Return of the Dredded Orange Bar

At FRED on the morning of the 22nd...

Uh-oh, I've seen the orange bar before. It means change is a-brewing at FRED.

That can't be good.

Hey, I clicked the thing. Got this:

Yeah, okay. Answered a few questions, skipped a few questions. Got this:

No good. I'm not allowed to skip questions. Even if I don't have an answer, I should make something up. That's what they want.

Oh, yeah, that's how you get valid responses.

//

I can't wait to see how they fuck FRED up this time...

It's not a glitch. It's just FRED.

From the FRED blog I took their Bank Failures graph and added total (TCMDO) debt (red):

Thought it was odd that the red area was floating up above the horizontal axis. But whatever, it's not my graph. It's FRED.

I didn't want an area graph. Changed it to a line graph. Here's what I got:

The red line sank down and now starts below the x axis.

Weird. The numbers shouldn't change, when I change the style of the graph.

Damn! It wasn't me!

Unsaid at FRED

Reblogged from my economics blog

Let's draw a conclusion

At FRED Blog: How much money is the Fed printing?

To answer that question they look at currency in circulation. Two graphs.

First graph:

Graph #1: Quantity of Money, in Natural Log Values

Before 1960, flat. After 1960, straight-line increase.

The graph uses natural log values to show growth rates. As the FRED post says, "if the slope is the same for two years, the growth rate is the same." A more upward slope is faster growth. A less upward slope is slower growth.

The FRED post says the values are "indeed increasing, but there is no indication that it is accelerating". In other words: The line goes up, yes, but after the early 1960s it doesn't curve up.

Second graph:

Graph #2: Quantity of Money Relative to GDP

Graph #2 shows the same money as Graph #1 but shows it a different way. No "log" numbers. Instead, the graph compares the quantity of money to GDP. On Graph #2 we have downtrend to about 1985, and then uptrend: The currency component of M1 grew more slowly than GDP in the early years, and more quickly than GDP in the late years. I want to say the change occured around 1985.

Leave out the years before 1960 where Graph #1 has the flat trend. Look at the years since 1960, where #1 shows a straight-line uptrend.

From 1960 to 1985 the trend is down on Graph #2. After 1985 the trend is up. And yet, as the FRED Blog says, Graph #1 shows no acceleration.

Graph #1 shows that currency growth did not change. Graph #2 shows that either currency growth or GDP growth did change. We have to put the two facts together in our head: Either currency growth or GDP growth did change, but currency growth did not change. Therefore, GDP growth changed.

Voila.

To confirm, I took a quick look at the numbers.

During the 25 years from 1960 to 1985, currency in circulation increased by a factor of 5.6. During the 25 years from 1985 to 2010, currency in circulation increased by a factor of 5.5. Almost exactly the same. So, a straight line increase it is. As FRED said.

During the first 25 years GDP increased by a factor of 8.0. That's faster than currency growth, so the line on Graph #2 goes down between 1960 and 1985.

During the second 25 years GDP increased by a factor of 3.4. That's slower than currency growth, so the line on Graph #2 goes up between 1985 and 2010.

Currency growth was the same for the two periods, even though the "Great Inflation" is entirely contained in the earlier period. Currency growth was the same, but GDP growth was different. GDP growth was much faster in the early period, and much slower in the second.

Thickness of lines and the drawing sequence

Sometimes I get two lines that run together. For example I want to see that UNRATE produces exactly the same graph as UNEMPLOYED divided by the civilian labor force age 16 and over, CLF16OV. I was doing that. But I lost my focus when I got the graph that should have shown what I wanted to see. Should have shown, but didn't.

When two lines completely overlap, you can only see one of them. Of course. So what I like to do is make one of the lines thinner and the other thicker. With the old FRED it worked every time: I made the first line 3 or 4 pixels wide, and the second line 1 or 2 pixels wide, and right away I would get a two-color line that showed either similarity or difference, and either way it answered my question and set my mind at ease.

But that was the old FRED. The new FRED saves a step when it refreshes the graph. Evidently, if you change the thickness of one line, it only redraws that one line... or anyway, it redraws the changed line last.

With the old FRED, it would draw Data Series 1 first, and Data Series 2 second, and I could depend on it. And I did depend on it. With the new FRED, I'm not really sure what the hell it does, and I have to fuck with it to get the graph I want.

And then, after I get it looking the way I want, if I create a link to the graph and use the link to see the graph, it often comes up different from what I want, again -- perhaps because it's now creating the graph from scratch.

Can't be done!

Two different measures of M1 money. The blue line starts earlier than the red. But then around 1994 the blue line numbers go bad.

I want to show the blue line till maybe 1990 and after that show the red line. The new FRED won't let me do that. The start-date and end-date values are established for the whole graph, not separately for each line on the graph.

That's not my idea of an improvement to FRED.

I like it.

Went to FRED just now & selected a series to look at. The graph appeared, and below it, thumbnails of graphs I was looking at recently:

That's a nice touch.

I don't know what that is, but it isn't GDP

I found a graph of "Total Assets of the Federal Reserve" somewhere. Wanted to know if I could get the same data from FRED, so I searched FRED and found two with that name. Only one line appears on the graph, so they're identical. The other one is too small to see (as I discovered after a second look).

I kept the first one, then added GDP to the graph. but GDP didn't show up because it's given in billions, and the Assets are given in millions. So I divided the assets number by 1000 to convert millions to billions. Here's the graph I got:

http://research.stlouisfed.org/fred2/graph/?g=GxR

I think the X-axis distances before 2002 got compressed. That's pretty strange.

//

Ohhh... The Assets data is "weekly" and the GDP data is "quarterly".

Yeah, that's it.

Funny glitch.

Nobody ever told me this...

I do drawings for a living. Oh, not artistic stuff. Drawings to make things of wood and steel, at work. AutoCAD, almost entirely.

Been doing drawings for a living for a long time. Usually, when you put text on a drawing, it is horizontal. Sometimes it is vertical. Here's the thing nobody ever told me, but I found out by having to cock my head from one side to the other to read my own drawings: When you do text vertical, you want to turn them all the same way, so you can tilt your head to the left and read them all. You don't want to have to swing that weight to the left to read one annotation, then to the right to read the next one, then back again for another. You don't want to do that.

See the text labels for the left and right scales on this graph?

You're swingin' that weight to read them.

Kinda like assembly language mnemonics

NonFederal Debt relative to GDP

For me it is not easy to read the title of the above graph:

(All Sectors; Credit Market Instruments; Liability, Level-Federal Government; Credit Market Instruments; Liability, Level)/Gross Domestic Product
It is much easier for me to know that "TCMDO" is the number that measures "All Sectors; Credit Market Instruments; Liability, Level"... that "FGTCMDODNS" is the number that measures "Federal Government; Credit Market Instruments; Liability, Level"... and that "GDP" is "Gross Domestic Product"... and then to read this:

(TCMDO-FGTCMDODNS)/GDP
It's just easier.

Maybe it's me.

I like this one!

FREDBlog: The peculiar Swiss unemployment rate, 2 June 2014. I customized their graph to show recession bars:

Those recession bars are probably the NBER dates for U.S. recessions -- not really related to the Swiss economy. Which probably explains why FREDBlog had the recession bars turned off in the first place. Oh well.

The reason I like this post so much? Because it shows something "peculiar", unique, and interesting, as opposed to the tiresome political complaints you can find all over the web.

That's what graphs are for, you know? To see the interesting things.

It's a really big jump in unemployment, between 1990 and 1994. Remarkable.

Want to know something odd? The timing of that rise in Swiss unemployment corresponds precisely with a unique fall in the U.S. debt-per-dollar ratio.

Now that's odd.

"New"

I like the FRED Blog. They present little explanations of economic topics by using graphs, much as I do. Only, they know their topic better than I do. I can feel it in their posts. And that's fine. It's why I go there.

But there is a problem. Here are the first two sentences from their latest post, dated 29 May 2014:

What’s new in FRED? Beyond the pie charts we saw on the blog a week ago, FRED also features scatter plots, like the one shown here.

What do they mean, new? Here's a FRED graph from mine of 12 October 2012:

Graph #1: GDP vs TCMDO, YoY% Change (annual data)
Click graph for FRED source page #bJZ

Here's the first line of the first comment on that post, from Jazzbumpa:

Dang! I didn't know you could do a scatter plot at FRED.

That comment is dated October 12, 2012 at 1:58 PM



The Rules of Exposition: 1. Tell the truth.


P.S.: FRED's pie charts aren't new, either.

FRED Blog's May 1, 2014 post

Screen cap:

Nothing fancy, but not bad.

Anyway I wanted to fiddle with their graph, so I clicked "Customize" there at the bottom left corner of the FRED graph... the ALFRED graph.



Got an error!

So I hovered the mouse over the "Customize" text, right-clicked, and copied the link address:

http://alfred.stlouisfed.org/fred2/graph/?g=wUh

The last three characters on the right (wUh) identifies the graph. The rest of the link gets you to it. I figured they identified the graph correctly, but just put the wrong path somehow. So I made my own ALFRED graph, copied the link address, pasted it into a textfile and changed the last three letters to wUh:

http://alfred.stlouisfed.org/graph/?g=wUh

That works.

Orange for St. Pattie's Day?

A Work-Around

Here's a work-around for the FRED problem described in the previous post. I took the graph just as FRED gave it to me (i.e., the previous post's image) and exported the link. Then I opened a new blank tab in my browser and pasted the link into the URL box. I got the same wrong graph. But this time, the "a" and "b" boxes both said Change, Billions of Dollars. And this time, when I changed the "b" units to Billions of Dollars, it took. Then, of course, I had to export the link again to get the graph I really wanted.

Unreliable

You can see right away that the red line on the graph is spastic. That's because it's a ratio of "change in" values. And the right-hand scale shows it. The units are Change, Billions of Dollars divided by Change, Billions of dollars.

However...

 

... See the white box at the bottom of the image? The one labeled "EDIT DATA SERIES 2..."

About in the middle of that box are the red letters "a" and "b" in parentheses. These are the two sets of data that I used to create the red line on the graph. The one labeled "a" correctly shows Change, Billions of Dollars as the units I want to use. The one labeled "b" correctly shows Billions of Dollars as the units I want.

But the graph doesn't show what I want. I want Change in Billions, divided by Billions. What the graph shows is Change in Billions, divided by Change in Billions.

This problem never came up before they "improved" FRED.