Am I right that for the MVO approach we use historical returns for the various asset classes?
Also, exam 2011 Q5 A ii) says that “The MVO approach does not allow a person to incorporate their own views, while the BL does”. I thought you also had variations of the MVO where you could incorporate your own views? Are you only allowed to use historical returns?
Interesting…that doesn’t make sense to me. MVO we use our forecasted capital market expectations as inputs for MVO, which is literally “incorporating our own views”. It is not always historical metrics that are thr inputs for MVO. Theres no reason you cant use whatever you want to make it more stable. Not sure if you’re taking that out of context or not but doesn’t sound right to me.
Googs,
Are you sure that what you use is traditional MVO approach? It seems like modified one. I agree for traditional MVO only historical data are used.
The books has a lot of stuff about MVO, but it is not really clear to me what kind of returns that is supposed to be used. But it seems to be historical mean returns?
The only place where you will see “Views” mentioned alongside returns is in BL (anywhere in the Level III curriculum) and they make a pretty big deal about that … (you see Views it is Black Litterman and returns are adjusted based on the investor’s level of confidence on his/her views).
MVO thus would be historical mean returns.
Actually I think MVO allows for own views because the book states that one of the BL advantages is that the result is generally better-diversified portfolios than those produced from a MVO based only on the investor’s views.
I agree with Jones 473. There are contradictions in the curriculum but for the exam I will just remember they want us to say BL has an advantage of incorporating investor views. Right now, I don’t really care about the “real life” aspect of all this. I only care about what they want me to say to pass. I don’t see why traditional MVO can not be adjusted for my own personal views though. Theres no reason it cannot be. cpk123 is solid with his knowledge though so I value what he says, if you can give me a legitimate reason I cannot change my inputs into a traditional MVO then enlighten me, please.
The MVO model is all based on historical data that is used to created asset class forecasts that go into the MVO model. If you were to incorporate your “own views” i.e. HY is a lot more riskier now because of blah blah blah then the MVO model literally won’t work. It literally takes the forecasts for a given period then spits what it thinks the EF should be. Most Asset Allocation software won’t even let you use MVO if you implement your own views. It’ll force you to use some type of scenario based optimization.
If you force your proprietary capital market assumptions into a MVO model to create a EF, that’s just goes against the whole concept of MVO and you can’t tell your clients you use MVO because that’s just not correct anymore.
sure you can implement your forecast all you want and tinker with the optimization model, but just don’t call it MVO anymore.
Id like to see that explained somewhere because our firms software allows us to conditionally adjust the inputs. Would if you see no way that the historical data is covariance stationary any longer? The inputs used would not be prudent. MVO, to me, is exactly that. It optimizes combinations of inputs based on a mean and the variance around that mean all while incorporating cross correlations. Again, for the test I know what to say. But I am discussing this for “real life” argument.
I’m assuming that’s why the text mentions the whole thing about small little changes being a disadvantage because all it looks at are a few historic data points without limits creating these big output swings from concentrated positions. I don’t feel the texts stresses that enough, but oh well. The answer will probably be multiple choice and an obvious choice anyways, so it’s all irrelevant.
If interested I know Morningstar talks about it because I’ve had to look into it before. But our other system wyndam or w/e allows you to input your own. But i’m pretty sure it’s using a scenario base analysis like monte carlos even though it says MVO. It’s definition guide wasn’t a detailed as morningstar was. Then there were a few others like BlackRock and Factset that don’t allow it either but makes you run a different report. In case you curious.
Ill worry about it on June 4th. For now I know enough.
for me it’s all about expected returns not historical
BL bases on historical and incorporates views to reach a new set of expectations