Suppose I calculated the “intrinsic value” of a stock from a certain valuation model. What is the expected time (as in statistical expectation) that the currently stock price will converge this “intrinsic value”. Are there any research about this, for example for the simple DDM model? Thanks
It may NEVER converge Sometimes the intrinsic value may converge with the market value.
virginCFAhooker Wrote: ------------------------------------------------------- > It may NEVER converge > > Sometimes the intrinsic value may converge with > the market value. Of course, it might never converge but I was asking for the statistical expectation. It seems to me this is a very legit question. I suppose there must be some research addressing it. But I don’t see it covered at all in CFA L2.
Probably never converge. Very slim level of confidence in any bet that it conforms within a given time period. First off, intrinsic value has assumptions built into it, therefore not all intrinsic values will be homogenous(Using DDM, my risk free rate, or my growth rate is not same as yours. No standard intrinsic value.
No research covering it at all in any of the CFA levels.
Rydex Wrote: ------------------------------------------------------- > Probably never converge. Very slim level of > confidence in any bet that it conforms within a > given time period. First off, intrinsic value has > assumptions built into it, therefore not all > intrinsic values will be homogenous(Using DDM, my > risk free rate, or my growth rate is not same as > yours. No standard intrinsic value. But does an expected convergent time exist for a particular valuation model? If it does exist for some models, how is it calculated? As a matter of fact, I am pretty familiar with GARCH models. They seems to be able to make volatility predictions that at least makes sense. Are there similar models for the underlying?
I’m not entirely convinced that your question is even legitimate in the sense that, from a scientific standpoint, it would be very difficult to test with any level of efficacy for reasons already mentioned. First, what intrinsic model are you using as your basis and why? Arguably, different companies valuation should be modeled differently (i.e. you wouldn’t value a bank in the same way you do a biotech firm). Second, on what time basis is your valuation model? Are you valuing the company with existing cash flows or are you projecting and does the time frame for which you are projecting intimate a time horizon for your analysis of value with the market? Lastly, all valuation models, in some form, express the value of a company as a present value basis, regardless of the inputs, so any deviation from market value should (in theory) converge immediately to your intrinsic value. The big assumption here is that every other market participant is using the exact same model you are using, so in essence the question of timing may in fact not be of that nature at all. It might be better described as being a question of how differently the same company is being perceived between market participants and what factors come in to play to eventually (if ever) have the separate views meet. From a purely observational standpoint, however, in almost all of the analyst papers I’ve ever read, the implicit or explicit time horizon has been 1 year. I really don’t know why the entire industry has focused on this figure. Maybe they think that that’s enough time for their predictions to be proven correct or not, or maybe they think that investors don’t have the patience to have longer time horizons (which is probably true and might be a more interesting question), I simply don’t know.
BTW, what does it mean by an “1-year target price”?
Are you serious?
ymc Wrote: ------------------------------------------------------- > virginCFAhooker Wrote: > -------------------------------------------------- > ----- > > It may NEVER converge > > > > Sometimes the intrinsic value may converge with > > the market value. > > Of course, it might never converge but I was > asking for the statistical expectation. Let’s see if there is a positive probability that X = infinity the expectation of X is, um, I used to know that one… > It seems > to me this is a very legit question. I suppose > there must be some research addressing it. But I > don’t see it covered at all in CFA L2. Can’t imagine why not
I think I know the answer now… Determining the intrinsic value is part of value investing. The next part is determining if there is enough margin of safety to trade (i.e. is your intrinsic value different enough from the market value to be “safe”) the last part, and the answer to your question… CATALYSTS. The catalysts could be dividends, buybacks, insider buys, roadshows, m&a activity, etc. If you want a specific time, then 3 years is the magic number. Bruce Greenwald at columbia probably studied the art of value investing more than anyone and he came up with that number. He says you want to look for a stock that is not just cheap, you want a stock that has been breaking the hearts of investors for 3 years. Investor exhaustion are the words he used.
You want statistical research discussing convergence of stock price to its intrinsic, yet can’t figure out what a 1 year price target is? Perhaps I’ve had a few too many tonight. I suppose I should read this again in the morning.
“you want a stock that has been breaking the hearts of investors for 3 years” That’s catchy. I’m going to steal it.
No, it’s covered somewhat in LIII. Here’s an example that they cited: Years ago, 2 oil companies (big names, but I forget which ones) decided to completely merge their operations. They split the earnings 40/60. Because they are identical, they should always trade at the same multiple and a 3/2 price ratio. Yet, inexplicabley, they traded at much different multiples for decades. Many investors - including LTCM - identfied the mispricing and took the proper positions, thinking the two stocks would converge. It didn’t happen. The main idea of the topic was that markets can stay irrational longer than you can stay solvent.
Joey: most distributions are unbounded and therefore you have a positive probability that X = infinity, yet they have a finite expectation… Ymc: Do you want to know how long it takes the market to find the fair price of a stock, ex-post, or ex-ante? Ex-post, the answer is easy, but would assume that all your pricing parameters are static. Therefore it doesnt tell you how long the market is ‘wrong’, but tells you how long it takes it to discover the information, regardless of if it was available or not. Ex-ante, you’re likely never to find an answer, as the market probably knows more than your model.
LTCM did a very good job at finding ‘mispricings’ by oversimplifying assumptions. I bet the company that took the 60% (company A) was trading at premium over the 3/2 ratio. The reason for that is that this company had legal control over the joint operation, although technically both companies were entitled to their respective shares of the profits. If a third party wanted to take over the operation, it just had to buy the shares of company A, and hence only pay a premium to the holders of the shares of company A. Because of that, it further creates liquidity issues. As people will tend to prefer the shares of company A, they will shift their holdings to these stocks. Most of the trading is likely to happen in the shares of company A, not company B. Lots of similar issues happen when you have two classes of shares, one which would allow you to take technical control of a firm, the other which would not. Significant liquidity premia can build up in those circumstances. Remember, an arbitrage is not an arbitrage until you can convert from A to B, and vice-versa. joemontana Wrote: ------------------------------------------------------- > No, it’s covered somewhat in LIII. Here’s an > example that they cited: > > Years ago, 2 oil companies (big names, but I > forget which ones) decided to completely merge > their operations. They split the earnings 40/60. > Because they are identical, they should always > trade at the same multiple and a 3/2 price ratio. > Yet, inexplicabley, they traded at much different > multiples for decades. Many investors - including > LTCM - identfied the mispricing and took the > proper positions, thinking the two stocks would > converge. It didn’t happen. > > The main idea of the topic was that markets can > stay irrational longer than you can stay solvent.
Intrinsic value is metaphysical idea. It’s something you try to find out and put your best guess on, but you never actually know when a stock is being priced at its true intrinsic value. That would imply that you perfectly understood every risk possible in the operation, which you can’t.
cosine Wrote: ------------------------------------------------------- > Joey: most distributions are unbounded and > therefore you have a positive probability that X = > infinity, yet they have a finite expectation… > Nope. “Most” distributions have infinite tails which is world’s different than saying they have positive prob of being infinity. A r.v. with the latter is called an “improper random variable”.
Yes, you’re totally right, my bad.
You will be the king of Wall Street if you know the answer.