The roundtable discussion with Stephen Penman yesterday was quite good.  Lots of good discussion in both voice and typed chat.  One thing I learned is that Stephen stresses subjectivity as a reason behind his views. While I am convinced that Stephen dislikes some forms of subjectivity, I think his views are more subtle than this.  Specifically, suppose that a machine used in the company’s operations experiences an increase in value during the year, and a highly reliable measure of this change in value exists.  The revised value measure may be less subjective that a measure derived from historical cost, expected useful life, and expected salvage value.  If subjectivity is the key element in determining desirable accounting, we should record the revised value of the machine.

However, what if the company creates value by synergistically combining the machine with a trained workforce and results of its research department?  I see nothing in the deliberations of the IASB or FASB to say they support recording increases in the value of the workforce or research this period, as increases in value of workforce training and basic research do not meet the definition/measurability criteria required for asset recognition.  Thus, anyone who believes that the so-called “balance sheet view” has a goal of equating a company’s book value and market value are complaining about a straw man that is not part of the FASB’s agenda.

But back to Stephen’s views.  I believe he would reject fair valuing the company’s plant asset even if a non-subjective level 1 fair value was available.  He expects reported operating earnings to reflect the enhanced value of a trained workforce or proprietary research over time as the company is able to earn higher margins than its competitors.  He also believes these margins are persistent.  I agree with these views.  Stephen’s concern is that if the company records fair value increases for the plant, the reported margins would include transitory gains and losses.  He would rather use some sort of allocated historical cost for the plant in the belief that using a comparable measurement approach for all operating assets produces operating margins that are easier for analysts to use in predicting earnings and valuing the company.

As a result, the debate seems to be less about subjectivity and more about how mixing measurement attributes for operating assets can reduce the ability of analysts to do their jobs.  I would think some of the propositions underlying Stephen’s views are empirically testable.  What do the other bloggers think?