Over the past 2 weeks, my colleagues and I have been thinking about how the next bear market will play out. These are just some thoughts, and they will not impact our trading. We trade the S&P 500 using our quantitative models.
*We don’t define “bear market” as the classic “20% decline”. That is an arbitrary definition that makes no sense (why 20%? Why not 21%, 22%, 23%…..). Based on our model, only the 1929, 1937, 1969, 1974, 2000, and 2007 market declines were bear markets. A 20% decline sits firmly within what we call “a significant correction”. These definitions are important because our medium-long term model predicts bear markets and significant corrections. There’s a massive difference between a 50% decline and a 20% decline.
Before we share our thoughts on how we think the next bear market (e.g. 40-60% decline in the S&P) will play out, I’d like to mention an article that economist Tim Duy wrote on Bloomberg.
Tim Duy’s key points
The title of his article says it all “This economic expansion will end with a fizzle, not a bang”.
We somewhat agree with this point.
In both of the last two cycles, there was a sizable imbalance in the economy that extended beyond financial assets themselves. So far, the current environment lacks such an imbalance. That suggests the expansion ends with more of a fizzle than a bang.
We somewhat agree with this point (and somewhat disagree). More on this later.
Investment activity illustrates the lack of imbalances in the economy relative to the last two cycles. In both cases, climbing asset prices filtered through to the real economy in the form of a growing share of investment spending. The information technology bubble of the late 1990s resulted in a surge of Nonresidential Investment Spending. Meanwhile, the housing bubble of the last decade shows up in Residential Investment.
Recall recent concerns that the tech industry was caught in the throes of another bubble. But if it was a bubble, it has been quietly deflating since 2015. If such a bubble is to be found anywhere, it would be in California, the epicenter of the technology industry. And note that job growth in the computer-design sector has now petered out.
The current expansion is most likely to end the same way. The economy doesn’t appear to be under the influence of a substantial imbalance that would end like the past two cycles. Unless such an imbalance arises, this expansion will end in a recession much like the mild one of 1990-91 brought on by a Fed that tightens too much or doesn’t loosen fast enough. Much more of a fizzle than a bang.
How we think the next bear market will play out
First of all, note that it is way too early to make any meaningful prediction as to how the next bear market in stocks and economic recession will play out. Based on our model, this bull market still has at least 2-3 years left. In other words, the next bear market will begin in late-2019 to 2020 at the earliest. This is not a fixed date. As new data comes in, our model’s target date will change as well.
Keep in mind that Tim Duy tends to be too optimistic. He successfully predicted the last housing downturn in 2005, but he predicted a soft landing for the economy. He was wrong. Like most economists, Tim tends to underestimate the domino effect in the economy.
What we agree with Tim Duy on:
We don’t think the next economic recession will be a severe one like 1980-1982 and 2007-2009. However, we think it will be more severe than the recession of 2001.
Tim Duy’s argument has 1 fundamental flaw, which he alluded to in the last paragraph. It’s basically saying “if the economy were to enter into a recession TODAY, the recession will not be a severe one because there is no economic excess”. On this front, he is right. Here’s a piece of economic and investment wisdom: it’s hard to die by jumping out the 2nd floor window.
However, we think the current economic expansion will continue for at least another 2-3 years. Historically, economic growth tends to accelerate in the final years of an economic expansion. So perhaps there will be a “sizable imbalance on the U.S. economy” (Tim’s words) by the time 2019-2020 comes around. This is data-dependent.
But on the other hand, we are less pessimistic than some of the ultra-bears. Some people argue “the next recession will be a Depression, because the government has run out of bullets and will not be able to save the economy”. We agree that the government has run out of bullets. Fiscal stimulus will be a very limited option next time because the federal government’s debt as a percent of GDP is much higher than it used to be. QE is becoming increasingly ineffective not just in the U.S., but around the world.
However, not all recessions need government-help to bottom! Only recessions that are accompanied by systematic crisis need government help. Systematic crisis tend to be caused by a massive stock market or real estate bubbles that collapse.
- Today, stock market valuations aren’t as high as 2000.
- Real estate valuations aren’t as high as 2006 either. The real estate “bubble” that pundits refer to are primarily confined to major U.S. cities. The real estate bubble was much more widespread in 2006.
Here’s a chart illustrating the current tech “bubble”. When you account for inflation, tech stocks aren’t extremely overvalued. And keep in mind that stocks go higher and higher over the long run, even when adjusted for inflation.
Some investors argue that bonds are in a bubble. But bond prices will not collapse unless interest rates soar to high-single digits. Interest rates will not soar unless inflation soars. And although inflation will go up over the next few years, we don’t think inflation will soar.
So if there’s no crisis, then the economic recession can end on its own, with or without the government’s help. Hence, the bear market in stocks can end on its own, with or without the government’s fiscal/monetary stimulus. Most historical bear markets and recessions ended without any MAJOR fiscal or monetary stimulus – 1970, 1974, 2003, etc. (Bush’s tax cuts and fiscal stimulus in 2001/2003 were relatively minor compared to Obama’s.)
The good thing is, our model can predict whether the recession will be a long one or a short one within a few months of the recession’s start date.
The next bear market’s driver.
Historically, every bear market in equities had a main driver.
- 1929: a financial crisis (like 2008) combined with very poor fiscal policies (Hoover raised taxes like crazy in the midst of an economic downturn).
- 1937: contractionary fiscal and monetary policy when the economy was extremely weak (unemployment rate was 15%).
- 1969: a massive contractionary fiscal policy when the U.S. was winding down Vietnam War-spending. (The U.S. went on a massive fiscal expansion via defense spending in 1966-1968).
- 1973: a massive energy/oil shock. (Manufacturing was still important in the U.S. at the time).
- 2000: tech bubble bursts
- 2007: financial system collapse (banks).
So the question is, what will drive the next bear market in stocks? We don’t think insanely high valuations will drive the next bear market. Valuations in 2019/2020 will not be as high as in 2000. Here’s Tobin’s Q. Notice how 2000 sticks out like a sore thumb.
Instead, we think the main driver for the next bear market in equities will be U.S. (and global) debt.
The federal debt-to-GDP ratio tends to become a problem around 120-130%. The problem is compounded if interest rates rise significantly (interest rates = the government’s cost of servicing that debt). Today:
- U.S.’ debt-to-GDP is 104%
- Italy’s debt-to-GDP is 132%
- Spain’s debt-to-GDP is 99%
- Greece’s debt-to-GDP is 179%.
- France’s debt-to-GDP is 96%
- UK’s debt-to-GDP is 89%
- Germany’s debt-to-GDP is 68%
U.S. debt (private and public) has surged since 2009. However, this has not caused a major problem thusfar because long term interest rates are low!
The Fed plans to start paring down its $4.5 trillion balance sheet in late-2017. Do not expect yields to soar instantly. Rates will rise slowly at first because the Fed will slowly sell bonds. The rise in interest rates will not result in a debt crisis until interest rates rise to a certain point:
- When the U.S. government’s servicing costs are too high. Interest on debt payments currently stand at 7% of the federal budget.
- When interest rates exceed corporate profit margins.
The Federal Reserve holds a lot of long term bonds (thanks to Bernake’s Operation Twist). The Fed will slow down rate hikes and sell its long term bonds. Rate hikes primarily impact short term bonds while a balance sheet reduction primarily impacts long term bonds. So it’s entirely possible that long term rates will go up faster than short term rates in the next few years. In this case, the yield curve might not invert before the stock market begins a bear market!
(In our previous post, we mentioned that the yield curve always inverts before the stock market’s bull market ends. But never before in history has the Fed needed to sell such a massive amount of long term bonds.)
Debt “crisis” tend to be different from financial crisis. Debt crisis do not result in a vertical crash like 2008. Instead, we think the next U.S. debt crisis will come in multiple waves. After all, the European debt crisis came in multiple waves from 2010-2012. Hence, a bear market like the 2000-2003 one with multiple waves is more likely.
*We do not think the entitlements “crisis” (social security, healthcare) will be the primary driver for the next bear market. The U.S. entitlements problem is a secular problem. It will slowly play out over years, if not decades. The next debt crisis is much more acute.
We have no idea how the next bear market and economic recession will play out. It’s way too early. Trying to make predictions for 3 years in the future is fun, but not useful for trading. When the time comes, our model will tell us:
- Has the bull market topped?
- Has the economic expansion ended?
- Has the bear market bottomed?
- Has the recession ended?
For more timely thoughts, read How bearish is the stock market’s low volatility and How the next stock market correction will play out.