The Money Growth Component

In mine of the 10th I quoted Antal Fekete:

... the trend in marginal productivity of debt has been most definitely down. Once the zero bound is reached, any incremental increase in aggregate debt will have a detrimental impact on the economy.

I pointed out that the last sentence of the excerpt is based on the view that "rising total debt caused the downtrend in GDP growth", the view that I hold. I wrote: If adding to total debt produces a smaller increase in GDP than it formerly did, then harm has already been done.

Next, I evaluated Fekete's analysis of the graph he provides. I rejected his view that going off gold caused the decline of the marginal productivity of debt, and I ended with a series of four graphs offered to clarify my view -- the four graphs I've been repeating every morning for the last few days.

Let me say now, how these graphs are relevant to Fekete's post and to the productivity of debt. Perhaps that wasn't adequately clear in the post. Based on Fekete's text, his graph, and the trends I see in his graph, I summarize the history of "the marginal productivity of debt" as one long downtrend with an upward countertrend during the years 1986-2000.

The rate of growth of "Potential GDP" -- best-case growth -- follows a similar trend: one long downtrend, with an upward countertrend from 1992 to the year 2000 -- a downtrend that I attribute to the growth of debt, and a countertrend that I attribute to a significant reduction in debt growth.

The gap in the dates noted here, between the 1986 start of the debt productivity countertrend and the 1992 start of the countertrend in Potential GDP, the dates 1986-1992 identify the time of the significant reduction in debt growth.

To summarize, then: The downtrend in Potential GDP and the downtrend in debt productivity result from accumulating total debt and the cost imposed by that debt on the productive sector of the economy. The uptrend in Potential GDP and in debt productivity resulted in large part from a slowing of debt growth.

In addition to the slowing of debt growth, to achieve the uptrend it was necessary to have an increased growth of circulating money.


According to Fekete's graph and discussion, the marginal productivity of debt has been declining since the 1950s. I accept that, but point out a persistent and regular increase lasting near 15 years. My graphs attempt to explain this time of increase in debt productivity.

My graphs show a decrease in debt growth which begins in 1986 with the start of the increase in debt productivity. This decrease is closely followed by an increase in the growth rate of spending-money -- essentially a low-cost substitute for borrowed money. Together, slower debt growth and more rapid money growth create a shift in the Debt-per-Dollar ratio from 1990 to 1994. This down-shift in Debt-per-Dollar was the key, I claim, that boosted real and potential GDP growth, a boost which lasted to the 2001 recession.

The medium-term increase in marginal debt productivity that appears on Fekete's graph was created, in my view, by the fall in debt growth, and was supported by an unusually rapid growth of money in the early 1990s. Here are three views of that money growth:

Graph #1: The Growth of Total Debt (blue) and Spending-Money (red)

Graph #1 shows a large red spike in M1 money growth beginning just after the 1991 recession. The blue line, showing the growth of total debt, peaked in the mid-1980s and fell dramatically until just after the 1991 recession. No other red spike rises significantly above the blue (debt) line, until the crisis in 2008.

Graph #2: The Growth of Non-Federal Debt (blue) and Base Money (red)

Graph #2 shows a strong red "M" pattern following the 1991 recession, indicating significant growth of base money as compared to debt growth (shown in blue). Except for that "M", the red (base money growth) line is almost completely submerged below the blue (debt growth) line.

Graph #3: A History of Federal Debt Held by Federal Reserve Banks

Graph #3 shows the amount of Federal debt that has been monetized -- in other words, the Federal debt held by Federal Reserve banks. The blue line here represents an older data series, and the red line a newer series. Both are shown relative to GDP for context.

All through the inflationary 1970s, Federal Reserve holdings of Federal debt were falling as a percent of GDP. The early 1980s hump in money growth can be seen on this graph, beginning during the 1982 recession. The late 1980s hump is clearly visible. And the early 1990s hump in money growth rises to the mid-1990s, and never falls until the crisis.

The overview of this graph is interesting, I think: From start-of-data to 1981 the trend is down -- from 7% of GDP to 5%, to 6%, and finally to 4% of GDP -- due, I suppose, to a lot of inflation-fighting.

The trend since 1981 is up and up. One, two, three humps of increasing size; a fourth, even bigger hump in the 2000s, and a monster fifth hump after the crisis.

It might be interesting to compare debt-to-GDP numbers of 1960 to those of the latter 1990s. Both times the debt-held-by-the-Fed number was around 5% of GDP, but there was an astounding difference in total debt, and in non-Federal debt, relative to GDP and otherwise. For another post.

Anyway, the point of these three graphs is to show multiple views of rapid money growth in the early 1990s, which caused a shift in the Debt-per-Dollar ratio and, I argue, was a key part of the economic change that led to near a decade of better-than-average economic growth.

More tomorrow.

Evidence that Excessive Debt is the Problem

From mine of 10 July:

A decline in debt growth begins in 1986:

Graph #3: Beginning in 1986 there was a strikingly unusual fall in
the growth of total debt that lasted into the early 1990s.

An increase in money growth soon follows:

Graph #4: Before that decline of debt growth had ended, an increase in the growth of
circulating money was under way. This increase lasted to the mid-1990s.

These patterns shift the "debt per dollar" ratio, 1990-94:

Graph #5: The declining debt growth, combined with the rising money growth, meant for people who use money that we were relying less on money with the extra cost of interest, and relying more on money without that extra cost. We were able to save money on our money. That's got to be good for the economy!

In response, best-case GDP improves almost immediately:

Graph #6: Potential GDP shows a full percentage point improvement between 1993 and 2000,
in response to the reduced ratio of accumulated debt to circulating money.

Interference Pattern

In an update to mine of 9 November 2011 there is this:

Quick and dirty, a graph showing both the debt-per-dollar increase and the fall in money relative to output.

Graph #1: DPD and MRTO

Interesting, the way the two seem to interfere with each other, there in the middle.

I want to take a closer look at that interference pattern.

Let me use the version of M1 that includes "sweeps" this time. It's a more accurate measure of M1 money, I think.

Note: The relative position of the red and blue lines differs on the two graphs because

1. the start-dates for the M1 money series differ,
2. the values for the M1 money series differ, and
3. the blue lines are on the left scale, the red lines on the right.

So the nifty interference pattern that appears on graph #1 doesn't even look like interference on Graph #2. But now at least you know why.

Graph #2: Debt per Dollar (red) and Money relative to Output (blue)

Oh, this is excellent. The mid-1980s peak of the second blue hump just kisses the end of the second red hump. Makes it easy to see that during these humps, the two lines tend to move in opposite directions.

It's pretty clear also that before the early 1980s also, the two lines moved in opposite directions. Let's look at that.

The blue line shows the quantity of spending-money, compared to the GDP we bought at the prices we paid for it. Usually when people show the "money relative to output" (MRTO) graph, it goes up. That's because they compare the quantity of money to inflation-adjusted output. They leave inflation in the numerator, and take it out of the denominator. That pushes the line up, right along with inflation.

But as you can see in Graph #2, when you compare the money we spend to the stuff we bought at the prices we paid for it, the line goes down.

The quantity of money fell from more than 21 cents (per dollar's worth of output) to less than 14 cents, between the late 1960s and the early 1980s. Those years coincide pretty well with the Great Inflation, when prices tripled. Evidently, it wasn't the quantity of M1 money that caused the Great Inflation.


The difference between the two versions of money used in the graphs, the difference arises in late 1993. But it becomes significant -- noticeable on my graphs, say -- in mid-1995.

On both graphs the blue line shows a small hump in the early 1980s, and larger humps in the late 1980s and early 1990s. After that third hump the difference arises. Graph #2 shows a hump in the 2000s, while on Graph #1 the blue line falls consistently from the third hump to the 2008 recession.

Let me zoom in on the second graph now, to look at the 1975-2000 period. The period includes all three humps that fascinate me so, plus a little extra on each end.

Graph #3: DPD (red) and MRTO (blue) 1975-2000

One simple way to read this graph is to notice that whenever the blue line goes down, the red line goes up. It's true from 1975 to the start of the 1982 recession. And it's true from peak to trough of all three humps visible on the graph.

When the blue line goes the other way, when it goes up, the red line tends to slow its upward pace (1980 recession) or even to fall a little (both 1980s humps) but falls significantly, strikingly, only with the 1990-1994 increase in the blue line.

Now look back at Graph #2 and consider the fourth hump.

Coming out of the third hump the blue line falls until the 2001 recession. During that time the red line rises. The two continue to move in opposite directions.

Then from 2001 to mid-decade the blue line rises. But so does the red line. This contradicts the "opposite movement" pattern.

By that point the growth of debt was dominating everything.


One more point about the second graph.

That third hump in the red line -- not the crisis-related peak there at the end, but the nicely rounded hump centered on the start of the 1991 recession -- just above that hump is the horizontal gray line with 0.17 at one end and 17.0 at the other.

The red line shows how much debt there was for each dollar of spending-money in our economy. That's what the 17.0 is: For every dollar in somebody's pocket, or in their purse or their checking account, there was just about $17 of debt. $17 dollars of debt in this country for every dollar we use for spending.

And that was in 1991. Things got worse.

At the same time, in 1991, the blue line shows there was only about 14 cents of spending-money for every dollar's worth of output.

Before the late 1970s, and going back to the late 1940s, the quantity of money had been much higher. But money-growth was restricted to fight inflation for all those years, and we ended up with not much money with which to buy the things we make.

We couldn't use the money we didn't have, to buy things. So we borrowed money, and spent that. The government did it, of course. But we all did.

That's how we ended up with so much debt.

How to Get Better Growth

From mine of 10 July:

A decline in debt growth begins in 1986:

Graph #3: Beginning in 1986 there was a strikingly unusual fall in
the growth of total debt that lasted into the early 1990s.

An increase in money growth soon follows:

Graph #4: Before that decline of debt growth had ended, an increase in the growth of
circulating money was under way. This increase lasted to the mid-1990s.

These patterns shift the "debt per dollar" ratio, 1990-94:

Graph #5: The declining debt growth, combined with the rising money growth, meant for people who use money that we were relying less on money with the extra cost of interest, and relying more on money without that extra cost. We were able to save money on our money. That's got to be good for the economy!

In response, best-case GDP improves almost immediately:

Graph #6: Potential GDP shows a full percentage point improvement between 1993 and 2000,
in response to the reduced ratio of accumulated debt to circulating money.

It's Hump Day

There are three significant humps in money growth -- in the early 1980s, the late 1980s, and the early 1990s:

Graph #1: Annual Growth Rates for M1 and Base Money, 1967-2007

Jim and I have been having some discussion about the early-1990s decline in Debt-per-Dollar, the growth of M1 money as a contributor to that decline, and the causes of M1 growth.

I want to say that M1 growth was supported by base money growth.

You (not Jim specifically) may tell me that the growth of base money was forced -- that the Fed had no choice -- because private-sector money is the driving force.

I will say: Of course the Fed had a choice. (See that thin blue spike between the 1980 and 1982 recessions? There was no matching increase in base money then, was there? Because Paul Volcker chose not to allow it.) There is always a choice, until it's too late.

I do agree that the Fed's power has waned in the decades since WWII, and that private sector money has become more and more the driving force. But this decay of Federal power happened because of the growth of finance -- because of the perpetual increase in the ratio of private-issue to Federal-issue money.

Further, if you are saying private-sector money is the driving force and that's final, then you are saying monetary policy is powerless, and life as we know it is over.

Four Thoughts

From mine of 10 July:

A decline in debt growth begins in 1986:

Graph #3: Beginning in 1986 there was a strikingly unusual fall in
the growth of total debt that lasted into the early 1990s.

An increase in money growth soon follows:

Graph #4: Before that decline of debt growth had ended, an increase in the growth of
circulating money was under way. This increase lasted to the mid-1990s.

These patterns shift the "debt per dollar" ratio, 1990-94:

Graph #5: The declining debt growth, combined with the rising money growth, meant for people who use money that we were relying less on money with the extra cost of interest, and relying more on money without that extra cost. We were able to save money on our money. That's got to be good for the economy!

In response, best-case GDP improves almost immediately:

Graph #6: Potential GDP shows a full percentage point improvement between 1993 and 2000,
in response to the reduced ratio of accumulated debt to circulating money.

Cost Hinders Growth: GDP-Debt Ratios

In this morning's post I showed the ratio of GDP to Total Debt as a blue line, and the ratio of the change in GDP to the change in debt as a red line:

In a comment on the post, Anonymous wrote:

The math dictates that over periods when the blue line is flat , the red line will cycle around the blue line so that its average value during the period is equal to that of the blue line.

When the blue line is flat at a value of , say , 1.0 , it means that $1 of gdp is matched by $1 of debt. If debt and gdp cycle up and down , but on average at the same rate of growth , then the ratio remains the same and the blue line is flat.

If the blue line is flat at a value of 0.5 , the percentage rate of change in gdp still happens at the same rate as for debt , the difference being that it is no longer at a 1:1 ratio , rather at a 1:2 ratio.

When the average value of the red line falls over time , it by definition means that the value of the blue line will fall as well, since the stock of debt is growing at a faster rate than gdp , which makes the gdp/debt ratio fall.

First, I want to say thank you for the comment. I couldn't see it till you said it.

Honestly, I couldn't even see it after you said it, so I made a spreadsheet so I could test out what you said.

Now I can see it. So, thank you.

Cost Hinders Growth

Graph #1: More and More Debt, Compared to GDP

If you look, you can find lots of people saying what a bad thing it is that we have so much debt, compared to GDP.

I agree. In my view, excessive debt -- total debt, I mean, not just the Federal debt -- is the reason we no longer get good growth. The cost of the debt hinders growth.

But good or bad, growth is measured as changes in GDP. GDP and changes in GDP and debt are knotted like tangled fishline. If the data and relations were less tangled it would be easier to show the thing I want to show: that excessive debt hinders growth.

On the graph below, the blue line shows a comparison of debt to GDP, as in the graph above. But this time the ratio is inverted, so it shows GDP relative to debt. Inverted: You can see that the blue line starts a little above 1 on the one graph, and a little below 1 on the other. You can see that it reaches a value of 2 around 1987 on the one graph, and a value of 1/2 around 1987 on the other, and that in both cases the blue line has a nearby "kink". And on the Graph #2 you may be able to see a second kink, just at the end of the 2008 recession, which corresponds to the high point on Graph #1.

Graph #2: Less and Less GDP, Compared to Debt

The red line also shows GDP relative to debt, like the blue line on Graph #2. But the red line is based on the change in debt and the change in GDP, from 12 months earlier. The blue line is the ratio of GDP (figured as an annual total) to total accumulated debt.

Does this graph show anything interesting? Or is it completely obvious that the ratio of changes, though jiggy, would follow closely the same pattern as the ratio of totals? I'm not sure. I've not stared at the graph long enough.

I am tempted to think there's an obvious reason for the similarity of the two lines, or for the similarity of their overall trends. After all, total debt is nothing more than the accumulation of the annual changes in total debt, and GDP is ... But no, come to think of it, GDP is *not* an accumulation of annual changes in anything.

For total debt, the annual increase adds to all that was there before. For GDP, none of what was counted before is counted in the next new total. And yet the choppy red line seems to follow closely the path of the smooth blue line. What explains this?

Could it be that some feedback emerges from the ratio of totals, feedback that affects the vigor of GDP growth? A cost, perhaps, that imposes itself upon GDP?

The cost debt, perhaps?

Excessive debt is the reason we no longer get good growth. Cost hinders growth.

The Marginal Productivity of Antal E. Fekete

So I googled debt productiv and one of the options was marginal productivity of debt -- comparing increases in GDP to increases in total debt. Exactly what I was looking for. The first result for that was Antal E. Fekete - The marginal productivity of debt... | Facebook. Fekete's got this graph:

Graph #1. Source: Antal E. Fekete

One nice thing. Antal Fekete shows his calculation:

Marginal Productivity of debt = (Change in nominal GDP)/(Change in nominal debt)

Unfortunately, he doesn't identify his data. I tried to duplicate his graph using FRED's TCMDO debt, but it didn't seem to match up. So I'll just go with Fekete's graph.

He writes:

The marginal productivity of debt measures the change in (dollar) economic output for a dollar change in the aggregate debt outstanding... What has been happening to this measure over the decades?

Nominal debt outstanding has, of course, been accelerating rapidly and since 2008, the government’s share of that has been soaring. Allowing for the short term variation – it can be seen that the trend in marginal productivity of debt has been most definitely down. Once the zero bound is reached, any incremental increase in aggregate debt will have a detrimental impact on the economy.

Agreed. Of course, implicit in that last sentence there is the idea that rising total debt caused the downtrend in GDP growth. Oh, I think that's true, certainly. But people who make policy don't seem to agree. And people who vote are iffy.

On second thought, I don't agree that once the zero level is reached, any incremental increase in aggregate debt will have a detrimental impact on the economy. I think debt already had a detrimental impact on the economy, since the 1980s when it slowed our GDP growth trend... in the 1970s, when monetary policy had to fight inflation... and in the 1960s, when financial cost gave the initial push to the cost-push inflation.

Would you wait until adding to total debt decreases GDP to say there is a harmful effect? Or would you say what I say: If adding to total debt produces a smaller increase in GDP than it formerly did, then harm has already been done.


We come now to Fekete's analysis of his graph:

It can be seen that the ratio was around 1½ during the mid-1970s – meaning that for each dollar in marginal debt, $1.50’s worth of GDP growth occurred. This ratio was as high as 3 in the 1950s. Deterioration in the ratio could be argued as deterioration in the quality of the debt. What would cause the quality of debt to decrease substantially? The removal of the ultimate extinguisher of debt: gold. Gold prevented & cleared the build-up of toxic unproductive debt.

I can almost see the 1½ he's talking about. (I'm glad Fekete uses the Alt-0189 version of one-half.) And I can believe that the ratio was higher in the 1950s. But it is odd he doesn't show the 1950s on his graph, as the decade adds so much to his argument.

His numbers -- as much as $3 in GDP growth for each dollar of unidentified debt growth -- don't ring a bell with me. I'm sure I looked at this before and I don't remember anything like $3 per. Maybe I had it wrong. But it sure would help if Fekete said what debt he was using, and linked to a data source. He doesn't.

I can live with that. That's not what drives me to write this post. It's the next thing he says that gets me. Fekete writes: "Deterioration in the ratio could be argued as deterioration in the quality of the debt." What is that -- moral philosophy? It is surely not economics.

When debt is going bad, it is because people are not making the payments. That's what causes deterioration in the quality of debt. When it happens, lenders are liable to stop lending, or set higher standards. You know, to keep the deadbeats out. And when the economy goes bad they think we're all deadbeats.

Is that what Fekete means? Clearly not. "What would cause the quality of debt to decrease substantially?" he asks. "The removal of the ultimate extinguisher of debt: gold."

Fekete says debt productivity is down because we pay off debt with money that isn't gold. The only way this can make sense is if you assume from the start that money has to be gold and if it isn't, then every problem we have is a result of not using gold for money. But that's really not an analysis of the problem.

Nor does Fekete's view explain the problems that arose in the past, that forced us off gold.

If we no longer use gold for money, one could possibly argue that the quality of money has deteriorated, the intrinsic value, say. But that has nothing to do with debt. Fekete takes advantage of people's confusion about money and debt to make his argument.

And there could be some vague connection between "what money is made of" and "how happy people are to get their money back" from borrowers. But really, if you lend paper you can't honestly expect to be repaid in gold.

Furthermore, there is no connection between "what money is made of" and "how productive the things are that are done with money". To assume that there is, is to say something like If money was gold I would invent a new technology for getting to Mars, but money is paper so I'm gonna just get drunk instead.

That's not "rational" behavior.

Fekete's analysis is terribly disappointing. But now, look at the timing. What caused the ratio to decline? Going off gold, he says. Going off gold in the 1970s. (President Richard M. Nixon "closed the gold window" on August 15, 1971.)

We can't see it on Fekete's graph, but he suggests that the marginal productivity of debt fell from $3 in the 1950s to less than $1.50 by 1975. Does he want us to think that all of this decline happened after 15 August 1971? I don't buy it. But even if that were the case, the decline could be due entirely to the very same cause that led to the closing of the gold window. Fekete does not discuss this.

Meanwhile, I don't see how any of this has any bearing on the things people do with the money they borrow, or how productive those things are.


I made a copy of Fekete's graph and eyeballed-in some trend lines:

Graph #2: Same as Graph #1 with Trend Lines Eyeballed In

I see a down-trend from 1975 to 1986, and a downtrend from 2000 to the crisis. And a downtrend overall, sure. But there is a stretch there, from 1986 to 2000, when the marginal productivity of debt increased with persistence and regularity. How is this possible? Did we go back on gold?

We did not.

Obviously it wasn't going off gold that caused the long-term decline.


Now, my version of the 1986-2000 phenomenon...

A decline in debt growth begins in 1986:

Graph #3: Beginning in 1986 there was a strikingly unusual fall in
the growth of total debt that lasted into the early 1990s.

An increase in money growth soon follows:

Graph #4: Before that decline of debt growth had ended, an increase in the growth of
circulating money was under way. This increase lasted to the mid-1990s.

These patterns shift the "debt per dollar" ratio, 1990-94:

Graph #5: The declining debt growth, combined with the rising money growth, meant for people who use money that we were relying less on money with the extra cost of interest, and relying more on money without that extra cost. We were able to save money on our money. That's got to be good for the economy!

In response, best-case GDP improves almost immediately:

Graph #6: Potential GDP shows a full percentage point improvement between 1993 and 2000,
in response to the reduced ratio of accumulated debt to circulating money.

As debt grows in size, the increasing cost of it increasingly hinders both production and consumption. As debt grows in size, therefore, the growth of output falls behind. Thus is the decline in the marginal productivity of debt explained.

The reverse is also true: As debt falls in size relative to circulating money, the cost of debt falls relative to the money we have. This reduces financial cost, boosting wage and profit income in the productive sector. Improved growth follows.

The downward shift of the debt-per-dollar ratio opened the door to improvement in real and potential GDP performance. We can repeat that success. We can do better.

We must do better.