They were. But now...
Hate Crimes Increase for the Third Consecutive Year, F.B.I. Reports
A memorial to the victims of the mass shooting in Pittsburgh in October. Hate crime reports have increased for the past three years. Hilary Swift for The New York Times
Trump's tough-guy talk has brought the vermin out of the sewers. It's going to get worse before it gets better.
Yea, I saw some of those articles, and since I pulled from the same fbi.gov source, you can see the uptick in the data that they refer to. However, I'm not sure the reporters are doing good analysis on the data. First, there were a couple of changes in reporting methodology that could explain the uptick. The data included two new sources for racial-based hate crimes: Latinos and Arabs. If you break out there data by just anti-black and anti-white hate crimes, it doesn't look like those are responsible for the uptick (which lends credence to the idea that it was the inclusion of two new categories).
(same sources as previous)
So I can't help but think that those changes in reporting methodology are responsible for the slight uptick.
Yet, even with those changes in methodology, it still appears that the changes in the data are probably noise since they are within a standard deviation and below the long run average.
To be fair too, and this will be a little more technical, I ran a quick Augmented Dickey-Fuller test (with anywhere from 1 to 9 lags) just to see if the data has the presence of a unit root, which indicates that it probably does. The long story short then is that the current observation isn't dependent on past observations. Trying to make use of the the mean will probably not be very useful as it won't be constant nor will the variance be constant (so me saying it's within a standard deviation and below the average may not mean much, if anything). So any impulse into the generating process could be persistent. So if there was, ever were or ever is an increase in hate crimes (as reflected by the process that generates this FBI data), there appears to be no mean reverting process in this data I have to indicate that downward trends are guaranteed to continue or that future observations will be below the average.
Also, even though there is a downward trend in the data, it shouldn't be missed that the data reflects consistently and significantly greater number of anti-black hate crimes than any other category. So that shouldn't be missed either.
http://www.fsb.miamioh.edu/lij14/672_2014_s6.pdf