State Space Models

All state space models are written and estimated in the R programming language. The models are available here with instructions and R procedures for manipulating the models here here.

Monday, April 25, 2011

How Could the Market Possibly be Overvalued After the Great Recession?













This would, on the face of it, seem to be an unusual question to ask after the Global Financial Crisis. How could the stock market possibly be overvalued? In the video above, Prof. Robert Shiller thinks that, based on longer historical analysis of P/E ratios, the market is overvalued. Prof. Siegel, on the other hand, thinks not. There are other players in the debate: David Bianco of Bank America Merrill Lynch (Shiller's data is questionable) and David Arnott of Research Affiliates (he likes Prof. Shiller's approach to earnings analysis). You can read the entire debate in the Wall Street Journal (here).

Friday, March 25, 2011

How Bubbles Get Inflated



Jeff Cox, a staff writer for CNBC.com, has written an interesting piece titled "For Investors, Missing Rallies, Not Taking Losses, the Biggest Fear" (here). Cox is also the author, with Peter Tanous, of the apocalyptic book "Debts, Deficits and the Demise of the American Economy" trumpeted in the video above.

Cox's point about the stock market, which has been made by other authors describing the Subprime Mortgage Crisis (here), is that since no analyst can afford to miss their comparative profitability numbers (they are all being evaluated against their peers), everyone has to aggressively follow the next bubble up. When the bubble pops, everyone has the excuse that the market has crashed and "I can't be expected to perform any better than anyone else in this environment."

In my brief, six year experience with the private sector, I have found this to be the pervasive business dynamic. When the company wasn't performing well, there was no pressure on my groups (IT and statistics). Also, when I was short staff, less was expected of my group (keeping a few open positions helped reduce pressure and kept the accountants happy with my budget performance). There was never any idea of building for the future during slow periods. It was always the right now and the "what have you done for me this week".

Such is American business. Cue the apocalyptic trumpets, crash the symbols, preview the next crisis!

Wednesday, March 9, 2011

Which Honeywell Five Year Forecast: Mine or Theirs?













Tonight on CNBC's Mad Money (here), host Jim Cramer interviewed the CEO of Honeywell. What caught my attention, listening to the interview (video above), was Cramer's statement that Honeywell (HON) "...is a cyclical business. In other words, its performance is dependent on the health of the U.S. and global economies".

To the ears of a Random Stock Walker, this means that HON should be neither a random walk nor a business-as-usual stock. It should be clearly linked to cyclical components of either the US or the World economies. The statistical estimation and comparison of state space models supports the assertion. HON is clearly linked to the world economy and especially to cyclical components of the world economy.
Jim Cramer actually used the term "hostage" to cyclical components of the US and World economies. He went on to imply that something had changed with current management and that the company had now become better able to resist cyclical downturns.

Let's compare the "bullish" Honeywell 5-year forecast to a 15-year Random Stock Walker forecast. The bootstrap forecast using the model linking HON to the world economy (this is the best model) does suggest that trends in the world economy (cycles generated by oil shocks?) will not be working in Honeywell's favor.

The graph above displays the long-run attractor for HON. The stock is very choppy, the instability being more a function of shocks rather than cycles in the world economy. The analyst opinion of HON is also not very positive with four out of ten research firms rating it a downgrade. From the Random Stock Walker perspective the price targets are still optimistic with the median target being 64.50, the high target being 71 and the low target being 40. The 98% confidence bands for the bootstrap forecasts never get above sixty and over time get well below 40. Unless something has actually changed, HON looks like a great candidate for the short sellers and their 5-year forecast looks very bullish indeed.

THEORY and METHOD: Honeywell is an interesting example of a company that might have changed with a change in management around 2006. There is not quite enough statistical data to determine whether there has been a break in Honeywell stock performance. This is a good example of a case where statistical forecasting may not be justified. It will be interesting to look back after a few years, test to see whether there was a break in 2006 and see whether the new Honeywell model is really different from the old.