Build a Climate-Finance Research Question That Can Be Answered

Climate-finance research often starts with a large concern: climate risk matters, transition investment is insufficient, or a portfolio should align with a temperature goal. These statements can motivate work, but they are not yet research questions. They do not identify the decision, unit, comparison, time horizon, or evidence needed for an answer.

A usable question is narrow enough to test and broad enough to matter. It tells the reader what changes when the answer changes.

Name the decision first

Start with the person or process that will use the result. A credit team can be deciding whether a borrower’s cash flow remains resilient under a transition scenario. A policy researcher can be comparing which financing constraint slows a class of projects. An asset owner can be deciding whether a metric is suitable for monitoring, engagement, or allocation.

These decisions need different evidence. A metric that supports portfolio monitoring may not estimate loss. A scenario that supports strategic exploration may not provide a probability forecast. State the use before selecting the tool.

Define the unit of analysis

The unit can be a company, facility, project, household, security, fund, loan book, sector, country, or financial system. Do not move between units without an explicit aggregation or attribution rule.

For example, facility emissions can inform a company total, but only after the researcher defines organizational boundaries and avoids double counting. A company score cannot automatically describe a security’s valuation effect. Write the unit beside every main variable.

Set the time horizon

Climate processes, policy changes, asset lives, and financial contracts operate on different clocks. A one-year risk view can miss an asset’s long exposure. A thirty-year scenario can be too coarse for a near-term refinancing decision.

Use a horizon that matches the decision, then show any mismatch. If the available data ends in five years while the project lasts twenty-five years, the missing period is a model assumption, not an observed result.

Choose a comparison

An answer needs an alternative: current policy versus an orderly transition, one technology versus another, disclosed emissions versus an estimate, or a project with and without a financing constraint. Define what remains constant between cases.

Avoid a baseline that quietly contains the conclusion. A “no climate action” case can include different demand, technology, and macroeconomic assumptions. List these elements instead of treating the baseline as neutral.

Write the evidence map

For each variable, label the source as observed data, company disclosure, official statistic, scenario input, model output, proxy, or judgment. Record the date and coverage. The IPCC explains the scientific evidence and assessed uncertainty behind climate findings, but a financial study must still show how that evidence enters its own question.

Use one sentence as a test

Complete this form: “For this decision, we compare these units over this period using this evidence, while changing this assumption.” If the sentence is impossible to write, the project probably contains several questions.

Split them. A small answer with a visible boundary is more useful than a broad conclusion that changes meaning from one table to the next.