HomeSolutionsInsightsNewsroomPeopleContactLogin

Acclaimed macro analytics
Portfolio Construction & Risk Management
Trading Solutions
Macro Models
Thematic equity baskets

Announcements
Press
Blog

1. If FCIs have seen their floor ITB is at risk of rolling over. Over the last 1yr, the ITB Home Construction ETF has delivered a 61% return – among the best across all US industry groups.

2. Macro matters for ITB – There is 62% correlation (39% RSq) between daily ITB returns and macro factor driven daily returns over the last year.

3. Over the last week, the drag from macro factors has been particularly large. Qi’s MFERM model dissects the return of ITB into its macro factor & idiosyncratic / specific return components. Note the recent divergence between the spot return and factor return in the chart below.

4. Where does this recent factor drag stem from? FCIs. USD TWI, forward growth expectations (5s30s), rates, HY corporate credit and CB QT expectations have all been drags for ITB over the last week.

5. From propellers to drags: Indeed, for rates, 5s30s and USD TWI the factor returns have switched from propellers to drags over the last 1mth

6. So what is ITB most exposed to today? HY Credit, the dollar, real rates and it would prefer bull curve steepening.

7. Across industry groups ITB is among the most sensitive to real rates, DM FX and corporate credit: Notable, as wait for inflation numbers this week - a beat would be troublesome.

By the end of last week, the S&P500 had rallied 6 of the last 7 week and sitting 4% above its 50d MA - the world’s most important central bank is easing, and the world’s most important government is stimulating! No surprise therefore to see this rally coincide with a sharp decline in the S&P500 macro beta impulse i.e. falling macro sensitivities. Recall, our hypothesis is that Mr. Market is a neurotic patient that would prefer a backdrop of macro factor stability / low beta to macro over a backdrop of increasing macro factor volatility / high beta to macro. Therefore, when macro betas are low it is a reflection of low fear on macro. The S&P macro beta impulse has hit the low end of its range post Covid. See the chart above.

In similar vein, we show the S&P500 Qi sensitivity to the average of VIX and HY credit spreads (ST model) as a fear gauge proxy. We are clearly in period of lower macro fear, relative to the last few years.

Consider also that SPY model price momentum has been fading over the last month. HY credit spreads are little changed while inflation expectations have been rising.

Looking ahead, if the market is increasingly comfortable on the data, we need to look to inflation expectations. This would have an upward impetus on a Trump surge, China stimulus and falling jobless claims.

While SPY’s sensitivity to High Yield spreads is towards the low end of the ranges, it remains a critical factor. Credit spreads account for 22% of model explanatory power so equity investors need to be aware of some early warning signs that credit is looking extended from a macro perspective.

LQDH is the ETF that provides exposure to US Investment Grade credit with the duration element hedged away, i.e. it is a “purer” credit play. A bearish divergence pattern is developing with the recent rally occurring despite macro conditions starting to deteriorate. If Qi model value falls below the early September low of 91.73, we would start to worry macro momentum is potentially turning.

The same bearish divergence formation is evident elsewhere too. Still in the US, the ANGL ETF captures firms that were rated investment grade at the point of issuance, but have since been downgraded to High Yield. It sits 1.4 sigma rich to macro with Qi model value approaching the September lows.

EMHY (JP Morgan’s EM High Yield ETF) and EMB (hard currency EM sovereign bonds) are over 1 sigma rich in FVG terms with model value rolling over.

The local currency Emerging Market bond ETF LEMB is also rich but there at least macro momentum is still trending higher. Ideally, a retracement would offer a better entry level, but soft currency EM bonds look the best bet to global asset allocators on our models.

Qi is firing warning signals on several China related plays. AIA (iShares Asia 50; 40% China weighting), AAXJ (iShares All Country Asia ex Japan; 29% China weighting) and VWO (Vanguard FTSE Emerging Markets; 23% China weighting) are all over 1 sigma rich to macro-warranted model value.

Copper is now 1.6 sigma rich to macro. LUXU (the Amundi Luxury ETF with Hermes, CIE Richemont and LVMH Moet Hennessy Louis Vuitton as the top 3 holdings) is 1.5 sigma above model value.

Yes there’s a large element of a positioning squeeze driving current price action. But, from a macro perspective, the tradeable bounce has already happened to a fair degree. The risk-reward suggests these are not great levels to chase.

Betting odds show that the Trump/Harris gap is narrowing at 50% probability for a Trump win and 48% for Harris. However, according to Qi, at least at the equity market level, a sudden Trump surge is clearly not priced. If all Trump’s policies were effected, it would be expected to lead to larger inflationary than growth impulse. Yet Qi shows the sensitivity of the S&P500 to Trump betting odds has falling to zero. See the chart below.

In contrast, the sensitivity to 10yr Treasury yields is deemed much higher. Higher Trump betting odds point to higher yields – no surprise given the heightened concerns on a ramp up in the fiscal deficit.

Crude may be close to 3yr lows but other commodities have been holding up better. Copper has risen ~8% from its August lows. Data at the end of last week showed copper inventories on the Shanghai Futures Exchange fall 14% on the week - the largest weekly decline this year – and the LME spot price may have now turned the corner after sinking 20% from a high in May. Today SXPP, the EU Basic Resources sector, is still close to its lowest level since late 2020. See the chart above.

Earlier this week, we note in our MacroHub publication, how depressed sentiment was in Europe (a quarter of the Dax stocks are trading below book value). On Qi’s model, the sector is trading -0.8 sigma (~5%) below Qi model value, hovering close to the 1yr range lows. It is one of the cheapest of the Stoxx sectors in Europe.

Global GDP growth is one of the top macro drivers of the sector. Since July, that sensitivity has accelerated higher firmly into positive territory. If there is belief in a soft landing any sign of a pick up in global growth confidence, owing upside in this space make sense given how depressed sentiment is.

Whatever the extent of the Fed rate cutting cycle, ultimately it is all about soft landing for the economy. The below chart shows the GS Qi long / short basket pair tracking US GDP growth. The dark blue line shows the deviation of that basket from its 200d MA i.e. the deviation from trend as measure of optimism / pessimism. Indeed, shown in this way, there is good relationship with the Citi US economic surprise index. Cyclical re-acceleration may well be a pain trade

Defensives have been among the best performing sectors recently and relative to their sector peers are more expensive on Qi’s macro fair value models.

Tech has been a laggard in the recent SPX rally. Positioning is now also cleaner in the space. Further, we know the strength of Q4 seasonality. SOXX is one of the cheapest sector ETFs at -1.3 sigma. The technology sector in general is a beneficiary of loose financial conditions. See the chart above highlighting the sector ETF sensitivities to FCIs (HY Credit spreads, USD TWI, real rates) vs. sensitivities to GDP. SOXX, COPX, XME, XRT, XLB, IWM, SLF, XLB are all sectors which relatively benefit from loose FCIs alongside trading below.

Above we show the Mag7 12mth fwd PE vs. Qi’s model FVG of the PE. On this basis, the 12mth fwd PE of the Mag7 stands ~1 sigma cheap to model value. That is close to the lower end of the valuation range of the last 1yr. A return to Mag7 leadership into year-end is a possibility if we see re-risking into YE.

Thus far, the recent rotation trade has favoured defensives and certain cyclicals, but not all. Energy has missed out because of the drag from crude oil prices.

But now we note US Energy XLE sits as 0.8 sigma (4.6%) cheap to Consumer Discretionary. Qi’s macro-warranted model value for the RV pair has been flat-lining since mid-August. Put another way, XLE has overshot to the downside relative to broad macro conditions.

Over the last year extremes in Qi’s Fair Value Gap has done a good job of marking local lows and highs in the actual XLE / XLY RV pair trade. If a 50bp rate cut from the Fed helps negate recession fears, maybe its time to be looking for cheap cyclicals that have lagged.

The European 2s10s yield curve was almost 60bp inverted as recently as May. Today it is almost back to flat.

There remain good reasons to think the risk-reward favours further curve steepening over the medium to long term. A hard landing means more rate cuts and a bull steepening; cutting when an economy is slowing but growing risks re-awakening the bond vigilantes who will bear steepen the curve.

But Qi’s model which has a high (72%) and stable R-Squared is suggesting this move may have too-far-too-fast. Relative to aggregate macro conditions, the curve sits 1.6 sigma (21bp) too steep.

The chart below shows there are times when the FVG becomes even more extended - last October around the Fed pivot, the gap reached 3 sigma. Even so we’re getting into thin air. Some consolidation could be healthy and offer better re-entry levels to re-set steepeners.

What have been the major XLF macro exposures ytd?


• Through 2024, XLF has benefited from a weaker dollar, falling 10yr yields and, since the Spring, falling energy prices and a steeper 5s30s curve; i.e. the focus ytd has been the easing in financial conditions.

• This has also occurred in a backdrop where there is confidence in the credit cycle – the XLF factor exposure to risk aversion and HY credit spreads have dissipated to cycle tights.

If the balance of risks shifts towards recession, what changes should we look out for?

With the balance of risks shifting towards to recession risks, we would expect to see in that backdrop:

1. the exposure to 10yr yields moving towards positive territory; i.e. higher yields / higher XLF
2. the positive exposure to a steeper 5s30s curve continue to rise; i.e. pressure remains high on the Fed to cut
3. the negative exposure to corporate credit and risk aversion to not be so sanguine but rather increase; i.e. credit cycle fears rise

This would also manifest itself back to a regime where macro dominates XLF predicted risk. Early July reflected complacency. Today, we may have returned back to a macro dominant regime to understand risk.

Macro has dominated both XLF returns and risk since early July

• Macro has also the dominant driver of returns since early July. See the chart below, showcasing the close relationship of XLF to the attributable factor return over the last month (reflecting XLF’s exposures to those factors and the underlying moves in the factors themselves). Like we have seen before, the XLF factor return seems to have peaked a few days before the actual ETF peak. If we suspect macro is to remain dominant, bulls want to see the XLF attributable factor return component trough and move higher.

• What explains the descent in the XLF factor return from its peak on 27th August? The table below showcase the attributable return by individual macro factor. Reflecting slowdown fears - the jump in credit spreads and risk aversion, alongside dollar strength and the large drop in inflation expectations capture the bulk of the decline.

Watch the exposure and risk attributable to HY credit spreads – that may well tilt the scales

• Looking ahead, the Fed has never started an easing campaign with 50 basis points when credit spreads were this tight and unemployment this low. There is plenty of uncertainty: This point in the calendar is hard for risky asset assets – Will the Fed begin the rate cut cycle aggressively from the get go? Will we see a proper rise in the UE rate? Who will win the US election? There are 7 long weeks between Sep FOMC and Nov FOMC.

We would say watch financial conditions – that will tilt the scales:

• Watch if the XLF negative exposure to credit spreads becomes larger

• Watch if the XLF predicted vol attributable to macro continues to rise and if HY credit spreads can explain a larger proportion of the risk. Currently, the factor vol attributable to HY credit spreads is in-line with its long run average contribution i.e. the market is still comfortable with the state of the credit cycle. We should stay alert if we see this start to lurch higher

This is a case study showcasing the power of Qi’s Macro Factor Equity Risk Model. Qi clients can train any single security or portfolio of assets on the same framework. Thereby identifying their macro factor exposures, and decomposing both their risks & returns into macro versus idiosyncratic. Finally by revealing which specific macro risks they are running, clients are in a position to consider which hedging strategies may be required.

#1 Everyone hates China. Time for contrarians to let smart machines do their work.

Qi recently launched the Qi China Alpha Index (KCAIQI on Bloomberg). This index seeks to outperform the CSI 300 Index by applying a systematic approach. Qi applies a proprietary, machine learning-led investment process which (i.e. systematically) identifying the appropriate combination of stocks and their weights to meet the investment objective.

Qi’s optimisation process screens all the possible combinations of stocks and weights to seek the most appropriate ~35-50 stocks that minimise tracking error in sample AND minimise over-fitting out of sample. No single constituent security can have more than a 5% weight. This process is repeated monthly.

The table below shows that the index has outperformed the CSI 300 in 7 out of 9 years between 2015 – 2023, with the lowest underperformance -3% and largest outperformance +65%.

Active funds are a small but fast-growing area within the ETF market. There is increasing appetite to take advantage of the ETF structure but use active strategies to generate alpha instead of relying on passive funds which simply track a target index or theme.

This week KraneShares have launched an ETF to track the Qi China Alpha Index.

#2. EURCHF – macro suggests decent upside

Two weeks ago, Qi flagged the Swiss Franc as looking expensive relative to macro conditions. The two crosses we highlighted (AUDCHF, GBPCHF) have bounced but only slightly (~0.3%) and both still show the Franc as rich.

Now we have a bullish signal on EURCHF which screens as 1 sigma (1.8%) cheap on Qi.

The bullish divergence has arisen because spot EURCHF has fallen back towards the early August lows, while Qi’s macro-warranted value has risen 1.6% since the Aug 5th ‘flash crash’ Monday.

Forward growth and interest rate differentials are key drivers but the factors themselves haven’t moved hugely in August. Rather, the bigger move has come via risk appetite channels – the fall in risk aversion (VIX < 20) and gains in EuroZone Sovereign Confidence (narrower BTP spreads) have driven the 1.6% gain in macro-warranted model value over the last 3 weeks.

In short, the Swiss Franc has ignored the recovery in risk appetite since Aug 5th.

A -1 sigma Fair Value Gap has a strong (71%) hit rate as a buy-the-dip signal. While over the last year the correlation between Qi’s FVG and spot EURCHF suggests the mean reversion has occurred via the market correcting back towards macro fundamentals.

#3. Gold vs. Digital Gold

For much of 2024 Bitcoin and Gold have traded in tandem. Until the August flash crash. Bitcoin has yet to fully recover from the hit to risk appetite, while Gold has benefitted from the uncertainty.

On Qi, the Relative Value model between the two shows Bitcoin as 1.7 sigma (26.5%) cheap relative to Gold. Such an extended Fair Value Gap is rare; it has only been seen 4x since 2009.

There’s a health warning. Model confidence of 50% means our macro factors can only explain half the variance in this pair currently. Other, non-macro factors (such as crypto entering a period of poor seasonals) are as important right now.

Still, it is interesting to note the tight correlation between Qi’s FVG & the spot price of the Bitcoin/Gold ratio. That doesn’t tell you whether this marks a buying opportunity in crypto, or a sign the air is looking pretty thin for gold. But it does suggest the elastic is stretched and some kind of mean reversion is pending.

#4. Amazon - a cheap Mag7 stock

Amazon is back in a macro regime for the first time in 2024. Its macro-DNA is pure Goldilocks: reflation (rising economic growth & inflation expectations), easy financial conditions (tight credit spreads) and healthy risk appetite (low VIX).

The latter has been instrumental in a ~20% rise in model value since August 5th & the flash crash low. And Amazon has lagged in this V-shaped recovery. It is now 1.1 sigma (9.1%) cheap to aggregate macro conditions.

There are other drivers to consider.

Amazon has yet to convince that they can compete in the AI arms race. Their Bedrock tool giving ASW customers access to third party AI models like Anthropic and Meta may turn the tide but, to date, they have lagged OpenAI, Microsoft and Google.

The charts remain damaged – technical analysis suggests the stock needs to close the gap around $180 which was the breakdown point after disappointing earnings at the start of August.

But, from Qi’s perspective, the stock is cheap to its macro environment and history shows these are significant levels.

Qi’s Macro Factor Equity Risk Model sheds light on the anatomy of equity returns and vol we have seen during the pullback and the subsequent rally

1/ The pullback and subsequent rally can be explained almost entirely by macro: We have decomposed the SPY total return from 1st July into factor (Qi’s macro factors) and specific (i.e. the idiosyncratic residual). From the 16th July peak through to last Friday, factors returns (red) best explain the index price action (blue). No surprise – macro dominates at turning points/

2/ Further, reflecting the dominance of macro, the recent correlation of SPY factor returns to SPY actual returns is at the upper end of this cycle’s range.

The RSq of SPY daily factor returns to SPY daily actual returns has been 77% over the last 4yrs. However, this relationship is changing over time with the factor return beta rising / falling over different periods.

Unlike style factors which represents a portfolio of stocks (i.e. risky assets which trend in the long run), macro factors do not always trend – across a large macro factor information set, some factors can be propellers and others drags, depending on the asset’s exposures.

Over long periods of time this can get lost in the aggregated, cumulative factor return of an asset. What we want to pay attention to is the rolling correlation of the factor return to the asset i.e. when high, macro is dominant

The second chart shows that on a 3mth rolling basis, the correlation is at the upper end of the range seen post Covid.

3/ The SPY factor return actually peaked on 12th July, ahead of the actual index peak on 16th July. Again, like we have shown above, we believe that at market turning points macro matters most, and factor returns can lead the actual market move.

4/ The dominant factor drags from the 16th July peak to 5th August trough were risk-aversion related as opposed to overt fears on fundamentals - credit spreads and risk aversion lurched higher more than offsetting the accustomed benefit from a drop in the dollar, 10yr yields and inflation expectations over this period.

to fade. See the chart below – a simple fear-gauge proxy

6/ The changes in factor exposures through this episode are telling – revealing just how sensitive risky assets have become to FCIs, with increasing prominence of the dollar.

But while VIX & credit spreads posted the biggest factor moves in August, under the hood the make-up of SPY’s key drivers are shifting. Lower inflation, bond yields and Dollar are becoming bigger factor exposures than corporate credit spreads. SPY’s exposures suggest equities see rate cuts as healthy, pre-emptive “risk management” from the Fed as inflation returns towards target; not because of fears about an impaired credit cycle.

7/ Where have macro factor returns recovered the most post the 16th July to 5th Aug drop? Quality & Value most; Growth least.

In order – Quality, Momentum, Value, SPY, Growth, QQQ. That QQQ macro factor returns have not kept up pace relative to say Value may well be indicative of the sector rotations we are seeing.

8/ There is a hierarchy across styles as to which factors matter most.

Momentum is most geared to FCIs, having the highest negative exposure to HY credit spreads and the dollar. Similar to SPY, we also note the sensitivity to higher inflation expectations and 10yr yields has also lurched more negative.

Value cares most vs. its peers on forward growth expectations (steeper 5s30s) but dollar weakness also matters – financials, homebuilders, consumer etc. all want lower FCIs.

Growth cares most vs. its peers on risk aversion and inflation expectations – presumably reflecting the cost of capital ramifications.

In conclusion – Aside from a sudden geopolitical vol shock, higher factor return momentum is highly dependent on the fate of the dollar and 10yr yields. Typically, we need a sharp decline in factor return momentum to signal a sell-off. However, a sudden, about-turn seems unlikely. Rather, a slowdown in factor return momentum would imply a more range-bound trading environment.