Configure your agent for economic research with ArXiv search, paper summarization, and research journal management

Economics Research Assistant Agent

Transform your general-purpose agent into a specialized economics research assistant that searches papers, summarizes findings, and maintains research journals.

Use Case Overview

Economics researchers need agents that:

Search academic databases (ArXiv, JSTOR, RePEc)
Summarize economic papers and extract key findings
Track research topics and literature reviews
Analyze economic data and trends
Maintain research journals and notes

Domain-Specific System Prompt

Research Assistant Persona

system_prompt = """You are an economics research assistant specializing in:
- Academic paper search and summarization
- Economic theory and empirical research
- Quantitative analysis and econometrics
- Literature review synthesis

When users ask research questions:
1. Search ArXiv/RePEc for relevant papers
2. Summarize key findings and methodology
3. Identify research gaps and future directions
4. Maintain research journal with citations
"""

Persona Design: System prompt shapes agent expertise, response style, and tool usage patterns. Economics agents prioritize academic rigor and citation accuracy.

Economics-Specific Tools

Tool 1: ArXiv Paper Search

Define ArXiv Tool

{
    "name": "search_arxiv",
    "description": "Search ArXiv for economics papers by keyword, author, or topic",
    "input_schema": {
        "properties": {
            "query": {"type": "string"},
            "category": {"type": "string", "default": "econ"}
        }
    }
}

Implement Search Function

import arxiv

def search_arxiv(query, category="econ"):
    search = arxiv.Search(query=f"{query} cat:{category}", max_results=5)
    return [{"title": r.title, "summary": r.summary} for r in search.results()]

Test Agent Usage

You: Find papers on platform economics
Agent: [Uses search_arxiv("platform economics")]
Agent: Found 5 papers:
1. "Two-Sided Platforms..." - Rochet & Tirole
2. "Platform Competition..." - Armstrong
...

Tool 2: Paper Summarization

Define Summary Tool

{
    "name": "summarize_paper",
    "description": "Extract methodology, findings, and contributions from paper",
    "input_schema": {
        "properties": {
            "paper_text": {"type": "string"},
            "focus": {"type": "string", "enum": ["methodology", "findings", "theory"]}
        }
    }
}

Implement Summarization

def summarize_paper(paper_text, focus="findings"):
    # Use structured extraction prompt
    prompt = f"Extract {focus} from this paper: {paper_text}"
    summary = client.messages.create(...)
    return summary

Tool 3: Research Journal Management

Define Journal Tool

{
    "name": "update_research_journal",
    "description": "Add entry to research journal with citations",
    "input_schema": {
        "properties": {
            "topic": {"type": "string"},
            "entry": {"type": "string"},
            "citations": {"type": "array", "items": {"type": "string"}}
        }
    }
}

Implement Journal Storage

def update_journal(topic, entry, citations):
    journal = load_journal()
    journal[topic] = {
        "entry": entry,
        "citations": citations,
        "date": datetime.now()
    }
    save_journal(journal)

Economics Research Workflows

Workflow 1: Literature Review

Goal: Find relevant papers on research topic

You: I'm researching network effects in platform markets
Agent: [search_arxiv("network effects platform markets")]

Agent: Found 8 relevant papers. Key themes:
- Direct vs. indirect network effects
- Multi-sided platform competition
- Critical mass and tipping points

Agent Actions:

Searches academic databases
Filters by relevance and citations
Groups by research themes

Goal: Extract key findings from papers

You: Summarize the Rochet-Tirole platform paper
Agent: [Retrieves paper, uses summarize_paper]

Agent: Rochet & Tirole (2003) key findings:
- Two-sided markets require different pricing strategies
- Platform profit = sum of both sides' surplus
- Network effects create coordination challenges

Agent Actions:

Retrieves paper text
Extracts methodology and findings
Identifies contributions to literature

Goal: Synthesize findings into research notes

You: Add this to my platform economics journal
Agent: [update_research_journal(...)]

Agent: Added to journal:
Topic: Platform Economics
Entry: Network effects literature review (8 papers)
Citations: Rochet & Tirole (2003), Armstrong (2006)...

Agent Actions:

Organizes findings by topic
Maintains citation format
Tracks research progression

Workflow 2: Empirical Research Support

Data Discovery

You: Find datasets for platform competition analysis
Agent: [Uses web_search + data repository tools]
Agent: Relevant datasets:
1. Uber ride-sharing data (Kaggle)
2. Platform pricing data (FTC)
3. Network metrics (Stanford Network Analysis)

Methodology Guidance

You: What econometric methods for platform competition?
Agent: For platform competition, common methods:
- Difference-in-differences for policy analysis
- IV estimation for endogeneity
- Structural models for demand estimation

Citation Management

You: Format citations for my platform paper
Agent: [Reads research journal, formats citations]
Agent: References (APA format):
Armstrong, M. (2006). Competition in two-sided markets...
Rochet, J. C., & Tirole, J. (2003). Platform competition...

Configuration Examples

Conservative Research Agent

system_prompt = """You are a rigorous economics researcher.
- Only cite peer-reviewed sources
- Flag methodological limitations
- Distinguish theory from empirical findings
- Use formal economic terminology"""

Behavior:

Prioritizes academic rigor
Questions data quality
Highlights research gaps

Exploratory Research Agent

system_prompt = """You are a creative economics research assistant.
- Explore unconventional connections
- Suggest novel research directions
- Identify emerging topics
- Connect across disciplines"""

Behavior:

Broader paper search
Interdisciplinary connections
Hypothesis generation

Testing Your Economics Agent

Test Paper Search

You: Papers on algorithmic pricing
Agent: [Searches ArXiv + RePEc]
Agent: 6 papers found. Recent work focuses on:
- Algorithmic collusion (Calvano et al. 2020)
- Price discrimination (Chen & Zhang 2022)
- Platform algorithms (Hagiu & Wright 2023)

Verify: Relevant papers, recent publications, correct citations

Test Summarization

You: Summarize the key mechanism in algorithmic collusion
Agent: [Uses paper summary tool]
Agent: Algorithmic collusion mechanism:
1. RL agents learn pricing strategies
2. Converge to supra-competitive prices
3. No explicit communication needed

Verify: Accurate extraction, clear explanation, key insights

Test Journal Management

You: Save this to algorithmic pricing research
Agent: [Updates journal with citations]
Agent: Research journal updated:
- Topic: Algorithmic Pricing
- 6 papers reviewed
- Key finding: Collusion without communication

Verify: Correct storage, citation format, topic organization

Success Criteria

Your economics research agent should:

✅ Accurate Search: Find relevant papers by keyword/author
✅ Quality Summarization: Extract methodology and findings correctly
✅ Citation Management: Maintain proper academic citations
✅ Research Journal: Organize findings by topic
✅ Econometric Knowledge: Suggest appropriate methods

Integration with Research Workflow

Daily Research Routine:

Morning: Agent searches new papers in your topics
Afternoon: Summarize 2-3 key papers
Evening: Update research journal with findings
Weekly: Generate literature review summary

Advanced Patterns

Multi-Paper Synthesis:

Compare findings across papers
Identify consensus vs. disagreement
Track research evolution over time

Research Gap Analysis:

Identify under-researched areas
Suggest future research directions
Connect theory to empirical needs

Automated Literature Reviews:

Periodic topic searches
Citation tracking
Research trend analysis

Common Use Cases

Platform Economics Research

Search papers, extract network effects findings, maintain literature review

Empirical Study Support

Find datasets, suggest econometric methods, organize results

Theory Development

Review existing models, identify extensions, track theoretical debates

Next Steps: Explore Software Engineering agent (Chapter 5) or Business Management agent (Chapter 6) for other domain applications.

Economics Research Assistant Agent

Platform Economics Research

Empirical Study Support

Theory Development

Table of Contents