36

I think you're overlooking a third explanation: Nobody that found a successful technique to generate alpha has published it. I can think of the following causes: If you're an academic, why share your brilliant idea? These techniques require a lot of data and financial data can be expensive, researches that work at firms that have access to this data don't ...


26

In the early 2000s I met the Quant Team at Barclays Global Investors in San Francisco and I tried to convince them to submit some of their research to the journal I was managing at the time, Quantitative Finance. This was met with some mirth and incredulity and they told me that they might consider submitting some of the work that went into v2 of their model ...


15

There are very many finance papers using machine learning One the of the top finance journals is the Review of Financial Studies (RFS). You find 87 published and peer-reviewed papers if you look up machine learning. You find a lot more recent research (working papers) 'published' by the National Bureau of Economic Research (NBER). Looking for 'machine ...


9

If one finds an anomaly relative to some asset pricing model, there are three possibilities: The anomaly you "found" actually isn't there: you're overfitting the data, found a spurious result. The efficient market hypothesis is wrong. The asset pricing model you're using is wrong. Any test of market efficiency is a joint test of market efficiency and an ...


8

The best explanation I have seen so far is the so-called Adaptive Market Hypothesis by Andrew Lo: The adaptive market hypothesis, as proposed by Andrew Lo, is an attempt to reconcile economic theories based on the efficient market hypothesis (which implies that markets are efficient) with behavioral economics, by applying the principles of ...


8

Adding to the answer of @BobJansen there are some additional worries with complex machine learning models (eg. Neural Networks of any kind and complex tree-based approaches) that you can encounter, within the setting of volatility forecasting (or forecasting in general). The mechanisms behind the results can be hard to explain: For forecasting purposes, ...


6

They are different concepts, and the relation between them can be described as a conditional: "if EMH holds (all available information about future price movements is already priced into the market), then future price movements will follow a purely random walk as new and unpredictable information emerges"


6

What you describe is known as the Equity Premium Puzzle - and it really is, as the name says, a real enigma: "The equity premium puzzle (EPP) is a phenomenon that describes the anomalously higher historical real returns of stocks over government bonds." Source: https://www.investopedia.com/terms/e/epp.asp#ixzz5HlCdHS2Z A good first introduction can be ...


6

The way I understand it is: In equation 2 $x_{j, t + 1}$ is defined as the change in of $p_j$ over the period $t$ to $t + 1$. The formula says that the expectation of the change is zero which is the same as saying that the expectation of the original variable at $t+1$ is equal to its current value.


6

Supply and Demand That's a great question and your question has a simple answer. @Daneel showed you the ``maths'' behind it. It boils down to: if an asset has too high returns than the CAPM predicts, then the asset is super attractive. An asset has a certainty riskiness (measured by its (market) beta) which warrants a certain future return (the CAPM formula, ...


5

On more than a few occasions, I have attempted to extrapolate the current trend towards passive allocation to its logical conclusion: more passive allocation means more inefficiency. I am not aware of any research which directly measures the correlation between market efficiency and active versus passive allocation. In general, the level of market ...


5

Any anomaly that can be phrased as a "mispricing" or "relative value" opportunity can be expected to disappear as more people discover it and trade on it. For example, say that stock movements over the last 15 minutes of the day are found to be strongly mean-reverting. That is, stocks which decline over the last 15 minutes of the day tend to be undervalued, ...


5

Historically the RWT (Random Walk Theory) came first, as empirical observations by for example M.F.M. Osborne (1959) and others in the 1960s. The EMH came about as a result of theoretical work by Samuelson in 1965 ("Proof that properly discounted prices...") and E.Fama (1969) as a general empirical/theoretical hypothesis that guided the field for many ...


5

As @skoestlmeier and @noob2 commented there's much research going on about the profitability anomaly. Firstly, there are different ways of measuring profitability. Novy-Marx (2013, JFE) uses gross profitability, Fama and French (2015, JFE) total profitability and Hou et al. (2015, RFS) return on equity. The $q$-theory model from Hou et al. claims to explain ...


5

For simplicity let us assume we are considering a single investment period, that is from $t$ to $t+1$. Let $S_i(t)$ be the price of the asset $i$ at time $t$. Then the return of the asset between $t$ and $t+1$ can be decomposed into its price return and any dividends paid: $$R_i=\frac{S_i(t+1)-S_i(t)}{S_i(t)}+D_i$$ where $D_i$ are the dividends paid during ...


4

Generally Kurtosis measures the degree to which a distribution is more or less peaked than a normal distribution. Positive kurtosis indicates a relatively peaked distribution. Negative kurtosis indicates a relatively flat distribution. In time series we can encounter high kurtosis which is caused by "fat tails" (higher frequencies of outcomes) at the ...


4

Perhaps an answer coming from a different angle and giving you some perspective: The typical approach taken by statistics is top-down: Just looking at the data and finding patterns and stylized facts (like excess volatility, volatility clustering, fat tails, no autocorrelation in returns but significant autocorrelation in absolute returns etc.) The problem ...


4

Adding to the excellent answer of vonjd: another, more cynical interpretation is that some of these "anomalies" never would have existed ex-post and that their discovery is the inevitable result of thousands of researchers looking for patterns in the same data set. Look at ". . . and the Cross-Section of Expected Returns" by Harvey et. al. for an ...


4

First, you might find this recent paper by Israeli, Lee and Sridharan (Review of Accounting Studies, forthcoming) interesting. This is the abstract: We examine whether an increase in ETF ownership is accompanied by a decline in pricing efficiency for the underlying component securities. Our tests show an increase in ETF ownership is associated with: (1) ...


4

Suppose markets are perfectly efficient and asset prices reflect all available information. Under this assumption one expects current prices to be non-biased estimators of future prices. It is a common mistake to think that market efficiency implies $P_t = E_t[P_{t+1}]$! In general, the correct statements are: $P_t = \frac{E_t^Q[P_{t+1}]}{R_f}$ where $Q$ ...


4

tl;dr– People don't publish trade-secrets. They're trade-secrets. Stuff like effective market-prediction algorithms tend to be trade-secrets: Trade secrets are a type of intellectual property that comprise formulas, practices, processes, designs, instruments, patterns, or compilations of information that have inherent economic value because they are not ...


4

Because it would not work. There was a very old paper several decades ago (dont recall but think it was in an IEEE journal on information theory) that showed that the stock market is RANDOM within an epsilon well smaller than the vig the brokers charge. The sad truth is that all AI is genuine stupidity that assumes that correlation is causation. And ...


3

What are some indicators that a given security might be inefficiently priced? What about efficiently priced (i.e., how can we estimate the degree of information already baked into price)? You would need to get lower level data than what was used in this paper that was referenced here. The MSF data is probably fine for the paper, but if you were to drill ...


3

You might find this paper interesting: "Does Finance Benefit Society?" It's a very complicated question and in my opinion the above paper provides a nuanced answer.


3

We know that: \begin{equation} R_{t+1} = \frac{P_{t+1} + D_{t+1}}{P_t} \end{equation} After some algebra and taking logs we can write the returns as: \begin{equation} r_{t+1} = k + \rho (p_{t+1} - d_{t+1}) - (p_t - d_t) + \Delta d_{t+1} \end{equation} where is constant $\rho = \frac{P/D}{1+P/D}$. or: \begin{equation} (p_t - d_t) = k + \rho (p_{t+1} - ...


3

At what scale do you see kurtosis? Daily data? Single stocks or indices? Let us not look at single stock data, because you always find crazy stocks whose price process breaks all rules. Talking about daily data of indices: they could be thought of the sum of hourly returns or other returns of high frequency (minute returns, milliseconds ...). What are the ...


3

I think there are a few conflating ideas here. With respect to the sum of logs idea, I think you're thinking about infinitely divisible distributions (https://en.wikipedia.org/wiki/Infinite_divisibility_(probability)). These ideas are indeed used to build more complicated models (i.e. Levy processes) for asset returns. With regards to the Efficient Market ...


3

Let's begin from the start. At its core, market efficiency is a statement about the compensation for risk embedded in asset prices. So, you can think of this issue as involving 3 quantities: (1) the price that you observe, (2) the price that you should observe and (3) the distance between them. The fundamental problem with trying to test for market ...


3

I upvote @BobJansen's points; and add one small incremental observation. Institutional investor scepticism about quant isn't so much scepticism about quant. It's more the fear that quant describes the current regime. Which investors believe (maybe correctly, maybe not) they already understand. Unless quant can tell them how this has changed, and do so ...


3

Was debating if I should even comment on this but then thought tonight I'm gonna have myself a real good time. JPMorgan Machine Learning in Financial Markets Conference, Paris 2019 offers a superficial summary. You simply do not have the data IMHO. To capture complex relationships you tend to have more parameters, which in turn leads to even more data ...


Only top voted, non community-wiki answers of a minimum length are eligible