How can cost decrease by more than 100 percentage points?

I saw the following graph today in the Atlantic (link), but I can't understand how the cost of televisions can go decrease by more than 100 percentage points in the last ten years. How is that mathematically possible? Or did the authors make a mistake?

-
The part where it says "change relative to a 23% increase in prices" is key. It means that the zero line is a 23% increase, a 100% is a 2*23%=46% increase and -100% is a -23% change in price. –  EngrStudent May 2 '14 at 22:55
I didn't notice that! If you write your comment as an answer, then I can accept it. –  I Like to Code May 2 '14 at 23:11
The title phrase "decrease by less than 100" is misleading. I suggest "decreases by more than 100" is actually the correct expression. Consider answering "*How many points of decrease is shown in the image?" for the lowest two items. For PCs it decreased by about 90, which is a decrease of less than 100. For TVs the number of points of decrease was about 105, a number greater than 100. –  Glen_b May 2 '14 at 23:16
Noted and changed. –  I Like to Code May 2 '14 at 23:19

The part where it says "change relative to a 23% increase in prices" is key. It means that the zero line is a 23% increase, a 100% is a 2*23%=46% increase and -100% is a 0% change in price.

I think the following is your y-equation.

$y = \frac {change - 23\%}{23\%}$

Per suggestion by user: whuber, the following might also be your expression.

$y = change - 23\%$

-
I doubt that's the right formula: division by 23% would not be useful or informative. –  whuber May 2 '14 at 23:50
It's correct in expressing a fold difference relative to the 23% increase, just like the misleading graph shows. Better yet would be fold-change for y, and percent change for change. –  leonardo May 3 '14 at 2:29
The top number is specified as percentage points, not percent of 23%, so I think whuber is right that the actual scale is just (change - 23)%, no dividing by 23. –  Kevin May 3 '14 at 3:37
@whuber - not in terms of science, but in terms of marketing, propaganda, or other media-driven FUD-dispensing it might be considered useful. –  EngrStudent May 3 '14 at 10:55