Two days earlier, traveling with the wife: A lot of traffic, enough that she commented on it. Mid-morning on a Wednesday, a week before Halloween.
Not particularly busy times of the day. Not weekends or holidays or special days. Just ordinary days. But roads and parking lots were busy. It happens once, you don't think twice about it. It happens again, you have to stop and look. To me, it looks like economic vigor.
Then too, there are three or four new houses going up a mile down the road from me. I haven't seen that kind of activity since the '90s. Looks like vigor to me.
//
Ten years back, every time you turned around somebody was saying there's gonna be inflation, bad inflation. Because of Ben Bernanke and "Quantitative Easing". It got to the point, it seemed they wanted inflation, bad inflation. Whatever the reason, I don't know, but you heard the prediction so often it seemed like they wanted things to go bad.
Lately, last couple years maybe, it's been like that with predictions of recession: There's gonna be a recession in 2018. Then: There's gonna be a recession in 2019. Then: 2020. Now? 2020 or 2021. It's like they want it to happen.
Desire influences the predictions, no doubt. That's the trouble with politics.
//
Back in March of 2016 I predicted economic vigor: "This is going to be the full tilt, rapid output growth, rapid productivity growth, high performance boom."
In April 2016: "In two years everyone will be predicting it."
And in August: "The stage has already been set for the vigor that will be attributed to our next President."
Well it's been over three years now, and I've been hearing nothing but predictions of recession. Oh, there was a flurry of media reports a year or two back, about how "good" the economy was doing, two percent growth and all. Two percent is not good growth. Vigor means four percent, and hints of more. We're not there. The economy is not "good". Some people set the bar way too low.
I was wrong about vigor starting to be obvious by 2018. And the repeated refrain of recession has recently raised doubt in my mind that we'll ever see vigor again.
The numbers aren't there. Private debt is still painfully high, by any measure:
 |
| Graph #1 |
And "Debt Service" continues to drift downward:
 |
| Graph #2 |
15 comments:
The above post may be seen as a follow-up to this one.
Looks like they do some significant revising of the debt service data. Maybe I'll do a post on that.
Maybe I will.
Hi Art
You might want to look at how debt service is calculated. It looks bogus to me. It appears they just calculate a percentage of debt owed.
Consider the debt service of the federal govt. The way I look at it, it has been zero for 20 years. Each year the govt borrows more to pay it. There is no debt service burden on taxpayers. All of ot is paid for by borrowing.
The same can be said for households. If the revolving debt is growing larger each year there is really no debt service burden. To be a burden the debt service has to reduce the ability to spend.
I would be interested in seeing what it would look like if you subtracted the change in debt from the debt service.
In other words when you borrow as much in a year as you payback in principal and interest you get to spend as much as your entire disposable income (debt burden is zero). Only if you borrow nothing does the debt service reduce your ability to spend by that much.
Hey Jim. I worked with a guy in the '90s who thought as you do about the debt burden. He said every month when he made a payment against his credit card debt, he'd go out and buy himself a new shirt, to get even.
That was in the '90s. By now he probably has to go out and buy himself a new car every month to get even.
Sneak Preview of the graph I'm showing tomorrow -- data from ALFRED.
Here is the graph i was thinking of
https://fred.stlouisfed.org/graph/fredgraph.png?g=poag
The graph is about what I imagined, During the housing bubble the burden of debt was close to zero. Even below zero at times.
And then when house prices stopped going up and extracting equity from houses became impossible the cost of debt jumped up to its real value (the burden of paying all the principal and interest that is due).
Even today the burden of debt payments is higher than it has been for most of the 40 years the graph covers. That means the money available for spending is lower than it has been for 40 years.
Jim, I get what you are saying about no burden:
"In other words when you borrow as much in a year as you payback in principal and interest you get to spend as much as your entire disposable income (debt burden is zero)."
You're saying that if the debt increase equals the debt service, aggregate demand is not reduced. But I don't think the debt increase and the debt service are as close as you think, because the debt increase is a net increase.
Suppose that in one year, the debt increase is 60 and the debt service is 50. So the difference is only 10; the numbers are pretty close. However, if the change in debt was 60 after paying back 50, the gross increase must have been 110, and that's NOT close to 50.
If change-in-debt runs close to debt service, what it means is that we borrow about twice as much as we pay back.
PS, if the 50 debt service was 20 principal and 30 interest, the 60 net debt increase means the gross increase was 80, not very close to the 50 of debt service. And there is still that 30 that has to be paid, the interest cost of the transaction.
Terminology: For me the words "debt burden" refer to the cost of repayment, apart from all else (even debt increase).
Art, You are assuming that the Fed has good data on how much is borrowed and how much is paid back. They admit that they have no such data.
They have pretty good data on how much is paid in interest and the debt totals at the end of each period but that is it. They don't know on how much is borrowed and how much is paid back. The DSR is just a made up number based on the assumption that a percentage of the current total is paid back each period.
In theory you would think that if you subtracted debt service from disposable income what is left is what you have to spend, but you would also think that if so much a month goes to paying debt then your debt would be going down. Since debt is going up the whole concept is bogus.
Debt service should be considered as equal to the change in debt. It is only when the debt is decreasing that debt service is taking away purchasing power from income.
Here is another graph:
https://fred.stlouisfed.org/graph/?g=poUS
The red curve is household disposable income.
The blue is change in debt added to the red.
The green is debt service subtracted from the red.
The black curve is retail sales.
The blue and black are pretty close to the same shape, but look at the green! The green curve makes it look like nothing at all happened in 2007-2012. Obviously the DSR tells you nothing meaningful about what households have available to spend out of disposable income.
Jim: "...but you would also think that if so much a month goes to paying debt then your debt would be going down. Since debt is going up the whole concept is bogus."
I can't accept that at all! People are borrowing all the time, and people are paying down debt all the time. Both numbers are generally increasing, but the "borrowing" number is generally bigger than the "paying down" number. This is what the FRED data says, and that's what it says whether the numbers are "good" or not.
Analyses of the DSR that I have read say they don't know how high the numbers really are, but they think they have a good picture of the overall shape of the DSR curve. This view is supported by the revisions to DSR, which put the curve sometimes high and sometimes low on the graph, but consistently present the same basic shape.
Your argument that "if so much a month goes to paying debt then your debt would be going down" can only be true if the monthly borrowings are smaller than the monthly paybacks. You leave the monthly borrowings out of your evaluation and THAT, I think, is what's "bogus" here.
I think what you aren't getting is that I am looking at this from the perspective of what is left of income after debt service is paid.
I leave monthly borrowing and payback out of it because both are unknown quantities.
The only known quantity is how much debt is changing each period. It should be obvious that when debt is increasing that means more is borrowed than is paid back.
What I am declaring as bogus is the idea that if DSR is at let's say 10% that means that households have 90% of their disposable income available for spending. I am saying anyone who is thinking that statement is correct is mistaken.
The problem is that when you set up a ratio of debt payments to income you are implying that those payments come from income. That is wrong.
I posted a graph (with the green curve) that shows what it looks like to subtract debt service from disposable income. The shape of that green curve alone is pretty good evidence that the idea that represents what is available to households for spending is invalid. The blue curve is obviously a much better representation of what households have available for spending.
For the last 40 years there have been more and more financial innovation that has allowed debt service to be paid not from income but from borrowing more.
I challenge the notion the the 4 years preceding the financial meltdown was a period of high debt service payments. it was the exact opposite. It was a period that due to to innovation such as loan consolidation, sub-prime mortgages, negative amortization, home equity loans, reverse mortgages, etc the borrowers paid very damn little towards debt service out of income. Sure there was significant amounts of principal and interest paid off but it did not come out of income , it came from more borrowing. And for the most part the people who benefited from not having to pay on debt spent the money instead. This created a false and unstable economy that eventually caved in on itself.
This is what Minsky called Ponzi lending and that is why Paul McCulley called the crash the "Minsky moment".
Yes, I think you may be onto something with that green curve that "makes it look like nothing at all happened in 2007-2012". But that's the kind of thing I have to think about for a good long while.
Meanwhile, if I borrow a dollar, that dollar has to be counted. And if I pay off a dollar of debt, that dollar has to be counted, too. You say "It should be obvious that when debt is increasing that means more is borrowed than is paid back." It *is* obvious. But they still have to be counted; both the borrowing and the paying off have to be counted. Otherwise, the arithmetic contradicts the obvious.
Money is fungible. The money I earn and the money I borrow both go into a pool of funds I have available for debt service and other uses. To point at that pool of funds and claim that ONLY THIS PART, THE BORROWED-FUNDS PART of that pool was used for debt service... makes no sense at all to me. And it's certainly not true.
Out of the pool of funds comes money for debt service and money for spending (and money, say, for saving). If N% of the pool of funds is used for debt service, then the remaining "100% minus N%" is available for other uses. And that is true no matter the source of the N% that goes to debt service.
RE: that green line and your "poor us" graph poUS... I get it now. That's a good graph!
Retail sales (black) fell when new borrowing (blue) fell. And debt service (green) fell with a lag, and fell only a little because debt service is a small percentage ... no that's not right. Debt service in billions is several times larger than the change in household debt. Unless FRED's "annual rate" thing is messing me up again...
"If N% of the pool of funds is used for debt service, then the remaining "100% minus N%" is available for other uses."
That assumes the money "used " for debt service is burned or disappears into a black hole. At the macro level what comes out of the pockets of one household goes into the pockets another.
But that is not what I'm talking about.
I am looking at this at the micro level.
In 2004 you could have got a loan for one million dollars to buy a house and you could also have borrowed the N% to service the loan. Then a year later you could sell the house for 1.1 million and settle up your debt. The lender has made a handsome 10% profit and you got to live in a mansion for free. If you were lucky you could unload the house for more than 1.1 million and make a little profit yourself. That is just one example of how debt service can be paid without constraining household spending.
Needless to say with deals like that there are a whole lot of borrowing and a whole lot of houses sold. House prices and debt were rising 10%-12% as a consequence of all the borrowing.
But then house prices started to fall and then there were a lot of loans in default. At default the borrower still does not pay The principal that is owed. It comes out of the lender's pocket.
" both the borrowing and the paying off have to be counted."
Absent any accounting fraud what is borrowed is always paid back one way or another. That is true. But I don't think you can say anything meaningful about household ability to spend on other things based on the DSR. I would be curious if it correlates to anything households are doing.
"Debt service in billions is several times larger than the change in household debt. Unless FRED's "annual rate" thing is messing me up again..."
It is the "annual rate" thing.
Debt service is a ratio of annual income even when the frequency is quarterly, so you either need to change the change in debt to annual frequency or make it change in debt from a year ago
"Debt service is a ratio of annual income even when the frequency is quarterly, so you either need to change the change in debt to annual frequency or make it change in debt from a year ago"
THANK YOU! I remember you pointed out the annual rate problem in one of my graphs before, long ago, but I understand it better now.
Is there some principle that limits the increase in debt so that it does not go above debt service? If so, maybe I could use high points on the red line to approximate the blue line before 1980.
Hi Art
I just read again the Fred page describing how DSR is estimated. They don't claim this is the amount actually paid. It represents an estimate of the minimum required amount that should be paid. For example, for past due credit card debt that is around 2%-3%.
I made a graph of the ratio of total Household debt to household debt payments. It looked like it correlates to interest rates for the past 40 years, so that might help with estimating what DSR was before 1980
https://fred.stlouisfed.org/graph/?g=ppBJ
Interesting. I did a graph very much like that one the other day:
https://fred.stlouisfed.org/graph/?g=pmdr
The blue line is household debt service as a percent of household debt, billions as % of billions.
The red line is the interest rate, and like yours, my graph shows much similarity between red and blue.
But I wasn't sure what interest rate to use so I used the "effective" rate, calculated as interest paid by households as a percent of household debt.
When I looked at what I had done, I saw that the two calculations were almost the same: the one looks at debt service, the other looks at the interest portion of debt service, in both cases relative to debt. So it's no wonder the lines were similar.
Comparing the two vertical scales, you can even see that the blue line runs about five percentage points higher than the red line. That difference is the principal component of debt service!!
//
I was thinking a bit about when I go to the bank to get a loan. My repayment options are limited to a "window". The upper limit is set by my ability to repay, and the lower limit is set by what the lender is willing to accept. Both my ability to repay and the lender's willingness to accept are strongly related to economic conditions, I think. So debt service (both before 1980 and since) must be related to economic conditions...
Yes, it looks like it is calculated as 5% of income plus interest.
Since debt has been mostly growing more than income that suggests that the amount households are required to pay back each year has been shrinking as a percentage of total debt they owe.
Post a Comment