Monday, May 18, 2015

Challenge to self #1bCq

The challenge: Duplicate this:

Graph 1: from Marcus Nunes: "The chart describes in ‘phase space’
the degree to which growth ‘spreads out’ (is volatile)."

What I think I'm looking at:

X-axis: Percent Change from Year Ago of RGDP, at time t (quarterly data)

Y-axis: Percent Change from Year Ago of RGDP, at time t+1 (quarterly data)


Where I'll start:

Data Source: FRED GDPC1 in various formats I considered for the "What I think I'm looking at" part of this post.

Source Data: The "Download as XLS" file from the "Data Source" FRED graph.

Now, how do I get to Zoho? It's been a while.

I just type ZOHO in the Firefox URL field. ZOHO DOCS is on the list. I go there and sign in.

Loading... Please wait.

There it is.

I upload the file I got from FRED and check the box to convert it to Zoho format.

To start, I'm just gonna use the one column of data: "GDPC1_PC1: Real Gross Domestic Product, Percent Change from Year Ago, Quarterly, Seasonally Adjusted Annual Rate".

So far, so good.

On Marcus's graph the dates run from 1992 to 2007, then nothing, then from 2010 to 2015. Makes sense; that takes the crisis data out of the picture. You'd want to look at that separately I'm sure. (That was the first thought I had while looking at the FRED graph: Why doesn't Marcus's graph show the negative three, four, five percent growth we had during the crisis? And my answer: Oh! He omitted those years.

Okay. The last data I got from FRED is for 2015Q1. I can't get t+1 for that. So my most recent t data will be 2014Q4.

This is May. We're still in 2015Q2. No data yet.

I made a copy of the original XLS worksheet, deleted the data I'm not using, and inserted some columns for the data I want to graph.

One of the first things that strikes me while duplicating the graph is that the dots suggest a trend line that runs from lower-left to upper-right. In other words, when growth is pretty good one year, it is likely to be pretty good the next. And if growth sucks one year, it is likely to suck the next. (But I think we knew that already!)

I ended up doing the graphs in LibreOffice Calc. I think it gives me more options than I get with Zoho. (Or maybe I just didn't find the options in Zoho.)

It was easy enough to duplicate Marcus's graph. I didn't dress it up with circles and ovals and annotations. But the dots group the way Marcus says they do:

Graph #2: Duplicating Marcus's Dots

It's easy to imagine a circle around the group of orange dots, and an oval that overlaps it at lower-left but reaches way out to the upper-right to  surround all those blue dots. And there are those three crappy-economy dots in the extreme lower left, two of which Marcus identified as from the year 2001.

So next I added the years Marcus left out, 2008 and 2009, the crisis years. These dots (dark red) reach way down to the lower left, and push the graph's origin to mid-image.

This confirms our lower-left-to-upper-right trend line hypothesis , and confirms what we already knew: that when times are tough one year, they are liable to be tough the following year:

Graph #3: Same as Graph #2, but with 2008-2009 added into the picture.

Next, I went back to Graph #2 and split the blue dots into two groups. The first years, 1992 thru 2000 now show in green. The remaining years of the period, 2001-2007, remain blue:

Graph #4: 1992-2000 (green), 2001-2007 (blue), & 2010-2015 (orange)
So most of the good years that Marcus was talking about were the Clinton years. The "macroeconomic miracle" years. The "Goldilocks" years.

For what it's worth.

// The LibreOffice (spreadsheet with graphs) is yours for the taking.

No comments: