Unions had a pretty good year in 2017: they didn’t lose any ground.

According to the latest edition of the Bureau of Labor Statistics annual survey, released this morning, 10.7 percent of employed wage and salary workers were members of unions, unchanged from last year. There was a mild uptick in the share of private-sector workers represented by unions (aka union density), from 6.4 percent to 6.5 percent. Density was unchanged at 34.4 percent for public-sector workers — mildly surprising, given the war on labor being conducted by Republican governors and legislatures across the country.

As the graph below shows, 2017’s flatness is a bit of stability amid a long-term decline; we saw similarly unchanged density readings in 2013 and 2015. While the rate of decline has slowed from the rapid rates of the 1980s and 1990s, density for private-sector workers was down 1.0 percentage points from 2007 to 2017; for the public sector,  it was down 1.5 points. Declining density is not new to the public sector — it peaked in 1994 at 38.7 percent, 4.3 points above last year’s reading — but the pace of slippage has accelerated over the last five years or so.

Declining union density is bad for at least two reasons: it’s bad for working-class power and for living standards. Yes, most unions are weak subordinates of the Democratic Party establishment, serving as little more than ATMs and get-out-the-vote operations for boring candidates who do little for their labor supporters once elected. But they do serve as a brake on the political class’s most barbaric instincts, which is why politicians like Scott Walker and moneybags like the Koch brothers (who funded Walker’s anti-union drive in Wisconsin) have it out for them.

And, as the graph below shows, unions bring higher wages — especially for workers who are neither white nor male.

For example, black men who are not in unions earn 71 percent as much as all white men (union and nonunion); with a union, that rises to 89 percent as much. For black women, the numbers are 65 percent for nonunion and 81 percent for union. For what the BLS calls Hispanic or Latino workers the union boost is even sharper: from 69 percent of white men for nonunion men to 98 percent for unionized ones, and from 60 percent for nonunion women to 94 percent for union.

Unions also narrow gender gaps. Nonunion women of all races/ethnicities earn 82 percent as much as men; that rises to 88 percent for unionized women. Unions add 21 percent to the average weekly wage for men, and 30 percent for women. In other words, unions reduce inequality along all the familiar demographic axes — and make it harder to pit workers against each other.

Admittedly, the nonunion/union figures compare workers across a variety of occupations, sectors, and regions; a more rigorous analysis would control for those differences to isolate the union effect alone. (There’s a version of that here.)  But even after controlling for those things, you’d still find a substantial union wage advantage of 10 percent or more — not to mention far better health, retirement, and vacation benefits and greater job security, things not captured by weekly wage stats.

There are good reasons employers hate unions, which is why we need to get those union density graphs pointing upwards. A major threat to that: the forthcoming Supreme Court decision in the case of Janus v. AFSCME, which would make public-sector union dues optional. Should the court decide against the union, which is almost certain given the configuration of the cast of robed ghouls, many workers would stop paying dues and not bother to join the union, and the “public” line will, as they say on Wall Street, accelerate to the downside.