There is a persistent perception that oestrogens have an adverse effect on carbohydrate metabolism. It might therefore be expected that their use would result in a corresponding increase in the incidence of diabetes. Recent evidence from clinical trials suggesting that women on postmenopausal oestrogen hormone replacement therapy (HRT) have a reduced incidence of type 2 diabetes therefore appears paradoxical. Short-term supraphysiological oestrogen administration has an adverse effect on glucose tolerance, resulting from suppression of first-phase insulin secretion and increased insulin resistance. Oestrogen-induced increases in glucocorticoid activity could account for these effects. Oestrogen-induced deterioration in glucose tolerance is, however, accompanied by a reduction in fasting glucose, an effect that could be accounted for by glucagon antagonism. These short-term effects contrast with long-term preservation of insulin secretion and glucose homeostasis by oestrogens. In animal studies, ovariectomy is associated with decreased insulin secretion and increased risk of diabetes, whereas oestrogen administration protects against diabetes and increases the insulin response to glucose. The mechanism is uncertain, but direct effects on the pancreas via steroid receptors or indirect effects via oestrogen-induced glucagon antagonism and subclinical increases in glucocorticoids and growth hormone could all contribute. Recent evidence that HRT increases the risk of cardiovascular disease suggests that it should not be used for the prevention of diabetes, but the mechanism responsible for this benefit merits further investigation and might lead to new therapies.