Thursday, March 1, 2012

Watching the Dodo Bird Fly


In a post entitled Fear of inflation, Char Weise wrote:

The WSJ blog reports that companies are worried that inflation will rise in coming years...

Of course, raising inflation expectations is precisely the point (or should be, if the Fed is doing its job properly) of the Fed's program of quantitative easing...

[T]he fact that the businesses that were surveyed believe inflation will be higher than the Fed's target 5-10 years from now is an indication that the Fed's statements about its inflation target are not credible. And this lack of credibility should actually be helping to stimulate the economy.

Okay, I accept all that. But then Char says

[I]t would be preferable if the Fed were to loudly proclaim that yes, it does intend to let inflation rise above 2 percent in the future.

Char wants inflation.

I want to know at what moment it was that the inflation problem was magically transformed into the inflation solution. I responded to Char's post:

But are economists not embarrassed and ashamed to call for inflation?

Defending economists, Char said

I'm not ashamed to call for more inflation. Higher expected inflation would lower real interest rates and help get us out of the recession. An inflation rate of 3-4% would not have significant harmful economic effects relative to the current 2%.

No significant harmful effects? Depends who you ask, probably. I let that go.

I replied:

I understand that.... But I think the problem is excessive private debt, not excessively stable prices.

Defending a principle, farm land investment said:

[T]he private sector absolutely must deleverage. However, isn't it also true that a bit of inflation above trend would lessen the real value of these debts and contribute to the private sector deleveraging process, essentially helping to "inflate debt away"?

It's funny, you know? Change the phrasing a little bit, and it makes you think. "A bit of inflation above trend," FarmLand said.

What trend?

I think FarmLand was talking about the inflation trend. But I think for his plan to work, for inflation to inflate debt away, inflation has to be above the trend of debt growth. Otherwise prices increase but debt increases faster, and we get further behind.

I pointed this out in the Microbes post:

Debt relative to income is high. Sumner says we can solve the problem by increasing income via inflation. Did that work, say, in the 1970s?

Not really.

Maybe it did work for the latter half of the 1960s. Inflation pushed NGDP up, and at the same time inflation reduced the burden of existing debt: Inflation increased Sumner's denominator and decreased his numerator simultaneously. And all it took was a 23% increase in prices in the five years between 1965 and 1970. Meanwhile, total debt increased 44.6%, almost twice the rate that prices went up.

Prices couldn't go up fast enough to keep up with the growth of debt. After 1970, inflation got worse, but Sumner's "debt to income" ratio WENT UP ANYWAY.

As long as debt continues to increase, inflation cannot solve the problem of excessive debt.

And I pointed it out in Dancing with Steve:

Keen still fails to see the cost of debt in general as a cost issue for the economy. A factor cost issue.

I quoted Steve Keen:

[W]hen debt rises faster than income, and finances not just investment but also speculation on asset prices, the virtuous cycle gives way to a vicious positive feedback process...

My reaction?

debt ALWAYS "rises faster than income" -- except at the end

and I showed this graph:

Graph #1: Total Debt relative to GDP (GDP is Income!)

Debt ALWAYS rises faster than income, except at the end.

Jazzbumpa disagreed with that, of course:

Have another look at Fig. 1 of the current post; no growth from 1960 through the mid 70's. Very little growth over the first three decades shown.

And where is the break point? right after 1980. Hmmmmm.

Jazz sees everything in terms of 1980. Sure, debt increased faster after 1980. Sure, the uptrend was gentle before 1980. But still it went up!

If you look at that first half of the graph, debt increased more before 1965 than after. After 1965, we had the "Great Inflation" which (as farm land investment suggests) helped to reduce the burden of debt by inflating it away.

Still, between 1950 and 1980, debt increased substantially as a multiple of GDP:

Graph #2: Total Debt relative to GDP, through 1982

Between 1965 and 1980, despite accelerating inflation, debt increased relative to GDP.

Invert that ratio and take another look. On the graph below, the blue line shows GDP relative to total debt. See if you can find the spot where inflation reduced debt enough to let the blue line show an up-trend:

Graph #3: GDP relative to Total Debt, and Prices

Oooh! There it is! From 1965 to 1970, the blue line goes up.

Whoop-de-do.

The green line shows the price level as measured by the CPI. The red line shows the price level as measured by the GDP deflator. Both measures show a slight, straight-line uptrend to about 1965, then definitely rounded upward curves as inflation accelerated from 1965 to 1980, and then continued inflation after 1980 but again as straight-line uptrends.

Focus on the years from 1965 to 1980, when inflation was raging. To see how inflation affected debt and income, look at the blue line.

A reminder: GDP is "output" but it is also a measure of "income" because transactions are two-sided. When somebody buys some of the output, somebody receives income. The one number is recorded in two places: as income, and as output.

The blue line shows output (or, income) relative to total credit market debt owed. The general trend is downward, meaning that output (or, income) grew more slowly than total debt. The steeper the downtrend, the more slowly output increased. The flatter the blue line, the closer output growth was to debt growth.

Between 1965 and 1970, the blue line actually trends upward. For those few years, income grew faster than debt.

For the rest of the years of the Great Inflation, 1970 to 1980 or '84  maybe, the blue line trends down again -- slowly at first, then faster as time goes by. As inflation got worse during the 1970s, the plan to inflate debt away proved itself to be more and more a failure.

Why? Because -- as Char says -- inflation stimulates spending. And since we use credit for money, when we spend more we create more debt. And if debt hinders economic growth (which it most certainly does), then creating more debt slows the growth of GDP and pushes the blue line down.

You cannot reduce debt by controlling only one variable unless that variable is debt.

If you control only income, and you inflate income, and you do not stop the commensurate increase of debt, you have done nothing to reduce debt relative to income: You have done nothing to solve the debt problem.

Debt increases more quickly than prices:

Graph #4: Total Debt and the CPI

Debt always increases more quickly than prices, except during severe recession:

Graph #5: Total Debt relative to the CPI

We cannot inflate debt away. In order to solve the debt problem we must deal with debt directly.

13 comments:

Jazzbumpa said...

Too much here to digest at the moment. For now I'll just say that it's not that I see everything in terms of 1980. It's that every economic measure you can think of really changed around 1980. And I keep finding new ones. The number of vehicles per 1000 people has an inflection point close to 1980 - because that figure is related to per capita income.

It aint a figment.

Cheers!
JzB

Jazzbumpa said...

Ok. What does this suggest in terms of policy?

Have a look at your graph 1, adjusted for inflation. Surprisingly, it hardly changes the curve at all. But it does flatten the plateaus, and sharpen the break points.

http://research.stlouisfed.org/fredgraph.png?g=5oV

What you have are 6 distinct realms. I'm going to suggest - just as an opinion - that in the first realm, up to the mid 60's there was not enough leverage. Hope that didn't make your head explode.

Then we had a flat spot for about 15 years, ending in 1981. Imagine that.

After a sharp rise thought the mid-late 80's, there is another flat decade, or so.

Then in this century, the fastest growth ever - until the crash.

Here it is on a YoY % change basis.

http://research.stlouisfed.org/fredgraph.png?g=5oX

Note many individual years below 0.

The excessive debt growth took place during the Reagan and Bush II administrations. Supply side economics, and Supply side economics on steroids.

I suspect financial leverage caused the late 90's bumps.

Debt does not always go up faster than income. It does so mainly when the official policies are voodoo economics. And a bit of inflation really does help to attenuate the trend.

Cheers!
JzB

The Arthurian said...

jzb: "Have a look at your graph 1, adjusted for inflation. Surprisingly, it hardly changes the curve at all."

You took my TCMDO/GDP,
divided TCMDO by the CPI, and
divided GDP by the GDP Deflator.

The difference between your graph and mine is the difference between the CPI and the Deflator. That is all.

I took your
TCMDO/CPIAUCSL/GDPC1
and divided it by my
TCMDO/GDP
and compared it to
GDPDEF/CPIAUCSL

The patterns are identical.
The red line is lower than the blue on my graph only because the base years are different for the CPI and the Deflator. That is all.

Your graph is a very confusing bit of noise. That is all.

And hey, what is the conceptual significance of the result from dividing accumulated debt by the Consumer Price Index? That is not clear to me.

Jazzbumpa said...

You took my TCMDO/GDP,
divided TCMDO by the CPI, and
divided GDP by the GDP Deflator.


Nope. That would be adjusting for inflation twice, which doesn't make any sense to me. I didn't do anything with the GDP deflator.

I simply adjusted your Debt/GDP for inflation, which is pretty relevant to the topic of your post.

Visually, it makes a bit clearer what your graph 1 already shows - that the vast majority of debt accumulation occurred during the Reagan and Bush II regimes. There were two + decades in there when inflation adjusted debt growth was nil. Those facts are pretty important to understanding, developing and presenting your thesis.

Which graph is a confusing bit of noise? The YoY % change? That's a significant protocol in time series data analysis, and I suggest you become familiar with it.

Seriously, Art, I think you reject things that I say simply because you don't like hearing them from me.

If you want me to go away, then just say so.

Cheers!
JzB

Anonymous said...

As Jazz says , there was a real change - whether it started in 1980 or a few years earlier is beside the point. The total nonfinacial debt/gdp was remarkably stable for decades , at about 150% , as private sector leveraging was ofset by public sector deleveraging ( in debt/gdp terms , which is what matters , not absolute levels ).

Since 1980 , about half of the increased leverage has been financed by foreign borrowing - surely not a good sign.

The graphs in this paper illustrate the trends well :

http://voxeu.org/index.php?q=node/4206

Also see this , from Ben Friedman in 1987:

http://www.sciencemag.org/content/236/4800/397.abstract

Anonymous said...

Jazz,

This graph from a post by Rebecca Wilder illustrates the "voodoo economics" effect nicely:

http://4.bp.blogspot.com/_Et4TQ-a0gGU/TIzcEfwsS6I/AAAAAAAADOw/F6gH6RFZvRc/s1600/leverage_chart.png

Bush I , a voodoo non-believer , is an outlier in the post-1980 era .

The Arthurian said...

JzB: "Nope. That would be adjusting for inflation twice, which doesn't make any sense to me. I didn't do anything with the GDP deflator."

Jazz, my denominator is GDP.
Your denominator is GDPC1.
Your denominator is equal to GDP/GDPDEF.

You removed inflation from the numerator via CPI, and removed inflation from the denominator via the Deflator.

That's what your FRED graph shows.

The Arthurian said...

Anonymous: "As Jazz says , there was a real change - whether it started in 1980 or a few years earlier is beside the point."

I associate 1980 with the start of Reaganomics. It is perfectly reasonable to expect to find plenty of changes in the numbers, with 1980 (or so) as a turning point, because of the major shift in policy at the time.

However, Reagan was elected to change things for a very specific reason: there were already problems in the economy BEFORE Reagan. Any analysis that begins with Reagan is incomplete and likely flawed.

I do not deny that debt/gdp increased much faster after 1980 than before. It would be silly to do so, for the increase is obvious.

What I do not understand is why anyone would deny that debt/gdp also was increasing before 1980. Clearly it was.

The VOXEU link shows a graph that excludes financial sector debt, which increased all through the 1960s and 1970s. This graph also excludes the 1950s, when nonfinancial debt shows significant increase.

The Rebecca Wilder leverage chart associates a significant decline with George H.W. Bush; this is the decline that I identify as providing "slack" in the financial system, a necessary precursor to the "miracle" years of the latter 1990s. Note that this graph also excludes financial debt.

The ScienceMag link points out that slow growth of government debt mostly counterbalanced private debt growth before 1980, an important point. This link also says "total outstanding debt remained steady relative to nonfinancial economic activity." I'm not sure what "nonfinancial economic activity" is, but again finance is being excluded -- from the debt numbers, I suspect.

These links exclude financial debt, and then claim debt didn't grow. I don't buy the argument.

Anonymous: "The total nonfinacial debt/gdp was remarkably stable for decades..."

Again, the exclusion of financial debt. But the cost of financial debt is borne by the non-financial sector. And financial debt adds to the demand for credit. And financial debt is not immune to cascade failure. And finance is part of our economy. Financial debt must be counted, not ignored.

The growth of finance relative to the growth of production is the central problem of our age. Excluding finance from the picture is refusing to face the problem.

The Arthurian said...

JAZZ
I noticed a change in the way FRED behaves lately. Editing a graph, I added a data series by typing GDP into the search box and hitting ENTER. What I got was GDPC1, I don't know why, possibly because it was the first one on the list that came up.

Perhaps the same thing happened to you, because your graph has GDPC1 in the denominator and you apparently didn't know it.

Good, because it didn't make sense to me either that you would remove inflation from numerator and denominator both.

I checked this morning; FRED doesn't seem to do that any more.

Jazzbumpa said...

Jazz, my denominator is GDP.
Your denominator is GDPC1.
Your denominator is equal to GDP/GDPDEF.


Shit! I pulled the wrong series.

I just tried to compare three different GDP series at FRED, and it wouldn't do it. It overlaid the same series 3 times.

I tried again, using a different series first, and it did work. WTF?

Then I did what I intended to do he first time:
TCMDO/CPIAUSL/GDP and got this astounding result.

http://research.stlouisfed.org/fredgraph.png?g=5qA

I need a martini.

JzB

Jazzbumpa said...

What I do not understand is why anyone would deny that debt/gdp also was increasing before 1980. Clearly it was.

I certainly never said that. What I did say, and will say again, is that there were periods of meaningful duration when debt/gdp did not rise: 1963 - 73.

http://research.stlouisfed.org/fredgraph.png?g=5qB

I was also flat from '91 to '94 and mid '95 to mid '98, about 5 out of 7 years.

JzB

Jazzbumpa said...

Forgot the last link.

http://research.stlouisfed.org/fredgraph.png?g=5qC

The Arthurian said...

Jazz,
The flat at '95 to '98 (or more like 1997-98, based on your graph) is flat because GDP growth was really good then. This was the time of the "macroeconomic miracle".

The flat at '91 to '94 -- visible in my DPD graph here -- is related to a debt growth slowdown that began around 1986, which I discussed here.

But all of this misses the point, which is that the relatively slow increase in TCMDO/GDP, 1965-1980, was due to a very serious inflation. NOT to "a bit of inflation" to use your words.

Finally, I continue to reject the "correction" of TCMDO for inflation. Neither TCMDO/CPIAUCSL nor TCMDO/GDPDEF gives a useful measure of debt. To be useful, you would have to MULTIPLY debt by the inflation series, not divide.