[Tex/LaTex] How to narrow the width of the table notes

tablenotes

The problem is that the whole table notes come as wide as the normal text in my document, which looks bad.

Could you please help me to get the simple solution to this, with the smallest change from the original code?

Thanks so much.


 \documentclass[12pt, a4paper]{article}
\usepackage[english]{babel}
\usepackage{amsmath,amsthm}
\usepackage{amsfonts}
\usepackage{cite}
\usepackage{array}
\newcolumntype{P}[1]{>{\centering\arraybackslash}p{#1}}
\usepackage[pdftex]{graphicx}
\usepackage[top=2.5cm, bottom=2.5cm, left=2.5cm, right=2.5cm]{geometry}
\usepackage{setspace}
\usepackage{float}% for changing positions of figures
\ifx\pdfoutput\undefined
% we are running LaTeX, not pdflatex
\usepackage{graphicx}
\else
% we are running pdflatex, so convert .eps files to .pdf
\usepackage[pdftex]{graphicx}
\usepackage{epstopdf}
\usepackage{booktabs} 
\fi
\usepackage[hang,footnotesize,bf]{caption}%special captions
%%% Labels after captions!!!!
\usepackage{natbib}
\usepackage{color}% define color
\definecolor{MyDarkBlue}{rgb}{0,0.08,0.45}
\usepackage[hyperfootnotes=false]{hyperref}% set color links
\hypersetup{
    colorlinks,%
    citecolor=MyDarkBlue,%
    filecolor=black,%
    linkcolor=MyDarkBlue,%
    urlcolor=black,
}
\usepackage[all]{hypcap}
\usepackage{placeins}
\usepackage{makeidx}
\usepackage{appendix}
\usepackage{multirow}
\usepackage{multicol}
\usepackage{graphicx}
\parindent 0.5cm
\usepackage[flushleft]{threeparttable}
\usepackage{url}
\usepackage{dsfont}
\usepackage[affil-it]{authblk}
% THEOREMS -------------------------------------------------------
\newtheorem{thm}{Theorem}[section]
\newtheorem{cor}[thm]{Corollary}
\newtheorem{lem}[thm]{Lemma}
\newtheorem{prop}[thm]{Proposition}
\newtheorem{aspt}[thm]{Assumption}
\theoremstyle{definition}
\usepackage{tikz} 
\usetikzlibrary{snakes}
\usepackage{verbatim}
\newtheorem{defn}[thm]{Definition}
\theoremstyle{remark}
\newtheorem{rem}[thm]{Remark}


\begin{document}



The reason why I am doing the sequential sorting with the institutional holding, is that the institutional holding is extremely endogenous in this case - fund managers choose to hold some specific stocks for some reason. So it is more reasonable to condition on the same institutional holding level first, and then explore the effect of concentration. 


\begin{table}[htbp]
  \centering
  \caption{Sorting with Institutional Holding level and Concentration level}

 \begin{tablenotes}
\item\small The table reports the equal weighted six-month return with a half-year lag after the assignment of groups for each small-sized stock. At the end to each quarter of year $t$, I sequentially sort all the stocks using market cap (at the end of year $t-1$), the reported institutional holdings and the concentration index. For example, at the end of March in year $t$, the groups are assigned, and the returns are counted from October in year $t$ to March in year $t+1$.
\end{tablenotes}

    \begin{tabular}{ccccccc}
    \toprule
          &       & \multicolumn{5}{c}{insti after 1990} \\
    \midrule
          &       & 1     & 2     & 3     & 4     & 5 \\
    \multirow{5}[0]{*}{concent} & 1     & 6.55\% & 4.87\% & 2.98\% & 3.11\% & 2.73\% \\
          & 2     & 1.11\% & 5.83\% & 3.72\% & 4.07\% & 4.00\% \\
          & 3     & 10.61\% & 4.89\% & 6.76\% & 4.64\% & 4.63\% \\
          & 4     & 21.42\% & 3.26\% & 6.47\% & 4.01\% & 4.91\% \\
          & 5     & 6.81\% & 7.05\% & 5.83\% & 6.21\% & 5.14\% \\
    \bottomrule
    \end{tabular}%


  \label{concent55}
\end{table}


\end{document} 

Best Answer

You should use tablenotes after the tabular so that the width of the tabular is available and the notes follow that width:

\documentclass[]{article}
\usepackage{showframe}
\usepackage{multirow,booktabs}
\usepackage{threeparttable}

\begin{document}
\begin{table}[htbp]
\centering
\begin{threeparttable}
  \caption{Sorting with Institutional Holding level and Informed Change level}
    \begin{tabular}{ccccccc}
    \toprule
          &       & \multicolumn{5}{c}{Institution Holding} \\
    \midrule
          &       &  1     &   2     & 3     & 4     & 5 \\
    \multicolumn{1}{c}{\multirow{5}[0]{*}{Informed}} & 1     & 0.085772 & 0.06621 & 0.053334 & 0.050598 & 0.043458 \\
    \multicolumn{1}{c}{} & 2     & 0.093572 & 0.05932 & 0.023613 & 0.025465 & 0.009882 \\
    \multicolumn{1}{c}{} & 3     & 0.068156 & 0.042421 & 0.067455 & 0.047022 & 0.032287 \\
    \multicolumn{1}{c}{} & 4     & 0.088821 & 0.062215 & 0.064485 & 0.055428 & 0.068727 \\
    \multicolumn{1}{c}{} & 5     & 0.177559 & 0.132563 & 0.108866 & 0.103303 & 0.114933 \\
    \bottomrule

    \end{tabular}%


  \label{informed55}

  \begin{tablenotes}[flushleft]
 \small
\item The table reports the equal weighted six-month return after the assignment of groups for each small-sized stock. At the end to each quarter of year $t$, I sequentially sort all the stocks using market cap (at the end of year $t-1$), the reported institutional holdings and the informed holding position change index. For example, at the end of March in year $t$, the groups are assigned, and the returns are counted from April in year $t$ to September in year $t+1$.
\end{tablenotes}

\end{threeparttable}
\end{table}

\end{document}

enter image description here

Related Question