Saturday, December 19, 2015

"only then can we really see if rising debt is something we should be concerned about"


Simon says

Figures for aggregate personal debt should always be normalised with respect to household income, because only then can we really see if rising debt is something we should be concerned about, or just the result of growing incomes.

Lets look at that.

The first problem is to pick some appropriate data. I'm just gonna use "disposable personal income" -- personal income after taxes -- and CMDEBT, FRED's measure of household debt. I don't have good reasons for choosing these two particular series. I'm not an economist; I don't know these things. Usually I try to use the same data that the guy I quoted was using. But Simon is making a general statement and doesn't get into specific data. I'm on my own here.

At FRED I put CMDEBT over DPI, turned off the recession bars, and took five snapshots of the graph, with the mouse highlighting five different turning points in the plotted line. I cut and pasted them to get all five turning point indicators in one picture:

Graph #1: Household Debt relative to Disposable Personal Income, 1952-2015
Approximate dates of the turning points:
1 = 1964Q4 ... 2 = 1984Q3 ... 3 = 1987Q2 ... 4 = 2001Q1 ... 5 = 2007Q4

Before turning point 1: rapid increase
From turning point 1 to 2: not much change at all
From turning point 2 to 3: very rapid increase, faster than the years before #1
From turning point 3 to 4: gradual increase (slower than the years before #1)
From turning point 4 to 5: very rapid increase, like 2 to 3, for a longer period
After turning point 5: decline, at first very rapid but then gradual.

So -- what does all this mean? Can you see from the graph whether "rising debt is something we should be concerned about, or just the result of growing incomes"?

I can't.

From what we know about the economy, we can look at the graph and say wow, the very rapid, sustained increase from point 4 to 5 was something we should have been concerned about.

Funny thing, though. From what I know about the economy, I look at the graph and say we should have been concerned about debt from 4 to 5... and from 3 to 4... and from 2 to 3... and from the start to 1, too. The way I look at it, debt was going up all the while.

But, you say, but debt obviously was not going up from point 1 to point 2. I think that's what you say. And then you might add: so the debt problem could not have started before turning point 2.

Many people think that. It is true that the plotted line on Graph #1 runs pretty flat from turning point 1 to 2. But that does not mean debt wasn't going up. All it means is that debt wasn't going up relative to disposable personal income. It means debt and DPI were going up about the same, from point 1 to point 2. In the other periods, 2 to 3 and 3 to 4 and 4 to 5 (and also before point 1) debt was going up faster than DPI. That's what makes the line go up, debt going up faster than DPI.

Oh, and after turning point 5, debt is going up more slowly than DPI. That's what makes the line go down. Yeah. And the line going down is how we know rising debt is something we should have been concerned about before the peak at turning point 5.

Sadly, you don't know it's going to be a peak until after the line starts going down. And then ... well, by then it's too late.


What about that flat spot between turning points 1 and 2? I say debt and income were growing at the same rate, give or take. To see if I'm right we can separate debt from income and look at them separately.

Graph #2 shows year-on-year increase in household debt:

Graph #2: Growth Rate of Household Debt (blue)
The numbers are all over the place, of course. But you can see the numbers are lower in the 1960s than the 1950s, and seem to drift downward till 1970. Then there are three large jumps, two in the 1970s and one in the 1980s.

Knowing that much, we can say that household debt growth in the 1960s was mostly between 5% and 10%, with an average about halfway between. Then in the 1970s the plot line races up and down, but seems fairly well centered on the line that indicates 10% growth, up from 7.5% or so the decade before.

The next graph, Graph #3, shows year-on-year increase in disposable personal income, in red. Again the line is all over the place. But it looks to be centered on the 5% line in the 1950s and the early 1960s. Then it runs a little higher, centered maybe a little above the 7.5% line for most of the 1960s and early 1970s. Then it runs higher yet, averaging somewhere close to the 10% line for most of the 1970s and early 1980s.

Graph #3: Growth Rate of Disposable Personal Income (red)
Despite all the variations visible on the two graphs, for about ten years beginning in the early 1960s the red and blue lines both average out at about 7.5% growth or a little higher. And for about ten years beginning in the early 1970s both lines average out around a 10% growth rate. For that 20-year period, the growth rate of household debt and the growth rate of disposable personal income run neck and neck.

That's why we see a flat spot for that 20-year period on Graph #1.


There is another line on Graph #3, a very faint gray line. That line also shows disposable personal income, like the red line, but with inflation stripped out of the numbers. You can see that before the mid-1960s the red and gray follow the same pattern and run close together. Then the two lines separate for about 20 years; during those years there was rather a lot of inflation. Then after the early 1980s the two lines come together again.

As you can see, the inflation of that 20-year period pushed disposable personal income up. This is the reason debt and income grew at comparable rates in those years. It explains why the debt-to-income ratio on Graph #1 runs flat in those years.

If not for that inflation, there would be no flat spot on Graph #1. The graph would show increase from inception to crisis. At what point in that counterfactual would you say rising debt becomes a concern?

Rising debt is always a concern, because policymakers don't yet know enough stop it before it brings the economy to ruin.

8 comments:

jim said...

Hi Art,
In the last year you have revealed that you took out a loan to buy a car and that you are close to retirement. I suspect that means that you have money in retirement accounts (401K maybe) that you could have converted to cash to buy the car but you chose not to. Maybe you didn't even consider it. But you undoubtedly did consider your future income in weighing the decision to increase your debt. Everyone measures debt against income - I suspect even you do.

The issue of whether debt is an excessive burden (or a concern as you put it) is not really due to the absolute magnitude of debts but to how realistic the analysis of anticipated income was.

What changed at point #5 on your graph#1 was the private sector's analysis of income versus debt. As a result of that sudden change in analysis of many agents at the micro level, at the macro level private debt suddenly went from increasing by $1 trillion every 3 months to decreasing at a rate of 1$ trillion per year.

The Arthurian said...

Hi Jim. You wrote "Everyone measures debt against income - I suspect even you do."

I had a laugh over the "even you" part of that. Yes, even I stop and think about whether I can afford to buy a car before deciding whether to buy one.

In your third paragraph you wrote:
"As a result of that sudden change in analysis of many agents at the micro level, at the macro level private debt suddenly went from increasing by $1 trillion every 3 months to decreasing ... "

Glad you brought up macro and micro. I want to say macro is NOT equal to the sum of the micro parts. It is appropriate for individuals to measure debt against income. It is not appropriate for macro policy. Macro is (quite literally) the "big picture" view.

In your second you wrote:
"The issue of whether debt is an excessive burden (or a concern as you put it) is not really due to the absolute magnitude of debts but to how realistic the analysis of anticipated income was."

Yeah, "concern" was Simon Wren-Lewis's word. I had "burden" or "cost" or something, until I changed it to "concern" to remind the reader of the quote I started with.

It's not all one way, Jim. It's not all "how realistic the analysis of anticipated income was". Because debt affects income and growth. When debt is at a low level, income and growth may be so good that people underestimate anticipated income. When debt is at a high level, income and growth are dragged down by debt and then the estimate of anticipated income may unrealistically high.

An analysis of anticipated income may be too low or too high because the level of existing debt (in the economy as a whole) is left out of the analysis.

Jazzbumpa said...

Art -

Interesting post. I question your conclusions though. I look at your graphs and wonder - what narrative does that suggest? I think you come to data with a narrative, and try to make it fit.

If debt can be too great, either absolutely or by reference to some benchmark, then that suggests that a lower level of debt is OK. That seems to be the meaning of the debt/ DI comparison. And if some lower level is OK, perhaps there is some still lower level that is not enough.

Don't scoff. Debt has many good and useful purposes, and commerce could not run without it.

Now let's put graph 1 in a historical context.

Coming out of WW II there was a great deal of pent up demand, and personal debt levels were very low - in part because people had few uses for credit, and few opportunities to get it, even if they wanted it. During the 50's there was a very erratic expansion as millions of service men reintegrated into the working economy, started families, bought houses and cars, etc. etc. I don't think this should in any way be surprising.

One may posit that in 1950 debt levels were too low, and a post war correction took place. Too low for what? is a reasonable question. I'll suggest that the answer is - for the economy to operate at an optimum level, recognizing that that answer is both vague and facile.

Then for 20 years the household debt burden was constant. I think your reference to comparing real disposable income to nominal debt is unfair and invalid. Nominal to nominal, where you started, is more reasonable.

At the onset of those 20 years, the economy had finally stabilized from the massive disruptions of first the Great Depression and then the ensuing World War. In the first decade, starting in the early 60's, there were no significant economic upheavals. In the second decade there was noticeably growing inflation, resulting from [among other things] oils shocks, and a large number of women entering the workforce, along with the first wave of now grown boomer babies, like me.

Despite all that, the debt/DI ratio remained remarkably stable. Where I worked we got frequent cost of living adjustments in the pay envelope. inflation was high, but not devastating because we were able to keep up with it. I think it would be hard to argue that debt levels were too high in that period.

So why that sharp increase in the late 80's? I think answering that would require a deep dive into the specifics of those few years. Graph 2 shows debt growth in rapid decline over the decade. But in a close up view we can see debt growth balloon from Q3 '83 to Q4 '87, peaking in Q4 '85. Meanwhile the growth of D I is in a bumpy decline for the entire decade. It looks a lot less dramatic in a close up along with debt growth than it does in your graph 2.

https://research.stlouisfed.org/fred2/graph/?g=2Xow

Jazzbumpa said...

[Continued]
After that, per graph 1, debt/ D I never recovered. Growth of the ratio was slow through the 90's. DPI growth was choppy, but trendless, while debt kept growing from '92 on. I think in through there, attitudes changed, and people became more comfortable with a higher debt burden. This might have been generational. People who lived through the depression were largely scared of debt, and they were dying off. My generation hadn't had that experience - yet. And times were good, so why worry?

I always say policy matters. But it occurs to me that in this discussion you blame policy while I am looking at the behavior of consumers, and also the willingness of lenders to extend credit that they shouldn't have. This becomes more apparent in the first decade of the 21st century. But it's not an either-or situation.

After 2000, DI growth was remarkably stable at around 5% until the crash. Meanwhile, debt growth was running above 7.5% the entire time and well over 10% for a good chunk.

Arguably, somewhere along the line, debt/DI got too large. But another part of this was issuing loans to unqualified borrowers. Maybe the former couldn't happen without the latter. I'm not sure. I am sure, though that another aspect of debt - leverage to speculate in arcane financial instruments with no rational way of determining their value - played a major role in the crash. This is the big policy contribution [lack of regulation.]

That's my narrative. A long period of stable debt use seemed to be just fine. After 1980 your graph 3 shows D I growth to be substantially less than before 1980. To maintain a continuously improving standard of living - the promise of the American dream - people needed and were encouraged to borrow so they could live beyond their means.

All the buzzards came home to roost in 2008.

Debt use hasn't increased much since then. Have we earned anything?

Probably not.

Cheers!
JzB

Jazzbumpa said...

i think the idea of people basing behaviors on anticipated future income is largely bogus. Isn't that one of Friedman'a ideas that have been debunked?

Look at how people actually behave. I think the operative algorithm is - If I make this purchase, can I still afford to feed my cat?

People live in the present, and are rarely forward looking in the way that Jim's argument describes.

Look at how corporations behave - is the focus long term on on this quarter's stock holder report? Look at all the stock buy backs to increase shareholder value now, rather than invest in product development to increase value long term.

And most people are not even that financially sophisticated.

Cheers!
JzB

jim said...

Jazz wrote "Look at how people actually behave. I think the operative algorithm is - If I make this purchase, can I still afford to feed my cat?"

It's not a question of whether you make a purchase or not. The question is whether you going to borrow. that is a different calculation. People are always forward looking when they borrow. Look at the guy who borrows from a loan shark to play the horses. Or look at Art's loan to buy a car. If Art believed the stock market was about to crash and his 401K was going to be wiped out do you think he would borrow instead of converting 401k funds to cash? Ir maybe if he thought a lot of his future income would disappear.

If money is borrowed both the lender and the borrower analyze whether there is income available to make the payments. The whole point of borrowing is to bring future income into the present. A borrower has to think about future income. If he doesn't he's probably a thief who is going to take the money and run.

The reason many corporations are not putting borrowed money into product development is they don't think it will produce the income that you seem to think it will. It' may be a mantra of supply-side economics that all you have to is funnel the money into development and it will create it's own demand, but nobody in business really believes that nonsense.

The Arthurian said...

Hi Jazz.

"If debt can be too great, either absolutely or by reference to some benchmark, then that suggests that a lower level of debt is OK. ... And if some lower level is OK, perhaps there is some still lower level that is not enough."

Yes Jazz, exactly. Some level of debt is neither too much nor too little.

By the way, my conclusion is not that any debt is too much but rather that debt is a problem "because policymakers don't yet know enough stop it before it brings the economy to ruin".

I put the "yet" in there, thinking that policymakers will one day be willing to seek the level of debt that provides the most benefit and the least harm to the economy.

"Too low for what? is a reasonable question. I'll suggest that the answer is - for the economy to operate at an optimum level"

Exactly so.

//

"I think your reference to comparing real disposable income to nominal debt is unfair and invalid."

Really?? I'm not comparing real income to nominal debt. I'm comparing real income to nominal income, to show the increase of income that was due to inflation. I wrote: "As you can see, the inflation of that 20-year period pushed disposable personal income up. This is the reason debt and [nominal] income grew at comparable rates in those years. It explains why the debt-to-income ratio on Graph #1 runs flat in those years."

In Graph #3, I'm not comparing real income to nominal debt. I'm showing the effect of inflation on nominal income!

"In the second decade there was noticeably growing inflation ... Despite all that, the debt/DI ratio remained remarkably stable."

Stable, yes. But not despite the inflation. Stable because of the inflation. Inflation pushed income up so that income and debt were growing at about the same rate for that 20-year period. It is no coincidence that the only time the debt/income ratio was stable was the time known as the Great Inflation!

"Arguably, somewhere along the line, debt/DI got too large. But another part of this was issuing loans to unqualified borrowers. Maybe the former couldn't happen without the latter."

Agreed. I think that "issuing loans to unqualified borrowers" is the sort of desperate thing lenders had to do, when debt was already too large and the qualified borrowers were all full up. Lending more to qualified borrowers who are all full up turns them into unqualified borrowers anyway, I imagine.

//

Hey -- I'm retiring this week. My last official act is to go to the Christmas party on Thursday.

Jazzbumpa said...

Hey, Art - congrats on the retirement!

I already can't keep up with your posts. Can't imagine what will happen now.

Cheers!
JzB