There's several reasons why the quit/layoff data may be misreported, e.g.
- Quits that anticipate layoffs, will bias up the number of quits
- Firms might not care / not know exactly, and assume it was a quit/layoff (direction ambiguous)
- Responder may purposefully or subconsciously try to defend the firm's image, and misreport
Importantly, there's no reason that the aforementioned reasons should average out to the true data, or that the bias is constant over the business cycle.
Since there are many studies that actually rely on layoff versus quits - is there a study on the reliance of this data?