no framework for — a union whose membership can decrease.

The immediate market consequences are covered in the UK Investment Report 2016. The European consequences were different in kind and longer-lasting, and they concerned what the union's remaining assets were worth.

Three effects compounded through the year.

Redenomination risk returned. Sovereign spreads within the euro area partly reflect the market's assessment that a member's debt will continue to be denominated in euros. If membership is reversible, that assessment becomes a probability rather than a certainty, and the spread widens to compensate. This had been a live issue during the earlier sovereign crisis and had substantially subsided; 2016 reintroduced it.

Negative rates broke the bank funding model. As the Europe Investment Report 2015 establishes, European monetary policy transmits through banks. A negative policy rate compresses the spread a bank earns, because it can rarely charge retail depositors a negative rate without losing them. Policy was compressing the margin of the intermediary it depended on — which is close to self-defeating.

Political risk became a developed-market category. The reclassification described in the Global Investment Outlook 2016 applied with particular force to Europe, where the entire investment case for several asset classes rested on institutional stability.

What "irreversible" was worth

The pricing of an assumption is worth setting out precisely, because it explains why a vote in one member state repriced assets across the whole union.

What integration provided, and what its assumed permanence was worth:

• A single currency with no exchange rate risk between members. A German investor holding Italian assets faced no currency conversion. That is a genuine and large benefit, and it depends entirely on both remaining in the currency.
• Legal harmonisation across an increasing range of commercial activity, which reduces the fragmentation cost described in the 2015 report.
• A backstop assumption. Markets had come to assume that the union would act to prevent a member's failure — an assumption reinforced during the sovereign crisis and priced into spreads.
• A one-way ratchet. Because reversal was not modelled, the option value of exiting was priced at zero for every member.

When reversibility is demonstrated, each of these repriced:

• Currency risk between members becomes non-zero. Not large, but no longer zero — and a small probability applied to a large redenomination is a material expected cost.
• Legal harmonisation becomes contingent on continued membership, which is a longer-dated risk but a real one for a company structuring across borders.
• The backstop becomes conditional on political will, which has been shown to be finite.
• The option to exit acquires value, which means every member's debt carries a component that was previously priced at nothing.

An assumption that has never been tested is priced at certainty. The first test does not have to be adverse to be expensive — it only has to demonstrate that the thing was an assumption.

This is the same mechanism the China Market Report 2015 describes for the August currency adjustment: a small move that reclassifies a constant as a policy choice matters more than a large move that confirms an existing view.

Negative rates and the intermediary problem

The Global Investment Outlook 2016 describes the zero lower bound proving soft. In Europe the consequence was specific and close to perverse.

The mechanism. A bank earns the spread between what it pays for funding and what it charges for lending. Its principal funding source is deposits.

The problem with a negative policy rate. A bank can be charged a negative rate on reserves held at the central bank. It generally cannot pass that to retail depositors, because a depositor charged for holding money will withdraw it — physical cash has a storage cost but the cost is low for a household.

So the deposit rate is floored at approximately zero while the lending rate falls with the policy rate. The spread compresses from both sides.

The consequences run directly against the policy's intent:

• Bank profitability falls, which reduces retained earnings, which is the primary way a bank builds capital.
• A bank with weaker capital lends less, not more.
• The lending that does occur skews toward lower-risk borrowers, because riskier lending requires more capital which the bank cannot generate.
• The businesses most in need of credit — smaller, younger, higher-risk — get less of it, which is precisely the segment where company formation happens.

The uncomfortable conclusion is that the policy's transmission mechanism and its cost fell on the same institution. Easing was intended to increase lending and operated by compressing the margin of the entity that does the lending.

This is not an argument that the policy was wrong — the counterfactual of not easing had its own costs, and sovereign borrowing costs fell substantially, which mattered. It is an argument that the transmission channel in a bank-based system has a limit that a market-based system does not have, and Europe was operating well inside that limit.

Redenomination risk and the incomplete union

The return of redenomination risk to sovereign spreads exposed an incompleteness that specialists had discussed for years and markets had largely stopped pricing.

What a complete monetary union requires, on the standard analysis:

• A single monetary authority. Present.
• A banking union — common supervision, common resolution, and common deposit insurance — so that a bank's health is not tied to its sovereign's health. Supervision and resolution existed by 2016; common deposit insurance did not.
• Some fiscal risk-sharing, so an asymmetric shock to one member can be absorbed. Largely absent in 2016.
• Labour mobility sufficient to adjust to asymmetric shocks. Present in law, limited in practice by language and qualification recognition.

The missing deposit insurance matters most and is worth explaining. Without it, deposits in a bank are guaranteed by that bank's national government. So the safety of a deposit depends on the fiscal capacity of the state where the bank is located.

That produces the doom loop: a weak sovereign makes its banks weaker, because their deposits are less credibly guaranteed and because they hold that sovereign's bonds. Weaker banks make the sovereign weaker, because they may require support. Each reinforces the other.

Common deposit insurance breaks the loop by removing the link between deposit safety and national fiscal capacity. Its absence meant the loop remained live, and 2016's reintroduction of exit risk made it relevant again.

For an investor the practical consequence is that European bank exposure and European sovereign exposure are not independent positions. A portfolio holding both is more concentrated than it appears, and the concentration is largest in exactly the member states where the individual holdings look cheapest.

What repricing political risk actually required

The reclassification of political risk from an emerging-market to a developed-market category, described in the Global Investment Outlook 2016, required specific changes to European analysis.

What the old framework assumed: policy in developed Europe changes gradually, through predictable processes, with outcomes bounded by institutional constraints. Political risk was therefore a low-variance input that could reasonably be ignored.

What 2016 required instead:

• Modelling discrete outcomes rather than continuous drift. A referendum has a binary result. A portfolio exposed to it needs scenarios, not a trend.
• Identifying which assets are exposed to which political variable. Currency, sovereign spread, sector regulation and cross-border operating structure each respond to different political questions.
• Recognising that the timelines differ. As the UK Investment Report 2016 describes, the market reaction resolved in weeks while the policy consequences ran for years. A portfolio can survive the market reaction and still be impaired by the policy outcome.
• Treating electoral calendars as a risk input. This was already standard practice for emerging markets and had not been for Europe.

The most useful practical change was the widening of scenario analysis to include outcomes previously treated as tail events. That change proved durable and correct — the subsequent decade produced several developed-market political outcomes that a pre-2016 framework would have assigned negligible probability.

What venture capital did while this happened

Beneath the macro repricing, the European venture market continued to develop, and its trajectory was largely independent of the political question.

Formation continued. The talent, the cost advantage relative to the United States and the improving availability of early-stage capital were unaffected by referendum outcomes.

The growth-stage gap persisted, as the 2015 report describes. Political uncertainty made it slightly worse by giving US investors one more reason to prefer domestic opportunities, but the gap's causes were structural rather than political.

Location decisions became live. For financial services in particular, the question of where to be regulated became urgent, prompting contingency arrangements — subsidiaries, licences, staff relocations — that cost money regardless of the eventual outcome. This is the deferred-decision cost the Asia-Pacific Investment Report 2018 describes for trade policy, operating in a different domain.

The ecosystem's centre of gravity began to distribute. Where technology activity had concentrated in a small number of European cities, uncertainty about one of them accelerated a distribution that was happening anyway — which is a genuine long-term positive for the continent's ecosystem, arrived at for an unwelcome reason.

What an allocator could act on

Check whether European bank and sovereign exposures are independent. They are not, while the deposit insurance leg of the banking union is missing. A portfolio holding both in the same member state is more concentrated than the line items suggest.

Model reversibility as a probability, not as excluded. An assumption that has never been tested is priced at certainty. Once tested, it is a distribution, and pricing it requires scenarios rather than a discount.

Watch bank profitability, not just bank capital. A bank whose margin is compressed cannot generate retained earnings, which is how capital is built. The ECB publishes net interest margin data free, and it is a better leading indicator of lending capacity than the capital ratio it eventually feeds.

Separate the market timeline from the policy timeline. The 2016 market reaction resolved in weeks; the policy consequences ran for years and were incurred in deferred decisions that no price series shows. Observing a market recovery is not evidence that the consequences were small.

Set currency hedging policy in advance. For a foreign investor in European assets, the currency decision was again the largest single determinant of outcome, and it is only available before the move.

Treat the venture trajectory as separate from the macro one. European company formation continued through 2016 largely unaffected. The constraint on European venture was structural — the growth-stage gap and the domestic capital equilibrium — and political events changed it only at the margin.

Why the doom loop is a concentration problem, not a credit problem

The sovereign-bank feedback described above is usually presented as a credit risk. It is more useful to an investor as a concentration problem, because that framing makes it measurable.

The standard framing. A weak sovereign weakens its banks; weak banks weaken the sovereign. Both credits deteriorate together.

The concentration framing. A portfolio holding a euro area bank and that country's sovereign debt holds two positions with a common exposure. The exposure is the member state's fiscal capacity, and it sits underneath both.

Why the second framing is more useful:

• It is measurable. ECB data on bank holdings of domestic sovereign debt is published free, and it quantifies the linkage directly rather than requiring it to be inferred from credit spreads.
• It identifies the diversification failure. An investor who believed they held two positions held one, levered. The failure is in the correlation assumption, not in either credit assessment.
• It shows where the concentration is worst. The linkage is largest in the member states where both assets look cheapest — which means a value-driven allocation process concentrates the exposure rather than diversifying it.
• It points at the fix. Common deposit insurance breaks the link between deposit safety and national fiscal capacity. Until that leg exists, the concentration is structural rather than a matter of individual bank quality.

The general principle applies well beyond Europe. Two assets that appear independent may share an underlying exposure that neither's own analysis reveals. The check is not whether each holding is sound but whether a single variable, if it moved, would move both — the same question the Global Investment Outlook 2020 poses about portfolios that looked diversified by sector and were concentrated on the discount rate.

What 2016 established for Europe

• Integration was demonstrated to be reversible, repricing an assumption that had been carried at certainty across every European asset class.
• Negative rates were shown to compress the intermediary that bank-based transmission depends on — a limit a market-based system does not have.
• The incomplete banking union became a priced variable, with the missing deposit insurance leg keeping the sovereign-bank doom loop live.
• Political risk was reclassified as a developed-market category requiring discrete scenarios rather than continuous drift.
• The venture trajectory was shown to be structurally rather than politically determined, which is why the growth-stage gap persisted regardless.