Working on a project yesterday the question of Sarbanes Oxley costs came up. I had modeled in some of the additional 404 costs in order to measure some of the potential "going private" benefits in a transaction. Then I got curious about something. Are Sarbanes Oxley costs reasonable risk-return adjusted investments?
The real way to do this would be to plug in the difference in fraud "beta" pre-Sarbanes and post-Sarbanes and see if the present value of future Sarbanes expenditures meets the value of the reduced risk. Of course, modeling the "beta" of fraud and determining how much fraud risk is actually eliminated via Sarbanes is not a trivial task. Instead, I took audit expenses pre-Sarbanes, ran them forward 7 years with a 4.00% growth rate per year (this is quite low probably, given that in 2000, 2001 and 2002 audit fees rose an average of 10% each year) and then discounted them back at the S&P 500's rolling 10 year average return of 9.10% (which isn't really the discount rate to be using, but it is easier than computing the average cost of capital for similar firms). The present value of those expenses for a $5 billion - $10 billion revenue firm turns out to be $16,744,097. (This isn't really right either since the firms aren't going to stop spending on audit in year 7, I suppose I could use a perpetual growth model here but I'm just bouncing rubble at that point).
I then took the discounted value of 7 years of post-Sarbanes expenditures ($36,837,013, same growth rate issues as above however) and adjusted the discount rate until it equaled the present value calculation of normal audit expenses. The discount rate required to even the two out was 37.95%
Maybe it is just me, but Sarbanes seems a little steep by this measure.