When large inductive or capacitive loads are applied, the power meter will not indicate and bill you as the electricity gets lost in the lines. Why don't power companies just charge for this?
Sometimes the power company will charge you. The issue isn't that the power companies need to make more "power", rather poor power factors mean that various equipment which transfers the power from the station to your house has to be bulkier, and thus more expensive.
In the past traditionally this was only a major concern for industrial customers who had very large loads and/or poor power factors. However, as our homes are filled with more and more switch mode power supplies (particularly those without active PFC) real and reactive loading becomes a larger issue for power companies. Traditionally households didn't use much electricity and had pretty high power factors to begin with so the companies don't bother.
As of late, though, it's become more of an issue with residential customers because we have more gadgets and are using more switch mode supplies which can have really low power factors. It's such a concern that in the EU they've passed regulation requiring larger SMPS's to have some sort of power factor correction to reduce the reactive loading.
Incidentally, there are some power companies with "smart meters" which charge higher peak usage rates, and offer free nights/weekends. They can do this because the actual amount of electricity you use is relatively cheap, but all of the "grid" part is quite expensive, especially if everyone wants to turn their gadgets on at the same time.
In Australia, New South Wales, if a consumer is drawing excessive power outside the normal power factors, they can be asked to pay. It is in the contract. Detection isn't by the meter though, but through occasional audits. Newer meters might increase the chances of an audit.
Poor power factor causes other problems though, so it isn't a good idea to use it as a means to avoid paying.
I'm going to say that question's assumption isn't universal enough!
One of my power meters does in fact measure both active and reactive power and I get charged for both kWh and kvarh measured by the meter. In addition to that, I also get charged for bad power factor and peak consumption. In my country, using this power charging scheme results in lower costs for consumed energy and additional charge for peak power consumption, making it more suitable for consumers which have constant loads with no large peaks.
I'm going to speculate a bit now, but my guess is that incentive to develop meters which measured reactive power wasn't great enough up until recently.
I don't have many devices in my home which are from this new high power factor era and despite that, my power factor is around 98%. In fact, only time it was lower than that was when the power company guy entered random numbers when reading power consumption.
Back in the era of electromechanical meters, measuring reactive power would have required a separate meter in addition to the meter for active power, which would result in increased costs for initial electricity installations, increased maintenance costs, since there would be one more meter to look after and calibrate from time to time, longer measurement times, since power company guy would have to read two meters per each consumer instead of just one and so on.
So I'm going to agree with helloworld922 here and say that households did have high power factors to begin with. In general, it would be cheaper to just assume a power factor for consumers and appropriately form the price of active power to take it into account.