Limited mobility bias correction
Abstract.
A description of what limited bias correction is. And how to compensate for it.
1. Introduction
We assume we have an OLS model in matrix form
| (1) |
where is a -matrix, is a -matrix, and is a -matrix. It is assumed that and are matrices of dummy variables and that both and are very large (as in ). The lfe-package estimates and as OLS coefficients. Some applications study the variances and , as well as the covariance . It was shown in [1] that if one uses the estimates and for this purpose, the resulting variances are positively biased, and the covariance is typically negatively biased. The biases can be large enough to change the sign of the covariance. They also provided explicit formulas for the size of the bias, in terms of the trace of some very large matrices.
In short, the bias for can be computed as
| (2) |
where is a large matrix depending only on , and . The unbiased variance can then be estimated as
| (3) |
In some typical instances, the matrix will be too large to handle directly, being of size . The bias for and the covariance, are of the same form.
2. How lfe handles this
In [2] a method for estimating was outlined. It uses the fact that
| (4) |
if is a vector of independent random variables with and . In fact, each can optimally be choosen as drawing from with uniform probability. This works because even if is too large to be handled directly, it is possible to compute for a vector . Thus, lfe’s fevcov routine uses a sample mean to estimate .
3. Another method
Formula 3 suggests another method for estimating . We rewrite it as
| (5) |
Now, and are constants, so we can write
| (6) |
If we resample and scale the residuals , and do the estimation of over again, we will obtain a series of different observations of formula 6. We may then do an OLS on model 6, and the estimate will be an estimate for which can be used in equation 3. The is not normally distributed around its mean, this typically skews more than , so is a better estimate. lfe does not implement this method.
References
- [1] M. Andrews, L. Gill, T. Schank, and R. Upward, High wage workers and low wage firms: negative assortative matching or limited mobility bias?, Journal of the Royal Statistical Society(A) 171(3) (2008), 673–697.
- [2] S. Gaure, Correlation bias correction in two-way fixed-effects linear regression, Stat 3 (2014), no. 1, 370–390.