Partisanship and perception

In early 2017, I had a post on perceptions of economic conditions, which noted that they were related to party identification--if your party was in office, your assessment was more positive--and that this relation seemed to be getting stronger.  I suggested that this put a floor under the ratings, and pointed out that they were never as bad during the 2008-9 recession as they were in the late 1970s or early 1990s.  I went on to say that partisanship should also work the other way:  "ratings will never get all that favorable, because supporters of the other party would be slow to acknowledge that the economy was doing well.  However, in order to test that, we'd need to have a real economic boom, which hasn't happened in this century and is unlikely to happen in the near future."  The question I was discussing involved perceived change in economic conditions, and in that sense there hasn't been a boom--economic growth is good by recent standards, but not that unusual by historical standards. 

However, the Gallup Poll has asked a question which goes back to 1992--"How would you rate economic conditions in this country today -- as excellent, good, only fair or poor?"--and the unemployment rate at the lowest level it's ever been in that whole time.   In principle, I think that this question is easier to interpret than the one about change, because it simply asks about current conditions, and doesn't require people to compare with what they remember about the past.  I had jut used the change question because it goes back farther.  If we look at the average rating of current conditions (excellent=4, good=3, fair=2, poor=1): 

Even though the unemployment rate is lower than it was in the late 1990s, average ratings are less favorable than they were then.  They are also more polarized--more "excellent" ratings and more "poor."  Both of these are what would be expected from a bigger partisan split--people are less likely to accept that the economy is good if the "wrong" party is in power.  Also, although I've said this before, it's worth saying again:  perceptions were not especially negative at the time of the 2016 election.  The vertical lines indicate elections, and the 2016 rating was about what it was in 2004, and only a little lower than 1996.  It was considerably higher than in 2008 (no surprise), but also than in 1992. 

A natural follow-up question is whether the partisanship effect is about the same for both parties, or is stronger for one than the other.  The breakdowns aren't available for all of the surveys (or at least I haven't seen them) but I may investigate sometime in the future.