Résultats de recherche pour « Optimal ETF Selection »

This paper considers the problem of isolating a small number of exchange traded funds (ETFs) that suffice to capture the fundamental dimensions of variation in U.S. financial markets. First, the data is fit to a vector-valued Bayesian regression model, which is a matrix-variate generalization of the well known stochastic search variable selection (SSVS) of George and McCulloch (1993). ETF selection is then performed using the decoupled shrinkage and selection (DSS) procedure described in Hahn and Carvalho (2015), adapted in two ways: to the vector-response setting and to incorporate stochastic covariates. The selected set of ETFs is obtained under a number of different penalty and modeling choices. Optimal portfolios are constructed from selected ETFs by maximizing the Sharpe ratio posterior mean, and they are compared to the (unknown) optimal portfolio based on the full Bayesian model. We compare our selection results to popular ETF advisor Wealthfront.com. Additionally, we consider selecting ETFs by modeling a large set of mutual funds.

The problem of optimal allocation of voltage-reactive regulatory resources has been widely considered in the field of planning and management of the power systems in order to maintain the voltage profile, as one of the most important aspects of the operation of a network, within the given limits. During the installation of these devices, the topology of the network is changed, which is why it is essential to pay sufficient attention to the sizing of these devices in the planning phase, from the point of view of selecting adequate power, and to choose suitable locations for their location.This paper proposes a method for determining the minimum required number of reactive power compensation devices, their locations and powers, with the aim of improving the voltage profile. The difference between this work in comparison to existing research work is that the number of devices, as well as their locations, will not be defined in advance, so that the locations and power of the devices will be treated as variables. The effectiveness of the proposed method will be tested using IEEE test systems.

The goal of Feature Selection - comprising filter, wrapper, and embedded approaches - is to find the optimal feature subset for designated downstream tasks. Nevertheless, current feature selection methods are limited by: 1) the selection criteria of these methods are varied for different domains, making them hard to generalize; 2) the selection performance of these approaches drops significantly when processing high-dimensional feature space coupled with small sample size. In light of these challenges, we pose the question: can selected feature subsets be more robust, accurate, and input dimensionality agnostic? In this paper, we reformulate the feature selection problem as a deep differentiable optimization task and propose a new research perspective: conceptualizing discrete feature subsetting as continuous embedding space optimization. We introduce a novel and principled framework that encompasses a sequential encoder, an accuracy evaluator, a sequential decoder, and a gradient ascent optimizer. This comprehensive framework includes four important steps: preparation of features-accuracy training data, deep feature subset embedding, gradient-optimized search, and feature subset reconstruction. Specifically, we utilize reinforcement feature selection learning to generate diverse and high-quality training data and enhance generalization. By optimizing reconstruction and accuracy losses, we embed feature selection knowledge into a continuous space using an encoder-evaluator-decoder model structure. We employ a gradient ascent search algorithm to find better embeddings in the learned embedding space. Furthermore, we reconstruct feature selection solutions using these embeddings and select the feature subset with the highest performance for downstream tasks as the optimal subset.

A U.S. investor holding a large, low-basis single stock can move into a diversified fund without triggering the embedded capital-gains tax by contributing securities in kind under Internal Revenue Code Section 351, provided the basket she contributes clears a statutory diversification test: no single issuer above 25% of its value, and the five largest issuers together at or below 50%. The regulation tests each contributor's own basket, so the other contributors cannot rescue a concentrated investor. This note formalizes the seeding decision the test creates and quantifies its economics. We write the 25%/50% test as linear constraints on contribution weights, cast the choice of seed basket as a linear program that maximizes deferred gain, and derive three closed-form consequences of the caps. Contributing a position of value C at the 25% ceiling requires at least 3C of eligible ballast from the investor's own holdings, so the share of a concentrated position that can go in is her eligible ballast divided by 3C; a seed that places one issuer at the cap can be no more diversified than sixteen effective names; and deferral loses to an immediate sale only if the capital-gains rate at the horizon exceeds a breakeven set by today's rate, the growth factor and the step-up probability alone. We then compare four ways to handle a $10M low-basis position (351 seed, exchange fund, hold-and-borrow, and sell-and-pay) on after-tax terminal wealth and on a risk-aversion-adjusted certainty equivalent, over horizons of 10 to 30 years and across the probability of a basis step-up at death. In the base calibration the 351 seed has the highest certainty equivalent at every horizon, because it pairs the deferral of a sale with real diversification at a low fee. Hold-and-borrow posts the highest expected wealth and the lowest certainty equivalent, as borrowing against an undiversified position would predict. All figures are reproducible from a single script with a fixed seed.

We discuss how to build ETF risk models. Our approach anchors on i) first building a multilevel (non-)binary classification/taxonomy for ETFs, which is utilized in order to define the risk factors, and ii) then building the risk models based on these risk factors by utilizing the heterotic risk model construction of https://ssrn.com/abstract=2600798 (for binary classifications) or general risk model construction of https://ssrn.com/abstract=2722093 (for non-binary classifications). We discuss how to build an ETF taxonomy using ETF constituent data. A multilevel ETF taxonomy can also be constructed by appropriately augmenting and expanding well-built and granular third-party single-level ETF groupings.

<p>Better forecasts do not necessarily produce a better global equity portfolio. The question is whether those signals improve the final portfolio once benchmark-relative risk, overlapping holdings, trading, and costs are taken into account. This paper examines a global equity ETF portfolio that combines machine-learning selection, benchmark-aware regional allocation, and a small allocation across themes, sectors, and styles. The components are evaluated together against the iShares MSCI ACWI ETF (ACWI).</p> <p>The common daily evaluation covers 31 January 2019 through 31 August 2026. In this reconstructed research history, the portfolio compounded at 18.9% annually versus 13.8% for ACWI, with 6.15% tracking error, a 0.819 information ratio, and a maximum relative drawdown of 8.10%. The advantage was uneven: the portfolio lagged ACWI in 2022, rolling active returns turned negative in some periods, and primary selection alone had a better recent trade-off between active return and tracking error than the complete portfolio.</p> <p>Longer component histories include additional market environments but are not a longer backtest of the integrated portfolio. Because the model combination was selected during research, the evidence does not establish an optimal mix. The relevant test is what each component contributes after the portfolios are combined.</p>

The growth of the exhange-traded fund (ETF) industry has given rise to the trading of options written on ETFs and their leveraged counterparts {(LETFs)}. We study the relationship between the ETF and LETF implied volatility surfaces when the underlying ETF is modeled by a general class of local-stochastic volatility models. A closed-form approximation for prices is derived for European-style options whose payoff depends on the terminal value of the ETF and/or LETF. Rigorous error bounds for this pricing approximation are established. A closed-form approximation for implied volatilities is also derived. We also discuss a scaling procedure for comparing implied volatilities across leverage ratios. The implied volatility expansions and scalings are tested in three well-known settings: CEV, Heston and SABR.

Investments in cryptocurrencies (CCs) remain risky due to high volatility. Exchange Traded Funds (ETFs) are a suitable tool to diversify risk and to benefit from the growth of the whole CC sector. We construct an ETF on the CRIX, the CRyptocurrency IndeX that maps the non-stationary CC dynamics closely by adapting its constituents weights dynamically. The scenario analysis considers the fee schedules of regulated CC exchanges, spreads obtained from high-frequency order book data, and models capital deposits to the ETF stochastically. The analysis yields valuable insights into the mechanisms, costs and risks of this new financial product: i) although the composition of the CRIX ETF changes frequently (from 5 to 30 constituents), it remains robust in its core, as the weights of Bitcoin (BTC) and Ethereum (ETH) are robust over time, ii) on average, a portion of 5.2% needed to be rebalanced at the rebalancing dates, iii) trading costs are low compared to traditional assets, iv) the liquidity of the CC sector has increased significantly during the analysis period, spreads occur especially for altcoins and increase by the size of the transactions. But since BTC and ETH are most affected by rebalancing, the cost of spreads remains limited.

A high liquidity, low expense ratio and the possibility to conduct arbitrage allow exchange-traded funds (ETFs) to be used for short sales. Bearish investors can also buy inverse ETFs. This paper aims to outline two investment approaches for bearish ETF investors and the differences between these two approaches; it also aims to examine the relationship between price and an indicator of volume and evaluate the final positions in selected ETFs in selected periods. Short ETFs dominate in simplicity, flexibility, paying out dividends and especially in the limited size of the loss. On the other hand, their structure, which demands daily rebalancing, causes substantial deviation from the benchmark in the long-term and leads to a higher expense ratio, and lower liquidity increases bid-ask spreads. Negative aspects of ETF short selling lie in unlimited loss, high borrowing costs, the need for margin accounts, variability of loan fees and the possibility of a transaction recall by the lender. On the contrary, margin operations enable potentially higher appreciation of capital by generating rebate rates. Our results show that with the decrease in value of the most used ETFs, short interest is growing for those funds where there is a very strong negative correlation implying hedging tendencies. Short selling proved to be a more advantageous strategy in the observed period of market downturn, as well as in 2011–2017, due to negative returns, however, by applying margin trading inverse ETFs turned out to make less losses. Sector-oriented inverse ETFs are the exception, where the largest differences between these two strategies are recorded. However, the final conclusion of the suitability of one of the analyzed strategies depends on the market volatility and the direction of the market itself.