A comprehensive meta-study from UCLA of 31 other studies of dieters found that 83 percent of people who go on diets eventually put on more weight than before they started. What's more, the wear and tear associated with yo-yo weight loss and gain makes them much less healthy for trying. This would include the low-fat, high-fiber diet recommended by the U.S. government.
I'm generally skeptical of meta studies, but this one carries the ring of truth. Obesity critics like Paul Campos have pointed out that if you look at epidemiological tables, you'll see that black women tend to skew higher on the obesity-mortality curve than white women. That is, black women can carry more weight without much of any additional risk to their health. In fact, black women can score well into the "obese" levels of the BMI with no effect whatsoever on mortality.
Studies also show that black women don't have nearly the body image problems that white women do--they obsess less about weight, and have much, much lower incidence of eating disorders, for example. All of which strongly suggests that the (slight) increase risk in mortality that comes with moderate obesity may well be more related to constant dieting and fretting over body weight than the weight itself. Of course, this hasn't stopped hysterics like the American Obesity Association (a front group for pharmaceutical companies pushing anti-fat drugs) from trying to scare black women into dieting, anyway.
All of which could mean that all of these calls from ant-fat activists and PR campaigns from the U.S. government encourage people to lose weight aren't just meddlesome. If 83 percent of people who try to lose weight fail, and are less healthy for trying, these sorts of messages could well be doing harm. As the dietitian in the Guardian article suggests, you're far better off just trying to get some cardiovascular exercise several times per week and not worrying so much about weight.