Liquidity Risk

As calculate the market risk and liquidity risk need the market returns and the innovations of liquidity, Lee use the equally weighted averages of individual stocks’ returns and zero-return proportions in a given country (across countries) through the auto-regression filtering to the first-differences of illiquidity to estimate them.

There are three advantages and one disadvantage that he uses the individual stocks to estimate the beta. The advantages are to avoid the potentially spurious results, to minimize the potential loss of information, to increase the power of test; the disadvantage is that it will have higher level of noise. For this reason, Lee estimate market risk and illiquidity risks at the portfolio level and uses the estimated risks present to the individual stocks to perform the cross-sectional regressions at the individual stock level.

The data are over years t-5 to t-1 with respect to local and global factors. The 5 years starts at either January 1988 or the first month the stocks are present in the sample. Then, the pre ranking betas are obtained. However, there is constraint that the stock should have 36 monthly returns and innovation of illiquidity within five years. Based on the pre ranking beta, Lee put them into ten equally weighted portfolios in a given country. To form a portfolio, Lee forms portfolios on the basis of the pre-ranking beta instead of the size and pre-ranking as using the basis of size will include the bias. Those countries those total stocks less than one hundred are dropped. With all four beta obtained, the liquidity net beta are also been calculated. The cross-sectional regression for each month is performed by using the stock returns and post-ranking betas.

Finally, the cross-sectional regressions are performed over economic or geographic regions because the data in other country may not as guaranteed as US and the sample period is relatively short. Also, Lee use dummy variable which is mean the within-country...