Here we go again...

FRED is redesigning their graph system, again.

Couple links with announcements

https://news.research.stlouisfed.org/2016/05/freds-new-look/

https://news.research.stlouisfed.org/2016/05/do-i-look-ok-weigh-in-on-freds-new-suit/

//

For the record, they still have not fixed bugs in the previous new system (officially launched on 17 March 2014). Bugs in color selection and date selection. just off the top of my head.

Color selection works UNTIL you add a second series, like when you want to show a ratio. After you've got the two series and the calculation entered, you can't change the color of that line any more.

Date selection -- use the drag bar to set start- and end-dates, and it shows up correctly on the graph. Then get the link for your graph and open it in a new window, and it knows nothing of your date settings!

Eh, no matter. What matters is to make it "new". Keep changing the way the graphs look, and keep changing what you have to do to create them. Oh, and keep moving the controls so we have to hunt for them all the time.

The case of the missing seasonally adjusted Employment level

When I search for employment level at FRED, it finds 1244 series. Third on the list is

Employment Level, Thousands of Persons, Not Seasonally Adjusted (LNU02000000)

but I can't find a "seasonally adjusted" version of it. Oh, there's plenty to look thru, broken out by race and gender and marital status and god knows what-all, and various combinations of those. Plus, it looks like half or more of the 1244 results are for the unemployment level, in case I want exactly the opposite of the thing I asked for.

Thanks, guys.

I know they have the numbers, they must have the numbers. They have Employment level: Men, seasonally adjusted and not. And they have Employment Level: Women, seasonally adjusted and not. So, unless FRED is counting more than two genders these days, they can add up the men and women series and get Employment Level [Total] for "seasonally adjusted" and for "not". But I'll be damned if I can find it.

I got a screen shot of the first page of results, most of the first page. Maybe you can see the third one down is "employment level" with no subsetting conditions. That one's not seasonally adjusted. But I'll be damned if I can find its seasonally adjusted mate.

The best part of that screen shot is that I used Internet Explorer for the search. Couldn't find FRED's search box on the IE display. The search box is always on the right. This time, it's on the left and it overwrites some other text. In addition, the whole lower part of the display is pretty well garbled. And down at the bottom, I love that Done, but with errors on page message. Internet Explorer, what a piece of shit.

I just used IE for the screen shot. Normally I use Firefox, which creates no such errors.

In the end, I gave up and just did a Googlesearch for

"employment level" "seasonally adjusted" "FRED"

The second hit showed me ...... FRED series name. First ... and so I tried it. The PDF opened up in Firefox. I did a quick CTRL F search for FRED, and there it was: FRED series name CE16OV, the employment level. The third item down in Table A2.

Holy cow. Can it be I missed it, among those 1244 FRED search results?

//

If you do a FREDsearch for CE16OV -- that's letter O, not number 0... Civilian Employment, age 16 and Over -- there is no ambiguity and FRED shows the graph of the CE16OV series.

If you do a FREDsearch for CE1, just the first three characters, you get four items, including CE16OV. But it doesn't really help, because if I know to search for CE1 then I probably remember the series name and don't have to do any search at all.

If you do a FREDsearach for employment level and get 1244 results -- 25 pages of results, thank you very much -- and then add an sa tag to filter out everything that's not "seasonally adjusted", you're down to five pages: 235 series.

Using Firefox's CTRL F search then, searching those five pages for CE16OV, I get nothing on the first page... nothing on the second page... nothing on the third page... nothing on the fourth page... and nothing on the fifth page. Hm.

Looks like a glitch in the matrix.



// Edit: 5:20 AM 1 April ... They fixed it already! 1245 results now. And look what's right at the top of the list: CE16OV.

I won't even try to explain this one

Screen capture of FRED graph source page

Started with three data series. Deleted the red one. Now there are only two data series -- as you can see from the two "EDIT DATA SERIES" lines at the bottom of the image.

However, there are three lines on the graph, and three series identified in the blue upper border.

Screwy.

FRED's new Terms of Service

UPDATE 20 Feb 2016: shortened the intro. added images and notes at the end.

Here's the copyright notice as copied from FRED's new Terms of Service:

FRED® Graphs ©Federal Reserve Bank of St. Louis. . All rights reserved. All FRED® Graphs appear courtesy of Federal Reserve Bank of St. Louis. http://research.stlouisfed.org/fred2/

I noticed that after the ©Federal Reserve Bank of St. Louis there are two periods where I expected one. So I went looking through the HTML. The text of the required statement has an error.

See  how some of the words and letters are purple? Like TD at the top and TABLE at the bottom of the image... Almost all the text within angle braces is purple. That's because the "view-source" interpreter thinks that each text within angle braces is an HTML statement of one sort or another. Well, in the text of the legal statement, the word YEAR is in angle braces. The view-source interpreter thinks the word YEAR is supposed to be an HTML statement because of the angle braces. So the interpreter made it purple. No doubt your typical browser will interpret it the same way.

I think what FRED really wants is that we should enter the current year (like "2016") when we show the credit and legal notice. I pointed this out to them via their "send feedback" tab. If they change the legal notice, you'll know I was right.

Oh! Now I know why the slash i in angle braces in that image is red instead of black with a purple "i". It's because the view-source interpreter expected to find in the HTML a closing YEAR statement (slash YEAR in angle braces) before it found the slash-i statement.

When it found slash-i instead of slash-YEAR, it discovered the error, and it made the slash-i red to indicate the error.

//

Saturday morning first thing, I see they have revised their copyright notice to include the date:

Yeah, that's more like it.

Not sure if it's still the case, but it used to be that to copyright something you had to claim the copyright and you had to put the date on it. Invalid without the date. Without the date, I think, rights are not reserved.

Here's the HTML code again. I highlighted the changed part in blue:

(The slash-i in angle brackets is no longer red.)

My work here is done.

a method to their madness

This caught me by surprise:

FRED's Nonfinancial Corporate Business debt series is discontinued? First I'm hearing of it. But that's what the add-a-data-series box shows. Disturbing. First TCMDO, now this? Yeah, they brought TCMDO back, but still...

All I get on the Notes page for this discontinued series is: "The source no longer includes this series on their release causing the discontinuance of the series in FRED."

Thanks, guys.

So I dug around a little, looking for nonfinancial corporate business debt. Found this one:

Debt securities; liability.

Sounds familiar.

I dug up my The Replacement for TCMDO from last October. The two series they brought in to replace TCMDO are All Sectors; Total Debt Securities; Liability and All Sectors; Total Loans; Liability.

Debt securities; liability. And that series ID is ASTDSL. DSL. And the "Total Loans; Liability" series ID is ASTLL. Hmm.

And the "nonfinancial corporate business debt securities liability" series ID is NCBDSL.

Hmm.

Maybe there's a method to their madness.

Yup, there is an NCBLL series: "Nonfinancial Corporate Business; Loans; Liability".

Okay.

It's like finding a four leaf clover

My favorite FRED data series, TCMDO, is no longer discontinued:

I noticed when I went back and looked at the last graph here. My screen capture from back in October shows TCMDO (red) running slightly below the sum of the two replacement series. But when I followed the link below that graph to get back to FRED, the red line now runs right on top of the blue.

So they revised the TCMDO data and brought it back from the land of the discontinued. Nice.

It also has a new name now.
Old name: All sectors; Credit Market Instruments; Liability, Level
New name: All Sectors; Debt Securities and Loans; Liability, Level

The new name makes sense. The two series that I had to add together back in October were Total Debt Securities and Total Loans.

Talk it up

At FRED Blog: The evolution of employment costs in the private and public sectors. They show "wages and salaries" for public- and private-sector workers --

and they say

The graph above tells us that wages and salaries have increased faster in the private sector. But, as both series are normalized to 100 in 2007, we can’t say whether wages had been better in the public sector and the private sector is just now catching up.

I take their graph, change the units to "Percent Change from Year Ago" --

and I wonder why they are talking about increase.

//

They show "benefits" for public- and private-sector workers --

and they say

Again, this graph says nothing about the levels. However, the two graphs in combination do tell us something: Relatively speaking, compensation in the public sector is increasingly in the form of benefits, while compensation in the private sector is increasingly in the form of wages and salaries.

I take their graph, change the units to "Percent Change from Year Ago" --

and again, I have to wonder why they are talking about increase.