# Tag Info

28

I would offer the distinctions are i) pure statistical approach, ii) equilibrium based approach, and iii) empirical approach. The statistical approach includes data mining. Its techniques originate in statistics and machine learning. In its extreme there is no a priori theoretical structure imposed on asset returns. Factor structure might be identified thru ...

14

There are few things to consider. Trading moves the price, to minimize market impact and maximize return it is generally optimal to split an order in several child orders. See the Kyle model. Splitting optimally dependents on specific assumptions that you make. The simplest (and first) approach is that of Berstsimas and Lo (Optimal Control of Execution ...

13

I think you have the correct dichotomy here. Things started in the late 1980s and through the 1990s with analytical approaches particularly to derivative pricing (as in "hey, let's create yet another exotic option we can sell to the buy side"). The risk modelling "fashion" of the 1990s (when regulated entities such as banks needed to beef up reporting) ...

9

Interest rates in general are far from independent and identically distributed. A high interest rate observation is quite likely to be followed by another high observation, and the volatility is likely to be higher as well. Interest rates are also mean reverting, as in most real-world situations (at least for developed markets) interest rates rarely rise ...

8

MCMC can be used for Bayesian inference of other models with hidden variables. Gibbs sampling, for example, is used in Hidden Markov Models. Here is a paper that discuss the differences between MCMC and the more classical approach using the EM algorithm. The question is: Are HMMs a useful model in finance? Some academics argue that they have predictive ...

8

As far as I know MCMC and also (PMCMC) can be usefull for (bayesian) estimation of parameters of some Hidden process like in the Heston Model case based on observations of the Stock (filtering). But the problem here is that those estimates are not matching those based on calibration of vanilla options of the Risk Neutral measure. So as an econometric tool it ...

7

There is a huge difference between R (and Matlab, SAS, or other statistical languages) and relatively low-level languages such as C/C++/C#/Java in exactly this regard. The latter category is used more often for stable end-products, where speed and performance can be crucial, whereas the former category is used more often for model testing and prototyping. ...

7

The main component of that option premium is (forward-looking) volatility $\sigma$. The very simplest formula you could use for ATM options is the Bachelier model $$\text{Call}_T = \sigma S \sqrt{\frac{T}{2\pi}}$$ where the time to expiration is $T$ and $S$ is the current underlying price. This formula is "wrong" strictly ...

7

The investor's holdings is a consequence of an investor's utility function interacting with the investor's perceived trading opportunity subject to constraints. (Indeed, the Kelly criterion is also utility maximizing.) We produced trades by re-balancing -- that is to say, we have new expectations of alpha or risk and the optimal portfolio net of these ...

7

Attilio Meucci does some very interesting things with PCA. See e.g. his paper on managing diversification which makes heavy use of it (and explains it very intuitively along the way): Managing Diversification by Attilio Meucci

7

GARCH will work if volume has memory with some decay. AR will work if volume has mean reversion properties. Both of these are empirical questions and depend on the market. You should also consider if there are seasonal (day-of-week, monthly, quarterly effects) in which case you would want to add dummy variables. MA models will work well if volume behaves ...

6

There are certainly (short-rate) models which assume bounded interest rates. I suppose I should clarify - the design of the model prohibits negative interest rates. Further, some models asymptotically reach some target, or mean rate which is considered mean reversion, the most famous perhaps the Vasicek. Short rate models where rates cannot go negative: ...

6

I assume you're using returns to compute beta, not the prices. And yes, remove the "jumps", though this should happen automatically since you're looking only at intraday returns. One final piece of advice: you'll get more meaningful results if you smooth the returns via a moving average.

5

The model you assume for the interest rate process is a Geometric Brownian Motion. As strimp099 highlights in his comments it is mainly used to model equities because you most of the time want your interest rate models to be positive and mean reverting. A few models have been developed: Vasicek, CIR, HW. You could have a pick in there. As for the ...

5

High VIX arguably leads to less predictability of the market factor (i.e. market timing), but high volatility does lead to greater predictability of the cross-section of returns. Indeed, linear risk factor models have higher explanatory power during bear markets. However, your goal is to build a better market timing model where the forecasts (and perhaps ...

5

Is your question more about approaches taken on the buy side vs. sell side? If so, you may want to read Attilio Meucci's paper, P vs. Q, on this topic. He breaks down the dichotomy as derivatives pricing (the "Q" world), which uses a lot of very sophisticated modeling involving Ito calculus and PDEs, and portfolio management (the "P" world), which makes ...

5

From an academic viewpoint you do not have a lot of choices: The Rosenbaum-Robert approach, the price model with uncertainty zones is a model of trades and duration between trades (implicitly). It is worthwhile to try it. You can also use an Hawkes process, it will have the nice effect of capturing clustering effects on trades. if you want to use ...

5

Yes, there is a software application that you can purchase for $39.99 which stores all your tick data in a highly compressed format while still allowing maximum throughput and lowest latency data queries that I have ever seen. The package provides APIs to all languages under the sun but because they have a special sale going on it comes with the complete ... 5 Let's first restate the formula of the beta of a portfolio$P$relative to a benchmark$B\$: $$\beta_P=\frac{Cov(r_P,r_B)}{Var(r_B)}$$ As chrisaycock said in his comment, the key thing to understand is that the beta is a statistical measure computed relative to a benchmark. Hence, I believe that the real question you should be asking is: Which benchmark ...

5

The best paper is probably Relative Volume as a Doubly Stochastic Binomial Point Process - James Mcculloch. In this paper the volume is modelled via a Point Process, and theoretical laws are derived (with confident intervals, etc). And if you can wait few days (it will be available very soon), we put elements about this in Market Microstructure in Practice, ...

4

Have you looked into "noise trader" models? This seems like a market that is mostly noise. A few betters may have some information on or real knowledge of who might win, but certainly nothing like equity markets where there are a lot of people who really know the ins and outs of the firms they're trading. The classic model is Pete Kyle's, which should give ...

4

Liquidity Since this is an asset class which is so tightly coupled with interest rates - it makes good products for clients inherently complex. It also makes good sense to make wider markets for more exotic products than the plain vanilla ones - in which razor-thin spreads rule (and trading huge notionals is not everyone's cup of tea)

4

Since you mention beta, I assume you're familiar with the capital asset pricing model (CAPM). The concept is that an asset's expected returns are linearly correlated with the market's returns. Of course, there are other ways "normalize" returns, as you put it. We can extend CAPM with Fama-French, which adds market cap and relative value to the equation. ...

4

“Make things as simple as possible, but not simpler.” The problem you want to avoid is (near) multicollinearity. The tip-off will be that adding/removing a regressor will significantly change the coefficients on the other regressors. In practice (well, in the research that I read) I rarely see this explicitly tested. If you think that you have ...

4

I mainly speak as market practitioner when I say that I believe in the end all models that are applied to data and real life pricing issues are discretized. Think about it, even the BS hedge argument is in the end just a "theoretical continuous time overlay" of actual discrete time steps and re-hedges. Thus some of the limiting assumptions re BS. You do not ...

4

You question is quite strange: so you do not want to use methods inspired by bioinfo and genetics (neural networks, GA, geometry of folding, etc) but methods that are used in these fields? In terms of modeling, the problematics in bioinfo and genetics are mainly: tree or graph matching (to build metrics in the space of molecules), like in SIGMA: a ...

4

To create such a model, you'd start with some data, and then start fitting curves to it. For example, let's take a company where there are reasonable consensus forecasts about the next few years' earnings; and let's assume you've got some time-series data on changes in those consensus forecasts, and changed in the price. You could then fit a model based on ...

4

I have honestly not come across a good book (or good enough review to make me buy the book) on Fund Transfer Pricing. While it is not my career focus, I had to familiarize myself a bit with the topic because of certain requirements involving funding trading operations and the performance of funding specific operations. Personally I would recommend the ...

3

You could try using the Gaussian Affine Term Structure Models (GATSM), with the right boundary conditions to stop rates being negative (in the style of their Black implementation). See, for example, Monika Piazzesi, the "Affine Term Structure Models" if you want to enter/modify the basis or the work of Krippner, for example "Measuring the stance of monetary ...

3

Exponential distribution, although it's a good distribution for modeling non-negative numbers, doesn't make sense here since it's mode is 0. From a pure statistical point of view, without any knowledge of interest rate, I'd recommend log-normal as in modeling stock prices and inverse-gamma or gamma distribution which are used to model variance or other ...

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