When the NBER’s Business Cycle Dating Committee draws the boundaries on the current recession, it’s unlikely to stand out as an especially long one. In fact, by the time the committee publishes the official start date, it could be past its end date.

Why?

Because it’s front-loaded. Spending has dropped so sharply in such a large portion of the economy that many types of activity have nowhere to go but up. And once activity starts increasing, even from nothing, that’s expansion, not recession.

But the eventual business-cycle dates tell us little about our current situation. We could hit bottom in 2020 but then expand so weakly that we don’t restore vitality for several years. So let’s consider how the economy might unfold over a fixed horizon—say, three years from 2020 to 2022—rather than fixating on business-cycle dates.

First, I’ll look at the reasons why our situation is really, really bad, and then I’ll consider why it might not be that bad, after all. I’ll benchmark my calculations against the post–World War II period, but especially against the economic destruction from 2008 to 2010.

Why this is the worst economy since the Great Depression

I have six reasons.

Reason #1: This is a “double-recession.”

Consider that our last ten recessions were shaped mostly by four categories of spending: business equipment, commercial real estate, home building and consumer durables. If you isolate only the ups and downs of those four categories (but throw in changes in inventories to account for milder, inventory recessions), your partial business-cycle history would be almost indistinguishable from the actual history.

Moreover, those four categories typically amount to less than a quarter of the economy. In 2019, they composed 19.5% of GDP, as shown below:

• Business equipment:  5.8%
• Commercial real estate:  2.9%
• Home building:  3.7%
• Consumer durables:  7.1%

So a 19.5% chunk of the economy explains the first part of the double-recession, and we know it’s currently recessionary because the usual precursors are back—collapsing business profits, tightening loan standards, widespread job losses and rising delinquencies. With the usual precursors in place, we can expect a sharp contraction in all four categories noted above.

But the second part of the double-recession is separate. Consider the 2019 GDP weights for the additional spending categories below, totaling 11.3%:

• Transportation services:  2.2%
• Recreation services:  2.7%
• Food services and accommodations:  4.8%
• Gasoline and other energy goods:  1.6%

Now we’ve reached the piece that’s completely new—it has no precedent in past business cycles. Each of the four categories shown above has contracted far more than ever before. In each case, activity is only a fraction of what it was just three months ago—probably less than half, and maybe even less than a quarter. Considering the severity of the contraction, together with the GDP weights, the destruction in these items alone is enough to establish a recession, even without the usual fixed investment and consumer durables categories discussed earlier.

So that’s what I mean by a double-recession. The first part includes the fixed investment and consumer durables categories (totaling 19.5% of GDP). The second part includes the additional categories (totaling 11.3% of GDP) that imploded during the last two months, even as they’re normally only bit players in the business cycle.

Reason #2: Pandemic-related business costs could last for years.

Businesses will have to manage through some combination of the following:

• Migration to more secure but higher cost suppliers (in response to supply chain fragilities exposed by the pandemic)
• Measures to facilitate social distancing, including larger business premises in some cases
• More frequent and thorough cleaning of business premises
• Personal protection equipment for employees and, in some cases, customers
• COVID-19 testing costs
• Potentially greater contributions to employee health insurance (when insurance companies build COVID-19 into their cost structures, premiums can only go up)
• Potentially greater absenteeism (employees being told to stay home with even mild illnesses, employees relying on public transportation facing greater challenges getting to work safely)
• Potential work stoppages when employees test positive
• Potential hazard pay

We can only guess how widespread and persistent these costs will be. But across the whole economy, they’ll surely add to a significant, positive number. They’re bad news for business profits, inflation and probably both (more on inflation in a moment).

Reason #3: The Fed only had two bullets in the interest-rate chamber (the two March rate cuts).

After cutting the fed funds rate in March from 1.6% to just above zero, the Fed can’t reduce it further without entering the Twilight Zone of negative rates. (Sure, other countries have tried negative rates, but it’s still the Twilight Zone.) By comparison, here are the fed funds rate changes during the last five recessions, from business-cycle peak to business-cycle trough: –4.8%, –9.8%, –2.0%, –3.2% and –4.1%. So this year’s change of –1.5% is only a fraction of the interest rate stimulus we normally see in recessions.

Reason #4: Bankruptcies could be more severe than in any other post-WW2 recession.

I wrote “could be” because we don’t know for sure, but record bankruptcies seem consistent with three things we do know. First, business shutdowns within the 11.3% of GDP noted above (the second part of the double-recession) will surely result in record destruction in that particular portion of the economy.

Second, activity has already contracted more sharply than at any time since the 1933 national bank holiday, and in that instance, widespread business stoppages only lasted a week.

Third, nonfinancial businesses are loaded up with record amounts of debt. As of Q4 2019 and relative to GDP, nonfinancial businesses were more indebted than ever before on a gross basis (74% of GDP), and they also carried more debt than in any prior expansion after netting out interest-earning assets and cash (55% of GDP). In short, nonfinancial businesses could hardly have entered this crisis with a riskier aggregate balance sheet.

Reason #5: Meet the zombies—next generation.

Stimulus programs are helping forestall economic destruction, but they’re also propping up companies that wouldn’t be viable without cheap financing backed by the Federal Reserve and Treasury Department. Some of those companies will still go bust, despite public support. Others will become zombies, dependent on loans that can only be paid back by obtaining more loans. To those who pointed to the zombie companies of the last decade as one reason for a less-than-vibrant global expansion, you haven’t seen anything yet.

Reason #6: Inflation risks are unusually high for a recession.

As noted above, the pandemic has lifted business costs by adding procedures and complexities that didn’t exist before. Rising business costs damage profitability, at first, but should eventually have some effect on inflation. And that’s not all. Inflation is normally a policy choice (either intentional or inadvertent), and policy makers are more inclined to risk it than at any time in the last four decades. Notably, current policies include direct injections of Fed-financed spending power into the Main Street economy. Moreover, those injections appear to be augmenting rather than just supplanting spending power supplied by commercial banks. (I’ve shown several times that past QE programs merely substituted Fed financing for commercial bank financing, without having a significant effect on the total.)

Note that I’m not using the flawed logic of monetarist economists who predicted rising inflation during the Fed’s earlier QE programs, nor did I join those predictions (just the opposite, as shown here, for example). Also, the inflation outlook is hardly one-directional, since certain items, such as housing costs, are now less likely to inflate than they are to deflate or remain stable.

But the factors discussed above should threaten the benign inflation of recent decades. After remaining below 3% for the last 24 years, an increase in core inflation to just 4% would be a major event. And if we get there, fiscal and monetary policies would become more challenging, to say the least. After many years of disinflation, policy makers would again be forced to choose between snuffing out inflation and sustaining growth.

Why this isn’t the worst economy since the Great Depression

I have six reasons, once again, the first three of which compare 2020 to 2008.

Reason #1: The big-4 “home” risks—home prices, home mortgage debt, home building and home equity extraction—are relatively nonthreatening.

The mid-2000s housing bubble brought unsustainable prices alongside unsustainable growth in mortgage debt, home building and home equity extraction. Just before the pandemic, by comparison, house prices and housing activity appeared sustainable. Here’s a rundown of 2019 data versus “peak” housing boom data:

• Home prices:  Grew 3% in 2019 versus –19% in 2008 (after peaking in mid-2006)
• Home mortgage debt:  49% of GDP in 2019 versus 72% in 2007
• Home building:  3.7% of GDP in 2019 versus 6.6% in 2005
• Home equity extraction:  1% of DPI in 2019 versus 8% in early 2006 (according to Bill McBride’s calculations)

We can link each of the items above to a significant drop in household spending power or housing activity in the 2008-9 recession and the years that followed, whereas the data show much lower risks today. Clearly, the big-4 home risks are unlikely to wreak as much destruction in the current recession as the destruction caused by the housing bubble.

Reason #2: Sterilization? What’s that?

In 2008, the FOMC fretted for months before dropping long-established central banking orthodoxies. But such lengthy deliberations have long since gone out of style. The committee now crams money without hesitation into every financial-sector crevice that appears to be leaking. The new policy “normal” invites both moral hazard and zombification of wide swathes of the economy, as noted above. But the immediate upside is significant—the Fed’s interventions short-circuited the financial crisis that appeared to be unfolding in March.

Reason #3: Banks have more capital than they did in 2008.

We’ve all heard the story about the better capitalized banking system, and it’s true. But higher capital ratios won’t stop banks from slowing or even shuttering their lending operations. (They’ve already done that.) So the capital cushion is larger, and that’s nice to have, but it won’t save the economy. The main benefit is that measures to bail out the banks won’t need to be as large as they would otherwise be.

Reason #4: Some areas of the economy are seeing stellar demand.

I noted above that spending has evaporated like never before in portions of the economy that total 11.3% of GDP. Now consider three other types of spending:

• Food and beverages purchased for off-premises consumption (4.8% of GDP)
• Other consumer nondurables (5.6% of GDP)
• Health care (11.5% of GDP)

Solid spending in these areas, which total 22% of GDP, doesn’t negate the destruction in the transportation, recreation, restaurant, hotel and energy sectors. But it’s important to recognize that some of the spending lost through health fears and business shutdowns is being redirected, not extinguished. It’s flowing strongly into other parts of the economy. And the jobs market demonstrates that point—many recently jobless workers are finding new positions at Amazon, Instacart, CVS or one of a smattering of other companies whose outlook has brightened. So the double-recession I noted above might net out to more like a recession-and-a-half.

Reason #5: Furloughs, not layoffs.

Of the newly jobless workers who don’t find jobs elsewhere, many remain on their employers’ payrolls, retaining certain benefits but not working or receiving wages. One survey shows that 78% of employees who lost their wages due to the coronavirus expect to return to their former jobs. That might prove more hopeful than realistic, but it’s a less bearish recession story than the more typical story of companies slashing labor unconditionally.

Reason #6: Helicopter money!

Now for the elephant in the room. Fiscal policy makers are intent on providing “what it takes” to overcome the crisis. For that, they’re tapping into the Fed’s unlimited capacity to finance government spending with newly created money. They’re tapping it like never before. To highlight just two data points:

• Roughly half of unemployment claimants will have more income than they had while working (through July, at least), thanks to an extra $600 weekly of CARES Act benefits on top of their normal state benefits.
• Millions of working Americans will also make more than they would have without COVID-19, thanks to CARES Act stimulus payments.

It shouldn’t be surprising that the survey linked above shows people feeling better about their finances than they did a month ago, despite weekly unemployment claims averaging over five million between the two survey dates.

So helicopter money gives us yet another surreal and unprecedented development to ponder. In the short-term, it’s certain to blunt the pandemic’s economic impact. In the long-term, we’ll face consequences, but I won’t delve into that in this article. I’ll only suggest tuning out pundits who claim that “advanced” nations with their own currencies can drop helicopter money without repercussions. In fact, the advanced nations of today reached their advanced status long ago after enduring tumultuous periods of fiscal profligacy, learning from those experiences, and then maintaining relatively sound finances thereafter. And if they didn’t learn from experience? Well, that’s one of the biggest reasons that many countries fail to advance.

(My book supports that argument with an examination of every recorded instance of governments accumulating a higher debt-to-GDP ratio than America’s debt-to-GDP as of 2018. For anyone interested in the general idea without the historical detail, I published an excerpt here.)

Conclusions

The eventual COVID-19 wreckage pivots on many unknowns, and future policies are among them. But the biggest unanswered question—at least when it comes to the economy—is this:

For how much longer will the pandemic prevent vulnerable businesses from operating profitably (or operating at all)?

Optimistically, new COVID-19 cases will descend downward until they hit bottom in a few months, allowing businesses to restore profitability. That’s the scenario the President’s task force includes in its slideshows—it shows virtually no new cases by July. And the more political voices on the task force have reinforced that message, playing up the idea that we’ll be back to normal by this summer. If their prediction proves accurate, the economy should perform better in 2020–22 than it did in 2008–10, for these reasons:

• With a COVID-19 resolution in the summer, the housing market would be in far better shape than it was in 2008.
• The financial sector would recover relatively quickly—banks would still be cautious but not as cautious as they were during and after the Global Financial Crisis.
• Many depressed businesses would bounce back at least partially and rehire furloughed employees.
• Some businesses boosted by the pandemic would continue to thrive.
• Exiting the recession, household spending power would be unusually strong, thanks to the recession’s short duration as well as generous government handouts.

So that’s the outcome we’re hoping to see, but it has an obvious weakness. That is, it presumes the coronavirus remains dormant after the economy restarts. A different theory says the virus revives whenever it finds an opening. Evidently, that’s a common feature. Epidemics tend to attack in waves. Until vaccines become available, the challenge in snuffing out this epidemic is that it only takes a handful of infected people going about their normal lives to reseed it.

In other words, a future resurgence of COVID-19 seems the most likely outcome. It’s the scenario many experts warn us to expect, and not just any experts but the ones who’ve been most accurate to date.

Where does that leave the economy?

The worst case combines a historically deep recession with a disappointing recovery that feels more like continued recession. If future COVID-19 waves prove as dangerous as the first wave, the recession could be an early 1980s–style double-dip. But other possibilities are less severe. For example, medical discoveries could make the virus less risky, restoring confidence in normal business activities. (Note that Dr. Fauci was citing “quite good news” on remdesivir trials as I write this.)

So the possibilities run from one extreme to the other. We need to be ready for anything, unfortunately, from a mid-2020 rebound to a prolonged crisis more severe than any since the Great Depression.

With hoops “out” and exponentials “in” (referring to March Madness, the 2020 pandemic definition), there’s a new, customary disclaimer on economics and financial sites. Mine says that I, too, knew nothing about infectious disease modeling only two months ago. But I’m catching up, just like everyone else. By now, I might have reached a “Dummies Guide” standard, and I’ll keep this article at about that level.

So with that preface out of the way, I’ll first offer a health warning of sorts about a type of COVID-19 chart that’s popular with market bulls. Here’s a version that appeared in the New York Times:

And here’s a second version that uses the same axes but with different data:

The second chart bounced around Twitter and then appeared on at least one financial site—to support a breezy message that people should stop panicking and start preparing for a surprise melt-up in stocks.

And why do the charts need a health warning? The problem is the horizontal axis—showing a statistic that epidemiologists call R0—which is presented as though it’s the only information you need to have to understand how widely a disease spreads.

In fact, modelers who actually use R0 stress that you can’t summarize the spread of an epidemic with that single input. They say that R0 changes with time and place—it’s not actually a constant. Also, it can approximate the spread of a disease only if the entire population is vulnerable. Critical factors such as prior immunities are handled through other model inputs, not R0.

More to the point, we already know how widely COVID-19 is spreading, and we can readily compare that information to every other disease that’s already come and gone. So let’s stop pretending that R0—an incomplete and imprecise contagion statistic—somehow does a better job than ripping off the blindfold and watching what’s really happening, in real time.

The reality-based approach tells us that SARS, to pick one comparable, peaked at about 8000 cases in 2003. By contrast, COVID-19 is rocketing through the hundreds of thousands, certain to hit millions and by some forecasts billions. That’s a stark difference that seems highly relevant, and yet you won’t find it on the R0 chart, no matter how many versions you conjure up.

The eyes-open approach also says to be wary of claims that “we’re very close to a melt-up” or “the bottom may be closer than you think” or “we’ve probably seen the worst.” (Those are just a few of the bullish stock market predictions we’ve been seeing lately.) To explain why, I’ll first extend the analysis I shared in my last article.

Case Trajectory Update

For background, my March 2 article showed how the confirmed case trajectory could evolve if the most bearish epidemiological forecasts prove accurate. This time, I’ll set the forecasts aside and examine how the trajectory has changed in recent weeks.

I’ll exclude both China and Iran, because their reported numbers fail to match the on-the-ground evidence. I’ll then divide everyone else into two groups:

• South Korea and Singapore, which I’ll call the “most prepared” group (MPG). These countries implemented a variety of effective disease-containment strategies after suffering from other epidemics in recent years.
• The rest of the world, which I’ll call the “least prepared” group (LPG). That naming might be unfair to a few other well-prepared countries, but further regrouping wouldn’t materially change the conclusions.

Here are the confirmed cases for each group through March 21:

As of today, the MPG has successfully “flattened the curve,” whereas the LPG continues to see exponential growth. Here are the best-fit exponential curves for the LPG, calculated on a weekly basis from February 15 until March 21:

The curve steepened sharply in March, probably due to increased testing. But let’s take a closer look. The next chart shows the daily evolution for the “steepness statistic,” which is the part of the exponential function that determines the trajectory:

According to the chart, the case trajectory has passed through two stages since late February.

1. It steepened almost continuously as testing activity increased.
2. It then peaked on March 14 and started to flatten.

So that’s the past, but what about the future? How quickly, if at all, will the steepness statistic fall?

I doubt anyone can answer that with precision, especially with countries such as the U.S. only recently ramping their testing. But it sure seems as though the overall trend should continue downwards. With the extreme containment measures currently in place, it’s hard to imagine any other result. Or maybe I just don’t want to imagine a different result.

Either way, let’s be optimistic. Let’s agree that the March 14 peak delivers good news. It says the first derivative of the daily new case count is no longer rising day-after-day-after-day. That might be hard to conceptualize, but it’s exactly how these types of processes change direction. The first derivative turns downwards, and then the daily new case count turns downwards, and then the inflection point becomes apparent to all.

Implications for Stock Prices

Now for the bad news, which is that I’ve just exhausted my supply of good news, at least when it comes to stock prices. As the case trajectory flattens, we should see bursts of optimism that push prices higher, but probably not a market bottom. The problem is the two elephants that seem in no particular hurry to leave the room.

1. Businesses have shut down more suddenly and completely than ever before.
2. We have no idea when the business shutdowns will end.

Let’s say restaurants, bars, stores, factories, offices, schools, entertainment venues and transportation hubs reopen in about eight weeks. That seems possible, but it’s probably not sustainable. If political leaders remain consistent, any newfound infections would then send businesses right back out of work. In my county, for example, all nonessential businesses were told to close just as the confirmed case count climbed from 1 to 2.

In the meantime, we’re about to be bombarded with the worst economic data releases we’ve ever experienced. The economy is probably contracting even more sharply than it did during the 1933 national bank holiday, which is the only comparable I can think of. At a time when people withdrew cash from banks to make purchases, the bank closures brought much of the economy to a sudden halt. But the bank holiday only lasted seven days.

Others point to the 1980 and 2008-9 recessions, but in those instances the economy shrank organically. This time, it’s literally being locked down. I fail to see how those past recessions—or any past recessions, for that matter—can shed much light on the present one. And I can say the same thing about past pandemics, which have never before triggered worldwide, government-mandated business closures.

So instead of looking at, say, stock market returns in 1980, 2008-9 or during any other recession or major calamity, I expect the key to the market’s performance to lie in the answer to this straightforward question: At what point does our capacity to treat COVID-19 patients allow businesses to reopen?

Notably, at least one policy maker has pledged not to restart the economy until fatality risks reach zero. That approach might not be universal, but it doesn’t seem that far from consensus within the political class. Elsewhere, though, commentators are challenging that consensus, stressing connections between economic decline and public health. See, for example, this editorial by the Wall Street Journal and this commentary by a contributor to the CFA Institute’s Enterprising Investor blog.

Clearly, my question doesn’t have a single “correct” answer, and yet it might be the most important piece of the current policy mix. To stray briefly from financial to societal commentary, I hope policy makers develop their answers carefully. The WSJ might have summarized the predicament best when writing that “no society can safeguard public health for long at the cost of its economic health.”

Now bringing the discussion back to stock prices, the answer to my question determines when the cloud of uncertainty finally lifts. It determines when investors can resume normal financial analysis, and when commentators predicting the next bull market might finally be onto something.

Bottom Line

No doubt, market volatility brings opportunities. Long-term investors are finding value as the market falls, and even short-term traders are finding buying opportunities as prices bounce wildly up and down.

But risk tolerance is paramount. For those unable to stomach losses on new holdings, my nickel’s worth of advice is to be cautious. I’ve shared my most optimistic chart—showing a declining case trajectory—but I don’t expect it to turn the tide in the stock market. For that, we’ll need more clarity about when we can once again claim “the business of America is business.” That’s at least weeks and possibly many months away.

As recently as two weeks ago, it wasn’t clear which infectious disease experts had the best handle on COVID-19’s likely path.

Among the optimists, one operation ran a model in February that showed the following maximum case counts by the “end of the epidemic.” (I’ve included the running case tally as well, using data from the Johns Hopkins dashboard.)

With the benefit of hindsight, these well-intentioned forecasts were clearly too low. But I’ve shared them anyway because they support the premise for this article—that it seems time to listen to those experts who were telling us all along that a pandemic is inevitable. In comparison to the misplaced forecasts above, the experts who expected a global health crisis can point to the chart below as validation:

So recent events seem to settle the question of which experts have the best grasp on what’s happening.

And recent data also provide another way to look into the future, one that doesn’t require complex epidemiological models. I’m suggesting the following:

• Ignore the case data from China, which falls somewhere between glitchy and grossly incomplete.
• Fit an exponential curve to the daily number of confirmed cases outside China.
• Create a “base case” by following the curve’s trajectory until the pandemic reaches proportions expected by those experts who’ve been most accurate to date.
• Experiment with flatter trajectories to map the interplay between the pandemic’s intensity and its duration.

This approach offers three advantages. First, it produces plausible scenarios that are easy to rerun and calibrate to the latest data. Second, it requires only a few assumptions. Third, those assumptions are transparent and easily understood.

More to the point, while the approach doesn’t substitute for complex modeling, it gives reasonable answers to a few important questions. Essentially, I’m trying to answer the following:

• How many new cases could appear each day (the intensity) at the pandemic’s peak?
• How long will the pandemic last?

The endgame, in numbers

So let’s crunch the numbers. I’ll start with the predictions of Harvard epidemiologist Marc Lipsitch, who said in mid-February that he expected 40% to 70% of adults worldwide to be infected. I’ll apply that prediction to the world ex-China, using the following initial calculations:

1. Subtract 1.8 billion under-15s from a world ex-China population of 6.4 billion. That’s 4.6 billion adults.
2. Apply the 40% lower end of Lipsitch’s estimated infection rate to those 4.6 billion adults. That’s 1.8 billion infections.
3. Allowing that many cases will never be recorded, reduce the 1.8 billion infections to an even 1.0 billion. (Imperial College recently estimated that only about a third of infections outside China are being recorded. That proportion—should I say optimistic estimate?—should increase with more testing, but it should still fall well short of 100%, considering that 80% or more of cases are mild.)
4. Assume further that the growth rate slows after the number of cases reaches somewhere between 10% and 50% of the eventual one billion. Mathematically, the growth rate has to slow, since a growing number of infections means a shrinking number of people susceptible to new infections. Applying 10% and 50% to one billion, we’re looking for the growth rate to slow after it reaches somewhere between 100 and 500 million recorded cases.

(Note that I’m assuming all infected survivors build immunity. If that’s wrong—and according to various reports, it could be at least partly wrong—that calls for a completely different analysis. It means the pandemic probably continues until a vaccine becomes available.)

Follow the curve

Now for the easy part. Here’s an exponential curve that fits the daily case data:

The next chart extends the fitted curve until late April, when it reaches the 100–500 million case count I estimated above. I’ve highlighted the point where the case count exceeds 500 million, which I’ll call “Half a Billion Day”:

To be sure, the first projection is a base case, not a likely one, and probably not even a realistic one. It simply answers the question: If the closest-fit exponential curve continues towards a 40% infection rate, when does the pandemic begin to run out of steam?

Moreover, the curve will need refitting as time passes. The actual case count might soar above it, and it will certainly fall below it, eventually.

One reason for the case count to overshoot the curve is that test kits are becoming more widely available and actively used. So the percentage of actual cases included in published statistics should rise.

On the other hand, the trajectory should eventually flatten, for at least four reasons. First, the propensity to test suspected cases should level off and might even decline in some countries as resources become stretched. Second, governments, businesses and other organizations are likely to impose stricter containment measures. Third, the virus will eventually run out of unprepared communities to attack while their defenses are low. Fourth, the pool of uninfected people will increasingly tilt towards those who take extreme precautions against infection.

To show how the eventual flattening could affect the pandemic’s path, I created a second projection with an inflection point. As of April 30, I toggled the trajectory from exponential to linear, freezing the number of daily new cases. And between March 16 and April 30, I gradually flattened the trajectory to make the transition smooth. In other words, the projection now has three phases:

• Up until March 15: Exponential growth along the fitted path
• March 16 – April 30: Gradual transition to linear growth
• After April 30: Linear growth, meaning the number of daily new cases remains at the April 30 peak

(Note that the March 15 timing for the transition to linear growth is subjective. The fitted curve steepened over the past week as testing activity increased, and testing activity continues to increase, so a transition towards linear growth seems unlikely during the next few weeks.)

The new curve pushes Half a Billion Day back to May 19. Labeling April 30 as “Peak Intensity Day,” here’s the picture:

Finally, the crux of the analysis is to examine the interplay between intensity and duration. I looked at eight possibilities for Peak Intensity Day, calculating the implications for both peak intensity (new cases per day) and Half a Billion Day.

Note that Scenario 1 fails to deliver a 40%–70% infection rate over the next year, which is how Lipshitz framed his prediction.

At the other extreme, Scenario 8 raises a question of whether it’s realistic for the pandemic to grow exponentially beyond Half a Billion Day.

So Scenarios 2 through 7 appear to be the most realistic 40% infection scenarios. They show the pandemic reaching peak intensity in the second half of April or the first half of May, with daily new cases running anywhere from a couple of million to just over fifty million.

Conclusions

At a first pass, if a 40% infection rate proves accurate (and I haven’t yet seen a convincing case that such an extensive global diffusion—whether 40%, 30%, 20% or in that general vicinity—can be prevented), the pandemic should continue to intensify for another couple of months. Daily new cases could reach into the tens of millions. Alternatively, daily new cases will flatten after reaching a peak of between two and fifteen million. But an early flattening would have the undesired effect of prolonging the pandemic. That’s just a mathematical observation—if COVID-19 passes through the population at a slower pace, it’ll remain active for longer before it runs out of targets.

As far as the economic implications, a 40% infection rate would mean COVID-19 persisting long enough to cause a global recession, and probably not a mild one. Call it somewhere between moderate and severe. In that scenario, we can only hope the shoes have already dropped—that people don’t get infected more than once, that the mortality rate doesn’t worsen, and that the health care sector moves quickly towards a vaccine.

More optimistically, those who stress that “this, too, will pass” will surely be proven right. The problem is that we don’t know when. We only know that the optimistic estimates of just two weeks ago are clearly wrong. But we can use that information—together with a month and a half of data—to explore the possibilities. We can cast a cold light on the pandemic’s sudden exponential growth.

“Ah, excuse me. Oh, will ya excuse me. I’m just trying to find the recession. Has anybody seen the recession?”

Ask that question in a roomful of forecasters, and you’ll hear plenty of reasons why the next recession is dead ahead: the inverted yield curve, the tariff war, weak PMIs, the global manufacturing downturn.

Events might eventually prove those recession forecasts to be correct, although I would say not until mid-2020 at the earliest, and a recession at that time remains just a possibility. I say that because we haven’t yet seen enough cause for alarm in the three areas that most reliably predict recessions. Before every recession, we see at least one, usually two and often every one of the following three precursors:

• Deterioration in the housing sector
• Restrictive public policies
• Significant damage to the real spending capacity of households, businesses or both

In other words, when trouble emerges across some combination of housing activity, public policies and real spending capacity, we’ll know to expect a recession. Trouble in one of those areas should put us on alert, whereas two or three would mean we should bank on it. So what’s missing from today’s popular recession narratives is adequate support from the “Big-3” precursors, and without that support, it’s probably too soon to bet on a recession. The U.S. economy always expands when the housing sector is stable, public policies are growth-supportive and real spending capacity is increasing. Simply put, no sign of the Big-3 means no recession.

But isn’t there a first time for everything? Can it really be so simple?

There is, and I don’t expect to convince anyone the economy is that simple without first providing some evidence, so I’ll continue. I’ll focus mostly on spending capacity, which is where I stray furthest from traditional, mainstream methods.

Why spending capacity?

Behavioral research, empirical data and casual observation all point towards households and businesses increasing their spending for as long as they have the capacity to do so. Changes in spending capacity predict changes in spending with remarkable accuracy, notwithstanding the Keynesian idea that spending follows the mysterious ebbs and flows of “animal spirits.” In fact, the spirits described by Keynesians might not be all that mysterious—they’re always present in some degree, they just happen to flow in proportion to spending capacity. They don’t disappear for no particular reason and then later reappear.

So I suggest closing your Keynesian textbook and looking instead to natural human behavior for clues about spending. Behavioral research tells us we’re naturally overconfident, believing our ventures will succeed with a certainty that defies the true probability of success. It also tells us we’re at least partially blind to certain obstacles to success, such as basic randomness. We’re naturally wired to have an illusion of control and an optimism bias alongside hindsight and confirmation biases, all of which encourage us to spend for as long as we have the capacity to do so.

But that’s not all. We’re also prone to a lack of self control that researchers have termed present bias and a tendency to spend like drunken sailors whenever in the company of other free-spending drunken sailors, thanks to our natural herding bias. I could go on, but you get the idea—once we consider human nature, it’s easier to appreciate why spending capacity is the economy’s driving force.

What Exactly Is Spending Capacity?

All that being said, I still need to define spending capacity, and my definition is broader than you might think. It starts with earnings—both household and business earnings—which of course help determine the resources available to be spent. It also includes risky asset prices, because spending depends partly on house price cycles and investment portfolio values. In fact, spending is more exposed to asset price volatility than ever before, with assets owned by households and nonprofits currently valued at 608% of annual GDP, compared to averages of 385% in the 1970s, 407% in the 1980s, 454% in the 1990s and 537% in the first decade of the 2000s.

Finally, there’s a third piece that’s usually overlooked, and I blame the economics profession for that. Spending depends not only on what households and businesses earn and own, as noted, but also on what they can borrow. And it depends not just on what they can borrow, but on what they can borrow from banks, in particular.

Why banks as opposed to other types of lenders?

Because banks are the only lenders that create spending power from “thin-air.” That’s not something you’ll learn in mainstream economics, which mangles the mechanics of money and banking, but if you’re trying to understand business cycles, it’s an essential fact. The key insight is that new bank credit expands the circular flow of income and spending (see the chart below), whereas other types of credit mostly sustain the existing flow by passing spending power from one party to another. 

In other words, only banks increase spending power on a net basis, because they can make loans without requiring prior savings from past income. Apart from a small allocation of bank capital, banks conjure loan proceeds from thin air—that’s the crux of what their charters allow them to do. The monetary expansion authorized by bank charters explains why new bank credit is 69% correlated with spending in the same period and 58% correlated with spending in the next period, whereas corresponding figures for credit financed by prior domestic savings (not banks) are negligible (see the chart below).

(As obvious as the charts above might be to those who’ve either worked as bankers or studied banking from up close, mainstream economists tell a different story, teaching that banks are mere intermediaries and only central banks determine the money supply. Those mainstream claims have been called out repeatedly by practitioners, heterodox economists and even central bankers—the Fed, Bank of England and Bank for International Settlements have all provided information refuting textbook money and banking theory—but to no avail. Schools continue to teach money and banking incorrectly at both undergraduate and graduate levels.)

What Do the Big-3 Say about the Next Recession?

So spending capacity depends on real earnings, asset prices and the availability of new bank credit. Getting back to the Big-3 precursors that always materialize in some combination (again, not necessarily all three at once) before recessions, we can expand the third precursor using the determinants of spending capacity. Here’s the expanded list:

• Deterioration in the housing sector
• Restrictive public policies
• Significant damage to the real spending capacity of households, businesses or both, which could mean any of the following:
  • Earnings that fail to keep pace with inflation
  • Falling real asset prices
  • Restrictive bank credit

We can then expand the list further with more subcategories and by adding the foreign sector, which I include as a less influential input but one that could contribute to a recessionary process. The expanded list includes separate readings for:

• fiscal and monetary policies
• household and business earnings
• house and stock prices, and
• bank credit conditions for households and businesses

Below are my assessments for each item, with comments on at least one of the indicators that support each assessment. (My email subscribers know that this is the same list that drops out of the 6-cycle forecasting approach described in my book, I’ve just tweaked the terminology to match the language used here.)

Housing sector: Not recessionary. Indicators such as new home sales and the NAHB housing market index show housing activity recovering nicely from a period of weakness in 2018.

Fiscal policy: Not recessionary. Government spending is growing at a decent clip in 2019, while taxes and net transfer receipts dropped to a new six-year low as a percentage of GDP in the first half of the year, which can only help household and business spending capacity. Also, July’s debt ceiling deal freed up more federal spending in the 2020 fiscal year.

Monetary policy: Not recessionary. I often use the yield curve slope as a guide to monetary policy, and today’s inverted curve might suggest that monetary policy is recessionary, but I’ve overridden that for three reasons: 1) the Fed barely lifted the inflation-adjusted fed funds rate in the 2015–18 tightening cycle, and therefore, fell far short of the typical recessionary tightening, 2) we’re now 11 months removed from the last rate hike, and 3) we’re three rate hikes into an easing cycle.

Business earnings: Slightly recessionary. With the Q3 earnings season over 70% complete, S&P projects GAAP earnings for the S&P 500 to be 2% lower than the matching year-ago figure, compared to increases of 3% in Q2 and 6% in Q1. S&P 500 operating earnings by I/B/E/S from Refinitiv tell approximately the same story—lower by 0.8% in Q3 (as of Nov. 5) after increasing by 1% in Q2 and 3% in Q1. Factset, by comparison, shows three consecutive year-over-year declines, but the changes are small (-0.3% in Q1, -0.4% in Q2 and likely to be somewhere between -1% and -3% in Q3). So however you look at it, the Q3 earnings dip is shallow. It’s significantly less severe than the 2015–16 earnings recession, although not as easily explained away as being a reflection of oil price volatility—this time the global manufacturing downturn is also part of the story. Factset expects the dip to continue in Q4, whereas projections from S&P and Refinitiv show positive year-over-year growth. All things considered, I have to call business earnings slightly recessionary, but the numbers aren’t yet convincing. Stay tuned.

Household earnings: Not recessionary. Using average hourly earnings as a guide, household earnings growth outpaced inflation by 1.2% over the past 12 months, while real disposable income increased by 3.2% over the same period. By either measure, household earnings are growing strongly enough to support continued gains in consumer spending.

Business credit conditions: Not recessionary. Although demand for C&I loans has stalled of late, ample credit remains available. The Fed’s Senior Loan Officer Opinion Survey (SLOOS) shows that business lending standards haven’t significantly changed in either direction.

Household credit conditions: Not recessionary. Again, lending standards haven’t significantly changed in either direction. Also, bank balance sheets are expanding at a healthy pace—recent data shows banks adding enough real estate loans and mortgage bonds to grow thin-air spending power despite the dip in C&I loan demand.

Stock prices: Not recessionary. Record stock prices have boosted investment portfolio values and should help to support spending.

House prices: Not recessionary but on watch for a possible downgrade. House price growth is decelerating but remains slightly above the CPI inflation rate according to the S&P Case-Shiller 20-city index. It would have to drop another 2% or 3%, depending on changes in consumer inflation, before I would call it a recessionary reading.

Foreign sector: Not recessionary. Although slowing exports have weighed slightly on GDP, imports have dropped as a percentage of GDP (a measure of import penetration) in the first three quarters of 2019. On balance, data fail to support a “recessionary” assessment, although that could change in 2020 with either a steeper drop in exports or a rising propensity to import.

As reminder to regular readers and a heads-up to new readers, I’ve documented the predictive value of indicators discussed above in my 6-cycle forecast articles, my TSP (thin-air spending power) articles and my book Economics for Independent Thinkers.

Conclusions

All Big-3 precursors considered, the near-term outlook is weaker than normal but not yet recessionary—the expansion appears to have enough policy support, spending capacity growth and overall momentum to continue through at least the first quarter or two of 2020.

Deeper into the year, the outlook could darken as many forecasters predict, especially if corporate earnings continue to slide. But we could just as easily see more of the same—an economy that grinds slowly higher as real incomes grow, asset prices trend upwards and bank balance sheets expand. To gauge which of the scenarios is becoming more likely, I suggest watching the Big-3 and tuning out most everything else.

“Have you seen the recession? I ain’t (yet) seen the recession!”*

*In the actual lyrics I’ve based this on, Robert Plant was looking for a bridge, not a recession, although in a song that never had a bridge. So am I saying we’ll never have a recession? Nope, not saying that.

(To receive more frequent updates to the indicators we track, including TSP and our 6-cycle forecasts, send an email with “indicator updates” in the subject line to queries@nevinsresearch.com. Note that our distribution list for indicator updates is different to our blog subscriptions—we only update indicators on the blog when we find the time to write an article about them, meaning the blog doesn’t report most of our research.)

Limit of Liability/Disclaimer of Warranty: Although the author has used his best efforts in preparing this article, he makes no representations or warranties with respect to the accuracy or completeness of the contents and specifically disclaims any implied warranties of merchantability or fitness for a particular purpose. The author shall not be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential or other damages.

“If I can’t picture it, I can’t understand it.”
—Attributed to Albert Einstein

“In order to properly understand the big picture, everyone should fear becoming mentally clouded and obsessed with one small section of the truth.”
—Xunzi, Confucian philosopher

There’s hardly a more hackneyed expression than “painting the big picture,” so I avoided using it in the first article in this series, even though the series is partly about, well, painting the big picture.

In this article, though, I’ll throw in the towel. (Yup, I’m explaining my use of one hackneyed expression by using another.) As the popular Confucian rhyme goes, “If it’s good enough for Xunzi, it’s good enough for me.”

So we’ll follow Xunzi’s advice, along with that of Albert Einstein, and paint the big picture. From the painting advice I offered in the intro article, we’ll apply the following principles to the U.S. economy:

1. Just as your overall health depends on a complex web of interacting bodily systems, the economy’s health depends on a web of interacting cycles, which I’ll call component cycles.
2. The most important component cycles are the core business cycle, business credit cycle, stock price cycle, household credit cycle, home building cycle and house price cycle.
3. The component cycles demand our attention both individually and collectively—they help us determine the economy’s balance of positive and negative forces at a point in time.
4. At a threshold defined as the momentum of negative forces (think vicious loops) exceeding the momentum of positive forces (think virtuous loops), a recessionary process is underway and probably irreversible.
5. A well-constructed checklist—one that separates leading from coincident and lagging indicators within each of the component cycles—can be an effective tool for weighing positive and negative forces and predicting recessions.
6. That checklist should assign predetermined roles to predetermined indicators (each has a specific job to do), and it should encompass all of the potentially leading areas of a “six-cycle map” of the economy.

In my experience and in backtests, these principles lead to better Business Cycle forecasts (capitalizing to avoid confusion with the component cycles) than the more conventional approach of pouring a giant bag of indicators into a giant pot and statistically blending them.

I’ll put the principles to use in just a moment, but first I’ll repeat the diagrams included in the intro article that help demonstrate how the approach works. The first diagram is a map combining key pieces of each component cycle:

The second diagram collapses the map into a stylized template for tracking the economy:

The last diagram uses the template to show the economy’s descent into recession from 2006 to 2009:

Is a U.S. Recession Imminent? (Part 1)

I included the diagrams because they can help you “picture” the six-cycle approach, but they’re not especially predictive without refinements, and that’s where the aforementioned checklist comes in. My primary checklist, designed to screen for recession risks, consists of ten assessments—two for the core business cycle, one for every other component cycle and one each for the foreign sector, fiscal policy and monetary policy. For each of the ten, my assessments take one of only three readings—“risk-on,” “risk-off” or “neutral.” Checklist items that are risk-on are booming, those that are risk-off are weak enough to help push the economy into recession, while a neutral reading means that the particular item signals neither acceleration nor deceleration from the economy’s long-term trend.

As I argued in the intro article and in greater depth in my book, checklisting can provide the discipline we need to carefully weigh the economy’s positive and negative forces. Although sometimes written off by the pretentious as not sophisticated enough, checklist-based methods are time-tested and easy to understand. In a macro world bulging with madcap mathematical theories and short-lived econometric models (they habitually “blow up”), longevity and simplicity are nice features to have.

So without further review, here are my checklist assessments for the U.S. economy, which I’ll keep short by glossing over the more detailed analysis. I’ll then use the assessments to gauge the risk of an imminent recession.

• Core business cycle (business earnings): Neutral. Although the perpetually revised national accounts data show five years of falling corporate earnings on a pre-tax operating basis, and Factset shows S&P 500 earnings declining on a year-over-year basis in Q1 and possibly also Q2, the more transparent GAAP earnings grew strongly through Q1 and also Q2 with the reporting season just more than halfway over. For most purposes, I prefer GAAP, but in this case I’ll be conservative and call it a stalemate.
• Core business cycle (household earnings): Neutral. Household earnings growth is a bright spot, but it isn’t quite strong enough as of this writing to record a risk-on reading.
• Business credit cycle: Neutral. Business credit demand is weaker than usual, but lending standards haven’t significantly changed in either direction, while bank balance sheets are expanding at a moderate pace.
• Household credit cycle: Neutral. Household credit demand is also weaker than usual, but once again lending standards haven’t significantly changed and banks continue to expand.
• Stock price cycle: Neutral. Although the 12-month change in real stock prices is strong, my research suggests that market levels haven’t climbed above last year’s highs by enough to merit a risk-on reading.
• House price cycle: Neutral. Although house price growth is decelerating, it’s still slightly above the CPI inflation rate. It would have to drop another 3% or so, depending on changes in consumer inflation, before I would call it a risk-off reading.
• Home building cycle: Neutral. The home building cycle is weakening, although the rate of change is glacial and the effects on GDP almost imperceptible. It remains to be seen if the flat-to-downward trend will continue at the same pace or develop into something more jarring.
• Foreign sector: Neutral but on watch for a possible downgrade. Although slowing exports have weighed slightly on GDP, imports have dropped as a percentage of GDP (a measure of import penetration) in the first half of 2019. On balance, data fail to support a risk-off assessment, although that could change with either a steeper drop in exports or a rising propensity to import.
• Fiscal policy: Neutral. Government spending is growing at a decent clip in 2019, while taxes and net transfer receipts dropped to a new six-year low as a percentage of GDP in the first half of the year, which can only help household and business spending power. So fiscal policy continues to support growth, although it falls short of a risk-on assessment. Also, the mini–fiscal cliff that many forecast for late 2019 and 2020 now looks as though it might be more like a mogul, if anything at all, thanks to this week’s free-spending, drop-it-on-the-kids debt ceiling deal.
• Monetary policy: Risk-off. The (mostly) inverted yield curve suggests that monetary policy is restrictive.

*For all four credit and asset price cycles, I also rely on a measure of “thin-air” spending power, which my indicator-update subscribers know as “TSP.” See “TSP Indicator Update: Criss-Cross, Flip-Flop and Remembering 1966.”

Overall, the checklist assessments are unusually, well, neutral. And neutral isn’t especially good—the checklist looks worse than it has since early 2012 at the depths of the European debt crisis. The first quarter of 2012 was the last time that risk-off assessments outnumbered risk-on assessments, and the absence of any risk-on assessments in the current checklist is also a rarity. So the economy comes through my checkup looking far from vibrant, which you might have expected considering recent buzz about a potential global recession.

All that said, the paragraph above is the gloomiest one I’ll write, because my checkup falls short of a recession prediction for two reasons. First, recessions normally follow a few quarters with four or more risk-off assessments, which would tell us that vicious loops are cutting across component cycles and threatening to overwhelm the economy’s positive forces. Second, recessions normally begin with significantly more risk-off than risk-on assessments, which would also tell us that negative forces are beginning to prevail. Using a mechanistic version of the checklist and calculating a composite indicator as the number of risk-ons minus the number of risk-offs, here are the readings at the last nine Business Cycle peaks: -1, -3, -6, -5, -8, -8, -5, -2 and -5. And here are the readings one quarter before the last nine BC peaks: -3, -1, -6, -3, -6, -6, -2, -4 and -1. (For reference, the economy peaked in Q3 1957, Q2 1960, Q4 1969, Q4 1973, Q1 1980, Q3 1981, Q3 1990, Q1 2001 and Q4 2007.)

But that’s enough numbers, because the approach is easier to grasp in pictures. Here’s a template that shades risk-on and risk-off assessments, first for the current checklist and then for the checklist readings before past recessions:

The charts approximate the economy’s balance of positive and negative forces at a point in time, helping to expose impending recessions. They map the contours of each recessionary process, which can then be compared to the present, demonstrating the ease, discipline and balance of the six-cycle approach. They can also help to manage any debate—if you disagree with this or that assessment, you can substitute your own view without building a whole new “model.” And as far as the debate goes, in a future chartbook I’ll expand both the historical analysis and the rationale behind my current assessments. For now, though, let’s take a deeper look at today’s outlook and get subjective.

Is a U.S. Recession Imminent? (Part 2)

Here are three points to consider about the current expansion, building once again on the six-cycle map and checklist but subjectively this time, meaning without the composite indicator’s arithmetic.

First, the expansion continues to draw strength from rising after-tax real incomes, steady bank credit growth and record (or near-record) asset prices, and those three forces usually boss the Business Cycle. In fact, over the last 65 years the economy has never fallen into a recession with all of the following circumstances in place: (A) after-tax real incomes rising for households and businesses, (B) bank credit growing at a decent pace and (C) risky asset prices also growing or at least flat. Since A, B and C feed into spending capacity, that historical fact only says that spending is unlikely to fall (as it would in a recession) if spending capacity is growing.

In other words, neither households nor businesses will slash spending without having a good reason to do so, and historically that reason is that they’re either earning less money or they have less access to money via bank credit or asset sales. So history aligns perfectly with how you might expect households and businesses to decide how much to spend. And as of this writing, I’m not sure what would cause spending to fall today, since A, B and C are all true according to my analysis, as shown by the first six checklist items being neutral.

Of course, not everyone looks at it in the same way, and some might say the risk of a spending contraction lies in a fear of the unknown, such as the uncertain effects of a tariff war. But if fear alone drags spending lower, I would expect to see it in bank behavior and risky asset prices before we see it in the broad economy, and that brings me right back to waiting for a false reading from A, B or C.

Second, and notwithstanding my point above, another argument holds that some combination of the housing, foreign and government (think fiscal policy) sectors could combine with the lagged effects of monetary tightening and push the economy into recession. With at least some reasons for concern in each of the three sectors mentioned, this seems a valid risk to watch for, but I don’t see it in the data. In fact, the last few quarters of national accounts data show that the government sector is still growth supportive, as noted in the assessments above, and the housing sector is hard to get excited about one way or the other—it hasn’t boomed during the current expansion nor does recent weakness qualify as a bust.

Third, the checklist is relatively sanguine in comparison to other popular indicators that have stalled or pointed downwards in recent months, such as the yield curve, PMIs, industrial production, construction, small business hiring and new home and car sales. If these indicators have convinced you the end is nigh, you’ve probably already stopped reading and may have jumped straight to the comment section of whatever site you’re on to tell everyone what a dunce I am. That’s fine, conflicting views are what makes a market, and you might be proven right. I’ll just point out that many of the weakest indicators are coincident or lagging indicators that tend toward mini-cycles of strength and weakness within the overall Business Cycle, which makes them unreliable as recession signals. The biggest exceptions are probably the housing sector and the yield curve, but you’ve already heard my thoughts on housing, and as alarming as the recent curve inversion might seem, historical correlations suggest that other indicators might be more effective. (For example, scroll near the end of this chartbook for a comparison to TSP.)

Bottom Line

All things considered, the economy’s positive forces appear strong enough to keep the negative forces in check through probably the rest of the year, so the expansion should continue into 2020. But the balance of forces has weakened in recent months, and that weakening trend could continue, so stay tuned.

(If you’re interested in more frequent six-cycle updates—to either diversify or replace any econometric inputs you might use—join our mailing list for indicator updates. You can do this by sending an email with “indicator updates” in the subject line to queries@nevinsresearch.com. Note that this is different to our blog subscriptions—we only update indicators on the blog when we find the time to write an article about them, meaning the blog doesn’t report most of our research.)

It’s usually a bad idea to stand too close to something—whether an object, a problem you’d like to solve or any number of other things—which could mean seeing all of the pixels but none of the patterns. That’s why we populate albums, frames and holiday cards with bird’s eye views and sweeping vistas. It’s why every city that aspires to “destination” status advertises this or that Tower, Arch, Needle or Eye.

But if we look from too far away, we run a different risk of missing important information. That’s why we send probes, ships and occasionally scientists into outer space. It’s why we don’t Facetime our doctors, we hop on the examination table and show them exactly what’s bothering us.

Now hold those thoughts for a second except the last one—when I segue to economics momentarily, it’ll be from the medical field. But instead of being the patient lying on the examination table, imagine you’re the physician standing alongside it. That means you’re an expert on bodily systems—respiratory, circulatory, nervous, digestive, immune and so on—and you use that knowledge to diagnose the symptoms your patient describes. In other words, the key to your diagnosis is your understanding of the systems and how they interact.

So the connection to economics is this: the economy, too, consists of a mish-mash of interacting systems, or if we’re interested in forecasting, we should think of cycles. But I’m not referring to the capital-B, capital-C Business Cycle as a uniform force—that would be the too-far-away view. I mean the component cycles that are easier to monitor and conceptualize. Component cycles relate to the overall economy in the same way that bodily systems relate to overall health. And like bodily systems, component cycles interconnect while also operating with a degree of autonomy. Therefore, we should choose vantage points that allow us to see the patterns mapped by each one. Our ideal positioning isn’t too close (all anecdotes, no distinguishable patterns), but it isn’t too far away, either.

The Map

I’ve argued that six component cycles demand our attention and should be in clear sight both individually and collectively. Here are my big six:

• Core business cycle
• Business credit cycle
• Stock price cycle
• Household credit cycle
• Home building cycle
• House price cycle

When two or more of these cycles push in the same direction, they reinforce each other and strengthen the economy’s momentum. When they pull in different directions, their relative strength determines the economy’s eventual path. And the potential push–pull combinations are numerous, meaning that patterns change over time. In the 2008–9 recession, for example, the home building and house price cycles began to deteriorate well before the core business cycle. By comparison, the 2001 recession began in the business credit and stock price cycles.

But let’s narrow it down—let’s pick one recession and paint the broadest view, which should help demonstrate how the approach works. Glossing over interactions within each component cycle (not because they’re unimportant, but to keep it at a high level), we can show how the pieces fit together using the map below:

We can then collapse the map into a stylized template for tracking the economy, as follows:

Then the next figure uses the template to show the economy’s descent into recession from 2006 to 2009:

The Threshold

Scanning through the economy’s descent—as above for 2006–9 although we could do the same for any recession—you might believe as I do that there’s a threshold beyond which a recession becomes unavoidable. Needless to say, that threshold is a critical juncture in the Business Cycle, maybe the most important of all junctures. The better we understand it, the better prepared we’ll be for the next recession, and here’s my nickel’s worth of advice. (Note that I’ll continue to capitalize the overall Business Cycle to distinguish it from the component cycles.)

First, imagine a world of virtuous and vicious loops that continually battle for supremacy. You might be more accustomed to the term circles than loops, but the idea is surely familiar. In a virtuous loop, growth in one part of the economy spills into other parts, and these growth forces then continue to strengthen in a mutually reinforcing way. Vicious loops describe the same process in reverse. Although both types of loops can encompass any of the component cycles listed above, they usually connect the real economy (the core business and home building cycles) to the financial economy (the credit and asset price cycles). Typically, the financial economy supplies the accelerant.

So the recession threshold, then, is a tipping point in the perpetual tug-of-war between virtuous and vicious loops—probabilities tip toward recession when vicious loops gain more momentum than that of the remaining virtuous loops. And then at that point, the Business Cycle is likely to peak within six to nine months. In a nutshell, that’s the context for economic forecasting using the six-cycle approach. It’s about having the discipline to carefully weigh the forces acting within and between component cycles.

Easier said than done?

Maybe so, but as part of that nickel’s worth of advice I’ll tell you it’s achievable—the approach has a solid record in real life and backtests.

Sensible enough but vague?

Again, maybe so, so let’s expand. I think it’s important to start with the philosophy, but I can be as specific as you’d like.

The Checklist

Readers familiar with my book know that I’ve augmented the six-cycle map with a variety of rules to make my approach as predictive as possible. Let’s skip over those rules, for now. You’ll be left without the arguments I use to justify each piece of the analysis—and the absence of the pieces I leave out—but the end result should be easy enough to follow. After all, it’s just a checklist. If you don’t mind pretending to be a doctor again, it’s no different to the list of questions and diagnostics you might use when your patient comes in for a check-up. In other words, whether we’re talking about the economy, the human body or any other complex system, constructing a checklist is an obvious way to screen for problems. In an economic context, a simple checklist can provide the necessary discipline discussed above—it enforces a broad view even as public attention swings from one topic-of-the-day to the next.

So add the checklist to the map and the threshold, and we have the makings of a disciplined approach to evaluating economic risks. The checklist needs to be well-constructed, but again, I’m skipping over the research and reasoning behind the construction, for now. I’ll share one of the end results—my primary checklist, which consists of ten assessments. I make two assessments for the core business cycle (one each for households and businesses), one for every other component cycle and the following three extras:

• The foreign sector
• Fiscal policy
• Monetary policy

These last three items are mostly external to the six component cycles but no less important to the economy’s path.

Before I go further, notice that I’m not using the most common economic forecasting tool: econometrics. I don’t statistically or numerically optimize anything. I don’t regress anything on anything else. I don’t vector autoregress or stepwise regress or cointegrate. There are no principal component derivations, Markov switching models, Kalman filters or Grozerian cross-dimensional reversals.

Okay, the last one isn’t real, I took it from Ghostbusters. But when it comes to everyday volatility, the others are fictional, too, because they presume the economy sits atop a stable mathematical structure, typically a linear structure with Gaussian distributions. That makes them fundamentally unrealistic before you’ve even booted up your stat software. Real-world economic volatility isn’t stable, linear or Gaussian, to name just three conditions it fails to satisfy from a list that’s longer than a Grozerian mile.

To be sure, I test my methods against history, I’m just unimpressed by stacks of econometrically derived coefficients—they give me the same sense of futility that I get from certain political ideologies. So like everyone else, I stick to what makes sense to me. You might call my approach back-to-basics. The things I try to do well include choosing the best vantage points, separating leading from lagging or coincident indicators, uncovering stellar indicators that aren’t readily available because of substantial flaws in economic theory, and borrowing pragmatic methods that are as old as dirt, such as checklists, which the pretentious regard as overly simplistic even though we all know they’re indispensable.

(As a final clarification, none of the above is meant to suggest that basic econometrics can’t yield some insights if limited to appropriate questions—think micro, not macro—but the unmistakable if rarely spoken truth is that the world abounds with madcap econometric models that overreach and underdeliver.)

Conclusion

So now that you know me better, I’ll summarize and then preview the next article in this series.

To recap, I’m suggesting that the best recession screen begins with a conceptual view of how the component cycles and other economic forces fit together. That might seem an innocuous thing to say—who would put their hand up and insist the conceptual view doesn’t matter?—but it’s different to the typical leading economic index. In the typical approach, forecasters pour a giant bag of indicators into a giant pot and then econometrically blend them, which means letting their chosen statistical methods do the cooking. My rebellion, in comparison, holds that it’s better to do our own cooking, by using the six-cycle map and a checklist to assign predetermined roles to predetermined indicators, such that each has a specific job to do. We can then assess those indicators with a goal of weighing positive versus negative forces, once again informed by the map.

That’s the basic idea, and I’ll put it to use in my next “recession watch” article, which I’ll publish later this week. I’ll share my checklist assessments for the U.S. economy, and then I’ll shape those assessments into an outlook. If you’re willing to accept that I might be onto something, check back for a practical application of the six-cycle approach.

“The fact that financial markets responded in very similar ways … lends credence to the view that these actions had the expected effects on markets and are thereby providing significant support to job creation and the economy.”
—Ben Bernanke defends the idea that markets and the economy respond significantly to quantitative easing

“… it will be like watching paint dry, that this will just be something that runs quietly in the background.”
—Janet Yellen refutes the idea that markets and the economy respond significantly to quantitative tightening

It doesn’t take much calculation to see that the Fed’s position on quantitative tightening (QT) is blatantly inconsistent with its position on quantitative easing (QE). You only need to notice that the excerpts above, taken together, violate the following pair of postulates:

1. When A and B are opposites, the effects of A should be opposite to the effects of B.
2. QT is the opposite of QE.

So financial markets and the economy should respond significantly to both QE and QT—although in opposite directions—or they should respond to neither QE nor QT. To claim otherwise, as in the excerpts above as well as other similar communications, is like arguing that one of the two postulates is wrong in the context of the Fed’s bond portfolio. That seems unlikely, but not impossible. In particular, the first postulate falls short of an absolute truth, reality sometimes being more complicated than we’d like it to be. Consider that Newtonian physics seemed absolute enough until Einstein came along.

But former Fed chairs Bernanke and Yellen aren’t relativity theorists and haven’t framed it like that. In public comments and speeches that I’m aware of, they haven’t acknowledged the postulates in the first place let alone explained why they reject the logic of opposite actions yielding opposite effects. In fact, they don’t have to acknowledge or explain, because major media outlets—those with invites to FOMC press conferences and “sources” close to the key decision makers—rarely challenge the Fed’s narrative. If you’re established in the mainstream, there’s no upside to investigating inconsistencies, only the downside of being seen as impolite while jeopardizing your coveted invites and sources.

Awkwardness Alert: Here Comes a Chart that Really Excites Me

So let’s agree that the Fed’s positions on QE and QT are incompatible, at least on the surface, and it’s only decorum and incentives that prevent, say, the Wall Street Journal from calling bullshit. Since I’m only stating the obvious, you agree so far, right? Yes? Good. Now let’s see what the data say. And by “data,” I mean the Fed’s own Z.1 report. Regular readers might recall that I used the Z.1 to test claims about QE, posing two questions that seemed reasonable to ask:

• How rapidly do banks and broker–dealers (BDs) expand credit during QE periods (QE1, QE2 and QE3) compared to QE pauses (all other times)?
• How does bank and BD credit expansion compare to the Fed’s credit expansion during the same periods?

Here are the answers in chart form:

Call it TMI, but the chart still excites me five years after I first produced it. (You can find the latest iteration here and then follow links to earlier versions.) Two of the most obvious questions yield the cleanest imaginable outcome—the private financial sector and the Fed almost perfectly offset each other! The Fed grabbed the credit-growth baton for QE laps and then passed it back for QE pauses, but whoever didn’t have it more or less stood still.

In other words, whereas the Fed expected to increase the amount of credit supplied to the nonfinancial sector, the Z.1 suggests it only substituted one source of credit for another. That’s not to say QE had no effects at all—it clearly jolted market psychology, and therefore, influenced market prices. But the Z.1 refutes the primary mechanism claimed by Bernanke, which required QE to increase total credit supply. The hypothetical chart that fits Bernanke’s expectations would have been less argyle and more Charlie Brown zigzag, with total credit creation rising to new highs during QE periods and then falling back during pauses.

All of which brings us from QE to QT. As of this month, the Z.1 shows four consecutive quarters of the Fed reducing its net lending, thanks to QT. So the argyle effect is necessarily over, because the Fed no longer alternates between only two options—QE and not-QE. The pattern has to change, but to what? Well, here’s the answer:

Not as clean this time, right?

On one hand, the private financial sector appears to have neutralized some portion of QT by increasing its net lending after the Fed’s lending began to shrink. On the other hand, the financial sector’s increased lending didn’t exactly offset the Fed’s decrease. The change in the first is about half the magnitude of the change in the second.

Three Possibilities

So what exactly does the new QE–QT pattern tell us? I’ll suggest three possibilities, ordered from my least likely to my most likely:

1. QT might have caused the total amount of credit supplied to the nonfinancial sector to fall, a result that would be intriguing largely because of what it would say about the central bank’s expectations. If the private financial sector fully offset QE’s credit creation but not QT’s credit contraction, which is the conclusion you might reach if relying on nothing but the chart, then Bernanke and Yellen got it completely backwards. You’ll never see that conclusion in the Wall Street Journal, for the reasons noted above, but it actually fits the data.
2. We might not have enough QT data to reach a firm conclusion just yet. In other words, we might find that the picture changes with a few more QT quarters, especially as the volatility of these figures is quite high. That’s why I waited for four quarters of data before publishing my chart, but four still might not be enough.
3. Total net lending might have declined over the last four quarters even without QT, such that any QT effects were negligible, notwithstanding my chart. Instead, debt-saturated borrowers might have decided to take a rest after nine years of expansion and independently of monetary policy. This view lines up nicely with the Fed’s loan officer survey, which shows weakening demand for most types of loans during the last four quarters. It’s also consistent with the idea that monetary policy makers neither manufacture nor extinguish creditworthy borrowers, who travel to loan and underwriting desks from all corners of the economy, just not from the front steps of the Eccles building as the Fed adjusts its bond holdings.

For what it’s worth, my last point above ties into what I believe to be one of the biggest flaws in the Fed’s make-up. That is, scholars like Bernanke and Yellen reason from theories that require people to respond slavishly and robotically to every adjustment in public policy, however odd, circular or obscure the adjustment might be. In Bernanke and Yellen’s world, it’s no exaggeration to say that central banks really do manufacture creditworthy borrowers, as if the term open market account describes a freakish assembly line, not a simple bond portfolio.

Bottom Line

You’ll form your own views on all of the above, but my advice is to filter the Fed’s narrative with a healthy skepticism, however deferentially the media chooses to endorse it. I’m not saying the narrative is always wrong, just that monetary policy is notoriously difficult to evaluate. But not impossible—the Fed’s data can help separate fact from fallacy. Dig into the Z.1, and you might find that the best reality-check is a single chart—call it the Burberry truth—modest in ambition but scintillating nonetheless.

In November, we argued that the business cycle rests heavily on a certain type of incremental spending—namely, spending that doesn’t require prior savings. We used the term thin-air spending power (TSP) to describe spending that’s financed by external “injections” instead of prior savings.

As part of our argument, we shared the chart below, which compares TSP-derived spending on the left (financed by fresh bank credit) to spending that merely recycles savings, such as the prior domestic savings category on the right.

We also shared a diagram that puts the argument in pictures, illustrating how bank credit boosts economy-wide spending.

As shown, bank credit stands apart because it doesn’t require savings from prior income. Rather, it creates fresh TSP, which again represents spending power that materializes from “thin air.” (See articles here and here for further explanation.)

Measuring TSP

But banks aren’t the only source of TSP. Investment gains and losses can also boost or dampen spending in ways that leak into the circular flow depicted above. Recognizing the similarities between these TSP sources, we can build a highly predictive indicator from only two components:

• Real new bank credit. Inflation-adjusted new bank credit aggregated over four-quarter periods and expressed as a percent of final domestic demand in the prior period.
• Real holding gains. Inflation-adjusted holding gains (household and nonprofit gains from equities, mutual funds, real estate and pensions) aggregated over four-quarter periods and expressed as a percent of final domestic demand in the prior period

The chart below shows the indicator’s average path during the last nine business-cycle expansions. Note that we’re mapping a path through two dimensions—one for each of the two primary TSP sources—by connecting data sequentially.

And here’s the early 2000s expansion on its own.

As shown in the last chart, TSP tends to cycle through three phases: recovery, financial inflation and financial deflation:

• Recovery. TSP meanders upwards and rightwards as the financial economy heals from the prior recession.
• Financial inflation. TSP enjoys the big air of the upper-right triangle.
• Financial deflation. TSP completes the cycle by becoming scarce once again, dropping below a diagonal recession warning.

In other words, TSP normally triggers a recession warning shortly before the onset of a recession, anywhere from one to five quarters before. (Note that we use a ten-to-one ratio for our recession warning, meaning we expect a dollar of additional bank credit to have about ten times the effect on spending as a dollar of real asset holding gains—see the earlier TSP article for explanation.)

Remembering 1966

All of which brings us to our TSP indicator update, which begins with a look-back to 1966. After our November article, TSP carved a path that’s occurred only once before in the indicator’s 65-year history—during the 1960s expansion. As shown below, TSP breached the recession line in the midst of a 1966 credit crunch only to restore financial inflation one quarter later, completing a criss-cross that had much to do with monetary policy.

In the year and a half leading up the 1966 criss-cross, the FOMC was actively tightening policy—raising bank reserve requirements, pressuring credit growth and lifting the discount rate in December 1965 in an action that famously earned Fed chief Martin an invitation to LBJ’s Stonewall, Texas ranch.

But not for the first time and certainly not for the last, the financial sector responded to the monetary tightening in some ways that were expected and others that weren’t. On the expected side of the ledger, banks sold Treasuries to meet reserve requirements. But by the summer of 1966 they had few Treasuries left to sell, so they began dumping municipal bonds while removing their usual support for new municipal issues, an unexpected response that weighed heavily on municipalities while rattling confidence throughout the financial sector. A mini–financial crisis with convulsions in the muni market was hardly the result the FOMC was seeking.

And how did the Fed react?

On September 1, 1966, the Fed’s regional bank heads sent the same letter to all member banks in their regions—they asked those banks to restore the flow of credit to municipal issuers. The letter also encouraged banks to replenish reserves with borrowings at the discount window, and the FOMC then sharply increased its Treasury purchases during Q4, again with an aim of replenishing reserves and reversing prior policies.

In other words, the Fed’s 1966 flip-flop explained the TSP criss-cross. Perhaps not surprisingly, the economic expansion then continued for another three years. (Besides the Fed’s archives, you can find an interesting account of the 1966 credit crunch in Hyman Minsky’s Stabilizing an Unstable Economy.)

Fast-Forwarding to Today

With that story as background, here’s a chart showing TSP’s path during the current expansion, using last week’s flow-of-funds data for Q4 2018 and our preliminary estimate for Q1 2019.

And here’s the same data in just one line, equating $1 of new bank credit to 10 cents of asset holding gains as in the rationale behind the recession-warning line. (Again, see the earlier TSP article for explanation.)

You can think of the Q1 estimate as like a GDP tracker—the final figure will surely be different but possibly not much different.

(As an aside, if you’re interested in more frequent TSP updates along with further detail on the underlying inputs, join our mailing list for indicator updates. You can do this by sending an email with “indicator updates” in the subject line to queries@nevinsresearch.com. Note that this is different to our blog subscriptions—we only update indicators on the blog when we find the time to write an article about them, meaning the blog doesn’t report most of our research.)

So TSP grazed the recession line and then almost certainly jumped back into financial inflation in Q1 (barring a market crash in the next two weeks). This most recent instance is less criss-cross than ricochet, but it’s still reminiscent of 1966. In both cases, TSP reversed direction just after the Fed sharply changed course. It’s almost as though the Powell Fed waited for TSP to touch the recession-warning line before unleashing a flurry of statements that walked back a prior commitment to policy tightening.

Conclusions

It’s not hard to spot risks to the U.S. economy in early 2019—most obviously: tepid auto sales and housing data, a sharp drop in retail sales, and continued weakness in Europe and Asia—but the Fed’s policy flip-flop has helped restore a degree of financial inflation.

So as weak as Q1 GDP is likely to be (see, for example, the Atlanta Fed tracker), we expect the expansion to muddle through, reaching the third quarter to become the longest on record at over ten years.

That said, we can’t rule out another TSP reversal, especially as stock prices remain vulnerable, the housing market grinds ever slower and the recent resurgence in bank credit appears to be leveling off. As we argued in November, TSP bears watching. It might once again prove to be the best indicator for timing the next recession.

For the full TSP chartbook, click here. For a deeper dive into the underlying philosophy, see Economics for Independent Thinkers.

In “You Might Like to Watch This Picture as Asset Prices Fall,” we discussed a composite indicator called thin-air spending power (TSP). We also promised that the accompanying chartbook (the post that you’re reading) would track thin-air spending power during every business-cycle expansion since 1954, among other charts. Here are the business-cycle expansion charts.

TSP in an Average Expansion

Now for the other charts. In the next one, we constructed an average TSP path for the expansions shown above. Our average path includes five expansions (all of those that lasted four years or longer) for the first data point, which is sixteen quarters before the onset of a recession, and then it blends in the shorter expansions as soon as they fit completely into the remaining window.

The average TSP path becomes especially interesting when we compare it to our estimate for Q4 2018. As noted in the main article, we estimated Q4 TSP using high frequency data (including weekly data from the Fed’s H.8 report, home price data from the Case-Shiller Index and the S&P 500’s close on November 27). Here’s the Q4 TSP estimate alongside our final readings for 2018’s first two quarters.

The chart shows the Q4 estimate lying directly on top of the average TSP path at a point that’s just two quarters before the end of an expansion. That’s not to say we’re predicting a recession in two quarters time (other key indicators haven’t yet deteriorated alongside TSP as in past expansions), but the chart certainly has our attention.

(As an aside, if you’re interested in watching the Q4 estimate unfold in real time, join our mailing list for indicator updates. You can do this by sending an email with “indicator updates” in the subject line to queries@nevinsresearch.com. Note that this is different to our blog subscriptions—we only update indicators on the blog when we find the time to write an article about those indicators, meaning the blog doesn’t report most of our research.)

TSP versus the Yield Curve

For any readers who’ve decided to dismiss TSP only because they think the yield curve is all they need to forecast the economy, we have a special chart to show you. To create it, we reduced TSP to a single dimension using our recession-warning line’s ten to one ratio between the spending effects of real new bank credit and the spending effects of real holding gains. (See the main article for explanation.)

More exactly, we combined real new bank credit with 10% of real holding gains to calculate TSP in a single dimension. Then we calculated the correlation between TSP and changes in economy-wide spending (real final domestic demand) and compared it to the correlation between the yield curve and changes in spending. Here’s the comparison.

To be clear, we’re not disputing that the yield curve can help predict the economy—we think it can. In fact, if we had a longer history for other curve measures (especially for the short end of the forward curve), we would possibly see higher correlations with spending. But the measure we used (the 10-year minus the 1-year, which we chose because it has a history that goes all the way back to 1953) sits right in the middle of the mix of most popular curve measures. And based on the results for that measure, anyone who bets on the yield curve should know there’s another horse—TSP—that run laps around traditional curve measures when it comes to predicting spending.

TSP versus Spending and the Business Cycle

So the correlations tell us that TSP is closely related to spending, and the next chart shows the same relationship in line form.

Our last chart compares TSP to the last nine recessions, and once again, we can see that recent developments carry a strong warning. When TSP drops as severely as suggested by our latest estimate, a recession virtually always follows in five quarters or less, and typically less. As noted in the main article, we suggest watching this picture.

“It is high time we rediscovered the role of the financial cycle in macroeconomics.”
—Claudio Borio, Bank for International Settlements

In May, we queued up the b-side of a record describing America’s balance sheet—we looked at the mix of lenders instead of the usual “a-side” analysis of the borrowers.

We showed that the balance sheet includes four types of lenders—banks, the Fed, foreigners and prior domestic saving—as in the updated chart below. And the “prior domestic saving” category, since you asked, is mostly households, pension funds and insurance companies investing in bonds and bond funds.

Then we showed why the b-side is so important, even as it gets little attention. That is, the four types of lenders are fundamentally different from one another—lending by banks is highly correlated to spending (same-period and next-period spending), whereas the other lenders show no such correlation.

But economic theory says all lending is the same—how can banks be different?

Finally, we shared a diagram that explains the previous result.

The diagram shows that bank lending is unique because it creates fresh spending power from “thin air.” We’ll leave further explanations aside for now, but you might check our articles here or here to review how banks create spending power from nothing, and why that process invalidates entire libraries full of mainstream thinking. (Or see our book for more detail.)

Bank balance sheets are also highly predictive, as we showed when we used bank credit to construct a business-cycle indicator. (Again, see the b-side article linked above.) Considering the connections—empirical and conceptual—between bank credit and the business cycle, our indicator might be the best first step to business-cycle forecasting.

Okay so banks conjure spending from thin air—does anything else do the same?

Now we’ll take a second step by asking: What else materializes from thin air? How about the gains and losses in your investment portfolio? It sure seems as though investment gains and losses pop up from nowhere. And by combining them with new bank credit, we’ll create a highly predictive composite indicator that we’ll call thin-air spending power (TSP). Here are the two inputs to the composite:

• Real new bank credit. Inflation-adjusted new bank credit aggregated over four-quarter periods and expressed as a percent of final domestic demand in the prior period.
• Real holding gains. Inflation-adjusted holding gains (household and nonprofit gains from equities, mutual funds, real estate and pensions) aggregated over four-quarter periods and expressed as a percent of final domestic demand in the prior period.

And the chart below provides an example, using data from 2002 to 2008, of how we can track real new bank credit and real holding gains through a business cycle by placing one on each axis. Note that we’re mapping a path through the two dimensions by connecting data sequentially. Although the path shows just a single cycle (the last decade’s housing boom), the pattern is similar to that of the previous eight cycles, which you can confirm by reviewing the chartbook we’ll link at the end of this article.

But what exactly does TSP tell us?

So TSP is a creature of habit, and it has a habit of cycling through three phases: recovery, financial inflation and financial deflation.

• Recovery. TSP meanders upwards and rightwards as the financial economy heals from the prior recession.
• Financial inflation. TSP enjoys the big air of the upper-right triangle.
• Financial deflation. TSP completes the cycle by becoming scarce once again, dropping below a diagonal recession warning.

In other words, TSP typically triggers a recession warning shortly before the onset of a recession, anywhere from one to five quarters before. But you might wonder where the recession-warning line comes from—how do we determine its slope and position? Here’s the rationale for our choices:

• Slope. We consider the additional spending that could result from a dollar of real new bank credit versus a dollar of real holding gains. We expect a dollar of real new bank credit to result in up to a dollar of additional spending, but probably not a full dollar due to the portion that banks invest in securities rather than loans—security purchases don’t always flow into the real economy as directly and reliably as loans do. And for real holding gains, there’s a substantial literature suggesting that each dollar of additional wealth boosts spending by anywhere from three or four cents to a little more than ten cents. So weighing up real new bank credit against real holding gains, we see a ratio of about ten to one as far as the effects on economy-wide spending, and that determines the slope of our recession-warning line.
• Position. We draw the line through the origin to keep it as simple as possible. That choice won’t be optimal in every cycle, but we don’t believe it’s realistic to think we can “engineer” a substantially better choice, especially as cycles change from one to the next.

Note that we’re cognizant of the risks of false precision. We didn’t fit the recession-warning line using regressions or other statistical techniques—we chose nice, round numbers that seemed reasonable, conceptually, and then we stopped there. Our choices may or may not hold up in the future, but we’d rather focus on whether current dynamics could be different to the past than on data mining the past to the fifth decimal point.

What can we say about the next few years?

And since we mentioned it, are we expecting the dynamics to be different this time? Or, will they be the same as usual?

You’ll form your own views, but our nickel’s worth of advice is to expect the usual. After eight rate hikes (and counting) and nine years of expansion, it’s natural for bankers, borrowers and asset markets to anticipate slower growth, and that’s exactly what we think we’re seeing in 2018. We’re seeing the financial economy lead the real economy. Or, to use a term that’s become popular in some circles of economics, the financial cycle is leading the business cycle. As a next step, we expect the financial cycle to fall into a more definitive contraction.

So once again, the financial cycle should drag the business cycle lower, and our TSP chart offers clues about the timing. Most importantly, the diagonal recession warning provides a decent tripwire for the countdown to the business cycle’s apex. We haven’t triggered the tripwire just yet, but we get an interesting result when we use high frequency data to estimate where TSP might fall at year-end. That is, our year-end (Q4) estimate sits only just above the recession-warning line, as shown below.

Conclusions

To be clear, we won’t know TSP’s actual Q4 reading until the Fed’s “flow of funds” data becomes available. But for now, we suggest watching the high frequency data, as above, especially as financial markets appear to be losing their nine-year-long buoyancy. (See this article for interesting research on how the struggle to regain buoyancy might play out.)

More generally, we’ll continue to promote the following beliefs:

1. There is such as thing as a financial cycle (skeptics notwithstanding).
2. The financial cycle explains a significant portion of the business cycle.
3. To properly account for the financial cycle, you have to first reject a handful of the most pervasive and deeply held tenets of Keynesian, Monetarist and New Classical theories.
4. The most predictive financial-cycle indicators are those that measure spending power created from thin air, as in our TSP chart.
5. Other methods decompose the financial cycle into component cycles. (See our book, Economics for Independent Thinkers.)

To demonstrate the fourth point, in particular, we’ve published a chartbook with more history. The chartbook tracks TSP through every business-cycle expansion from 1954 onwards, among other charts, and shows that the two-dimensional approach has fewer anomalies than real new bank credit alone. Of course, that doesn’t necessarily make TSP better than other approaches—there are plenty of data-mined models that show ‘A’-grade back-tested results. But those models often require checking your intuition at the door, whereas our b-side approach is built mostly on intuition. In other words, we aim for indicators that extend our intuitive beliefs about how stuff works. If you’re willing to entertain that we might be onto something, check back for updates and further discussion.