@Bipolar: based on the example you outlined if you are valuing the company at time zero (i.e. before the the 5% growth for two years starts) this would be a three stage DDM not an H model. The way to value this company is to start at the end of the first two years and value the remaining dividends using the H-model since at that time the growth begins to decline linearly as is necessary for the H-model to be applied. This give you a value of the company at time 2. You discount this value back two periods along with the time 1 and time 2 constant high growth dividends to get the company value.
Here’s my 2 cents… The H-model is a form of a 3-stage DDM, with the Transition stage being a LINIAR decline from “Super Growth” initial stage, to the “Terminal Growth” stage. The way I remember it is that H (in my mind) stands for “Half-life” – a LINIAR decay of the Super Growth rate to the Terminal Growth rate over a given time period. So for example, if the super growth rate of 8% lasts for 10 years and the terminal growth rate is 3% which is reached in 5 years after the end of the super growth period, the H model assumes that in years 11-15, the growth rate will decline by 1% => (8%-3%)/5. – year 11, growth will be 7%, year 12 will be 6%, year 13 will be 5% and so on until we reach the 3% constant terminal growth rate. However, in order to standardize this methodology, the H-Model formula was derived => D1/(r-g) + ((D0 x H x (gS - g))/(r-g) where H = duration of the super growth period, and where gS = super growth rate. The reason for your confusion (in my opinion) is the fact that the H-Model is a form of a 3-stage DDM model. However, in a regular 3-stage model, there is no requirement for the transition period to have a liniarly declining growth rate – in fact the transition period may have whatever growth rate the anlalyst sees fit. While in the H-Model, by definition, the transition period MUST decline LINIARLY from the super growth rate to the terminal growth rate over a given time period.
Oops, the H in the H-model is the duration of the super growth period devided by 2. Sorry, the above is missing the “devided by 2” part.
@ylager: Again this is not a question of opinion to me as the CFAI Book 4 pg. 319 states verbatim: “A second variation of the three-stage DDM has a MIDDLE stage similar to the FIRST stage in the H-model…In the second stage, dividends decline linearly as they do in the H-model.” Hence, the H model cannot be said to be a form of the 3 stage model (what are the three stages?). In an H-model the linear decline starts immediately (CFAI Book 4 pg. 315), not after a first stage in which you have a high stable growth rate.
No more thoughts? Anyone? If not I’m chalking this as another blistering fail of Qbank.
The question isn’t wrong. The first stage of the H-model is a given level of growth for X amount of years. Then, the second stage is a linear decline. Finally, the last stage is a constant growth. In the Three-stage DDM, the linear-decline second stage is replaced by an immediate drop to a given level of growth for X years. Then another immediate drop to a constant level of growth. I feel your pain though about bad Qbank questions. For the most part, they are okay. But there are some AWEFUL questions.
@DC: I am sorry to say but I think the question is wrong. I don’t mean to be an a$$hole either bu your statement: “The first stage of the H-model is a given level of growth for X amount of years. Then, the second stage is a linear decline.” is incorrect. I say this not based on opinion but based on the authoritative text on the issue (i.e. the CFAI): CFAI Volume 4, Level II 2009 states on pg 315: “Fuller and Hsia (1984) developed a variant of the two-stage model in which growth begins at a high rate and declines linearly throughout the super-normal growth period until it reaches a normal rate at the end.” Based on this statement and the numerous examples that follow, the CFAI demonstrates that the first stage in an H-model sees a linear decline from a high growth rate to the stable growth rate. The second stage is the company growing at the stable growth rate in perpetuity. What you describe is a three stage model, since there is clearly a first stage in which the super-normal growth rate is constant and not linearly declining. This is exactly the same thing that is happening in this question, making it a variant of the three stage model. Again don’t mean to be an a$$, but I think a lot of people are missing the point of what I am trying to say and the fact that I am not basing my comments on opinion but rather on the word of the CFAI, which should be final.
it is confusing. Report this question of schweser and ask them to provide you an explaination
You’re not being an a$$. No worries there. I’ll have to study the H-model more. It was my understanding that the H-model had a first stage of constant growth for X years. I think the “linear decline” in the question is the thing that gives the answer away. But if the H-model is really a form of two-stage DDM, then yes, the question sucks.
Well the thing is that the linear decline doesn’t really give it away, because a three stage model can have a second stage of linear decline. Check out CFAI Volume 4, Level II 2009 states on pg 319: “A second variation of the three-stage DDM has a MIDDLE stage similar to the FIRST stage in the H-model…In the second stage, dividends decline linearly as they do in the H-model.”
linear is the key word!
@ampnooan: well that was helpful
I return to my underlying answer => the H-Model MUST have a liniar decline in growth rate from super growth rate to terminal growth rate, while a 3-stage model does not, it can have a liniar decline, an exponential, or whatever growth rates you see fit during the 2nd stage. The CFAI instruction for taking the test is to SELECT THE BEST ANSWER. It almost always uses terminolgy like “best describes” or “best fit” or “most/least likely”. So although you are totally right that a 3-stage ddm can have a linear decline to the terminal growth rate from the super growth rate, the way that this question would be asked on the exam is “which of the below DDM’s MOST LIKELY assmues…” so the best fitting answer out of the 3 is H-Model. If H-model wasn’t one of the choices, you would be correct in picking the 3-Stage model. So to conclude, this is a poor quality question by the Q-Bank, but the underlying piece of information that you should take away from it, is whenever you see LINIAR decline, think H-Model (although there is a way to create a 3-stage model with a liniarly declining 2nd stage).
@ylager: the H-Model must be a two stage model, the question describes a three stage model (as I have outlined repeatedly). Hence, regardless of what the second stage of the model is this is a three stage model, making it impossible for this to be an H-model which by definition has to be a two stage model. Please note that even though the linear decline is a crucial assumption of the H model I don’t need to rely on this feature of the problem to show that the answer cannot be H-model. The argument is as follows: Fact: The H model must have two stages. Premise (follows from the question and can be demonstrated, and has been demonstrated, at length previously in this post): the question describes three separate stages. Conclusion: The model described in the question cannot logically be an H-model. See, I don’t need the “linear decline” at all to show that this model can’t be an H-model, even though many have suggested that just because it says linear decline it must be H-model, which I am sure is what Schweser wanted the answer to be, but couldn’t be bothered to write the question properly to elicit the desired response. Yet another example of Schweser incompetence: The cheat sheet states to use VaR as a downside measure of risk for assessing hedge fund risk. This is wrong, hedge fund returns are not normally distributed and so VaR cannot be used. The CFAI prescribes using loss st. dev., downside deviation, and/or the Sortino ratio. More proof that Schweser fing sucks!!!
Adav, kudos for a good topic. I initially read the question the same way you did and wanted to pick A. But after thinking about it, I think the key is that a three-stage DDM does not require you to have three distinct growth phases, it requires you to need to calculated three distinct PV calculations. The H-Model framework is based on two growth phases (the linear decline part and the stable growth part), but the equation simplifies to only one PV calculation. The present value of the stable growth area under the curve (rectangle) is built into the H-model value. 11 1 1 1 `_ _ _ _ _ _ _ _ 1 1 H-model Area under curve 1___________________ So in this problem you have a high growth period, and a H-modelable growth period, which totals two discounting calculations. In that sense it would be considered a two-stage model. I guess we could imagine a scenario with three stages where for example you have two subsequent windows of declining growth, but at different rates: Phase 1: 20% growth for 2 yrs Phase 2: linearly declines to 10% over 5 yrs (2% per year) Phase 3: linearly declines to 4% over 2 yrs (3% per year) and then stabilizes at 4%. That would be a three-stage model with an H-model component (and would make for a tricky valuation). Does this sound right or am I just lost?
What about 1.) 15% growth for three years (distinct dividend calculations using D1 / (1+r) + D2 / (1+r)^2 + D3 / (1 +r )^3 then decline from 15% to 5% linearly for four years (use the H model) Stable growth at 5% (part of the H model) Three, distinct stages
You know what, I must be high. Ignore whatever that jibber jabber I wrote above was. When it comes to all things finance, I trust Damodaran’s word as gospel. Here’s his explanation (see pp. 12-13) “Three-stage Dividend Discount Model The three-stage dividend discount model combines the features of the two-stage model and the H-model. It is the most general of the models because it does not impose any restrictions on the payout ratio and assumes an initial period of stable high growth, a second period of declining growth and a third period of stable low growth that lasts forever.” http://pages.stern.nyu.edu/~adamodar/pdfiles/damodaran2ed/ch5.pdf Adav, I think you’re right. The question is just straight wrong.
Hey adavydov7, after reading over both Schweser and CFAI, you are absolutely correct. This question is aweful and the answer should be A. Thanks for bringing up the topic. It was a nice little review of DDM.
I am so glad some of you opened up those CFAI texts to verify what I was saying. Thank you all for a good discussion.
No thank you for bringing this to light!