Shrinking the Cross Section
Citations
236 citations
199 citations
127 citations
Cites background or methods from "Shrinking the Cross Section"
...Kozak et al. (2017) use model-selection techniques to approximate the SDF and the mean-variance efficient portfolio as a function of many test portfolios, and compare sparse models based on principal 1Some of the factors proposed in the literature are based on economic theory (e.g., Breeden (1979),…...
[...]
...6For any matrix A, we use Ai,· and A·,j to denote the ith row and jth column of A, respectively. selection on top of PCs similar to Kozak et al. (2017))....
[...]
...…selection step corresponds closely to the approach taken in the current literature dealing with the proliferation of asset pricing factors (e.g., Kozak et al. (2017)): take a large set of factors (ht), apply some dimension-reduction method (LASSO, Elastic net, PCA, etc.), and interpret the…...
[...]
...Kozak et al. (2017) use model-selection techniques to approximate the SDF and the mean-variance efficient portfolio as a function of many test portfolios, and compare sparse models based on principal...
[...]
...…the cross-sectional LASSO, closely related to the dimension-reduction methods that recent papers in asset pricing have been using to tackle the factor zoo (e.g., Kozak et al. (2017)): the objective of this first LASSO is to select a parsimonious model that explains the cross-section of risk premia....
[...]
119 citations
Cites background or methods from "Shrinking the Cross Section"
...…onto quantiles of characteristics are exactly the input to PCA in Kelly, Pruitt, and Su (2019) or the elastic net mean-variance optimization in Kozak, Nagel, and Santosh (2020).12 The solution to minimizing the sum of squared errors in these moment conditions is a simple mean-variance…...
[...]
...Kozak, Nagel, and Santosh (2020) estimate the SDF based on characteristic sorted factors with a modified elastic net regression.5 Kelly, Pruitt, and Su (2019) apply PCA to stock returns projected on characteristics to obtain a conditional multi-factor model where the loadings are linear in the…...
[...]
...This is a standard transformation to deal with the different scales and has also been used in Kelly, Pruitt, and Su (2019), Kozak, Nagel, and Santosh (2020) or Freyberger, Neuhierl, and Weber (2020) among others....
[...]
...The linear approach with elastic net is closely related to Kozak, Nagel, and Santosh (2020) who perform mean-variance optimization with an elastic net penalty on characteristic based factors.21 In addition we also report the maximum Sharpe ratios for the tangency portfolios based on the Fama-French…...
[...]
...28Kelly, Pruitt, and Su (2019) and Kozak, Nagel, and Santosh (2020) construct factors in this way....
[...]
100 citations
References
40,785 citations
"Shrinking the Cross Section" refers background or methods in this paper
...…have noted that Lasso does not perform well when regressors are correlated and that ridge regression (with L2-norm penalty) or elastic net (with a combination of L1- and L2-norm penalties) delivers better prediction performance than Lasso in these cases (Tibshirani, 1996; Zou and Hastie, 2005)....
[...]
...Lasso is known to suffer from relatively poor performance compared with ridge and elastic net when regressors are highly correlated (Tibshirani, 1996; Zou and Hastie, 2005)....
[...]
...To allow for factor selection, we augment the estimation criterion with an additional penalty on the sum of absolute SDF coefficients (L1 norm), which is typically used in Lasso regression (Tibshirani, 1996) and naturally leads to sparse solutions....
[...]
24,874 citations
"Shrinking the Cross Section" refers background or methods or result in this paper
...We focus on the well known 25 ME/BM sorted portfolios from Fama and French (1993). We show that our method automatically recovers an SDF that is similar to the one based on the SMB and HML factors constructed intuitively by Fama and French (1993)....
[...]
...For instance, the CAPM predicts a single factor representation; the 5-factor model of Fama and French (2016) and investment-based asset pricing models represent an SDF in terms of size, book-to-market and/or investment, and profitability characteristics....
[...]
...We focus on the well known 25 ME/BM sorted portfolios from Fama and French (1993)....
[...]
...For example, Fama and French (1993) use two characteristics: market capitalization and the book-to-market equity ratio....
[...]
...To summarize, these results confirm that our method can recover the SDF that Fama and French (1993) constructed intuitively for this set of portfolios....
[...]
19,261 citations
16,538 citations
"Shrinking the Cross Section" refers background or methods in this paper
...Lasso is known to suffer from relatively poor performance compared with ridge and elastic net when regressors are highly correlated (Tibshirani, 1996; Zou and Hastie, 2005)....
[...]
...…have noted that Lasso does not perform well when regressors are correlated and that ridge regression (with L2-norm penalty) or elastic net (with a combination of L1- and L2-norm penalties) delivers better prediction performance than Lasso in these cases (Tibshirani, 1996; Zou and Hastie, 2005)....
[...]
...(28) we use the LARS-EN algorithm in Zou and Hastie (2005). to represent the value effect in an SDF, it may be advantageous to consider a weighted average of multiple measures of value, such as book-to-market, price-dividend, and cashflowto-price ratios....
[...]
...39 we use the LARS-EN algorithm in Zou and Hastie (2005), with few small modifications that impose economic restrictions specific to our setup....
[...]
...The approach is motivated by lasso regression and elastic net (Zou and Hastie, 2005)....
[...]
10,806 citations