Sunday, July 10, 2016


You thought it was gibberish didn't you, the title. But it snot.

f = Federal
d = Debt
h = Held
b = By
f = Federal
r = Reserve

Those six letters make up most of the series name for two data series at FRED:

I checked the check boxes for both series and clicked Add to Graph. Result:

I'm in a mood, so you get the default shape of the graph as demented by FRED.

There's that big up-thing on the right there, what with the crisis and all. I don't care about the big up-thing. I'm gonna cut it off like a finger in a gory movie.

Graph #2: FDHBFRB and FDHBFRBN thru 2007
Now there's not so much that's interesting to look at, except what I want you to look at. Bumps. One bump between 1985 and 1990, and one that seems to peak at 1995. Goose bumps: The Fed was goosing the economy.

I'm gonna go in and edit the graph and set the units to "Percent Change from Year Ago" for both series. So we can see the growth of Fed holdings of Federal debt.

Well shit. I did that, the graph changed, and the cut-off-finger years magically re-attached themselves to the graph. (The graph ends at 2016 again, not 2007 like I want.) So I cut them off again:

Graph #3: Percent Change from Year Ago for the Data Shown on Graph #2
There. Maybe you can see a hump centered between 1985 and 1990. And the second hump between 1990 and 1995. Yeah but there's humps all over the place. I have to move the data to Excel.

Odd, isn't it, how "bumps" changed to "humps" when the graph changed from "Billions of Dollars" to "Percent Change from Year Ago".

Number one, I want to take the two data series and average them together. That way I end up with one data series for the period shown.

Next I want to take and put a Hodrick-Prescott on it, to improve the visibility of the trend in the data.

Done. And it came out better than I expected:

Graph #4:Average of the Two FRED Series (blue) and the H-P Trend (red)
An H-P high around 1966, another high in late 1985, one in 1993, and one in 2002.


Okay. This being Saturday, the wife made me breakfast. I came back to the computer an hour or so later, ready to double check the dates of those four highs in the H-P trend line.

Nothing. No Excel file. It's gone. Where the hell is it?

Not on my desktop (where I put everything).
Not in my downloads folder. I didn't download the data from FRED.
Not on the list of recent files in Excel.
Not on the list of recent files in the Windows menu.

Where the hell is it? I know I created the file: I used it to make Graph #4.

Oh you know what? It probably got saved where Excel files go when you choose to open them rather than downloading them.

Like that. So, where does the file go when you "open with" like that? I have to do another one to see. Open it and check the path.

The Temp folder. It goes to the Temp folder in the Local folder in the AppData folder in my folder in Users. If I can remember all that.

So what's in the Temp folder?

Looking for an Excel file... not the first one... the second one I think.

Copy to Desktop.
Yeah that's it. The source for Graph #4 is there. Got it. Okay. I must have forgot to save the thing to the desktop immediately after I opened it, and it went to the default place. Temp.

You thought it was a virus, didn't you.


What I want to do now, I want to look at Real (inflation-adjusted) GDP and RGDP per Capita, percent change of these:

Graph #5: Percent Change from Year Ago for RGDP per Capita (blue) and RGDP (red)
I want to download the data, average the two together (as I did above) and figure the Hodrick-Prescott for it.

Graph #6: Average of the two RGDP Series (blue) and the H-P Trend (red)

Now we're getting to the good stuff. (I have not seen it yet, myself.) I want to take the Hodrick-Prescott from Graph #4 (Fed Holdings of Federal Debt) and put it on a graph with the Hodrick-Prescott from Graph #6 (RGDP). This will let us compare the growth-rate trends.

So this next graph shows the H-P for Fed holdings of Federal government debt, in blue. And it shows the H-P for RGDP growth, in red:

Graph #7: H-P Trend from Graph #4 (blue) and H-P Trend from Graph #6 (red)
I would want to argue that RGDP responds to changes in Fed holdings. Looks like there are some timing issues to deal with. Variable lags. But my day's energy is all used up and I don't want to get into it now. Maybe another day.

I will say only that the blue line shows policy, and the red shows the result of policy.

// (afterthoughts)

The constant used in the H-P calculations is 1600; all data are quarterly.

Here's my Excel file. Note, the file contains Kurt Annen's VBA code for the Hodrick-Prescott.


The Arthurian said...

Or maybe no causal relation between them. I have to look at it for a while.

The Arthurian said...

Note that the last peak in Fed Holdings occurs in 2002. And the big drop in Fed Holdings between 2002 and 2007 occurs entirely before the crisis.

Cause and effect? In my book, yes.

Come to think of it, I looked at this before.

The Arthurian said...

On Graph #7 the blue line (the trend of Fed holdings) falls from 1966 to 1981. So Fed policy is evidently *not* responsible for the "Great Inflation". Unless you want to say that all the effect of the blue-line increase (before 1966) was delated until the start of the Great Inflation in 1965. But I don't see how that could be right.

From a different perspective, the red line (the RGDP growth trend) starts to fall mid-1960s, just a little before the blue line peaks. Could it be that RGDP growth reached an upper limit at that time ... and that the additional increase in Fed holdings after that only burned itself off as inflation?

Yeah, I like that perspective. But then the blue line falls till 1981. Somewhere along that way it must have come back to a non-inflationary level that the economy should find acceptable. But that didn't seem to happen. Instead, small up-or-down changes in the blue line (1966-1981) seem to have been magnified in the red line, with a lag of about a year.

Then the blue hump around 1986 peaks slightly after the corresponding red hump. This is similar to what happened 20 years earlier. So a question arises: Was there an increase of inflation for a few years after 1986?