Annual Liquidity-Sorted Decile Portfolios with Missing Observations
Summary
The document describes a portfolio-construction task using monthly average quoted spreads for a broad set of companies over a long sample. The intended procedure is to rank stocks by a liquidity measure at the start of each year, divide them into ten equally weighted portfolios, and hold those assignments for the following year. The example data contain missing values, so the question also raises how unavailable spread observations should be handled when forming each annual cross-section.
The text is a request for guidance rather than a worked solution: it gives no sorting rule, code, portfolio returns, or evidence about the performance of the resulting portfolios. It does specify a lagged formation convention—using the prior year's measurements to assign portfolios for the next year—which helps avoid using future information. Any implementation would still need explicit rules for missing data, ties, changing stock coverage, and the direction of the spread ranking, since higher quoted spreads generally indicate lower liquidity.
Key ideas
- Rank companies by a quoted-spread measure to form cross-sectional liquidity portfolios.
- The proposed portfolios are equally weighted and divided into ten groups.
- Use measurements from the preceding year to form portfolios at the start of the next year.
- Missing spread observations require an explicit inclusion or exclusion rule.
- The document poses the method question but does not supply results or an implementation.
Tags
Full text
# How to form Decile Portfolios based on Liquidity measure with missing data in R # How to form Decile Portfolios based on Liquidity measure with missing data in R I have a dataframe with over 4000 companies data (as column) and have calculted their daily Quoted spread measure ( measures liquidity for each stock) for 15 years. And then from the daily have estimated the monthly average Quoted spread for each company( for 180 months). Now, I want is to form equally wieghted decile portfolios of these companies specifically at the beginning of each year in my sample period meaning I want to form 10 portfolios at the beginning of each year based on the liquidity measured. Below I illustrate small portion of my data ``` Month A B C D E F G H I J Jan-00 0.05 0.02 NA 0.00 NA 0.04 0.04 NA NA 0.03 Feb-00 0.07 0.02 NA 0.01 NA 0.04 0.08 NA NA 0.02 Mar-00 0.09 0.02 NA 0.00 NA 0.04 0.07 NA NA 0.03 Apr-00 0.11 0.02 NA 0.00 NA 0.03 0.08 NA NA 0.03 May-00 0.14 0.02 NA 0.00 NA 0.03 0.06 NA NA 0.02 Jun-00 0.15 0.01 NA 0.01 NA 0.04 0.05 NA NA 0.02 Jul-00 0.10 0.02 NA 0.01 NA 0.03 0.05 NA NA 0.04 Aug-00 0.11 0.02 NA 0.00 NA 0.03 0.03 NA NA 0.03 Sep-00 0.08 0.02 NA 0.01 NA 0.04 0.03 NA NA 0.02 Oct-00 0.11 0.02 NA 0.01 NA 0.03 0.04 NA NA 0.02 Nov-00 0.17 0.02 NA 0.01 NA 0.04 0.04 NA 0.01 0.05 Dec-00 0.18 0.02 NA 0.01 NA 0.04 0.03 NA 0.01 0.01 Jan-01 0.25 0.02 NA 0.00 NA 0.04 0.03 NA 0.01 0.01 Feb-01 0.23 0.02 NA 0.01 NA 0.04 0.03 NA 0.01 0.01 Mar-01 0.13 0.03 NA 0.01 NA 0.04 0.03 NA 0.01 0.01 Apr-01 0.19 0.03 NA 0.01 NA 0.06 0.03 NA 0.01 NA May-01 0.15 0.03 NA 0.01 NA 0.04 0.03 NA 0.01 NA Jun-01 0.33 0.03 NA 0.01 NA 0.05 0.03 NA 0.01 NA Jul-01 0.24 0.04 NA 0.00 NA 0.05 0.03 NA 0.00 NA Aug-01 0.29 0.04 NA 0.00 NA 0.03 0.02 NA 0.00 NA Sep-01 0.25 0.04 NA 0.00 NA 0.05 0.03 NA 0.01 NA Oct-01 0.20 0.05 NA 0.00 NA 0.05 0.02 NA 0.01 NA Nov-01 0.24 0.03 NA 0.01 NA 0.04 0.02 NA 0.01 NA Dec-01 0.33 0.02 NA 0.00 NA 0.04 0.03 NA 0.01 NA ``` It is hard to provide a reproduciable data.So,I want is that to form 10 portfolios at the beginning of year 2001 based the 2000. And then form decile portfolio at the beginning of 2002 based on 2001. I want to sort the stocks based on their liquidity. I would really appreciate your help even if you give me an idea how to go about this portfolio formation.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.