FCM and the Long View

Using Jazzbumpa's source,

I dug up productivity data from the BLS website, which took quit a bit of digging. Fortunately, Skeptic at Reality Base did the same thing, and identifies the data series - year-over-year non-farm productivity growth (% change) as PRS85006092.

we can now take a longer-term view of productivity and the Factor Cost of Money.

The BLS site offers "Major Sector Productivity and Costs" and provides numbers with units given as "% change quarter ago, at annual rate". I couldn't get "change from year ago" numbers but it probably doesn't matter.

I grabbed the data for all the years I could get -- back to 1947. (Annual numbers since 1948.) Popped 'em into a Google Docs sheet and fiddled with moving averages to get something fairly smooth.

Brought my "FCM per GDP" numbers into the file and threw a graph together:

Source: My Google Docs Spreadsheet
Edit 8 Feb 2019: The old link to the live graph is not working. But the graph is okay
in the spreadsheet. So I replaced the live graph link with an image.

In outline: Productivity goes down as the Factor Cost of Money increases; and productivity goes up as the Factor Cost of Money declines.

Looking at the early years, one could argue that productivity and the FCM moved up together when the FCM was very low, when finance was still facilitating economic performance. On my graph the two travel together until 1962, then veer apart. If I plotted the moving average at the last year of the period, the separation date would be 1966. So, somewhere in the 1962-1966 period is when finance became excessive and started hindering rather than helping economic performance.

FCM and the "Macroeconomic Miracle"

I still want to look at the Factor Cost of Money in relation to the "macroeconomic miracle" years.

Robert J. Gordon, 2002:

NBER Working Paper No. 8771
Issued in February 2002

This paper examines the sources of the U. S. macroeconomic miracle of 1995-2000 and attempts to distinguish among permanent sources of American leadership in high-technology industries, as contrasted with the particular post-1995 episode of technological acceleration, and with other independent sources of the economic miracle unrelated to technology.

Dean Baker and John Schmitt, 2007:

From the mid-1990s on, however, official productivity growth again accelerated rapidly, returning to a 2.9% rate reminiscent of the golden age. Quite suddenly, though, in the second half of 2004, productivity growth dropped sharply.

James Kahn and Robert Rich, 2011:

Following a resurgence of strong productivity growth in the late 1990s and early 2000s after nearly a quarter-century of slow growth beginning in 1973, the latest reading from a trend tracking model now indicates that slow productivity growth returned in 2004.

Mary Amiti and Kevin Stiroh, 2007:

Chart 1 Trend Labor Productivity Growth

Source: The Conference Board and Groningen Growth and Development Centre, Total Economy Database, January 2007, http://www.ggdc.net. Notes: Labor productivity is defined as real GDP per hour worked. Trend estimates are based on a Hodrick-Prescott filter with a smoothing parameter of 100.

That blue line on their chart shows US productivity dropping off continuously from 1961 to 1978, dragging the bottom till 1993, then rising during the "miracle" years.

Finally, I came upon BLS productivity numbers for 1990-2011. I figured the annual change in productivity, and plotted that against the Factor Cost of Money as a percent of GDP:

Source: My Google Docs Spreadsheet

I'm not making any bold claims for this graph. But it looks to me like productivity and the factor cost of money tend to move in opposite directions.

Keep in mind that productivity is not a policy; it is a goal. Monetary policy is policy. So if there is causality, it is the factor cost of money that influences productivity. Not that monetary policy presently pays attention to the factor cost of money. But maybe they should.

//

In evaluating my graph, I would ask that you not only point out its weaknesses but also examine its possible strengths. Ask, for example, Does it make sense?

For example: If our basis for comparison is Krugman's concern with the Zero Lower Bound and the need to have negative interest rates to revitalize the economy, then yes, it makes sense in that context. Negative interest rates would reduce the Factor Cost of Money, and reducing the FCM would enhance productivity and growth.

Given this basis of similarity, let me point out what the FCM captures. It is not only the rate of interest, but also the size of accumulated debt that affects the cost that inhibits growth.

Examine the strengths.

FCM: PAIR and the FedFunds Rate

I took Graph #2 from yesterday's post and renumbered it:

Graph #1: Federal Outlays for Interest relative to the Two Measures of Debt

The graph shows FRED's FYOINT ("Federal Outlays: Interest") as a percent of the Federal debt, for both the $10B and $14B version of the Federal debt. I'm considering the lower set of numbers as the "average" interest rate paid by the Federal government on its gross debt.

I'm in the early stages of using that set of numbers as a proxy for the average interest rate paid by everybody else on all the debt owed other than the Federal debt. I'm calling this interest rate the PAIR, or Proxy Average Interest Rate.

I took and slapped the FEDFUNDS rate onto the graph, for comparison:

Graph #2: PAIR and the Federal Funds Rate

PAIR and FEDFUNDS move in the same general direction, up together until the early 1980s, and down together thereafter. FEDFUNDS is a lot more variable, of course, but that's because the Fed jerks it around to influence economic activity. Then too, the PAIR is an average interest rate for debt accumulated over many years, so you might expect it to be slow-moving.

As the FEDFUNDS rate is essentially the lowest interest rate we have in the US, I have to say the PAIR is really a lowball estimate. Probably a good average for government debt, but even the red line is probably lower than the actual average interest rate on private sector debt.

FCM: The Factor Cost of Money -- Further Review

Turns out, the info I needed to figure the Factor Cost of Money is right there in FRED.

A quick search turned up four data series on the Federal debt:

Series	MaxValue	Units	Description
GFDEBTN	$14B+	Millions	Federal Government Debt: Total Public Debt
FYGFD	$14B+	Billions	Gross Federal Debt
FYGFDPUN	$10B+	Millions	Federal Debt Held by the Public
FGTCMDODNS	$10B+	Billions	Total Credit Market Debt Owed by Domestic Nonfinancial Sectors - Federal Government

The $4 billion difference between the first two series and the last two is the $4 billion of the Federal debt held internally by parts of the Federal government such as the Social Security Administration.

The two series that end up something over $14B look similar to each other on the graph. And the two that end up in the neighborhood of $10B look similar to each other.

Graph #1: The Federal Debt, four ways from sunday

One of the two series that passes $14B is given in Millions, and the other in Billions. Same is true for the two series that end up near $10B. I want to use the debt number as a denominator under the interest number, interest paid by the Federal government. And the interest number is given in Millions, so I will use the two debt series that use Millions, and ignore the two that use Billions.

Graph #2: Federal Outlays for Interest relative to the Two Measures of Debt

Graph #2 shows FRED's FYOINT ("Federal Outlays: Interest") as a percent of the Federal debt, for both the $10B version and the $14B version of the Federal debt.The red line shows the average interest rate the government pays, based on the smaller debt number. The blue line shows the average interest rate the government pays, based on the bigger debt number.

I'm assuming that the blue line is the accurate one... In other words, I'm assuming that "Federal Outlays: Interest" includes outlays for interest paid to the Social Security Administration and the other government agencies holding parts of the Federal debt. The government don't get to NOT pay the interest, just because of who holds the debt. I think.

The lower line on Graph #2, the blue line, shows the same peak as the vertical bar graph I showed a few days back, a peak around 7½% in the early 1980s:

Graph #3: The Proxy Average Interest Rate (PAIR)

So the interest rate I have figured is a lowball estimate.

Also, Federal debt has always been considered less "risky" than anybody else's debt, so the Federal government has typically paid lower interest rates than anybody else. So again, when I use that rate to estimate the total interest cost of the private sector, that number is a lowball estimate.

Estimating the Factor Cost of Money (2)

Picking up where I left off yesterday...

So I worked out some "average interest rate" numbers, uploaded the file to Google Docs, and cleaned it up a bit.

I'm calling this interest rate the Proxy Average Interest Rate, or PAIR.

Graph #1: PAIR, the Proxy Average Interest Rate

Now I want to look at total debt and use my average interest rate to figure a total interest cost for each year.

Then I'm gonna want to compare that number, the factor cost of money, compare it to wages and profits, see how things changed over the years, and see how those changes fit in with good times and with hard times.


Got annual TCMDO (debt) numbers from FRED, added them to my Docs file, and figured the interest owed each year at the PAIR rate. I just multiplied TCMDO debt times the PAIR rate. This series I'm calling the Factor Cost of Money, or FCM.

Graph #2: FCM, the Factor Cost of Money

Next, I got annual GDP numbers from FRED and added them to the file.

And I took a quick look at the Factor Cost of Money as a percent of GDP:

Graph #3: FCM as a Share of Total Income

Then, because it looked interesting, I compared FCM per GDP to the FedFunds rate.

Graph #4: FCM as a Share of Income, and the FedFunds Rate

You can see a similarity between the two data series. Sure, because the interest rate is one of the determinants of the Factor Cost of Money. But there is also a difference between the two series, and this is due to the other determinant: the level of total debt, which continued increasing from the early 1980s to the late 2000s while interest rates were falling. So the FCM has been more than the FedFunds rate, ever since interest rates started falling in the early 1980s.

As you can see, recently the FedFunds rate has been down almost to zero, but the Factor Cost of Money has not dropped below 5% because there is so much debt, public and private. And 5% is well above what the Factor Cost of Money was at the start of the graph, when our economy was having its Golden Age.

You might object, saying

Yes, yes it's true: The Factor Cost of money started out well below 5%. But it reached 5% before 1970, and the Golden Age continued on to 1973.

Right! It was the increase in the Factor Cost of Money that killed off the Golden Age.

In the early years, the Factor Cost of Money was less that five percent of GDP and was growing only slowly. In other words, the drain of income from the productive sector to the financial sector was at a minimum. Never again has that cost been so low. And never again have we had a Golden Age.

Estimating the Factor Cost of Money (1)

If you Google interest cost or interest cost data or stuff like that, it's pretty easy to find how to figure the cost of a loan, and pretty easy to find info on everybody's favorite fall guy, the Federal government.

But it is hard to find info on interest costs in the private sector.

So I had a thought: Why don't I use the total interest cost on government debt to figure an "average" interest rate for government debt each year, and then use that number as a proxy for the average interest rate of private sector debt.

I don't know how totally valid this approach might be. But it is representative. It will give me something to look at -- assuming I can find the relevant data for the Federal government!

//

Got it. The OMB site has a Historical Tables link with a good handful of Excel files available for download.

Table 6.1 shows the composition of Federal outlays for 1940–2016, including Net Interest. Table 7.1 shows Federal debt at the end of the year for the same period. I took the ratio to get an average interest rate for the whole Federal debt.

It's got the same 1982 peak that you might see in the Federal Funds rate. No surprise there, I guess.

See that odd short bar in the middle, around 1977? That's an accounting adjustment labeled "TQ" (Third Quarter) in the file. (It's in both tables.) If I use these numbers as a proxy for the average private-sector interest rate, I'm just gonna delete the TQ line.