At the same time, the sample contains all first differences of the firms from the levels analysis that meet sample selection restrictions. Berger et al. (1996) require that the first earnings prediction occur no later than the fourth month after the date liquidation value is calculated, which make sure that the changes in liquidation value and present value of cash flow are aligned properly in time for each firm in the sample. The change of percentage in equity value is for the purpose that captures the impact of operational decisions, not the impact of insurances and redemptions. So they delete the firms with insurances and retirements.The results for the changes is as expected, the fact that the latter estimate is significantly positive supports strong evidence, however, that the association they documented earlier between equity value and liquidation value was not affected by liquidation value and the present value of cash flow that both measure different part of true present value of cash flow. The constant component of any association between liquidation value and the omitted part of true present value of cash flow is removed by examining changes rather than levels. Therefore, Berger et al. continue to find the strong, positive association liquidation value and equity value of a firm.