VIX & yen carry trade unwind
Given the magnitude of the VIX we experienced several weeks ago, I figured I could spend some time understanding what triggered the kind of spike we are accustomed to seeing during crises (e.g., the dot-com bubble and the 2008 financial crisis). My reference country will be the US because VIX is derived from options on the S&P 500.
First and foremost, a VIX of 65.73 (the peak) translates to an implied volatility of the S&P 500 return over the next 30 days of 18.84%. So, it is expected that with a 68% probability, your return on the S&P 500, had you bought a contract of an ETF tracking the index today, would end up one month later anywhere in between [-18.04%, 19.64%], assuming a historical monthly return of 0.8% (annual of 10%). That is a wide gap for a month. So, what can justify such a magnitude of unpredictability, and what caused it in the first place?
The biggest catalyst, in my view, came from Japan. In short, investors used to borrow yen and exchange it for other currencies (e.g., USD) to purchase assets denominated in those currencies (i.e., carry trade). The BOJ increased the short-term policy rate to levels unseen in many years, unveiled how it plans to slow the massive bond-buying and considers another increase this year if inflation and financial markets allow. These are substantial changes to the BOJ’s monetary policy. As a result of these changes, the borrowing cost on all the carry trades increased. It is true that other things happened around about the same time, and I will cover some of them, but in my view, the BOJ’s decisions form the main catalyst. The change in monetary policy had ramifications worldwide and across many asset classes and industries. Now, in the short term, the Japanese yen non-commercial combined positions at CME show that nearly all net short positions in yen contracts have unwound. For perspective, the short positions were worth around $15.6B on the 2nd of July (i.e., 190,000 contracts) before the unwind. As the carry trade became less profitable, investors rushed to cash in the profits realised in the foreign market and settle the debts in Japan. This exacerbated the increase in the JPY/USD FX pair. For the medium term, the question is how much of the $1T Japanese banks' foreign lending in yen to non-banks (e.g., asset managers – not subject to margin calls) will be repaid in the coming months because that influences the unwinding of carry trades. For the long term, the question is how much of Japan's long-term investment position in the US markets will see a 180° turn.
I found it funny that some investors called in for a Fed emergency meeting as if the Fed is responsible for bailing their positions. Some retracted their impulsive statements in the meantime. Now, this doesn’t mean they can not make a case for a rate cut. The NFP number came in lower than expected (+114,000 jobs created by US employers in July below the expected 175,000, signalling a slowdown in hiring and a jump in unemployment) and put downward price pressure on WTI and oil byproducts. Furthermore, the 2s-FF spread (i.e., subtract the 2Y yield from the ff rate) is very low, and when it reaches well into the -1 levels, it is typically followed by a recession. 2.9% inflation is not incredibly great if the goal is 2%, and even at these high ff rates, it seems to decrease slowly since the US has been around 3% for a year now. The good news is that the PCE price trend across almost all industries is down. However, if you cut the ff rate rapidly, you ignite inflation. On top of this, you can exacerbate the unwind of the carry trades, leading to lower valuations and declining wealth, which manifests in reduced consumption, which can lead to layoffs or freezes if firms decide to reduce production. Overall, cutting the ff rate will have various effects on investors according to their portfolios’ positions. Because of the carry trade issue, I believe that a big cut is less preferred than a gradual one amounting to the same bps to avoid inducing a sudden unwinding of carry trades (i.e., a single 50bps cut is less preferred than 25bps cuts in 2 consecutive meetings). Nevertheless, I must understand how the magnitude of a change in the ff rate affects the economy over the “long run” from a quantitative perspective; otherwise, I would be rationally guessing at best. I’d have different ff rate decisions and different economic scenarios and would want to pick the ff rate decision that yields the best expected value of an equation or fraction (depending on how many things I care about, say inflation, unemployment, GDP growth, etc.), weighting in the likelihood of all scenarios. Scenario analysis (e.g., escalated Middle East war [yes/no] & another BOJ increase [yes/no] &…&...) + MC simulations (for the probability distribution of inputs for the respective scenarios) at the bare minimum.
As for the war in the Middle East, if it escalates significantly, crude oil and its byproducts will increase in price, resulting in the same cost-push inflation we experienced when Russia-Ukraine broke out. This will limit the Fed’s ability to cut the ff rate, and it is another blow to household consumption since disposable income will decrease following increased gasoline prices and the fear of the war spreading since the US is involved in it. Overall, wars are inflationary from multiple perspectives, and the US is about to be involved in 2. Powell says he can achieve 2% over the “long run” without a recession, but he would say anything to calm investors. The “long run” is also deliberately vague and leaves room for interpretation, but for the better part of the last 3 and a half years, it has been way above 2%, and at some point, you have to compensate by being well below 2%. How do you do that if you significantly cut rates now?
When it comes to the fears of a recession, I was taught that the Fed’s number one priority is to gain control over inflation and bring it to target, even if it means sacrificing economic growth. Furthermore, a recession would help decrease inflation. Some businesses will reduce their spending (i.e., CAPEX), cut costs (including through layoffs), and even cut their margins to lower their prices to maintain their revenue/growth and, most importantly, their valuation, given managerial incentives relating to compensation. Firms’ pricing power would fall because the consumer sensitivity to price increases is high during a recession. So, inflation will be positioned to reach the 2% target, and now you can think about 2% over the “long term”. The downside is that we’ll get to see what fierce competition can really do to a firm (this time around, without any bailouts in the many forms that came during the pandemic), which will result in bankruptcies, layoffs, and negative GDP growth but might also result in innovation which is welcomed.
All the catalysts that I mentioned (i.e., BOJ, NFP, Middle East tensions) timed too well in a relatively short timeframe, which influenced the S&P 500, the options on the S&P 500, and their implied volatility, and as a result, the VIX, but what can explain VIX’s magnitude? I believe the catalysts were the fire, and the high leverage was the petrol. Plenty of downward forces were applied to macro portfolios (FX pair losses, decrease in the prices of existing Japanese bonds, decrease in the prices of WTI and oil byproducts, lower Nikkei incentivising a change in asset allocation [see Warren Buffet’s Apple sell-off]), triggering a risk sell-off, whose magnitude was determined by the leverage level and perhaps a reassessment of where capital could be best deployed given the new conditions. In periods of considerable shocks (i.e., high volatility), the VaR underestimates the losses you could expect on a portfolio (i.e., the tail risk) if you assume a normal distribution of returns. Leverage amplifies that underestimated loss. Hence, some selling happened because of risk management; maybe some even happened because of margin calls.
Looking at past NFP releases, you’ll see that this year’s April underperformance (i.e., down 63k from the forecast) was about the same as the recent underperformance (i.e., down 62k from the forecast). However, the VIX, back on the 3rd of May, when the NFP (April) was released, was not impacted. Maybe the recent NFP release came in as more of a surprise (the recent lag in the plunge of WTI’s price might support this), but the magnitudes of the plunges in WTI’s price around the release dates are not that different (between 6% and 7%). Of course, one could argue that we are not comparing apples to apples here because the economic environment and expectations aren't the same, which is fair. You can justify selling your equity stakes based on a recessionary outlook, but why such a big sell-off? It wasn’t like Japanese banks were in deep trouble, which didn’t appear in the news. Q2 of 2022 was the last time the US GDP growth was negative (so you’d need 3 consecutive quarters on negative turf); back then, inflation was at 3%, and ISM Manufacturing PMI for July was 46.8 (the US GDP mainly relies on services). In the meantime, ISM Services PMI for July was 51.4, US retail sales were up 1%, and inflation came at 2.9%. So, was the US economic situation and its projections bad enough to justify a 65.73 VIX? I doubt it. Additionally, if the increase in VIX resulted from the economic conditions alone, weren’t we supposed to have a relatively high VIX in the short term? This was the case during the financial crisis. Now, if you looked at the VIX term structure a few days after the spike, it was in contango, which meant the market anticipated lower volatility in the future than the spot VIX at the time. This suggested that the increase in VIX was mainly due to the unwinds of leveraged positions and carry trades rather than the economic environment and its projections.
I refrained from being general on purpose. Please feel free to comment on this insight and correct me if appropriate.
Disclaimer