How is CAPM not technical analysis?

Modern Portfolio Theory is pretty mainstream in finance, whereas technical analysis is frowned upon by at least 40% of money managers. Furthermore, all sources I can find claim that MPT (specifically CAPM) are compatible with (actually, founded on) the Efficient Market Hypothesis. How does this make any sense?

The weak EMH states that technical analysis doesn't work because it's impossible to predict future price data from past price data.

Well, here's how the CAPM works:

>Input ((past returns))

>Calculate correlations

>Pick portfolio with maximum amoung of ((past returns)) at minimum possible ((past correlation))

>According to CAPM this portfolio will ((outperform)) all other portfolios in the ((future))


CAPM is technical analysis and cannot work according to EMH. Change my mind.

Attached: three_assets.png (810x578, 46K)

Every single thing you have in your post only helps mitigating risk, not actualy helping to predict the market.

You been sold into the idea of stay long on the game, so they can milk you more.

No, I know the CAPM doesn't work and I don't use it. This is theoretical.

Reducing risk at constant returns is equivalent to increasing returns (in a perfect capital market). The CAPM claims to be able to select a portfolio that will outperform other portfolios in the future, based only on past price (return) data. I don't understand how this is supposed to be based on EMH which forbids this exact thing.

Also, I understand that returns aren't necessarily based only on price action, but they could be, and in many cases are, which is enough to violate the EMH.

Do you have an exam on financial market today?

If I did, would I be on Jow Forums?

Any useful replies?

CAPM is a basic statistical model. While /biz TA is a bunch of meme lines

But fundamentally, they're the same. Both assume that you can increase returns based on past price information.

I mean theoretically sure if you deconstruct it enough then yeah both CAPM and TA are founded on a similar principle - that past data can be used to forecast future returns.

CAPM is a very basic model... at least use F&F three factor. Any analyst worth his salt understands however that the more factors a model has, e.g. the more complex it is, the more it can go wrong during uncommon market conditions.

TA differs in that while the EMH presupposes homo economicus is the only player in the market, TA presents a view of human fallibility. Mr. Market is prone to irrational caprice and you can capitalize on that.

But this is precisely my question: How can it be that CAPM blatantly disagrees with the EMH; yet, all the source say that it's based on it? The FF-model is actually an (improved) extension of CAPM, and therefore also violates the EMH, which is all the more ridiculous because Fama literally invented the EMH and has been its strongest supporter. It seems to me that the mainstream financial theories contradict each other/ themselves.

One of CAPMs various bullshit assumptions is that debt is always available to be taken or given at the risk free rate of return. Diversified portfolio, calculation of beta, market risk premium all very subjective. Might work in heaven but not on earth

Again, I already know it's a bad model. My point is that the CAPM, FF and the hundreds of other factor loading models based on it all violate the EMH but are used by the same mainstream money managers/economists who defend (( or literally created) the EMH. I'm confused about the fact that the entire field of finance seems to contradict itself on a basic level.

I am out of my depth here, but I think the way it works is
> pick a portfolio of least correlated assets to mitigate systemic risk
> aggressively monitor and limit draw-down from losers
> let winners run
>????
Profit!

In very volatile markets, you don't have to predict future prices, you have to cut lose assets that are going down, volatility does the rest

According to EMH, past correlation doesn't inform future correlation either. You shouldn't be able to draw any inference from past price movements. "Portfolio a will beat all others" is only ordinal inference, but inference nonetheless. In an EMH market, you shouldn't be able to do any better than to buy an ETF of all stocks out there and hold.

Read the new finance: the case against efficienct markets by robert haugen
Also read Nassim Nicholas Taleb

Understand that while the big players would have everyone believe that we live in mediocristan where finance can reduced to mathematical equations, in actual fact we live in extremistan where everything can change in the blink of an eye.
tl;dr it’s all bullshit lurk moar

CAPM just calculates what a security's EXPECTED return SHOULD be based on its risk (beta), not what it will be in the future. Securities can have positive or negative alpha.

Should is just another way of saying "will with more than 50% probability". It's still an inference forbidden under EMH.

It's a nice classroom tool because it's not too complicated and allows to construct interesting exams around it. Almost all bachelor business/finance students will understand the simple statistical intuitions behind it, after 10-30min.

Sadly has giant serious flaws..
>implying one can predict returns
>implying one can predict future volatility
>inb4 using past volatility and returns to predict future values
>bonus edition: Constructing portfolio only based on 1 or 2 years of historical returns and volatility

Attached: ydndtsee.gif (245x190, 525K)

No, it lets you construct the optimum efficient portfolio. Some assets according to the model are just not very good combinations of volatility and return (e.g. to much volatility for weak return)

Once again though, in EMH, you should never be able to determine a portfolio that outperforms other portfolios based on past price data. Correlaction is based on past price data. My question is: Why does half of the mainstream in fiance, including people like Fama, who espouse EMH, accept (create) models (such as FF and CAPM) that explicitly violate EMH?

I know that the CAPM is bad. That's not my point. My point is it violates EMH. All the less bad models based on it (Black CAPM, FF etc.) also violate EMH. BUT they're staples of the finance mainstream. BUT so is the EMH. You literally have the same people get Nobel prizes for papers on both. What's going on?

Since when were bizlets this intelligent? Back to your cagies!

My intuition is that we use past history because it's simply the best tool we have.

We assume that returns will fall in a normal or semi-normal distribution, which over the years they (largely) have

I could be completely wrong on this though

And is EMH proven? IMO, it has been shown that price movement is not like Brownian motion which would imply that EMH is not law

No, it's not proven. It's highly disputed. But the people who defend it overlap with those who use models such as CAPM and FF, which contradict it. Fama is the father of the EMH as well as the Fama-french model. This is what I'm confused about.

You don't seem to fully understand CAPM. CAPM doesn't violate EMH. CAPM states that the expected return on any asset is a fuction of its ->systemic

EMH doesn't state that past correlations and variances don't matter. You can use past correlations and variances to -->estimate

How doesn't it? Past correlations are computed exclusively from past price data. EMH, even in its weak form, states that no inferences as to future price data can be drawn from previous price data. According to EMH, there shouldn't be any stable correlations. If there is, I can sell a correlation swap for any two assets. Under CAPM, I implicitly assume the swap's price will remain constant, because it has in the past, which is nothing if not TA.

According to CAPM, there is an ex-ante determinable subset of the market (the market portfolio) that significantly outperforms a random selection (or a complete index) long-term. It's determined using only past prices.

> According to EMH, past correlation doesn't inform future correlation either
This is wrong
> You shouldn't be able to draw any inference from past price movements
This is wrong too
> In an EMH market, you shouldn't be able to do any better than to buy an ETF of all stocks out there and hold
This is true if markets truly are 100% efficient at all the time, but as I said, no one believes that to be true.

I think you're confused because you don't fully understand what EMH actually states.

>Efficient market hypothesis
if markets where efficient then traders wouldn't make money. Hedge funds wouldn't make money.

Ask me how I know you some eco student that hasn't spent more than 1 month in a market. Just look at coinmarketcap. Does that look like efficiency to you?

So, you're saying that the EMH states that no inferences to future prices can be drawn from past prices, but the same can be done with correlations? How is that possible in an efficient market where correlation can be securitized? If I can buy and sell correlation, are you saying that EMH permits me to conduct TA on correlation swap charts? I can't see any reading of Fama that suggests this.

EMH doesn't state that past price movements/correlations can't be used to infer how assets will move ->relative According to EMH, there shouldn't be any stable correlations
This is not true
>
According to CAPM, there is an ex-ante determinable subset of the market (the market portfolio) that significantly outperforms a random selection (or a complete index) long-term. It's determined using only past prices.
CAPM simly states that you only get rewarded for systemic risk (beta). Through diversification you can minimize your unsystemic risk (you don't get rewarded for unsystemic risk) and maximize your expected return at a given level of variance (variance = systemic risk + unsystemic risk).

> You shouldn't be able to draw any inference from past price movements

>>This is wrong too


That is literally what Fama states as the weak-form EMH.

ecomonists are the niggers of academic

EMH doesn't state that past price movements/correlations can't be used to infer how assets will move ->relative

Weak EMH states that past price data can't be used to predict which assets will outperform the market ->risk adjusted

I forgot I was on biz before this

No you can't according to CAPM. Here's an example:
Risk free rate = 0%
Asset A beta = 1.5
Asset B beta = 0.5
Market beta = 1.0
Market risk premium = 5%

Please tell me how you would use this knowledge to guarantee a winning trade. You can't.

And if you don't understand, this means that
> asset A is expected to appreciate 2.5*X%+0% when the market appreciates X%
> asset B is expected to appreciate 0.5*X%+0% when the market appreciates X%

The difference between actul returns and expected returns is net unsystemic risk (which the market doesn't compensate you for according to CAPM).

Asset A is expected to appreciate 1.5X%*.

Let me know if you have any final questions! I'm leaving in 20 mins.

when moon

>CAPM is technical analysis and cannot work according to EMH.

CAPM is quantitative analysis, and the numerical methods of quantitative analysis could be classified as technical analysis and that's precisely where the difference between quantitative and technical becomes fuzzy. However, the difference is that in TA you never check for correlations. In TA you have a symbol and all you want is to get buy and sell signals. Sure, you can pick 100 symbols and trade them all with TA, but when you are trying to pick these 100 symbols doing something like CAPM you are doing quantitative, not technical.

Also EMH is false.

If dubs the golden bull run will start this month.

I was away but if you're still here, I understand where you're coming from, but I don't understand why you don't consider nonsystemic risk reduction not an improvement in returns as per the EMH. The CAPM market portfolio supposedly has strictly superior returns to the market itself, therefore I can make a profit by selecting the optimal portfolio instead of an index of the entire market. This shouldn't be possible under EMH. I also still don't see how, according to game's original paper, price data processed into correlation data, does not fall under "past price data". If I can use past price data to create a portfolio that consistently beats the market (by consisting of a subset of the market that contains less risk =read:higher returns), then according to the EMH, the fact that everyone chooses this portfolio should lead to the stocks becoming more expensive.

In fact, according to EMH, this should eventually lead to all negative beta stocks being more expensive and all positive beta stocks being cheaper, until in the end, there is no portfolio that consistently performs better than a portfolio of the entire market. 2