Monday, February 18, 2008

Company Valuations -- short vs long term delimmas

One can argue that business and economic structures of Capitalism or pseudo-capitalism is in part based on proper valuation of companies and their assets -- present and future.

With start ups this almost becomes a black art for most people and even VCs. For Utilities sector of the market on the other hand it is almost like buying a bond.

I wrote this post a favor to a friend. He recommended that I look at a Business Intelligence company, and then their CEOs and CTO insisted that I look and comment about their blogs. I found their posts interesting but also too black and white. I guess in part that is the nature of the Business Intelligence market -- i.e., you try and pretend that everything is deterministic. The other part was the people's personality. In fact, this whole thing blew up into the discussion over Martin Friedman's philosophical arguments that I will spare this blog with.

But having read my reply to them, I thought it may be deserving of being its own post. So here it is.

Valuations are a key to successful investment analysis. Yes, if the long term shareholder value maximization and valuations are done correctly, then the opportunities for arbitrage are minimized. This is because no hidden value (or risk) is left on the table and we have complete transparency. In an efficient (perfect) stock market -- with complete and transparent knowledge -- the market valuation and the wall street valuations should be similar. (NB There are two side issues here that may need reflection: 1) I used the term long term shareholder value for several reasons. Martin Friedman's concept of maximizing the shareholder value is an amazingly powerful statement and one that raises many issues both financially and philosophically. Take for instance, tragedy of commonshttp://en.wikipedia.org/wiki/Tragedy_of_commons. Taken literally, a company will abuse common public resources to maximize shareholder value. The term long term is used to dampen the effect of short term greedy algorithms [decisions]. 2) There is the issue of perfect knowledge and efficient stock markets. Whether possible or not, a built in assumption into the security markets is that they are efficient and information is available to all. Arbitrage, is an example where this efficiency breaks down. I discuss this further in terms of time and uncertainty below).
For a computer program to be able to help investors, it has to be relevant, complete, and consistent. Completeness of information may not always be possible for real time trading. This is due to two factors: First, the emotional aspects of market trading (Footnote: MacKinsey among others did some study on this). Second, the dynamic and multi-variate nature of the market that make long term effects of a decision more uncertain -- specially as the time frame considered increases.
May be I should be a more clear by what I mean when I say dynamic and "multi-variate". Lets say that a firm A decides to acquire a firm B or to do a merger with B. The market will re-evaluate the value of A and B accordingly. But that valuation is dependent on many unknowns. For instance, a competitor of A, firm C, may have decided to acquire a company similar to B, but has not announced it yet. Or alternatively, C has a division very similar to B, that it is planning to spin off and sell. The re-valuation of A and B in both cases will be drastically different from each other and from C not having made any announcement yet.
One way of addressing these unknowns is to limit the temporal window of consideration and using a variable discount rate (forget about the long term, and build in the risk as into the wacc or discount rate as usual). That is to say, I will not sell my shares sometime between a month and 3 months from now, this way all the emotional aspects has been factor out (hopefully) and the effect of distant uncertainty is minimized. Second for longer term decisions, I increase my discount rate and its volatility accordingly, because we simply don't know that much into the future. The one nice aspect of long term investing -- besides the fact that big players like pension funds are interested in it and it has huge overall economic impact -- is that while it does include the long term unknowns that come with distant time, it reduces the short term volatilities that are associated with knee jerk variations in stock prices. But that also creates a dilemma. Do you invest in value companies long term, knowing that it may be subject to delta function type uncertainties, but lower short term volatilities due to noise? Or do you invest short term, hoping to catch the upside for greater riches and higher ROI's? :-)
I think studies and general market sentiments votes more toward the first option (unless you are an insider type!), but the general practice of the hedge -- and some mutual -- funds is geared to the later! :-)
Either way, the right criteria for a BI (or Business Intelligence) Software, or as some like to call themselves, a value maximization system is the one that captures the essential parameters, and is flexible enough to model/address the non-quantitative aspects as well. This goes back to the saying that: your model is as good/accurate as the parameters that it captures and utilizes properly, including the non-quantitative ones.

Personally, I am a firm believer in multi-variate Monte-Carlo simulations and sensitivity analysis and then strongly hedging ones investments anyway :-)
Cheers, E
-- Esfandiar Bandari, PhD, MBA
e.bandari@cantab.net, e.bandari@gmail.com
skype: ebbandari & gtalk: e.bandari
http://www.linkedin.com/in/ebandari