Building a Research Process for Global Macro Economics

What question is the research meant to answer?

Begin macro research by stating a question that could be challenged by evidence: for example, whether demand is losing momentum, inflation pressure is broadening, or credit conditions are tightening. Define the geography, period, and decision context before collecting charts. This prevents an appealing indicator from quietly changing the question halfway through the analysis.

Write down a starting hypothesis and plausible alternatives. A base case is a working explanation, not a forecast with certainty. Specify observations that would strengthen, weaken, or leave each explanation unchanged, and distinguish the expected direction from the confidence attached to it.

Assemble evidence with its provenance intact

Use a compact dashboard of indicators tied to the question rather than an indiscriminate list. Combine timely surveys with slower hard data where relevant, and note that they measure different things. For each series, record source, release date, reference period, units, seasonal-adjustment status where available, and whether the figure is preliminary or revised. Compare like definitions across countries; similarly named measures may not be directly equivalent.

• Growth: output, business surveys, production, or spending
• Inflation: price measures, wages, and relevant cost pressures
• Labor: employment, hours, vacancies, or claims as available
• Policy and conditions: rates, lending, credit, and fiscal updates

Move from observation to a transmission story

Lay out the proposed chain explicitly: what changed, which households, firms, banks, or public bodies are exposed, and how their choices might transmit the change to demand, prices, or financing. Mark each link as observed, inferred, or assumed. Several mechanisms may operate at once, and the timing can vary; an announcement is not the same thing as a completed transmission.

Check whether the evidence also fits rival accounts, including temporary disruptions, measurement changes, base effects, or external shocks. Market pricing can supply information about expectations, but it is not a direct reading of future economic outcomes or proof that a policy will be delivered.

Keep a dated decision and revision log

Save the data vintage used, the conclusion reached, open uncertainties, and the conditions that would trigger a reassessment. When figures are revised, preserve both the earlier and newer view instead of treating the latest value as if it had always been known. This makes it possible to evaluate the process fairly and identify whether a miss came from data, assumptions, or an unexpected event.

Review conclusions at the pace of the evidence

Use a scheduled review for slow-moving themes and event-driven updates for material releases or policy decisions. Update only the parts of the thesis affected by new evidence, and communicate uncertainty in plain language. A repeatable method improves comparability; it cannot eliminate forecasting error or turn an economic scenario into a guaranteed result. This is educational information, not individualized investment, tax, legal, or accounting advice.