Predicting Recessions The Easy Way: Monetarists, MMT, And The Money Stock
I have a new post up that has implications for stock-market investment, so I decided to try posting it over at Seeking Alpha, where they’re paying me a few tens of dollars for the post (plus more based on page views — not much luck so far).
The post argues that year-over-year change in Real Household Net Worth has been a great predictor of NBER-designated recessions over the last half century. (It’s either 7 for 7, or 8 for 7, over 50+ years, depending on the threshold you use.) If you were following this measure, you would have gotten out of the market on March 6, 2008, avoiding a 50% drawdown over the next twelve months.
But the post goes farther, offering a somewhat monetarist economic explanation but using total household net worth as the measure of the “money stock.” Short story: if households have less (more) money, they spend less (more). Not exactly a radical behavioral economic assertion.
If you’re wondering how recent days’ market events have caused billions (trillions?) of dollars to “disappear,” and are pondering how to think about that, you might find it an interesting read.
Cross-posted at Asymptosis.
This has been gnawing at me all afternoon. I spent some time building up the graph on my own hoping to find some flaw. The graph is impressive but I couldn’t believe it. But …
So, the question is, why does Household net worth decrease just before a recession?
It was not immediately obvious to me that the answer was that the recently unemployed either spend down savings or borrow more money immediately after losing a job.
This explanation fits neatly with my prior assumption that, recessions are brought on by employee layoffs. (The economy is humming along, manufacturers miss a drop in demand, inventories rise excessively and manufacturers layoff employees which reduces demand even further.)
Interesting graph. Of course there would be some delay in getting the data.
@JimH: “This has been gnawing at me all afternoon.”
Yeah, the pattern definitely raised my eyebrows.
> fits neatly with my prior assumption that, recessions are brought on by employee layoffs
I would suggest otherwise. The bottom 50 or 70% essentially have no assets or net worth.
This data series suggests to me that recessions are caused by declines in existing asset markets — mainly RE and the stock market. The top 50% has less money, so they start spending less. Employers sell less, and cut back on employment. Pretty straightforward…
Or, the clickbait version: All recessions are caused by financial crises.
I hear you.
But I never placed much credibility in the “wealth effect” or lack thereof.
I remember earlier recessions when inventories were huge. Old drive-in lots filled with cars. And now they tell us that inventories are up.
Also, look at the top 60% and you will find that it includes a lot of employees who can be laid off. They own stock in their 401Ks but they are not spending based on those accounts.
We have an economy which is laying off well paid individuals and they are shunted into lower paid jobs. Eventually that has to reduce their savings and thus net worth.
Our economy has more unemployed than the statistics suggest. It has more part time employees than in the past. Wages in this economy are barely going up. Put all that together and their net worth has to be reduced.
How are the part time and unemployed getting by? Those in the top 60% probably had savings which could be reduced, and slowly but surely it was. And so their household net worth is reduced.
This has been a gradual process that is continuing.
This deterioration began before 2008 and it has continued thru our supposed recovery. It was bound to reach a threshold where it could no longer be ignored.
Wages are more or less stagnant. Total household debt is extremely high and has been relatively stable since 2010. (In 2010 Q3 it was 11.844Trillion and rose to 11.853Trillion 2015 Q2.)
We have been witnessing a slow motion economic wreck.
And remember that your net worth statistic is an aggregated stat.
Of course it includes the paper profits and savings in commodities, stocks, and bond. But those belong to the “investor” class. They are not the “spenders” in our economy.
No one can seriously believe that those “investors” will spend a significant part of their net wealth. But it is included in your stat anyway.
I haven’t read it yet but it sounds like something I’ll find interesting. I wouldn’t be surprised if I clicked on it twenty or more times over the next few days.
BTW , I’ve got some ideas on money supply and such that I’ll run by you all as soon as I get some time to think it through.
Typically as the economy approaches a recession it causes rates to rise.
The rate rise in turn causes the stock market to fall that in turn reduces household net worth.
Something of that nature also happens in the housing market but the impact is not as big.
The more thing change the more they are the same.
I’m totally with you on the importance of (not your words, but…) rich vs poor’s marginal propensity to spend out of wealth and income. See for instance here:
And follow related-posts links at the bottom to see much, much more.
IMO our situation of wildly concentrated wealth and income both suppresses growth and creates the conditions for meltdowns. Think: an ever-collecting cornice at the top of an avalanche slope.
But as for proximate causes that happen fast, the rich are in a much stronger position to vary their spending quickly. I don’t think it’s crazy to suggests that big, fast gains, and especially losses (which tend to happen much faster), in the existing-asset markets could be the primary proximate driver of those households’ spending shifts. Sudden spending cutbacks by those with net worth could be sufficient, it seems to me, to trigger the avalanche.
The data series shown here at least lends some decent support to that theory.
JimH: “…We have an economy which is laying off well paid individuals and they are shunted into lower paid jobs. Eventually that has to reduce their savings and thus net worth…”
Plus reducing, indirectly, the net worth of their family and social circle, and their proximate community.
As individual households experience income degradation, at some point in time it will hit a break point where it falls to a lower wealth level. If the crisis is quick, as in a natural disaster or mine closure, I imagine households will adjust their lives quickly, to a much lower standard of living.
But if the crisis is slow, and appears to the individual to be random rather than systemic, their relatives and friends will likely offer assistance until they can get back on their feet. This spreads the weight of the slow crisis onto the rest of the community.
That’s a good thing, unless the pressures that created the crisis never return to normal.
I see your point, but the wealthiest among us do not each require 1000 vehicles, refrigerators, or water heaters. And your point is an invitation to ignore our worst problem leading up to the Great Recession. (And the Great Depression for that matter.) That is excessive consumer debt.
In the late 1920s and early 1930s small aircraft flying the mail would take off and fly thru terrible wet and cold weather. Ice would start to build up on the wing surfaces and eventually the pilot would be forced to trim the aircraft to increase lift. The flight would continue while even more ice accumulated on the wing surfaces and the pilot would be forced to trim the aircraft again. Eventually he is flying with the nose high and making some turning maneuver would be enough to cause the airspeed to decrease and the aircraft to stall and fall out of the sky. Pilot error would be blamed. Eventually heaters would be installed on aircraft wings and the aircraft is carefully deiced before takeoff. But even in modern times, inexperienced pilots fail to recognize the symptoms of icing and fall out of the sky.
Our economy can be compared to those flights.
Initially we are cruising along with a properly trimmed economy. (The Effective Fed Funds Rate is set at the norm.) Then consumer income raises start to lag or the consumer purchases more highly desirable consumer goods than he can legitimately afford, then must reduce spending and the economy slows. (The icing has begun.) The FED lowers interest rates and brings forward some sales and the economy picks up. Normally the FED would raise interest rates back to the norm but this time there are still signs of stress in the economy so they raise them but not back to the norm. All is well until the next lag in consumer income or excessive spending. Each time the FED repeats its earlier responses and now the interest rates are even lower than the norm. Consumers are going deeper and deeper into debt each time this happens and eventually consumers can not increase their debt. Finally interest rates can not be reduced further, spending must be reduced, and the economy stalls. In addition to the stressed consumers, fraud is discovered. There is plenty of blame to go around but no quick cures.
Note: IMHO the Great Recession was brought on by low wage growth after unbridled free trade was initiated, and the Great Depression was brought on by highly desirable consumer goods and no real life experience with the newly available installment credit.
You would think that all economists would recognize that reduced demand is a MAJOR problem in the economy. But you would be wrong.
Here is a quote from the January-February 1996 issue of the Harvard Business Review article by Paul Krugman (A Country is Not a Company):
“Moreover, beyond this indisputable point of arithmetic lies the question of what limits the overall number of jobs available. Is it simply a matter of insufficient demand for goods? Surely not, except in the very short run. It is, after all, easy to increase demand. The Federal Reserve can print as much money as it likes, and it has repeatedly demonstrated its ability to create an economic boom when it wants to.”
If our economy could have increased demand after 2008, it would have created good jobs provided that free trade was bridled to some extent. We should never lose sight of that.
I agree with pretty much everything you say, except this, which I think is a kind of silly statement:
“your point is an invitation to ignore our worst problem”
1. The post discusses one particular economic effect, out of zillions. (No post could cover all of them.) Paying attention to one effect is not “an invitation to ignore” others.
2. Consumer debt is directly involved in the effect being discussed — it’s the righthand side of household balance sheets (liabilities), which deducted from the lefthand side (assets) = net worth.
But again I agree with most everything else you say.
I disagree about the silly part, the supply-side trickle-downers will use anything to shore up their arguments. (Smiling here.)
I think the only thing we completely disagree about is the question of whether the top 50% of wage earners reduce spending and bring on recessions.
I believe that the effect comes on slowly from everyone buying fewer and fewer discretionary consumer goods. Until finally, the inevitable happens.
But your article is far more interesting than those considerations. I have overlaid the data for the DJIA from 2005 to try to gain some more insight. Still can not make up my mind about it. (Chicken or egg problem.) But you sure got my attention.
Unfortunately , I think our economy is becoming – more and more with the passage of time – one where ONLY the consumption of the rich will matter. When the bottom 90 or 95% only buys rice and beans , their portion of a shared rental on a converted shipping container , and maybe some fares on the driverless Uber-bus , then we can safely ignore them altogether , macro-wise.
QE and other such maneuvers aimed at artificially boosting asset values will only accelerate our approach to that neoliberal nirvana.
Maybe that’s what we need. Maybe we have to really hit bottom before things will change.
Let’s do it. Go ahead , Fed , buy up all of Roger Farmer’s stocks , at suitably inflated prices. Just tell Roger that when he trickles down on us , he should aim squarely at our faces. We’re having some trouble waking up , and that just might do the trick.
@JimH: “I have overlaid the data for the DJIA from 2005 to try to gain some more insight.”
Can you post or otherwise share that?? Including the spreadsheet/data if you’ve got it. Would be great to see.
I think JimH is probably working with this :
Adding the Wilshire and going to log scale gives a better perspective , and gets you data back to 1970 :
I was using the DJIA as Marko has pointed out.
Just add it to your graph as a separate Data Series as Marko has done. I used Percent Change from Year Ago for units. (No Transformation.)
I just tried adding the Wilshire, also using Percent Change from Year Ago for units. (No Transformation)
I have a technical question.
How do you get a global link address to a FRED graph? Have you established a user account?
My links all seem temporary.
Fred DJIA only back to 2015. Wtf.
If I use your original graph then:
Add the Wilshire 5000 to Modify the Data Series 1 and under transformation use (a-c)/b. (Display on left)
And add the Wilshire 5000 as Data Series 2. (Display on right)
I get Net worth excluding the stock market compared to the stock market. This revised net worth starts down before the stock market falls for the 1990, 2001, and 2008 recessions.
Steve Roth wrote @5:33: “Fred DJIA only back to 2015.”
Sorry, I missed your comment before. Probably a scale problem.
In my comment at 5:40,
In the Data Series 1, use “Index” as units for ‘c’ which is the Wilshire contribution.
In the Data Series 2, use “Percent Change from Year Ago” for units.
That should avoid scaling problems.
I have a user account but I’m not signed in. If you go to the share tab and click on the link button , it copies to your clipboard , which you then paste here or wherever. I’m not sure about the permanence of the links , however. I’ve gone back and checked a couple after a few weeks and they were still good. I have no clue how the system works , really. It all seems like magic to me.
I’ve played around like you’re doing – subtracting home equity and stocks from net worth to compare with debt growth rates. Basically , I came to the conclusion that stock booms are not necessarily associated with debt growth – the dotcom boom in particular. Real estate is associated with mortgage debt , of course , but interestingly the net worth after subtracting home equity and stocks is strongly associated with debt as well. I just did a quick and dirty study and need to do more on this.
Regarding your last point , I think if you look at the home equity component you’ll see what caused the early net worth (less stocks) downturns , certainly so for 2008.
About my comments at 5:40 and 6:11.
It would probably be better to look at where these graphs cross zero. Using that criteria, Data Series 1(Net Worth) always crosses zero before Data Series 2(Wilshire 5000 Total Market Full Cap Index©)
If you want market indices that go back before 1970 , here’s a selection that you can pick from :
Thank you for the tips.
I am only interested in when the ‘Percent change from Year Ago’ of “Net Worth (minus the stock market)” goes negative before the ‘Percent change from Year Ago’ of the stock market goes negative.
This would confirm my belief that “I believe that the effect comes on slowly from everyone buying fewer and fewer discretionary consumer goods. Until finally, the inevitable happens.”
Using Marko’s tip, here is a link to your graph modified to subtract the Wilshire from your Net Worth and provide a separate trace for the Wilshire for comparison.
Note that on this graph, zero for the left side scale is not aligned with zero on the right side scale.
Look at the recessions of 1990, 2001, and 2008.
I think you will agree that the stock market changes did not bring on those recessions.
I forgot to mention that your point about Net Worth is still valid, and still very interesting.
This just gets better and better.
Note that in June 1984 and August 1988 the Wilshire went strongly negative but there were no recessions!
But in September 1990 the Wilshire went only about as strongly negative as before and there was a recession.
The difference is in what “Net Worth minus the Wilshire” was doing.
The “Net Worth minus the Wilshire” between the recession is also interesting. Between 1990 and 2001 it was increasing and between 2001 and 2008 it was decreasing.
Better and better.
What was happening between the Qtr2 2011 and Qtr1 2012?
Whatever it was, it caused the change in “Net Worth minus the Wilshire” to be negative.
But the change in the Wilshire did not go negative.
And whatever it was, did not allow the negative change in “Net Worth minus the Wilshire” to cause a recession. It appears that “Net Worth minus the Wilshire” was overwhelmed by some other economic factor.
It was probably falling home prices. That should reduce Net Worth but would not decrease spending.
I like what you’re doing but I think you need to use the actual figure for the subtraction of stocks from net worth. I’ve substituted the flow-of-funds figure for HH corporate assets in place of the Wilshire index. Even this is not absolutely correct , since there may be debt associated with the stock acquisitions that would need to be netted out , but it’s probably pretty close.
Yeah, obviously we are making a rough estimate here, but that may be good enough. Otherwise the whole thing needs to be cleaned up.
I have wondered if removing home equity wouldn’t improve the result. (Subtract ‘total home values minus total home mortgages from Net Worth’) On the other hand, we went thru a period where home equity was converted to cash. That was certainly not the norm before the housing bubble.
But your change has broken the relationship between “Net Worth minus the stock market” and recessions.
Take a look at the 2001 recession in your graph. Your suggestion about the debt associated with the stock could be the solution since that would be a negative.
Anyway, I am just playing with Steve’s work. He might decide to wander down this road later.
Can’t just subtract the Wilshire index from HH net worth cause the index is only like 80 at its highest. Subtracted from trillions.
And, it’s a price index, not an index of total market cap $.
Am I misunderstanding?
Showing just the equity portion of HH assets is interesting. Obviously bonds (bigger quantity but smaller percentage moves) are important assets, as is RE (very big Q, much slower % moves).
But in any case that’s just the lefthand side of the balance sheet. Gotta include the right side to to get to NW. Liabilities/debt.
So *could* roughly view it as four different HH measures additively combined? Equities, bonds, RE, and liabilities.
You could calculate the total equities market cap time series using the wilshire, I suppose, starting at some known market cap on a given date.
On the Fed HH equity-ownership measure: I have no idea if that includes just direct ownership, or also mutual funds, pensions…
Point is that all equities are ultimately owned by households (or foreign holders). So take net worth and subtract total equity market cap, and you’ve got AllOtherAssets – Liabilities.
And that “Liabilities” is very precise: The HH Sector “aggregate liabilities to all other sectors.” Liabilities that HHs owe to each other (probably a pretty small number) have to be subtracted.
All of this is incorporated in HH NW.
Comparing the ‘Net Worth with the stock market’ to the ‘Wilshire’ is good enough to show which is going negative first before a recession.
But you are right that it can not be used to subtract the stock market from Net Worth. The DJIA would have had the same problem. I should have caught that myself.
This is starting to look a lot more complicated.
Yep. The only item that flow-of-funds data makes easy to separate from total HH net worth is home equity , which they provide as a separate item. That is , the liability is properly subtracted from the assets , so you can subtract home equity from total HH net worth and get something ~ ( HH net worth less housing ).
For corporate equities , anything that would be easy to do would be an approximation. One possibility is to deduct the “revaluation” component from the y-o-y change in total corporate equity assets figure. Since most of the gains in stock wealth have come from revaluation changes during the bubble era , this might give a decent estimate.
Here’s what the y-o-y asset change looks like compared to the annual revaluation change :
( Note : The revaluation data is already a “flow” so you need to use the “unchanged” value when comparing to the “change” value for assets. )