As discussed previously, it would be desirable to have additional models to verify the Altman. Research indicated that Ohlson's model had similar accuracy and yet very low correlation [.19] with Altman's indicating it looked at and saw other things. I found a copy of Ohlson's model in some research and e-mailed the author to get some clarification. Unfortunately what I ended up with has run up against my limited abilities to get it to work in a spreadsheet. So I'm posting it here in hopes those among you who are less mathematically challenged or more familiar with academic research modeling might be able to shed some light on whether it is "spreadsheetable" and how to do it:
In contrast, Ohlson's model (Model 4 in his 1980 paper, hereafter LGT)
is a conditional logit model which permits an estimation of the probability
of a firm's bankruptcy. This model does not consider any market variables
but it include size (logarithm of total assets) as an explicit variable. Since the LGT model measures the probability of failure (conditional on financial ratios), the values can only range from zero to unity, with low values indicating financial strength (low probability of bankruptcy) and high values financial distress (high probability of bankruptcy). We computed the bankruptcy probability for each sample firm based on his Model 4 as follows.
LGT = 1 /(1 +[exp{(-2.63-267 *Log(AT) +5.63 * TL/AT) - 1.43*(WCAP/AT) -2.35*(NI/AT) - 1.99*(OANCF/TL)1.56*D-.5092*(NI(t) -NI(t-1))/(AbsNI(t)+AbsNI(t-1))*(-1)}]) (2)
where AT = Total Assets, TL = Total Liabilities, WCAP = Working Capital, NI = Net Income, OANCF = CashFlows from Operating Activities and t represents sample years, "as usual in empirical research" [Prof's words]
"I think it was a sort of process related to natural logarithm from a logit analysis" according to the Prof who wrote the research from which this was extracted.
Source link
sbaer.uca.edu
Note: I'm a little puzzled at why the (2) at the end of the expression was separated from the rest of the expression. That is the way it appeared in the source, but perhaps in putting it in digital form they got the page number in there.
BTW: I've added a variation of Altman's aimed at non-manufacuring firms, the Springate model, and a market based model very loosely based on KMV's ideas to the basic Altman spreadsheet. I would be willing to email this if someone wants it...I'm at bobrgr@aol.com
Bob |
