However, there are various scenarios where either P/E or P/B will be favored over the another.
Firstly, where the company and/or its peers suffered losses in the trailing twelve months or in recent financial periods, P/E does not provide a meaningful comparison across peers or history. This is where P/B stands out; it allows for a comparison of firms across a broader section of peers and a longer period of history, since book value is rarely negative.
Secondly, stocks in sectors such as industrials and financials may be more appropriately valued using P/B than P/E. For industrial and financial companies, there is a stronger link between their assets on the books and their income generating ability. Industrials have the bulk of their assets in inventories and property, plant or equipment, while property and loans represent a huge portion of the assets of property companies and banks respectively. The fact that assets for such companies are comparatively more accurately valued on the books than others provides a strong case for the use of the P/B ratio. This provides a sharp contrast with computers on the balance sheets of consulting firms - the link between asset and income is much weaker.
The last scenario is a tough one. For companies which are serial acquirers, both assets and earnings are subject to significant distortions. If we use the P/B ratio, intangibles such as goodwill inflate the book value of a firm. An alternative is to use P/NTA and remove the effects of intangibles. If we use P/E, it may be necessary to separate the effects of organic revenue and acquired revenue, and recast net income. The quality of adjusted earnings is largely dependent on company disclosure of such information.