The concepts and methods used by NSSO to net work and workers are not able to capture the work of the poor, particularly of women, satisfactorily. Since that part of the workforce which is not captured by the NSS surveys is not likely to remain stagnant and is subject to increases and declines, depending on the specific situation, it is possible that an increase in this part of the workforce may explain the decline in the worker-population ratio (WPR) in the nineties. The workforce in these 'difficult to measure sectors', such as subsistence work, home-based work or informal work, can be better captured through time use surveys. Using data from the pilot time use survey (1998-99), this paper shows that (a) this survey technique is capable of getting more realistic estimates of workforce and (b) some of the work not captured in the NSS surveys but captured in the time use surveys is likely to explain the changes in the employment situation in the nineties to a considerable extent.