In the case of Seattle, average rents have gone up by 7.5% in the past year, which sounds pretty crazy. And as a recent resident, I can vouch for the fact that rents have definitely gone up significantly in the last several years. But the scope of these reports is so limited as to be almost useless.
To see why, consider the following example:
Imagine a city with 10,000 multifamily apartment units, each of which rents for $1,000 a month. Now build 10,000 more units, all fancy high-rises that go for $3,000 a month. Then build two more units and rent them for $2,000 each, just for funsies. In this scenario, assuming no old buildings are torn down to make way for the new, average and median rents both work out to $2,000 a month.
Looking at our imaginary city, notice that the average and median rents accurately represent just 2 of the 20,002 units in our city, but now we can make broad assertions like "rent has doubled since the construction of these new apartments." It's not wrong, strictly speaking, but not only does it tell us almost nothing about the actual character of the city or where and how people are living within it, it actually misleads us. This is a highly stylized example, of course, but it makes the point very clear.
Numbers like these give no indication of what kind of housing is available to someone earning 50%, 80%, or even 300% of the city's median wage. They don't say how much old housing stock was demolished or renovated to construct the new buildings. They don't tell us what the cost of housing is in the 10th or 90th percentile. They also tend to be limited to one type of housing, in this case one-bedroom apartments. But there are plenty of people who are happy with a studio, and many who need more than one room.
Perhaps worst of all, averages and medians often give the impression that new construction has a much larger impact on everyone else's rents than is realistic; in fact, much of the increase in "average" rents is just the result of adding high-end units to the total supply of housing stock—units that couldn't have been built in the dense neighborhoods and downtowns of twenty years ago, and are generally a boon for the revenue they bring to support city services. This kind of reporting reinforces the kind of zero-sum NIMBYism which supposes that the interests of residents, current and future, can never align with those of developers and property owners.
Better reporting on housing would account not just for the number of new housing units, but the number lost over that same time period; it might include average and median rents, but would also prominently address other parts of the market, like the 10th and 30th percentiles; it would include multiple housing types, from studios to two-bedrooms at least; it would relate the cost of housing to incomes in a broader sense, possibly with numbers on the share of people spending more than 30, 40, or 50 percent of their income on housing; and better yet, those numbers would also include the cost of transportation.