You could use the `tabularray` package. Here a short example to get you started:

```
\documentclass{article}

\usepackage{caption}
\usepackage{geometry}
\usepackage{tabularray}
\UseTblrLibrary{booktabs}

\begin{document}

\begin{table}[htbp]
\caption{Impacts of the demand and the supply side of corruption.}
\begin{tblr}{
  colspec={@{}X[2]XXXX@{}}
}
\toprule
Dependent variables: & \SetCell[c=2]{} Tax evasion && \SetCell[c=2]{} Bribing for government contracts &\\
\cmidrule[r]{2-3}\cmidrule[l]{4-5}
& (1) & (2) & (3) & (4)\\
\midrule
Demand side of corruption & 7.494*** & 15.93** & 3.964*** & 4.038*\\
& (1.231) & (7.489) & (0.452) & (2.146)\\
\bottomrule
\end{tblr}

\medskip
Notes: All regressions include firms' individual characteristics described above, country, sector, and year fixed effects. Robust standard errors clustered at the country level in parentheses.
\end{table}

\end{document}
```

![Screenshot 2026-03-03 at 10.17.28.png](/image?hash=f6a82b93646b0bdfc7c141dff0961aa6f573fc380696d26410ca5c40b9c0e25f)

In case you'd like to align all the numbers with respect to their decimal markers, you could use the siunitx package:

```
\documentclass{article}

\usepackage{caption}
\usepackage{geometry}
\usepackage{tabularray}
\UseTblrLibrary{booktabs}
\UseTblrLibrary{siunitx}

\begin{document}

\begin{table}[htbp]
\caption{Impacts of the demand and the supply side of corruption.}
\begin{tblr}{
  colspec={
    @{}
    X*{4}{
      Q[si={
       table-format=-2.5,
       table-alignment=left,
       table-space-text-post=***,
       table-align-text-post=false,
       table-space-text-pre=(,
       table-align-text-pre=false,
       input-open-uncertainty=,
       input-close-uncertainty=,
      }]
    }@{}
  },
  row{1,2} = {guard}
}
\toprule
Dependent variables: & \SetCell[c=2]{} Tax evasion && \SetCell[c=2]{} Bribing for gov.\ contracts &\\
\cmidrule[r]{2-3}\cmidrule[l]{4-5}
& (1) & (2) & (3) & (4)\\
\midrule
Demand side of corruption & 7.494*** & 15.93** & 3.964*** & 4.038*\\
& (1.231) & (7.489) & (0.452) & (2.146)\\
\bottomrule
\end{tblr}

\medskip
Notes: All regressions include firms' individual characteristics described above, country, sector, and year fixed effects. Robust standard errors clustered at the country level in parentheses.
\end{table}

\end{document}
```
![Screenshot 2026-03-03 at 11.48.38.png](/image?hash=fb6932b18c7d1513863b4bc3e05174bc82595007974c5361113f56833bfb0e8d)

Just the first few rows and the notes, using standard `tabular` inside of `threeparttable` and `booktabs` for the rules:

```
\documentclass{article}

\usepackage{booktabs,threeparttable}

\begin{document}
\begin{threeparttable}
\begin{tabular}{*5l}
  \toprule
  Dependent variables: & \multicolumn{2}{l}{tax evasion}
    & \multicolumn{2}{l}{Bribing for government contracts} \\
  \cmidrule(lr){2-3}
  \cmidrule(lr){4-5}
    & (1) & (2) & (3) & (4) \\
    & (OLS) & (2SLS) & (OLS) & (2SLS) \\
  \midrule
  Demand side of corruption & 7.494\tnote{1} & 15.93\tnote{2} & 3.964\tnote{3}
    & 4.038\tnote{1} \\
    & (1.231) & (7.489) & (0.452) & (2.146) \\
  \bottomrule
\end{tabular}
\begin{tablenotes}
  \item All regressions include firms' individual characteristics described
    above, country, sector, and year fixed effects. Robust standard errors
    clustered at the country level in parenthesis.
  \item[1] $p < 0.1$
  \item[2] $p < 0.05$
  \item[3] $p < 0.01$
\end{tablenotes}
\end{threeparttable}
\end{document}
```

![foo.png](/image?hash=f3bb5daa8d34afef9c4ddcf292eb07d045cb328e9ec332a92f07337390137995)