Adult height, as a marker of childhood health, has recently become a focus in understanding the relationship between childhood health and health outcomes at older ages. However, measured height of the older individuals is contaminated by height shrinkage from aging. Height shrinkage, in turn may be correlated with health conditions and socio-economic status from throughout the life-cycle. In this case it would be problematic to use measured height directly in regressions without considering such an e¤ect. In this paper, this problem is tackled by using upper arm length and lower leg length to estimate a pre-shrinkage height function for a younger population that should not have started their shrinkage. The estimated coefficients are used to predict pre-shrinkage heights for an older population, for which also upper arm and lower leg lengths are used. The height shrinkage for this older population is estimated and examine the associations between shrinkage and socio-economic status variables. [BREAD Working Paper no. 334]. URL:[http://ipl.econ.duke.edu/bread/papers/working/334.pdf].