Sweet!

I'm on a new project: trying to reverse engineer the average length of loans from debt and debt service data. So I rounded up some data, converted it all to annual (to match the "monetary interest paid" series) and changed the aggregation method to "end of period" (the same way debt series are figured), checked it quick, and downloaded the FRED data as an Excel file.

Happy as a pig in shit, I glanced at the file and right away noticed that FRED now shows the calculations as part of the data descriptions in the header area. Here's a clip from the file, where I convert debt service payments as a percent of DPI to debt service payments in billions:

At left, an identification tag. On the right, the formula in highly readable format, showing the unique series identifiers for the data I used in the calculation. Next, the formula again, this time showing the units I had chosen for the data. Finally, the data frequency -- "Annual" -- which is great, because I changed it from the default.

With this information, people can figure out what I did. People can duplicate my graph, if they want. This is perfect. This is how it should be. Pig in shit.

Then I glanced at the values for my calculated debt service numbers. Horrors! My numbers were smaller instead of bigger than the percent values I started with. After an "Oh, man" I took another look at FRED's documentation of my formulas.

Bingo: I divided by DPI-in-billions where I should have multiplied, and multiplied by 100 where I should have divided. And I figured out the problem from within the spreadsheet. FRED just made my life easier.

Thank you, FRED. Thank you, thank you, thank you.