From our review of the empirical evidence from a variety of studies in a range of countries, it should be evident to the reader that it is not possible to draw definite conclusions about the extent of earnings mobility. The different studies are very heterogeneous in sources and methods, and the differences may significantly affect the findings. Some empirical regularities suggest themselves: for example the degree of mobility appears to fall over the life-cycle, being especially high for the youngest age groups. But the range of values taken, for example, by the one-year correlation coefficients, is wide; and there is only limited evidence about correlations over periods of 10 years or more. Much of the evidence relates to men, and more evidence about the earnings mobility of women is needed. Considerable efforts have been made, in particular at the end of the 1970s, to estimate econometric models of earnings dynamics, but there are too few studies available at this stage for a cross-sample comparison to yield valid general conclusions. Time spans covered by those studies are often too short and it is unfortunate that there have been relatively few studies in the 1980s, since many of the data sources are now richer and permit more efficient estimation than ten years ago. As we have just seen, econometric techniques have developed which would now allow for a more rigorous treatment of some of the major biases affecting panel data such as that arising from attrition.