Guide to Latex - PhD (Generated using AI)

LaTeX for Economics Research: A Practical Guide

For PhD Students Working in Network Economics, Industrial Organization, Game Theory & Digital Piracy

This guide is meant to be plundered for snippets. Every example uses real economic notation and terminology instead of generic foo/bar placeholders, so the code is copy-paste-ready for actual papers, problem sets, and job market presentations. The package choices reflect general best practice across economics PhD programs rather than any single department’s house style — if your program or target journal supplies its own class/style file, defer to that over the generic settings shown here.

Using Overleaf? Everything below works unchanged, since Overleaf runs a full, current TeX Live distribution — every package mentioned in this guide (booktabs, natbib, siunitx, Beamer themes, etc.) is already installed, so there’s nothing to download or configure locally. The two places where Overleaf changes your workflow rather than your code are: (1) you never type compile commands yourself — Overleaf recompiles automatically as you type (or on clicking “Recompile”), and (2) the underlying compiler and bibliography tool are picked from menus rather than flags. Both are called out inline wherever they come up (Section 5.1 on bibliographies, and the boilerplate uploads in Sections 4 and 6).

Document Structure & Classes
Theoretical Economics & Math Formatting
Empirical Economics & Tables
Figures & Graphics
References & Bibliography
Presentations (Beamer)
Useful Online Tools

1. Document Structure & Classes

1.1 Choosing a Document Class

Class	Use Case	Typical Economics Use
`article`	Single, self-contained document with no chapters	Journal papers, working papers, referee reports, problem sets
`report`	Longer documents needing chapters but not a full book	Term papers, PhD qualifying exam write-ups, technical reports for funders
`book`	Multi-part, multi-chapter, front/back matter	PhD dissertation/thesis, edited volumes

Rule of thumb: use article for every stand-alone paper and job market paper draft — this is the structure journals like the RAND Journal of Economics or Journal of Industrial Economics expect, and it’s also what most conference submission systems assume. Reserve report for multi-chapter but non-final documents (e.g., a term paper with several loosely connected sections), and reserve book for the dissertation itself, where you genuinely need front matter (title page, abstract, acknowledgments), numbered chapters, and possibly appendices per chapter. Switching classes mid-project is painful because section-numbering depth and some package behaviors change, so decide the class before you start writing.

\documentclass[12pt, a4paper]{article}   % single papers, JMP drafts
\documentclass[12pt, a4paper]{report}    % term papers with chapters
\documentclass[12pt, a4paper, oneside]{book}  % PhD dissertation

1.2 Essential Preamble for Economics Papers

\documentclass[12pt, a4paper]{article}

% --- Encoding & fonts ---
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{lmodern}          % or newtxtext/newtxmath for a Times-like look

% --- Math ---
\usepackage{amsmath, amssymb, amsfonts, amsthm}
\usepackage{mathtools}        % extends amsmath (e.g. \DeclarePairedDelimiter)
\usepackage{bm}               % bold math symbols, e.g. \bm{\theta}

% --- Tables & floats ---
\usepackage{booktabs}         % professional horizontal rules
\usepackage{dcolumn}          % decimal-aligned columns for regression tables
\usepackage{siunitx}          % alternative/complement to dcolumn
\usepackage{multirow}
\usepackage{array}
\usepackage{threeparttable}   % table notes that respect column width
\usepackage{longtable}        % tables spanning multiple pages

% --- Graphics ---
\usepackage{graphicx}
\usepackage{subcaption}       % subfigures, e.g. network diagrams side-by-side
\usepackage{float}

% --- Bibliography (economics standard: natbib + author-year) ---
\usepackage[authoryear, round]{natbib}

% --- Cross-referencing & hyperlinks ---
\usepackage{hyperref}
\usepackage[capitalise]{cleveref}

% --- Spacing / layout ---
\usepackage{setspace}
\onehalfspacing                % most econ departments want 1.5 or double spacing
\usepackage[margin=1in]{geometry}

% --- Misc economics conveniences ---
\usepackage{enumitem}          % control spacing in lists (assumptions, propositions)
\usepackage{xcolor}            % highlighting derivations/comments during drafting

Why natbib and not biblatex here? See Section 5 — it remains the de facto standard in economics because AEA/Econometrica-style .bst files are built around it.

A note on package load order: hyperref should almost always be loaded last (or second-to-last, just before cleveref), since it redefines many internal LaTeX commands that other packages expect to modify first. Loading it early is one of the most common sources of mysterious compilation errors for beginners. cleveref specifically must be loaded after hyperref.

1.3 Document Hierarchy

\section{Introduction}
\subsection{Related Literature on Platform Competition}
\subsubsection{Two-Sided Markets}
\paragraph{Network effects.} Indirect network effects arise when...

Command	Numbering Depth	Typical Content
`\section{}`	1	Introduction, Model, Data, Results, Conclusion
`\subsection{}`	1.1	Equilibrium Concept, Identification Strategy
`\subsubsection{}`	1.1.1	Specific lemma derivations, robustness sub-checks
`\paragraph{}`	unnumbered, run-in	Short asides, definitions within a subsubsection

1.4 Table of Contents

\tableofcontents
\newpage

For dissertations, also generate lists of tables/figures:

\listoftables
\listoffigures

Note: most journal-style papers (article class, single-spaced or 1.5-spaced, under ~40 pages) don’t include a table of contents at all — reserve \tableofcontents for term papers, dissertation chapters, and long technical reports. Also remember to compile twice after adding or changing headings: the first pass writes the section titles/page numbers to an auxiliary .aux/.toc file, and the second pass reads that file back in to render the ToC correctly.

2. Theoretical Economics & Math Formatting

2.1 The `amsmath` / `amssymb` / `amsfonts` Trio

amsmath: the workhorse package. It gives you multi-line equation environments (align, gather, cases), improved fraction/operator spacing, and \text{} for inserting normal (non-italic) words inside math mode.
amssymb: extra symbols built on top of amsfonts’ fonts — \leq, \geq, \succsim (preference orderings), \subseteq, and dozens of relation/arrow symbols you’ll otherwise be unable to typeset.
amsfonts: the underlying font package that supplies blackboard bold (\mathbb{R} for the real line, \mathbb{N} for a player/agent index set) and Fraktur (\mathfrak{F} for, e.g., a filtration in a stochastic pricing model). In practice amssymb automatically loads amsfonts, so listing both is a common (harmless) redundancy for clarity.

\usepackage{amsmath, amssymb, amsfonts}

% Common economics symbol shortcuts
\newcommand{\R}{\mathbb{R}}
\newcommand{\E}{\mathbb{E}}          % expectation operator
\newcommand{\N}{\mathbb{N}}          % player/agent index set
\newcommand{\argmax}{\operatorname*{arg\,max}}
\newcommand{\argmin}{\operatorname*{arg\,min}}

2.2 Inline vs. Display Math

% Inline math: use for short expressions within a sentence
Firm $i$ chooses price $p_i \in \R_+$ to maximize profit $\pi_i(p_i, p_{-i})$.

% Display math: use for the "headline" equation you want the reader to pause on
\[
    \pi_i(p_i, p_{-i}) = (p_i - c_i)\, D_i(p_i, p_{-i})
\]

Prefer \[ ... \] over the legacy $$ ... $$ — the latter has spacing bugs under amsmath.

2.3 Multi-Line Equations: A Game Theory Optimization Example

A classic Cournot-with-network-effects best-response derivation using align:

\begin{align}
    \max_{q_i \geq 0} \quad & \pi_i(q_i, q_{-i}) = \left[ a - b\left(q_i + \sum_{j \neq i} q_j\right) + \gamma \, x_i \right] q_i - c\, q_i \label{eq:cournot-profit} \\
    \text{F.O.C.:} \quad & a - 2bq_i - b\sum_{j \neq i} q_j + \gamma x_i - c = 0 \label{eq:cournot-foc} \\
    \Rightarrow \quad & q_i^{*} = \frac{a - c + \gamma x_i - b\sum_{j \neq i} q_j}{2b} \label{eq:cournot-br}
\end{align}

Here x_i denotes a network-adoption externality term — natural in a network economics context. Two mechanics worth internalizing: the single & on each line marks the alignment point (here, right after the \quad, so every right-hand side lines up vertically down the page), and \\ ends each line. Every align line gets its own equation number by default — attach \label{} only to lines you actually plan to reference, and consider \notag on a line if you want it unnumbered without switching the whole block to align*. align* suppresses equation numbers entirely when you don’t need to reference any line later:

\begin{align*}
    U_i(\sigma_i, \sigma_{-i}) &= \sum_{s \in S} p(s \mid \sigma) \, u_i(s) \\
    \sigma_i^{*} &\in \argmax_{\sigma_i \in \Delta(S_i)} U_i(\sigma_i, \sigma_{-i}^{*})
\end{align*}

Use \label{} on any line you’ll reference with \eqref{eq:cournot-br} in the text (e.g., “Substituting the best response from \eqref{eq:cournot-br} into…”).

2.4 Integration by Parts (Pedagogical Notation)

For deriving expected consumer surplus or hazard-rate expressions in piracy-deterrence models, here is the classic pedagogical form of integration by parts (the version many intro courses teach before moving to the more common $\int u\,dv = uv - \int v\,du$ shorthand):

\[
    uv = u \int v - \int \left( du \int v \right)
\]

A worked example inside an align block, e.g. integrating a hazard function for time-to-detection of a pirated good:

\begin{align*}
    \int_0^T t \, f(t) \, dt
    &= \Big[ t \cdot F(t) \Big]_0^T - \int_0^T F(t)\, dt \\
    &\quad \text{(using } uv = u\!\int v - \int\!\big(du\!\int v\big) \text{ with } u = t,\ dv = f(t)\,dt \text{)}
\end{align*}

2.5 Theorems, Lemmas, and Proofs

Set these up once in the preamble with amsthm:

\usepackage{amsthm}

\theoremstyle{plain}
\newtheorem{theorem}{Theorem}[section]
\newtheorem{lemma}[theorem]{Lemma}
\newtheorem{proposition}[theorem]{Proposition}
\newtheorem{corollary}[theorem]{Corollary}

\theoremstyle{definition}
\newtheorem{definition}[theorem]{Definition}
\newtheorem{assumption}[theorem]{Assumption}

\theoremstyle{remark}
\newtheorem{remark}[theorem]{Remark}

A few mechanics worth understanding rather than just copying: \theoremstyle{plain} renders the body in italics (for results), {definition} renders it upright (for definitions/assumptions, which readers shouldn’t mentally parse as “claims”), and {remark} is similar but typically de-emphasized. The [section] argument on \newtheorem{theorem}{Theorem}[section] resets the theorem counter every time a new \section starts, so you get “Theorem 2.1” instead of a single running count through the whole paper. The [theorem] argument on the later \newtheorem calls (e.g., \newtheorem{lemma}[theorem]{Lemma}) makes lemmas, propositions, and corollaries share the theorem counter, so numbering runs Theorem 2.1, Lemma 2.2, Proposition 2.3 in document order — which is what most economics journals expect, rather than separate counters for each type.

Usage in the body:

\begin{assumption}[Homogeneous Piracy Cost]
    \label{as:homogeneous-cost}
    All consumers face an identical marginal cost of piracy, $\kappa > 0$, independent of network size $n$.
\end{assumption}

\begin{proposition}
    \label{prop:unique-equilibrium}
    Under Assumption~\ref{as:homogeneous-cost}, there exists a unique symmetric Bayesian Nash equilibrium in which the platform sets price $p^{*} = \dfrac{a + c + \gamma \bar{x}}{2}$.
\end{proposition}

\begin{proof}
    Suppose, for contradiction, that two distinct symmetric equilibria $p_1^{*} \neq p_2^{*}$ exist. By the strict concavity of $\pi_i(\cdot)$ established in \eqref{eq:cournot-profit}, the best-response correspondence is single-valued, a contradiction. \qedhere
\end{proof}

\qedhere places the ∎ symbol immediately at the end of the last line rather than on a new line — useful when a proof ends mid-equation.

3. Empirical Economics & Tables

3.1 The Philosophy of Publication-Quality Tables

Economics journals (AER, QJE, Econometrica) follow a shared house style:

No vertical rules. Vertical lines clutter the table and are considered amateurish.
Minimal horizontal rules — typically three: top, one after the header, and bottom. This is exactly what booktabs enforces.
Decimal alignment in numeric columns so that standard errors line up under coefficients.
Notes below the table, not above, explaining significance stars, standard error clustering, and sample restrictions.

Never use the default \hline-heavy tables from tabular. Always load booktabs.

The underlying logic: a table’s job is to let the eye compare numbers quickly, and every extra ruled line is visual noise competing with that goal. booktabs enforces sensible defaults — its rules are drawn at different weights and with built-in spacing, which is why \toprule/\bottomrule look noticeably better than a hand-drawn \hline even though both nominally “just draw a line.”

3.2 `booktabs` Basics

\usepackage{booktabs}

\begin{table}[htbp]
    \centering
    \caption{Summary Statistics: Platform-Level Network Data}
    \label{tab:summary-stats}
    \begin{tabular}{lccc}
        \toprule
        Variable & Mean & Std. Dev. & N \\
        \midrule
        Active users (000s)     & 542.3  & 118.7 & 1{,}204 \\
        Piracy incidence rate   & 0.184  & 0.091 & 1{,}204 \\
        Platform price (\$)     & 9.99   & 2.14  & 1{,}204 \\
        \bottomrule
    \end{tabular}
\end{table}

Notice 1{,}204 rather than plain 1,204: LaTeX by default adds a small extra space after a comma inside math-adjacent contexts, so wrapping the comma in braces ({,}) suppresses that spacing and keeps the thousands separator looking like ordinary text.

3.3 Regression Tables with `dcolumn` / `siunitx`

Option A — dcolumn (classic, works well with estout/esttab output from Stata):

Common pitfall: the D{.}{.}{3.3} column type aligns strictly on the decimal point, so a cell like 0.412^{***} will actually misalign slightly because dcolumn doesn’t know to ignore the superscript stars when computing alignment. In practice, most people either (a) let Stata’s esttab, star produce the whole table body pre-formatted and just wrap it in the d column type, accepting the tiny visual offset, or (b) switch that specific column to plain c alignment when stars are present and rely on booktabs rules alone for tidiness. Don’t spend hours fighting pixel-perfect alignment here — reviewers care about correctness, not sub-millimeter column alignment.

\usepackage{dcolumn}
\usepackage{threeparttable}
\newcolumntype{d}[1]{D{.}{.}{#1}}

\begin{table}[htbp]
    \centering
    \caption{Effect of Network Size on Piracy Take-Up}
    \label{tab:reg-main}
    \begin{threeparttable}
    \begin{tabular}{l d{3.3} d{3.3} d{3.3}}
        \toprule
        & \multicolumn{1}{c}{(1)} & \multicolumn{1}{c}{(2)} & \multicolumn{1}{c}{(3)} \\
        & \multicolumn{1}{c}{OLS} & \multicolumn{1}{c}{IV} & \multicolumn{1}{c}{Poisson} \\
        \midrule
        Network size ($\ln n_i$)  & 0.412^{***} & 0.389^{***} & 0.401^{***} \\
                                   & (0.087)     & (0.102)     & (0.093) \\
        Price ($p_i$)              & -0.215^{**} & -0.198^{**} & -0.207^{**} \\
                                   & (0.094)     & (0.088)     & (0.091) \\
        \midrule
        Firm FE                   & \multicolumn{1}{c}{Yes} & \multicolumn{1}{c}{Yes} & \multicolumn{1}{c}{Yes} \\
        Year FE                   & \multicolumn{1}{c}{Yes} & \multicolumn{1}{c}{Yes} & \multicolumn{1}{c}{Yes} \\
        Observations               & \multicolumn{1}{c}{1{,}204} & \multicolumn{1}{c}{1{,}204} & \multicolumn{1}{c}{1{,}204} \\
        \bottomrule
    \end{tabular}
    \begin{tablenotes}
        \small
        \item \textit{Notes:} Standard errors clustered at the platform level in parentheses. $^{*}p<0.10$, $^{**}p<0.05$, $^{***}p<0.01$.
    \end{tablenotes}
    \end{threeparttable}
\end{table}

Why threeparttable: the tablenotes environment used for the notes line below the table is provided by the threeparttable package, and only works when the whole tabular is nested inside \begin{threeparttable}...\end{threeparttable}. Its main benefit over just typing the notes as ordinary text below the table is that the note text automatically wraps to the same width as the table itself, rather than running the full width of the page — which looks wrong for a narrow table.

Option B — siunitx (more modern, more configurable, handles negative signs and scientific notation gracefully):

\usepackage{siunitx}
\sisetup{table-format=1.3, detect-weight=true, detect-family=true}

\begin{tabular}{l S S S}
    \toprule
    {} & {OLS} & {IV} & {Poisson} \\
    \midrule
    $\ln n_i$ & 0.412 & 0.389 & 0.401 \\
    {}        & (0.087) & (0.102) & (0.093) \\
    \bottomrule
\end{tabular}

Two things to note in the siunitx snippet: the S column type (from siunitx, distinct from plain S-as-in-“strong” text) automatically detects and aligns on the decimal point and correctly handles non-numeric content — but any cell that isn’t a plain number (like the header text OLS or an empty cell) must be wrapped in braces {...} so siunitx treats it as literal text rather than attempting (and failing) to parse it as a number. table-format=1.3 tells siunitx to reserve space for exactly 1 digit before the decimal and 3 after, so all columns align even before any content is typeset — adjust this to match your actual coefficient magnitudes (e.g., 2.3 if coefficients can run into the tens).

3.4 `\multicolumn` and `\multirow`

\usepackage{multirow}

\begin{table}[htbp]
    \centering
    \caption{Piracy Deterrence by Enforcement Regime and Market}
    \label{tab:multirow-example}
    \begin{tabular}{llcc}
        \toprule
        \multirow{2}{*}{Enforcement Regime} & \multirow{2}{*}{Market} & \multicolumn{2}{c}{Piracy Rate} \\
        \cmidrule(lr){3-4}
        & & Pre-Policy & Post-Policy \\
        \midrule
        \multirow{2}{*}{Strict (DMCA-style)} & Streaming & 0.24 & 0.11 \\
                                              & Software  & 0.31 & 0.14 \\
        \multirow{2}{*}{Lenient}             & Streaming & 0.29 & 0.26 \\
                                              & Software  & 0.38 & 0.35 \\
        \bottomrule
    \end{tabular}
\end{table}

\cmidrule(lr){3-4} draws a partial rule under columns 3–4 only, with slight left/right trimming — the booktabs-approved way to group sub-headers (never use \cline).

4. Figures & Graphics

4.1 Inserting Graphics with `graphicx`

\usepackage{graphicx}

\begin{figure}[htbp]
    \centering
    \includegraphics[width=0.75\textwidth]{figures/network_topology.pdf}
    \caption{Simulated Network Topology Under Preferential Attachment ($\gamma = 2.1$)}
    \label{fig:network-topology}
\end{figure}

Always compile figures as vector PDF (from R’s ggsave(..., device = "pdf") or Stata’s graph export ..., as(pdf)), never PNG/JPEG for line plots or network diagrams — this keeps them crisp at any zoom.
Reference figures in text with Figure~\ref{fig:network-topology} or, with cleveref, simply \cref{fig:network-topology}. The ~ (a non-breaking space) prevents “Figure” and its number from being split across a line break.
On Overleaf: create a folder named figures in the file tree (right-click → New Folder) and upload your PDF/image files into it, so the relative path figures/network_topology.pdf in \includegraphics resolves correctly — Overleaf’s file tree mirrors the relative paths you write in the .tex source exactly.

4.2 Side-by-Side Subfigures (e.g., comparing network structures)

\usepackage{subcaption}

\begin{figure}[htbp]
    \centering
    \begin{subfigure}[b]{0.48\textwidth}
        \includegraphics[width=\textwidth]{figures/network_core.pdf}
        \caption{Core-periphery structure}
        \label{fig:core-periphery}
    \end{subfigure}
    \hfill
    \begin{subfigure}[b]{0.48\textwidth}
        \includegraphics[width=\textwidth]{figures/network_random.pdf}
        \caption{Random graph benchmark}
        \label{fig:random-graph}
    \end{subfigure}
    \caption{Observed vs. Simulated Network Structures}
    \label{fig:network-comparison}
\end{figure}

\hfill between the two subfigure blocks pushes them apart to fill the available horizontal space evenly — without it, the two subfigures would sit flush against each other with no gap. Each subfigure also gets its own \label, so you can refer to \cref{fig:core-periphery} specifically or \cref{fig:network-comparison} for the pair as a whole.

4.3 Float Positioning Options

Specifier	Meaning	When to Use
`h`	“here”, approximately	Small figures that should stay near the surrounding text
`t`	top of page	Default preference for most journal figures
`b`	bottom of page	Good for wide tables that would otherwise break text flow
`p`	dedicated float page	Large figures/tables (e.g., a full-page regression table)
`htbp`	try all, in order	The safe, most commonly used combination

\begin{figure}[htbp]   % LaTeX will try h, then t, then b, then p
    ...
\end{figure}

Tip: If LaTeX ignores your [h] and pushes the float away, load \usepackage{float} and use [H] (capital H) to force exact placement — but use sparingly, as it can create ugly whitespace.

5. References & Bibliography

5.1 The BibTeX Workflow

Maintain a references.bib file with entries like:

@article{katz1985network,
    author  = {Katz, Michael L. and Shapiro, Carl},
    title   = {Network Externalities, Competition, and Compatibility},
    journal = {American Economic Review},
    year    = {1985},
    volume  = {75},
    number  = {3},
    pages   = {424--440}
}

@article{peitz2006piracy,
    author  = {Peitz, Martin and Waelbroeck, Patrick},
    title   = {Why the Music Industry May Gain from Free Downloading},
    journal = {International Journal of Industrial Organization},
    year    = {2006},
    volume  = {24},
    number  = {5},
    pages   = {907--913}
}

Compile.

On Overleaf: upload references.bib via the file-tree “Upload” button (or create it directly in the project), keep \bibliography{references} pointing at it, and just click Recompile — Overleaf detects the \bibliography{} command and automatically runs the BibTeX pass for you behind the scenes, no manual command needed. The one setting worth checking is under the project’s Menu → Settings → Bibliography Processor: it should be set to bibtex (the default and correct choice for natbib-based papers); only switch it to biber if you deliberately move to biblatex in Section 5.2 below. If citations show up as [?] or “undefined,” a plain Recompile a second time usually resolves it, since — same as local compiling — the citation list needs one extra pass to resolve after the .bib file changes.

Compiling locally (outside Overleaf), run in this order, whether by hand or via latexmk:

pdflatex main.tex
bibtex main
pdflatex main.tex
pdflatex main.tex

5.2 `natbib` vs. `biblatex`

	`natbib`	`biblatex`
Standard in economics?	Yes — matches AEA, Econometrica, RAND house styles	Rare in econ; common in humanities/sciences
Backend	BibTeX (or `bibtex8`)	Biber (more powerful, slower)
Citation commands	`\citet`, `\citep`, `\citeauthor`, `\citeyear`	`\textcite`, `\parencite`, similar but different names
Journal `.bst` file support	Extensive (most econ journals supply a `.bst`)	Poor — journals rarely provide `.bbx`/`.cbx` files
Recommendation	Use this for economics papers	Use only if a specific journal/template mandates it

Preamble setup for natbib:

\usepackage[authoryear, round]{natbib}
\bibliographystyle{aer}     % or apalike, econometrica, chicago
...
\bibliography{references}

5.3 `\citet` vs. `\citep`

% Textual citation — reads as part of the sentence
\citet{katz1985network} show that network externalities can generate multiple equilibria.
% Output: "Katz and Shapiro (1985) show that..."

% Parenthetical citation — for supporting evidence at the end of a clause
Network externalities can generate multiple equilibria \citep{katz1985network}.
% Output: "...multiple equilibria (Katz and Shapiro, 1985)."

% Multiple sources
The piracy-deterrence literature is well established \citep{peitz2006piracy, katz1985network}.

% Citing with a specific page/note
\citet[p.~430]{katz1985network} formalize this as a coordination game.

6. Presentations (Beamer)

6.1 Complete Boilerplate Template

\documentclass[aspectratio=169]{beamer}

\usetheme{Madrid}
\usecolortheme{dolphin}

\usepackage[utf8]{inputenc}
\usepackage{amsmath, amssymb}
\usepackage{graphicx}
\usepackage[authoryear, round]{natbib}

\title[Network Effects \& Piracy]{Network Effects and the Economics of Digital Piracy}
\subtitle{Job Market Paper}
\author[A.\ Researcher]{Your Name}
\institute[IGIDR]{Indira Gandhi Institute of Development Research}
\date{\today}

\begin{document}

\frame{\titlepage}

\begin{frame}{Outline}
    \tableofcontents
\end{frame}

\section{Motivation}
\begin{frame}{Motivation}
    \begin{itemize}
        \item Digital platforms exhibit strong network externalities.
        \pause
        \item Piracy interacts with network size in non-obvious ways.
        \pause
        \item This paper models the joint determination of price, network adoption, and piracy incidence.
    \end{itemize}
\end{frame}

\section{Model}
\begin{frame}{Setup}
    Firm $i$ sets price $p_i$ to maximize
    \[
        \pi_i(p_i) = (p_i - c)\, D_i(p_i, x_i) - \kappa \, \rho(p_i)
    \]
    where $x_i$ is network size and $\rho(p_i)$ is the piracy rate.
\end{frame}

\section{Results}
\begin{frame}{Main Result}
    \begin{block}{Proposition}
        Under mild regularity conditions, equilibrium price is decreasing in the piracy-deterrence cost $\kappa$.
    \end{block}
\end{frame}

\section{Conclusion}
\begin{frame}{Conclusion}
    \begin{itemize}
        \item Network effects amplify the welfare cost of piracy enforcement.
        \item Policy implication: uniform DMCA-style enforcement may be suboptimal across markets of different network density.
    \end{itemize}
\end{frame}

\begin{frame}{References}
    \bibliographystyle{apalike}
    \bibliography{references}
\end{frame}

\end{document}

You can use \citet{}/\citep{} inside any frame body exactly as in a paper (Section 5.3). On Overleaf, just upload references.bib alongside your .tex file and hit Recompile — Overleaf runs the necessary BibTeX pass automatically, same as for a regular paper (Section 5.1). Compiling locally, the same pdflatex → bibtex → pdflatex → pdflatex sequence applies — Beamer decks with citations need the extra passes just like papers do.

6.2 Best Professional Themes for Academic Economics

Presentation (structural) themes:

Theme	Character	Good For
`Madrid`	Clean header bar with section navigation dots	General seminar talks
`Boadilla`	Minimal, sidebar-free, understated	Job talks — very safe, conservative choice
`metropolis`	Modern flat design (requires separate package `beamertheme-metropolis`)	Talks aiming for a contemporary look
`CambridgeUS`	Traditional, high information density	Long technical seminars with many results

Color themes:

Color Theme	Palette
`dolphin`	Muted blue/gray — professional, low-key
`seahorse`	Soft gray-green
`whale`	Deep blue, high contrast
`default`	Beamer blue — safe but generic

\usetheme{Boadilla}
\usecolortheme{seahorse}

For metropolis, install via your TeX distribution’s package manager and use:

\usetheme{metropolis}
\usepackage{fontawesome5} % metropolis pairs well with icon-based bullets

6.3 Theme Galleries (Browse Before You Commit)

Beamer theme matrix (official-style gallery): https://hartwork.org/beamer-theme-matrix/
Overleaf Beamer theme gallery: https://www.overleaf.com/gallery/tagged/presentation
metropolis theme repository and preview: https://github.com/matze/mtheme

6.4 Overlays, `\pause`, and Frame Structure

\begin{frame}{Sequential Reveal of Assumptions}
    \begin{itemize}
        \item Assumption 1: Consumers are risk-neutral. \pause
        \item Assumption 2: The platform has full commitment power. \pause
        \item Assumption 3: Piracy technology is common knowledge.
    \end{itemize}
\end{frame}

% Finer-grained control with <> overlay specifications
\begin{frame}{Overlay Specification Example}
    \begin{itemize}
        \item<1-> This appears from slide 1 onward.
        \item<2-> This appears from slide 2 onward.
        \item<3> This appears only on slide 3.
        \item<only@2> This appears \emph{only} while on slide 2, then disappears.
    \end{itemize}
\end{frame}

\pause is the simplest tool: it splits everything before it from everything after it into separate reveal steps, and you can stack several in one frame. The <...> overlay syntax gives finer control when \pause isn’t expressive enough: <2-> means “visible from slide 2 onward and stays visible,” <3> means “visible only on slide 3,” and <only@2> additionally removes the item from the layout (rather than just making it invisible) outside slide 2 — useful when you want a later bullet to shift upward to fill the gap. Mixing \pause and <> overlays in the same list can produce confusing numbering, so pick one style per frame.

6.5 Title Page Fields

The title page is generated automatically from preamble metadata via \frame{\titlepage} (or \maketitle inside a frame) — see the boilerplate above for the full set of \title, \subtitle, \author, \institute, \date fields.

7. Useful Online Tools

A handful of browser-based tools are worth keeping bookmarked — they save real time on the fiddly, non-theoretical parts of writing a paper (retyping a table from a scanned source, tracking down a symbol’s command name, drawing a diagram) without needing anything installed locally.

Tool	What It Does	Link
SimpleTex	Converts a photo or screenshot of printed or handwritten math into LaTeX source — point it at an equation in a PDF, a textbook, or a phone photo of a whiteboard and it returns the corresponding `\[ ... \]` code	`https://simpletex.cn/`
Tables Generator	A spreadsheet-style visual editor for building LaTeX tables — paste data in from Excel/CSV or type directly into the grid, and it exports `booktabs`-ready `tabular` code	`https://www.tablesgenerator.com/`
TikZ Editor	A live, in-browser editor for TikZ diagrams (network graphs, game trees, causal diagrams) with instant rendering, useful for prototyping a diagram before pasting the final `tikzpicture` code into your paper	`https://tikz.dev/editor/`
LaTeX-OCR (pix2tex)	Converts an image of a single mathematical expression into LaTeX code; the same tool also has a “draw” mode for sketching an individual unfamiliar symbol to get its LaTeX command name, and can be pointed at a photo of a page of handwritten notes to extract the math into LaTeX	`https://lukas-blecher.github.io/LaTeX-OCR/`

Note on LaTeX-OCR: all three math-recognition uses above (expression-from-image, symbol-from-drawing, and notes-from-photo) point to the same tool — it handles all three from the one interface, so there’s no need to bookmark separate pages for each. For a dedicated symbol-lookup tool with a larger hand-drawn symbol database, Detexify (https://detexify.kirelabs.org/classify.html) is a common alternative worth having alongside it.

None of these replace understanding the underlying LaTeX — treat their output as a first draft to check and clean up (especially OCR’d tables and equations, which can silently misread a subscript or a decimal point) rather than something to paste in and forget.

7.1 SimpleTex — Image-to-LaTeX Conversion

Workflow:

Upload an image (drag-and-drop, paste from clipboard, or a file picker) containing one or more equations — a screenshot of a PDF, a scanned textbook page, or a phone photo.
SimpleTex detects each equation region automatically and runs recognition on it; a live preview renders the recognized math next to the source image so you can visually check it against the original before trusting it.
Click a recognized equation to get its LaTeX in a few interchangeable forms — inline ( $...$ ), display (\[...\]), and sometimes align-ready — plus a one-click Copy button. There’s no command syntax to learn on the tool’s own interface; you’re copying finished LaTeX, not typing anything.
Paste the copied code directly into your .tex source inside whichever math environment you’re using (Section 2.2–2.3 above).

What to double-check: OCR on math is good but not perfect on dense subscripts, nested fractions, or unusual operators (e.g., a piracy-deterrence hazard function with several nested sums) — always compile and visually compare the rendered output against the source image rather than assuming a one-shot correct transcription, especially for anything going into a \label{}led equation you’ll cite later.

7.2 Tables Generator — Visual Table Builder

Workflow:

Choose “LaTeX Tables” from the tool’s format menu (it also supports Markdown, HTML, and other outputs — make sure LaTeX is selected before you start, since switching later can reset formatting).
Build the table two ways: either type directly into the spreadsheet-style grid (right-click a cell for Insert row/column, Merge cells, etc.), or paste data copied from Excel/Google Sheets/a CSV, which auto-populates the grid.
Use the toolbar to set borders and alignment. For an economics-journal look, turn OFF vertical borders and keep only the outer top/bottom borders plus one line under the header row — this mirrors the booktabs \toprule/\midrule/\bottomrule convention from Section 3.1, though the raw export uses plain \hline rather than booktabs commands.
Click Generate (or the LaTeX tab, depending on the tool’s current layout) to get the \begin{tabular}{...}...\end{tabular} code, then Copy to clipboard.
Paste the output into your document and manually swap \hline for \toprule/\midrule/\bottomrule (Section 3.2) if you want true booktabs styling — the generator’s raw export doesn’t call booktabs on its own, so this one substitution is worth doing by hand every time.

What to double-check: the generator infers a column-alignment string (e.g., {lccc}) from your grid, but for a regression table with decimal-aligned coefficients you’ll still want to swap the relevant columns to d{3.3} or S (Section 3.3) after pasting — the tool has no concept of dcolumn/siunitx decimal alignment.

7.3 TikZ Editor — Live Diagram Prototyping

Workflow:

The editor opens with an empty document.tex-style pane on one side and a rendered preview on the other. Write ordinary TikZ commands directly inside the implicit tikzpicture environment — e.g. \node[circle, draw] (A) at (0,0) {$i$}; for a node representing player/firm $i$, or \draw[->] (A) -- (B); for a directed link in a network diagram.
The preview recompiles automatically (or via a Compile/Run button, depending on the editor) as you type, so you can iterate on node placement and edge routing interactively rather than recompiling your whole paper on every tweak.
Common commands worth knowing going in, since the editor assumes TikZ syntax rather than teaching it: \node (places a labelled point or shape), \draw (draws a line/arrow/curve between coordinates or named nodes), \foreach (loops, useful for generating a ring of $n$ players in a network diagram without typing each node by hand), and the ->/<-> arrow-tip options inside \draw[...] for directed vs. undirected links.
Once the diagram looks right, copy the full code block and paste it into your paper inside \begin{figure}...\end{figure} (Section 4.1), preceded by \usepackage{tikz} in your preamble.

What to double-check: the editor’s default preamble/packages may differ slightly from your paper’s — if you used a TikZ library in the editor (e.g. \usetikzlibrary{arrows.meta, positioning} for the positioning shorthand right=of A), remember to add the matching \usetikzlibrary{...} line to your actual document preamble, or the pasted code won’t compile there.

7.4 LaTeX-OCR (pix2tex) — Expressions, Symbols, and Notes

Workflow (image-to-LaTeX):

Upload or paste an image containing a single mathematical expression (crop tightly around just the equation for best accuracy — including surrounding text or multiple equations in one image degrades recognition).
The model returns LaTeX source in a text box, with a rendered preview above it so you can visually confirm the match before copying.
Copy the code and paste it into your .tex source inside the appropriate math delimiters ( $...$ or \[...\], Section 2.2).

Workflow (draw-a-symbol mode):

Switch to the tool’s drawing/sketch pad (a blank canvas you draw on with mouse or trackpad).
Sketch the unfamiliar symbol as closely as you can to how it appears in print — e.g. a preference-ordering symbol like $\succsim$ that you can’t otherwise search for by name.
The tool returns the closest-matching LaTeX command (e.g. \succsim) which you then add to your document, typically as one of the amssymb relation symbols from Section 2.1.

Workflow (handwritten notes):

Photograph or scan a page of handwritten derivations.
Upload the full page; the tool segments and recognizes each equation region much like SimpleTex does (Section 7.1).
Review each recognized line against your handwriting before pasting — handwritten OCR has a meaningfully higher error rate than printed-text OCR, especially for ambiguous handwriting (a hastily written $\gamma$ vs. $\chi$, or a subscript that could be $i$ or $j$).

What to double-check: across all three modes, treat the output as a suggestion rather than ground truth — re-derive or at least sanity-check any recognized equation against your own notes before it goes into a paper, since a silently misread sign or exponent in, say, the Cournot best-response derivation (Section 2.3) can propagate through several downstream results before anyone notices.

Quick-Reference Cheat Sheet

% On Overleaf: just click Recompile — it handles the bibliography
% pass automatically whenever \bibliography{} is present.

% Compiling locally instead:
pdflatex main && bibtex main && pdflatex main && pdflatex main   % paper with bibliography
pdflatex slides.tex                                              % Beamer deck (run twice if using \tableofcontents or citations)

Task	Package
Author-year citations	`natbib`
Decimal-aligned regression tables	`dcolumn` or `siunitx`
Rule-based professional tables	`booktabs`
Cross-references (`\cref`)	`cleveref`
Table notes that wrap correctly	`threeparttable`
Presentations	`beamer`

LaTeX for Economics Research: A Practical Guide

For PhD Students Working in Network Economics, Industrial Organization, Game Theory & Digital Piracy

Table of Contents

1. Document Structure & Classes

1.1 Choosing a Document Class

1.2 Essential Preamble for Economics Papers

1.3 Document Hierarchy

1.4 Table of Contents

2. Theoretical Economics & Math Formatting

2.1 The amsmath / amssymb / amsfonts Trio

2.2 Inline vs. Display Math

2.3 Multi-Line Equations: A Game Theory Optimization Example

2.4 Integration by Parts (Pedagogical Notation)

2.5 Theorems, Lemmas, and Proofs

3. Empirical Economics & Tables

3.1 The Philosophy of Publication-Quality Tables

3.2 booktabs Basics

3.3 Regression Tables with dcolumn / siunitx

3.4 \multicolumn and \multirow

4. Figures & Graphics

4.1 Inserting Graphics with graphicx

4.2 Side-by-Side Subfigures (e.g., comparing network structures)

4.3 Float Positioning Options

5. References & Bibliography

5.1 The BibTeX Workflow

5.2 natbib vs. biblatex

5.3 \citet vs. \citep

6. Presentations (Beamer)

6.1 Complete Boilerplate Template

6.2 Best Professional Themes for Academic Economics

6.3 Theme Galleries (Browse Before You Commit)

6.4 Overlays, \pause, and Frame Structure

6.5 Title Page Fields

7. Useful Online Tools

7.1 SimpleTex — Image-to-LaTeX Conversion

7.2 Tables Generator — Visual Table Builder

7.3 TikZ Editor — Live Diagram Prototyping

7.4 LaTeX-OCR (pix2tex) — Expressions, Symbols, and Notes

Quick-Reference Cheat Sheet

Further Reading

2.1 The `amsmath` / `amssymb` / `amsfonts` Trio

3.2 `booktabs` Basics

3.3 Regression Tables with `dcolumn` / `siunitx`

3.4 `\multicolumn` and `\multirow`

4.1 Inserting Graphics with `graphicx`

5.2 `natbib` vs. `biblatex`

5.3 `\citet` vs. `\citep`

6.4 Overlays, `\pause`, and Frame Structure