I live in the USA and I am concerned about the future. I created this blog to share my thoughts on the economy and anything else that might catch my attention.
Thursday, September 19, 2013
Real Retail And Food Services Sales per Retail Employee (Musical Tribute)
What's all this "per employee" stuff? I thought all we cared about was the size of the pie, not how many ways it has to be split.
Technical question: how do you get a trend line on a partial set of XL points, if you are using XL to create the charts. I tried to manually create a new series in the source data using a subset of the data, but I couldn't get it working.
What's all this "per employee" stuff? I thought all we cared about was the size of the pie, not how many ways it has to be split.
We're definitely not on the same page here. Once again, per employee matters a lot to me. It is an early indication if companies have hired too many people. When sales per employee begin to fall, it can easily imply that companies won't need to hire more and may have to let some go.
So yes, I am concerned both about the size of the pie *and* how the pieces are distributed. Which chart you get depends on the point I'm currently trying to make. This chart is all about potential future job creation relative to what it has recently been. I would argue that it should slow (or worse).
Technical question: how do you get a trend line on a partial set of XL points, if you are using XL to create the charts. I tried to manually create a new series in the source data using a subset of the data, but I couldn't get it working.
Here's my trick for this post's chart. You don't want all the series in the same column so it requires some cutting and pasting into new columns.
I started with the date in column A and the all the sales data in column B, just like it comes from FRED.
I then cut and pasted everything but the blue data (as seen in the chart) from column B into column C.
I then cut and pasted everything but the black data (as seen in the chart) from column C into column D.
So there are really three series in my spreadsheet. I then added trend lines on two of them.
One more thought just to be clear. All three series use the rows from 1992 to present even though the sales data for each column doesn't exist throughout that time frame (since each series is just a partial series based on cutting and pasting).
There is an alternate approach that I also use, especially when working with exponential trends and linear trends.
I use the "GROWTH" and "TREND" functions within Excel and fine tune their starting and stopping points. This works very well as long as I don't wish to extrapolate the data out into the distant future. If I do wish to extrapolate out into the distant future, then Excel's built in trends are easier to use (since I can't extrapolate a trend of my own without actually creating the data itself out into the distant future).
When I use this approach, I devote one column to the exponential and linear trend data and then simply chart it too as an additional series.
Please note in the chart that trouble was clearly visible long before the recession hit in December of 2007.
Companies were clearly building up too many employees compared to the sales they were getting.
Now look at where we currently are. As seen in red, it is really starting to flatten out. If the trend continues (and I see no reason to expect that it won't), then it will begin to roll over.
Since the economy is already weakened, I don't think it will take much for that rolling over to give us another recession. Sigh.
OMG, sarchasm! My initial comment was sarcastic about per employee, per capita, etc. Maybe that's my fault because in some cases, I don't think per capita tells us much, but in most cases I do and in this case I do.
What's all this "per employee" stuff? I thought all we cared about was the size of the pie, not how many ways it has to be split.
ReplyDeleteTechnical question: how do you get a trend line on a partial set of XL points, if you are using XL to create the charts. I tried to manually create a new series in the source data using a subset of the data, but I couldn't get it working.
Mr Slippery,
ReplyDeleteWhat's all this "per employee" stuff? I thought all we cared about was the size of the pie, not how many ways it has to be split.
We're definitely not on the same page here. Once again, per employee matters a lot to me. It is an early indication if companies have hired too many people. When sales per employee begin to fall, it can easily imply that companies won't need to hire more and may have to let some go.
So yes, I am concerned both about the size of the pie *and* how the pieces are distributed. Which chart you get depends on the point I'm currently trying to make. This chart is all about potential future job creation relative to what it has recently been. I would argue that it should slow (or worse).
Technical question: how do you get a trend line on a partial set of XL points, if you are using XL to create the charts. I tried to manually create a new series in the source data using a subset of the data, but I couldn't get it working.
Here's my trick for this post's chart. You don't want all the series in the same column so it requires some cutting and pasting into new columns.
I started with the date in column A and the all the sales data in column B, just like it comes from FRED.
I then cut and pasted everything but the blue data (as seen in the chart) from column B into column C.
I then cut and pasted everything but the black data (as seen in the chart) from column C into column D.
So there are really three series in my spreadsheet. I then added trend lines on two of them.
One more thought just to be clear. All three series use the rows from 1992 to present even though the sales data for each column doesn't exist throughout that time frame (since each series is just a partial series based on cutting and pasting).
ReplyDeleteHope this helps!
There is an alternate approach that I also use, especially when working with exponential trends and linear trends.
ReplyDeleteI use the "GROWTH" and "TREND" functions within Excel and fine tune their starting and stopping points. This works very well as long as I don't wish to extrapolate the data out into the distant future. If I do wish to extrapolate out into the distant future, then Excel's built in trends are easier to use (since I can't extrapolate a trend of my own without actually creating the data itself out into the distant future).
When I use this approach, I devote one column to the exponential and linear trend data and then simply chart it too as an additional series.
Back to the "per employee" concept:
ReplyDeletePlease note in the chart that trouble was clearly visible long before the recession hit in December of 2007.
Companies were clearly building up too many employees compared to the sales they were getting.
Now look at where we currently are. As seen in red, it is really starting to flatten out. If the trend continues (and I see no reason to expect that it won't), then it will begin to roll over.
Since the economy is already weakened, I don't think it will take much for that rolling over to give us another recession. Sigh.
OMG, sarchasm! My initial comment was sarcastic about per employee, per capita, etc. Maybe that's my fault because in some cases, I don't think per capita tells us much, but in most cases I do and in this case I do.
ReplyDeleteI was bitten by my own sarcasm!
Oh, thanks for the XL tips. I will give that a try.
ReplyDeleteMr Slippery,
ReplyDeleteOMG, sarchasm!
I was bitten by my own sarcasm!
I think we were both bitten, lol.
It could always be worse though!
Oh noes!
ReplyDeletehttp://tinyurl.com/mhcpx84
Watchtower,
ReplyDeleteD'oh! This deserves its own post!
The new post is up.
ReplyDelete