Global Investment Outlook 2017 — Synchronised Calm

2017 was, by most measures, an unusually good year — and the reasons it was good are more interesting than the outcome.

Growth accelerated in the US, Europe, Japan and major emerging economies at roughly the same time. This is rarer than it sounds. Economies normally move at different points in their cycles, so global growth is usually an average of expansions and slowdowns. In 2017 the components moved together.

Synchronised growth is favourable for risk assets for a straightforward reason: earnings improve broadly rather than in isolated pockets, so equity gains do not depend on a single sector or region carrying the index. There is less need to be right about which market — a rare condition.

Alongside this, volatility fell to historically low levels. Markets rose steadily with unusually few drawdowns.

The consequence of that calm is the most important thing about 2017, and it was not visible in returns. Modern risk management sizes positions partly on recent volatility: when measured risk falls, the same risk budget permits a larger position. So a period of low volatility mechanically increases leverage across the system, without anyone deciding to take more risk. That process was well advanced by the end of 2017, and its unwinding is what the 2018 report opens with.

Why synchronised growth is unusual

Economies are linked but not synchronised. Trade, capital flows and commodity prices connect them, while domestic policy, demographics, credit cycles and politics push them apart. The normal state is partial correlation — some expanding, some slowing.

Synchronisation occurred in 2017 for identifiable reasons rather than coincidence:

• The post-crisis repair was largely complete. Balance sheets across households, banks and corporates had been rebuilt over nearly a decade, removing a drag that had held back several economies simultaneously.
• Monetary policy remained accommodative nearly everywhere, even as the direction of travel began to shift. The 2015 divergence was in direction; the level of policy was still supportive in most places.
• The commodity correction had run its course. Producing economies stopped contracting, which removed a drag on global trade.
• No major shock occurred. This is under-appreciated: some years are good primarily because nothing goes wrong.

For portfolios the effect was that diversification produced less benefit than usual — when everything rises together, the diversifier rises too, and the payoff for having been thoughtful about correlation is small. That is a good problem, but it establishes a habit that costs money later.

What low volatility does to positioning

This is the mechanism that made 2017 consequential rather than merely pleasant.

Volatility is an input to position sizing. A risk-managed portfolio targets a level of risk, not a level of exposure. If an asset's measured volatility halves, the same risk target permits roughly twice the position.

This is true across a wide range of strategies — risk parity, volatility-targeting funds, many quantitative approaches, and the internal risk limits of banks and multi-strategy funds. None of them decided to take more risk in 2017. Their models observed less risk and permitted more exposure.

Low volatility also makes selling volatility profitable. Selling options collects a premium and pays out when large moves occur. In a calm market the premiums are collected and the payouts rarely trigger. The strategy produces steady positive returns — until it doesn't, and the loss arrives all at once.

Both effects compound into a specific fragility:

• Positions across the system are larger than they would be at normal volatility.
• Short-volatility exposure has accumulated because it has been profitable.
• Both are sensitive to the same trigger: a rise in volatility.
• When volatility rises, risk models require reducing positions — which means selling into a falling market, which raises volatility further.

Low volatility does not merely reflect calm. It manufactures the leverage that ends it.

This is the clearest available example of an endogenous risk: a condition created by market participants responding rationally to observed conditions, which changes the conditions themselves.

The first crypto cycle

2017 produced the first large speculative cycle in cryptocurrencies, and it is worth examining as a case study rather than as a verdict on the asset class.

The analytically interesting feature is that these assets have no cash flow. A stock can be valued on earnings, a bond on coupons, property on rent. An asset with no cash flow cannot be valued by discounting anything.

That does not make it worthless — gold has no cash flow and has held value for millennia. But it does mean the price is determined entirely by what the next buyer will pay, which makes price formation qualitatively different:

• There is no anchor. With no fundamental value to deviate from, there is no level at which the price is objectively "too high", and therefore no natural mechanism to stop a rise.
• Price becomes the primary evidence. In the absence of fundamentals, a rising price is itself the main argument for owning the asset — which is reflexive.
• The reversal is symmetric. The same absence of an anchor means there is no level at which the price is objectively "too low" either. Declines can run further than fundamental assets do.

The 2017 cycle also produced a wave of token issuance that funded projects on the strength of a document, with minimal disclosure obligations and no established investor protections. The subsequent regulatory response is a substantial part of what shaped the industry over the following years.

The general point is the useful one: when an asset has no cash flow, price discovery works differently, and the absence of a valuation anchor cuts in both directions.

Passive investing reaches scale

A slower structural development crossed a meaningful threshold around this period.

Index investing had grown for decades on a sound argument: most active managers underperform their benchmark after fees, so the fee saving is a reliable edge. Nothing about that argument stopped being true.

But scale changes the mechanism. When a small share of the market indexes, index funds are price takers — they buy at prices set by active participants. As the indexed share grows, several second-order effects appear:

• Flows become less price-sensitive. An index fund buys because money arrived, not because it formed a view. Large flows can therefore move prices without conveying information.
• Index membership acquires value independent of fundamentals, because inclusion guarantees a source of demand.
• Cap-weighting is momentum by construction. As a stock rises, its index weight rises, so index funds hold proportionally more of it. This amplifies concentration — a mechanism that became central to the 2023 report's analysis of narrow leadership.
• The active pool shrinks. Price discovery depends on active participants; a smaller pool doing that work has implications that are still not fully understood.

None of this argues against index investing, which remains rational for most investors. It argues that the market structure it produces is different from the one the original argument assumed — and that difference compounded significantly over the years that followed.

Why calm is harder to analyse than crisis

A structural observation about 2017 that generalises: periods of low volatility are analytically harder than periods of stress, and they are treated as easier.

In a crisis, the risks are visible. Prices are moving, correlations are breaking, and the questions being asked are the right ones. Attention is concentrated on exactly the exposures that matter.

In a calm period, the risks are accumulating and invisible. Nothing is moving, so nothing demands attention. The accumulating positions look prudent because their measured risk is low, and the measurement is low precisely because the accumulation has not yet reversed.

Three specific analytical problems arise:

• Backward-looking risk measures understate risk. A volatility estimate computed over a calm period describes the calm period. Using it to size positions for the future assumes the calm continues, which is the assumption most likely to be wrong at exactly that moment.
• Correlation estimates are unstable in the same direction. In calm markets, assets appear less correlated than they are, because the common factor — a shock — has not occurred. Diversification looks more effective than it will prove to be.
• The absence of losses is read as evidence of skill. A strategy that has not lost money in a calm period has not been tested. Distinguishing skill from an untested exposure requires a stress event, which by definition has not happened.

The practical consequence is that risk assessment in a calm period must be forward-looking and structural rather than statistical. The relevant question is not "what has this position's volatility been" but "what would this position be forced to do if it started losing money, and how many others would be doing the same thing at the same time."

The measured risk of a crowded short-volatility position in December 2017 was low. Its actual risk was that it would be forced to buy volatility into a rising market alongside everyone else holding the same position. No volatility estimate captures that, because it is a statement about the position's structure rather than about its history.

What an allocator could act on

2017's mechanisms had specific, checkable implications.

Ask what a strategy does when it loses money. This is the single most useful question about any crowded position, and it requires no forecast. A strategy whose loss-response is to transact in the direction that worsens the loss — selling as prices fall, buying volatility as volatility rises — has a risk profile that its historical returns do not describe. Volatility-targeting strategies, short-volatility strategies and leveraged positions all share this property.

Check whether measured risk fell because risk fell or because measurement did. If a portfolio's volatility halved while its notional exposure doubled, nothing about its risk improved. This is checkable from position data and is a different question from the one a risk report answers.

Treat synchronised growth as a rare condition, not a new baseline. The temptation in 2017 was to extrapolate — to conclude that the post-crisis repair was complete and that synchronised expansion was the new normal. It was not, and it has not recurred at that degree since. Rare conditions should be enjoyed and not extrapolated, and the discipline is to ask how often the condition has historically persisted rather than how good it currently feels.

Ask what an asset is worth if it never produces cash and can never be sold. This diagnostic, developed at length in the 2017 digital assets report, is worth applying broadly. It separates anchored from unanchored assets and determines which analytical framework applies. Applying cash-flow reasoning to an unanchored asset, or resale reasoning to an anchored one, reliably produces the wrong answer.

Recognise that indexing changes the market it operates in. The argument for indexing remained sound in 2017 and remains sound now. But the market structure it produces — momentum through cap-weighting, less price-sensitive flow, value in index membership itself — is different from the one the original argument assumed, and that difference compounds. It became central to the concentration described in the 2023 and 2025 reports.

What 2017 established

• Synchronised growth was demonstrated to be possible but rare, and rare enough that it should not be extrapolated.
• Low volatility was established as a risk factor rather than merely a description of conditions.
• Endogenous risk — the market creating its own fragility through rational individual responses — was set up for demonstration the following year.
• Cash-flow-free assets provided a clean lesson in how price discovery works without a valuation anchor.
• Passive scale began affecting market structure, with consequences that compounded for the next decade.

Cap-weighting as a momentum strategy

The observation that a market-capitalisation-weighted index is momentum by construction deserves fuller treatment, because it became the central mechanism in the concentration described in the 2023 and 2025 reports.

How the weighting works. An index holds each company in proportion to its market value. As a company's price rises, its market value rises, and its weight in the index rises with it.

The consequence for an index fund. The fund holds the index weights. When a company's weight rises, the fund's holding of it rises — not through a purchase decision but through the price move itself. When new money arrives, it is allocated at the current weights, which means proportionally more goes to whatever has risen most.

This is a momentum strategy. Not by design, and not as a criticism — it is a description of what proportional weighting does. An investor holding a cap-weighted index systematically holds more of whatever has appreciated and less of whatever has not.

Three consequences that compound as the indexed share of the market grows:

• Concentration is amplified rather than dampened. In a market where a few companies are appreciating faster than the rest, index flows reinforce the divergence. The concentration that defined 2023 and 2025 has this as a contributing mechanism.
• Index membership acquires value. Inclusion guarantees a source of demand independent of any assessment of the company, which means the inclusion decision itself moves prices.
• The price-setting pool shrinks. Price discovery is performed by participants who form views. As the indexed share grows, a smaller pool does that work — which does not necessarily make prices worse, but does change how quickly and how reliably new information is incorporated.

None of this argues against indexing. The original case — that most active managers underperform after fees — did not stop being true, and for most investors indexing remains rational. The argument is that the market structure indexing produces is different from the one its original case assumed, and an investor should hold the resulting concentration knowingly rather than as an unexamined consequence of a decision made on fee grounds.

The practical check is to look at the index's largest weights periodically and ask whether that concentration is the intended exposure. In 2017 it was moderate. By 2023 it was the defining feature of the market, and the 2023 report describes it having become a recognised risk rather than an academic observation.