Electric and magnetic fields are produced wherever electricity is used, in the home, office, or anywhere else. Electric fields are produced by voltage and magnetic fields by current. They come from the wires, not the pylons that hold them up.
Most of us get most of our exposure from distribution wires, either buried along the street or on wood poles, and from wiring in our homes. We also get short-duration higher field exposures when we pass close to electrical appliances such as fridges, microwaves, computers, etc.
In the population as a whole, not many people live close to (say within 100 m of) a high-voltage power line. But for those who do, this will also be a significant source of exposure.
Substations also produce magnetic fields, but you need to be very close to them to get an elevated exposure.
The UK Government set exposure limits for EMFs and the electricity system complies with these. The limits are designed to prevent all established effects of fields on the body.
However, there are suggestions that magnetic fields may cause other diseases, principally childhood leukaemia, at levels below these limits. The evidence for this comes from epidemiology studies, which have found a statistical association - an apparent two-fold increase in leukaemia incidence, from about 1 in 24,000 per year up to 1 in 12,000 per year, for the children with the top half percent of exposures. The evidence is strong enough for magnetic fields to be classified by the World Health Organization as "possibly carcinogenic". But because these studies only show statistical associations and do not demonstrate causation, and because the evidence from the laboratory is against, the risk is not established, it remains only a possibility.
See also WHO's 2007 short summary of the science.
A little bit more about the science...
The studies that give rise to the concerns about EMFs are "epidemiological" studies. They look at patterns of disease in real human populations and they report statistical associations. The trouble with these studies is you can never be sure what causes an association. It could be a genuine causal effect - the exposure causes the disease - or it could be produced by a bias in the study, such as which people chose to participate, or by some other factor at work in the population, such as, in the case of EMFs, the characteristics of the areas power lines pass through.
Studies over the years have found associations with childhood leukaemia, either with measured magnetic fields or with the presence of high-voltage power lines. But on the present state of knowledge, we really don't know if those reflect causation or one of the other possible explanations. We tend to think EMFs or power lines probably don't cause childhood leukaemia - but we can't rule out the possibility either.
With virtually everything else that causes cancer in humans, we can see the same effect in the laboratory - it also causes cancer in mice or rats, or we can pin down the mechanism. Magnetic fields have never been reliably shown to produce effects in mice or rats or any other laboratory test, and we've never been able to find a mechanism for them to affect biological systems. This is one of the stronger arguments against the idea that the epidemiological studies are finding a real causal relationship and in favour of it being some other factor.
The association is reported at different levels in different studies. If you hear the figure "0.4 microteslas" (0.4 μT), that's a cutpoint used in quite a few studies of magnetic fields, starting with the Ahlbom pooled analysis back in 2000. In the UK, it separates out the top half a percent of exposure. That doesn't mean that a risk suddenly appears at 0.4 μT, but it indicates the rough level of exposure where the statistical association seems to appear.
Likewise, if you hear "200 m" or "600 m", that's probably derived from the "CCRG" study which looked at proximity to high-voltage power lines in the UK and used those cutpoints for its analysis. That gives rise to an extra complication, because both of those distances are too far for the magnetic fields from the power lines to be significant.
The associations we're talking about are confined to childhood leukaemia. They don't show up for other childhood cancers, or for adult cancers, or for other diseases such as Alzheimer's, or heart disease, or pregnancy outcomes - in fact, scientists have looked at a whole range of diseases without finding any comparable associations.
The one thing we do know that EMFs can do is, at much higher levels, to have electrical effects on nerves. It is those effects on nerves that the exposure limits are based on. The authoritative bodies that set the exposure limits don't consider that the evidence about childhood leukaemia - the statistical associations found in epidemiological studies - is strong enough to base exposure limits on or to introduce other restrictions.
All power lines, substations, or any other part of the electricity system in the UK are compliant with the exposure limits. That means that it doesn't matter how close your home is to a power line or a substation, the fields will still be below the exposure limits.
This page gives a summary of the whole EMF issue in just one page. The other 300+ pages of this site give more detail on almost every aspect you could care to ask about - see the menu at the top of each page or the index/glossary.
Or download a pdf of our printed booklet on EMFs: EMF_The_Facts