21 March 2008

Can Accounting Tools Lead to, Rather than Prevent, Executive Mistakes?

For the last two weeks I have been lecturing on information systems controls and using the recent internal control failures at Societe Generale as an example. Issues of internal control generally lie at the heart of corporate failures be it deliberate fraud (as at Enron) or general incompetence.

Each corporate failure generates an avalanche of increased compliance regulation from the authorities. Yet the failures continue and the general public continues to suffer in one way or another, as do the staff who are required to undertake the additional compliance activity.

When thinking about better controls it is often argued that better, more timely and useful accounting data would help. That it is like the canary in the coal mine, giving early warning of potential problems.

However, there is an interesting article by Gavin Cassar over at K@W on Biased Expectations: Can Accounting Tools Lead to, Rather than Prevent, Executive Mistakes? He argues that:

"It's been shown in many studies that people are overly optimistic," saying that "What's interesting here is that, when you use the accounting tools, the optimism is even more extreme. This suggests that using the tools, which a lot of academics and government agencies say is good practice, can lead to even bigger mistakes."
Cassar then goes on to discuss how new entrepreneurs create storylines of their success that influence the spreadsheets upon which they rely for their business plans. The tendency for optimistic projections becomes part of the "objective" business plan. Given these tendencies for humans to think optimistically and to create narratives that support their optimistic views it is not surprising that companies and individuals can get into trouble.

The interesting question, is if we know how people think and the problems they have in being objective how can we control for that in business? I don't know the answer, but it is definitely something to think about.

By Carruthers via Aide-mémoire

1 comment:

alecthegeek said...

Programmers have been doing this for years -- we always underestimate the effort involved.

Clews' maxim for software estimation (my father's, not mine) is to ask 10 software developers how long it will take and then add the answers together.

Perhaps accountants should take a similar approach?