Construction Jobs Too High?

I have a question about how construction jobs are calculated in IMPLAN. We've had a few instances where we've been asked to assess the economic impact of what for our area are very large construction projects--think power plants, manufacturing plants, etc. greater than $1 billion. In each case, the number of jobs we get from the IMPLAN model is orders of magnitude higher than the jobs estimates put out by the company and/or the state government. In one case, our annual jobs estimate from IMPLAN was five times the jobs the developer was estimating!

I understand the basic calculation of employment based on worker productivity, and that construction jobs are supported over many years. But even using annual average employment we're still getting very large employment compared to the estimates provided by the company. My standard response is that IMPLAN includes all workers--permanent and temporary--over the course of the construction project and that construction is generally a low-productivity industry. But I would like to have a better answer as to why our numbers are more correct, if they are!

Thanks!