深度分析！手把手教你看懂此次金融危机背后的推手及是如何传导的The Hidden Hand Behind the Crash: How the 2020 Crisis Really Spread

"We don't know where the age is headed — a soaring climb, or a hurried final bow."

Piece No. 8

Before We Start

Lately the global financial markets have been raving like it's a nightclub — one circuit breaker after another, lurching up and down. On the stairwell up to the rooftop ledge, the crowd going up and the crowd coming down are jammed shoulder to shoulder. Since we're here to witness history, let's witness it properly. I'm going to walk you through the real logic of this upheaval and show you who the werewolf behind it is.

Everyone's an economist on social media these days, so you surely know the broad backdrop. But forgive me for saying so: most people's understanding still stops at the crude, blocky causal chain of "ugh, the pandemic spread; there was an oil war; so, ugh, the stock market crashed and a financial crisis broke out; ugh, the Fed opened the taps and started throwing money around."

Think about it for even a moment, though, and you'll spot how strange a lot of this is.

1. The US market's circuit breaker has existed since 1988. It was tripped once in the 1997 crisis — it didn't even trip during the 2008 subprime meltdown. So why has it been tripped so many times in a row lately? Is it really that bad right now?

2. Recently US stocks have been falling overall, but the action has been schizophrenic: crash one day, spike the next, crash again, spike again, ping-ponging daily between collapse and manic euphoria. Are they insane? Take a look at the month after Lehman Brothers went under in 2008. Back then, a drop was a drop — there were bounces, sure, but nothing like this ping-ponging. The capital markets' scythe usually keeps a far cooler head than we retail lambs do — so why is this run acting like a panicked retail crowd?

3. Gold and US Treasuries are the classic safe-haven assets; they're supposed to hold firm in a crisis. So why are they falling too? Is everything really selling off worldwide? Has the market genuinely wet itself with fear?

4. Everyone says a liquidity crisis means people scramble madly for cash. But the dollar is only rising in the FX market, while interbank lending rates aren't especially high. So where exactly is the dollar shortage? Who is hoarding dollars?

5. Most interesting of all: recently the Fed opened the floodgates, cut straight to zero, and launched unlimited quantitative easing. Yet on the very day it cut rates, the market crashed and hit a circuit breaker. There's a joke going around: it's because the Fed moved so aggressively that it scared the market even more. Investors figured the Fed must know something they don't, so they bolted. It's like fighting with your girlfriend: a couple of soothing words would have smoothed things over, but instead you dropped straight to your knees. At that point your girlfriend's composure snaps: what is going on? You didn't cheat on me, did you? That's the vibe. So plenty of people have blamed the Fed, saying it flooded the system so ferociously that it actually worsened the panic. Is that really what happened?

So many bizarre behaviors, and everyone concludes the Fed and the market are just being idiots. But that's when something feels off: every time I think the other side is a moron, I stop and mull it over, because there are only two possibilities — either I'm a lot smarter than them, or they're a lot smarter than me. Obviously, the Fed and Wall Street are a lot smarter than me. Once you accept that premise, you swallow your pride and get to work figuring out the logic behind all this — how these strange phenomena came about, and what the Fed saw that we didn't. That "what," it turns out, is the domino at the center of this crisis whirlpool. This is a story of being outsmarted by your own cleverness.

Let me first explain two technical concepts. Say I have 100 bucks and buy some stock. Buying stock obviously carries risk — but which risks am I taking on? First, the stock's own risk: user numbers, revenue performance, and so on. Call this alpha risk. But beyond the company's own performance, my stock is also swayed by the broad market: the economic cycle, monetary policy, all kinds of black swans — just like right now. That macro-market slice of the risk we call beta. So this one investment carries both alpha and beta risk — and, of course, both alpha and beta return. In one line: alpha is down to your own moves; beta rides the market.

An example. You pay attention in class, do the homework your teacher assigns, and score 80. Then you also go to after-school tutoring and buy practice books, and your score climbs to 90. Of that 90, 80 is beta — the baseline conditions the school gave you — and the extra 10 is your own effort's excess return. Alpha can be positive or negative, of course: grind games all day and your score slides to 70, and your alpha is –10.

Now everyone gets what alpha and beta mean. Back to that stock. Suppose I think the broad market looks bad right now and I don't want to bear beta risk or claim beta return anymore — I only want to keep the alpha part. What do I do? Hedge. Now, I add to this investment a Nasdaq index put option — a bet on the market falling, "–beta" — with a market value matching my stock. At that point I'm actually holding three slices of risk: the stock's own alpha; the market's beta; and my bet-against-the-market –beta. This achieves a hedge between beta and –beta. In other words, the market's swings no longer concern me — whatever the index does, I bear only the stock's alpha risk and enjoy the alpha return. Simple, right? Hedging is really just this: when you face a whole pile of risks, you cover off the parts you're unsure about and keep only the parts you're confident in.

And the world-class mega-funds, allocating assets across the globe, have hedging tools of every stripe: dollars, gold, stocks, futures, government bonds; short-term and long-term hedges; hedges between the US–European and Asian markets, and on and on — forming all sorts of complex quantitative trading models. Honestly, most people can't follow these complicated models, but that's fine. We just need to nail down what alpha and beta are and what hedging is. Good.

Now our protagonist enters. A few days ago there was a rumor that the world's number-one hedge fund, Bridgewater, had blown up. Its founder, Ray Dalio, immediately came out to explain: no such thing, that's a rumor, we're just fine — merely down huge, that's all. It hadn't blown up, but Bridgewater got shoved into the eye of the storm, and honestly not unfairly. Bridgewater has a famous macro quantitative strategy called the All Weather strategy. The name doesn't tell you much — but "all weather" means exactly what it says: spring, summer, autumn, winter, wind or rain, it can make money in any of them. Is it really that magical? It sounds like the miracle supplements they hawk on TV — but they actually pulled it off. At least, they pulled it off before this year.

In the 2008 subprime crisis, the S&P 500 fell nearly 40%, while Bridgewater's All Weather strategy rose 12% that year. Seriously impressive. So how does this strategy work? Perk up — the key part is here.

First, the All Weather strategy abandons alpha entirely and chases only beta. In other words, I don't rely on any of my own micro judgments; I only make macro money, only feast on the dividend of market development. Someone might say: isn't that just buying the broad market, buying an index ETF? No. In their eyes that's just a single asset, and any single asset flips over easily. So they allocate at a more macro level: stocks, bonds, commodities, gold, and more. And not just in Western markets, but in emerging markets too. Ultimately, the All Weather strategy chases not the beta of any one specific market, but the beta of overall economic development — it wants to feast on the dividend of the whole global economy.

Second, the All Weather strategy's other defining feature: it starts from the angle of risk, not return. Risk and return are two faces of one coin — but the two faces of that coin can give rise to two completely different logics. We all know not to put your eggs in one basket. But is scattering them across different baskets really enough to set your mind at ease? What if all those eggs come from the same hen? If that hen gets sick, doesn't every egg become suspect? And where did I buy these baskets from — the same factory? Might this whole batch of baskets have the same defect?

You get it now: seemingly, spreading your eggs across different baskets diversifies risk, but the risk isn't actually fully diversified. This is still a return-side strategy. The eggs may be split up, but there's correlation between the eggs — the same hen, and even if not the same hen, possibly the same farm; what happens if the farm gets hit by an epidemic? There's correlation between the baskets too — the same factory, and even if not the same factory, the raw material may come from bamboo in the same grove; what happens if that grove is overrun by borers? So the best approach isn't to scatter the eggs; it's to buy fewer eggs and allocate more into other things: some eggs in one basket, a few fish in a bucket, some vegetables in a bag, some meat in a fridge — and ideally these items, and the containers holding them, have zero relationship up and down their whole supply chains. So when you allocate assets from the risk angle, the core metric is the correlation between risks — and that correlation must be low.

Another example. You buy an apartment in Beijing and another in Guangzhou — are these two assets highly correlated? Actually, very: both are first-tier cities, both directly exposed to real-estate policy. But buy one apartment in Guangzhou and another in Australia, and the two assets have very low correlation. Domestic property policy has little effect on Australian home prices, and Australia's wildfires won't touch your Guangzhou apartment. So the All Weather strategy allocates across assets with extremely low correlation — when one part runs into trouble, the effect on the others is very limited, avoiding a "one falls, all fall" wipeout to the greatest degree possible.

Someone will ask: economies have cycles, every market fluctuates, assets always rise and fall — what if the falls outweigh the rises? This is where the All Weather strategy gets most brilliant. Remember the hedging we talked about? When we hedge, we slice up the risk in front of us, wash out one part and keep another. Bridgewater's All Weather strategy uses a higher-dimensional metric: the cycle. It hedges the economic cycle — hedging economic upswings against downswings, hedging high-inflation cycles against low-inflation ones. It wants to iron the economic cycle flat. How does it do that? The logic is risk parity. So on the market, the All Weather strategy is also called the risk-parity strategy.

Any second-grader knows that in an economic upswing stocks do better, and in a downswing bonds do better; in a high-inflation cycle commodities are the safer bet, and in a low-inflation cycle stocks and corporate bonds both do well. In one line: in any given economic cycle, there are always a few asset classes that outperform the rest. Bridgewater's method is to map out these four macro environments, allocate 25% of the risk to each — that's Bridgewater's classic four-square grid, with equal risk in each cell — and then, within each cell, allocate the assets suited to that particular environment. So one fund splits into four blocks, and each block holds different things.

And how do you allocate within each cell? Again, keep the risk equal. Say I buy 100 of stock and 100 of bonds; the stock's risk far exceeds the bond's. The same 100 dollars carries different risk, so at this point the assets in the cell aren't equal in risk. No problem — you can add leverage to the low-risk asset, bringing its risk and return up to a level comparable with the stock. What's the result of that operation? Every block within the same cell has the same risk level — and because the four grid cells each bear 25% of the risk, every single block across the whole strategy ends up at a consistent risk level. Let's see how this allocation actually performs.

(Figure in original.)

Suppose the economy shifts from upswing cycle 1 to downswing cycle 2, while inflation cycles 3 and 4 hold steady for now. Then the returns in cells 3 and 4 — beta3 and beta4 — stay unchanged, because the inflation cycle didn't move. Meanwhile beta1 in cell 1 falls — say stocks drop, dragging beta1 down by X overall — but beta2 in cell 2 rises, because in a downswing bonds rally, up by Y. Remember risk parity from before? Every asset in every cell has the same risk index; the bonds are leveraged, so their risk and return match the stocks' — which means X = Y.

And here's the magic: no matter how the economic cycle shifts, this strategy always steadily feasts on the market's beta without any big swings. In the face of market uncertainty, it holds remarkably steady. As the saying goes: however the winds and waves rise, sit calm as a rock in the fishing boat. We often hear analysts say: in an upswing we should do this, in a downswing we should do that. But Bridgewater's All Weather cuts across time and space, immovable — it deeply understands cycles without trying to time cycles. That's why Bridgewater's founder Dalio says: my view on risk is neutral, because the cycle is irrelevant to him — come who may, it's got nothing to do with me.

So how do these strategies perform? First, they're obviously far better than sitting on cash — a superb hedge against inflation — and they even beat the index most of the time. Excellent, in a word. In the 2008 subprime crisis, the market taught a lot of people a hard lesson, and the All Weather strategy performed flawlessly, making its name. From then on, more and more people realized that fund managers might put on a whole ferocious show of maneuvering and still fail to beat the market — so the people who can genuinely earn alpha money are few, since over the long run individual effort matters less than the march of history. And so the risk-parity strategy grew ever more popular, with all sorts of money — including pension funds and the Saudi sovereign wealth fund — pouring in, and more and more imitators and similar strategies appearing on the market. The total size of these funds grew enormous. By 2018, the US alone had reached US$1.5 trillion.

We set out to explain the logic of the crash, and now we finally get to the main event — though in fact all the clues are already buried in what I just laid out. Can you see where this strategy's problem lies? When only Bridgewater ran it and the scale was modest, there was no issue. But quantity turns into quality.

1. Identical asset allocation. The All Weather strategy and its many imitators run the same logic, and since everyone chases beta, and low-correlation assets are only so plentiful across the whole globe, these strategies are severely homogeneous — their asset allocations look nearly identical.

2. Enormous scale. With so many imitators, this class of strategy commands a huge footprint on the market. Big alone isn't scary — but big plus identical means everyone's moves are perfectly in lockstep: when it's time to buy something they all buy, when it's time to dump they all dump.

3. Heavy reliance on funding. Remember, it's leveraged — and leverage means heavy reliance on funding.

4. Programmatic trading. Quantitative trading, model-driven, and highly sensitive to volatility.

5. Risk parity. The model requires all assets to have comparable risk coefficients.

Let's feel out these last two points as we go. The risk-parity strategy performs fine when the seas are calm. But when we get to late February this year, the black swan flies out, the market starts shaking violently, and volatility spikes. Imagine these lumps of modeling clay are the risk-parity strategy, each strip representing one asset class.

1. The leverage problem. Where does the leveraged money come from? From banks. During a crisis, banks will surely pull liquidity back to protect themselves, so some of the leverage shrinks. Based on the risk-parity model — which requires all risk metrics to stay uniform — every asset has to take a cut. Chop. Sell.

2. Crude oil crashes, and commodities are again exposed to risk. Per the strategy, every asset has to take a cut. Chop. Sell.

3. Germany, Britain, Korea, Italy and others announce partial short-selling bans, so part of the strategy can no longer be used. Every asset has to take a cut. Chop. Sell.

4. When the market's in trouble, everyone is short of cash, so fund investors start redeeming. For example, rumor had it the Saudi sovereign wealth fund redeemed heavily from Bridgewater. The motive is anyone's guess, but if an investor wants to redeem, you have to do it — trimming positions proportionally across the board. Chop. Sell.

5. The Fed cuts rates sharply, and now the model can't take it. The model has variables pegged to the benchmark rate — the risk-free rate, say — and then a variable suddenly drops to zero, the balanced assets go unbalanced. Chop. Sell.

Terrifying, isn't it? One asset has to be sold, and all the rest have to be sold too. The more they sell, the more volatile it gets; every asset class, sold off worldwide. The All Weather strategy allocated so many low-correlation assets across the globe — assets that shouldn't affect one another — but it was outsmarted by its own cleverness. Once the strategy grew large enough to turn around and move the market itself, it tied uncorrelated assets together; it created the correlation. They thought they'd successfully dodged the devil, only to look in the mirror and find they were the devil.

Over the past few weeks, these risk-parity strategies collectively cleared out more than half their positions, becoming the single most important fuse of the global cross-asset crash. That's also the biggest reason not only stocks and bonds fell together, but gold fell too, and even some assets tucked away in a corner, untouched until now, thrashed violently. Various markets already risk a spiral downward in a financial crisis — and this strategy is basically a hard kick to the market's backside, deepening the drop: the more it falls the more they sell, the more they sell the harder it is to sell, and market liquidity dries up. When the music stops, everyone stampedes for the exit, and the crush begins.

Back to the Fed. This time the Fed surely learned its 2008 lesson: don't drag your feet, be decisive, because by the time the tower is about to topple it's too late. So its moves were very big. Specifically: one, it cut hard; two, it went wild.

First, cutting hard. On the very day the Fed announced a cut to zero, US stocks hit a circuit breaker. Did the market really wet itself? Emotion was surely part of it — after all, the Fed had pulled out the big guns, and the market gave a shudder and a nod of respect — but absolutely not to the exaggerated degree the joke suggests. There's another important reason: the quantitative models. The All Weather strategy I discussed today is just one kind of macro quant; the market holds a great many other quantitative hedging models of every stripe. These program-controlled quant hedge funds now account for a full third of US stock trading volume. Their strategies are similar, their assets are similar, and most of them hold heavyweight index stocks, so when they dump shares their moves are perfectly in lockstep — hence the phenomenon of everyone slamming the market in unison. That's why circuit breakers trip so easily now. The most textbook case was the day rates went to zero: a flood of models had their stability shattered by the drastic cut, and everyone trimmed positions at once, tripping the breaker. So why did it surge again a couple of days later? Same logic: the models had chopped their risk, so there was cash on the books, and add in the various stimulus measures — when they bought back in, they did it in lockstep, all at once. That's an important reason for the market's schizophrenic ping-ponging.

Second, going wild: bypassing the banks to inject blood straight into the market. After the 2008 crisis, the Obama administration rolled out a stringent financial-regulation package, the Dodd–Frank Act. Among other things, it strictly limited how much bank money could flow into hedge funds and private equity, and it also barred banks from proprietary securities trading. The rationale is simple: you're a bank, borrowing cheap money from the Fed, and then taking that money to play the stock market — that doesn't fly, does it? This effectively sealed off the banks' path to the secondary market. Good intentions, nothing wrong with them. But it's a double-edged sword. At the critical moment, when the Fed wants to open the taps, the water can't get out — the banks don't want the money, don't want it and don't dare take it. As the market deteriorates, with corporate bankruptcy risk and fund-blowup risk both climbing, banks will of course protect themselves first and pull liquidity back — and on top of that, the Act and the risk metrics constrain them. So even if you flood money to the banks, they can't pass it on.

Which lets us solve the dollar's strange behavior lately. There's a great dollar black hole in the US market. On one side, the money in the banking system can't flow out, so the interbank dollar lending rate is actually stable; on the other side, the market is left to yank dollars back from all over the world with even more force, driving a great dollar inflow and sending the exchange rate soaring. That's why the dollar has recently behaved so counterintuitively in both the money market and the FX market. The Fed saw this too, so it stepped onto the field itself, bypassing the banks — directly buying Treasuries, buying bond ETFs, buying MBS, pulling out the PMCCF, SMCCF, TALF and every other fancy tool like Doraemon reaching into his pocket. This kind of operation was only used during the Great Depression of the 1920s and the 2008 subprime crisis, which tells you just how anxious the Fed is. Ultimately it all comes down to this being different from 2008: the dollar shortage isn't in the banking system, it's in the market; not on Wall Street, but on Main Street. So this crisis's mode isn't a transmission from finance to the economy either — it's a transmission from the economy (the pandemic) to finance, and then feeding back into the economy.

The financial system is extraordinarily complex, and the cause is never one-sided. But today we can sum it up like this: the macro quantitative models, with the All Weather strategy as their poster child, were the single most important driver of the horizontal cross-asset crash in this crisis, accelerating and amplifying the transmission, and they were also one of the drivers of the vertical dumping and pumping. Together with ETFs and other quant models, they staged this historic drama of back-to-back circuit breakers and manic up-and-down swings — and amid this violent turmoil, the dollar's hegemonic position further accelerated the siphon effect, pushing financial markets the world over right to the edge of the cliff.

The financial system is forever more fragile than we think; even the most perfect tool has its day of failure; even the best regulation has its flaws. What now? Nobody knows. In an unprecedented sea change like this, no one has enough experience. As the line goes: the only lesson humanity ever learns from history is that humanity never learns from history. To witness history you needn't go get run over by the wheels of history — but the watching must go on.

Two keywords for what's next: debt. Debt has two meanings here. First: the most central problem of this US bubble — corporate debt. If this market doesn't hold, a financial crisis could very well turn into an economic one. The second debt: US Treasuries. The dollar flood and unlimited QE — where does all this money come from? Print it or borrow it. If it's printed, who ultimately foots the bill; if it's borrowed, who do you borrow it from. As the largest holder of dollar reserves and the second-largest holder of US Treasuries, we have to keep a close eye on this question.