Case Study: Mr & Mrs Market suffer a shock
They knew it could happen but they didn’t believe it would. Their income is likely to be cut by 7.5% says a letter from a Government department, just as a severe drop in sterling threatens to raise their outgoings. An emergency budget is called for. But some of the easy savings they thought they could count on don’t look so easy after all. Their children are in tears and blaming it all on their baby-boomer parents. And to cap it all, they think they may have lost their passports.
Seven weeks later, we revisit Mr and Mrs Market.
A shock? What shock?
Nearly two months after the Brexit vote, we can learn a lot from markets' actual response to what might reasonably have been seen as an exceptional shock to the economy, to exchange rates and to equities. If this places a question mark over whether markets are in fact treating this as a shock, we then have to think about whether the markets are wrong or whether all the shock rhetoric was wrong. That is the point of this article.
It's not about semantics, but meaning matters. A ‘shock’ could (strictly speaking) be defined as an external, unpredicted event with potential to change the equilibrium states assumed by economic agents before the shock. In this case, the referendum was hardly external (though we may blame the anti-reformers in the EU with whom David Cameron had battled) and was hardly unpredicted (although the consequences of Brexit may never have been priced into markets if they chose to price only the most likely outcome). But the economic position of the UK being either in or out of the single market could be radically different, in which case it could be a shock.
A shock to a market system might also be defined by the size of its price effect rather than the source, if it is outside or at the extreme end of the normal distribution of price changes for any unit of time. So how can Brexit be a shock, if equity markets barely flinched?
Trade effects and sterling
The answer is probably that the shock is to the exchange rate. This makes sense if the main impacts of Brexit are to be experienced via trade. This is certainly how the Treasury was modelling the effects. It was HMT that came up with the 7.5% loss of GDP after 15 years, on the assumption we left the single market and had to trade with the EU (and other countries) on a WTO basis.
We don’t know what offsetting exchange rate change the Treasury assumed in its 7.5% cost estimate. We do know that sterling did suffer something of a shock, in terms of the scale of the immediate reaction. Stable for several months before the referendum, it actually strengthened in the closing stages. The immediate response to the surprise of the vote was a fall of 9% against the USD, 8% against the euro and 14% against the yen. Allowing for the jump immediately before, we can shave about 3-5% off these movements to gauge the effect of the new information.
However, this was not the end of the adjustment. A few more percentage points were lost before a rally started which almost exactly reversed those additional losses, with all three key exchange rates then levelling out for the rest of July. Only in August has that level started to erode slightly. Projections are very common that sterling will continue to devalue. But forecasts would have had to allow for the possibility of a weak pound simply because of the scale (independent of Brexit) of the UK's current account and budget deficits.
What about the trade effects? Here’s what we said to clients in an email a week before the vote:
[Brexit] would be a far bigger issue if trade over the past quarter century had not become more open and tariffs had not halved or more. This does not mean to say that the effects on the UK as a whole of being forced to trade on WTO terms are trivial. But it may mean they are not an irresponsible risk for any voter for whom aspects of Britain’s relationship with the EU other than trade count for a lot. ‘Non-trivial but also non-decisive’ is a fair way to describe the Treasury’s predicted cumulative economic impacts over 15 years, bearing in mind the scale of the uncertainties about exchange rates and productivity that would normally dominate such forecasts.
So, is an adjustment of around 10% in the exchange rate enough to offset the possible effects of trading with the EU where trade-weighted average tariffs into would only be around 4% and where non-EU tariffs could fall when we are free to negotiate our own non-EU trade deals without the protectionist instincts of the EU? Are the non-tariff costs of clearing customs and paying duties that significant across our mix of high and low-value goods? We're not trade experts but on the data we can find relating to scale and volumes it looks plausible that the markets are right and the exchange-rate adjustment needed to reflect the change in trade arrangements is quite modest.
Lost passports are also an issue that looms large with portfolio investors. Are the currency markets overlooking the risks to our trade surplus in services or is the City just getting its lobbying in quickly? Without passporting, Brexit imposes frictional costs and inefficiencies on UK-based businesses but many of them already have EU locations. It is not necessarily a big threat to the level of service revenues. Away from the cameras and press releases, business managers appear more positive and creative - an attitude that is of course part of the explanation for our strong services exports.
Here’s what we found ourselves writing in an email to all our clients the morning after the Brexit news:
Following this morning's referendum result, a period of volatility ensued in the financial markets with UK and European indices opening down nearly 8% and sterling dropping to lows not seen since the 1980’s. Very quickly though equity markets settled back and the FTSE 100 is only down around 3% and remains above levels seen last week. Although down in local currency terms, sterling weakness meant the US and Japanese markets were up around 4-5% (based on ETF prices) and Europe was flat. Whilst projected returns from overseas equity markets have increased, this has been largely offset by a weaker sterling and as a result there has been little impact on model allocations between risky and risk free assets. In other words, sterling projected equity returns have not altered significantly overall. With the UK and Europe being weaker relative to the US and Japan there has been some intra-market movements but these are simply a reversal of the moves seen over the last week.
So here we are, running money with a contrarian 'quant' model that seeks to exploit long-term mean reversion in absolute returns, medium-term reversion in relative returns and (if it's not too quick for us) residual reversal in short-term returns; we're in the teeth of a shock, and the model can find very little to do. Taking one month before and one month after the referendum, our transactions were no higher than average. The reason is again sterling.
The chart below (source: Morningstar) shows the movements in each of local currencies (broken lines) and sterling-adjusted (solid lines) of the four major market blocks, UK, Europe ex UK, USA and Japan. Two factors would be expected to influence the UK market: Brexit and the independent movements in international equity markets. Of course, the movements in other markets, particularly Europe ex UK, are not entirely independent but anecdotally, from daily market reports, Brexit has not been a big influence on any of them. Note that the UK (the green line), though weak initially, is the best-performing of the four in local currencies in the seven weeks since the vote but when sterling weakness is taken into account it is the worst performer. Restarting the clock after the initial phase of sterling weakness, sterling-adjusted returns have been in lock step.
The nearest to a price shock is the FTSE 250 mid-cap index (not shown) which was down 12% in the first few trading days and is only just back to the level before the vote. FTSE 100, making up most of the FTSE All Share Index shown here, performed better, having more of the certain translation effects for overseas operations and fewer of the sector- or company-specific Brexit effects.
The shock to the EU
Investors holding EU investments have been exposed to very large uncertainties about growth and the risk structure (both volatility and correlations) associated with the course of European integration. Before the vote we suggested that these uncertainties could easily dwarf the impact on the UK of leaving the single market. With or without the UK, the EU would continue to be an experiment in integration that could lead in very different directions. We told clients: One possible direction of travel looks like the status quo, bumbling along without either further integration or disintegration, but it may have a lower chance than either of the alternative structural discontinuities.
If Brexit has increased the risk of disintegration, it does not show up in the behaviour of either the euro or the European exchanges. However, the idea that Brexit increases the chance of further integration is now being taken quite seriously, not because it might be politically popular (far from it) but because it may be the only way to save the single currency. Compared with such existential risks, Brexit looks like a modest affair.
We're backing the markets: Brexit is not a shock.
Because we are a quantitative manager, choosing to follow our own model rather than second guess it, backing the markets in this case can only mean deciding not to alter the assumptions used in the model. That was a decision we needed to take before the referendum. We did and we have not altered it since.
Brexit does not change the probable long-term real-return trend for UK equities or other markets, which is the product of a vast number of factors and about which there is anyway a huge amount of uncertainty. The risk structure matters. Our inputs for the risk structure of the markets, both equities and currencies, are drawn from long data histories that include incidents of stress far greater than Brexit; the dot com bubble and the credit crisis are even in the period running live. They do not need to change.