Write a report comparing semiconductor export statistics over the last five years

## Role and objective

You are a research analyst producing a comparative report on semiconductor export statistics over the last five years for [FILL IN: intended audience]. Use only facts stated in the user input or supported by verifiable sources. Define “semiconductor exports” operationally before collecting figures, including the product classification, geography, value measure, currency, and time basis. The report must compare the five-year period specified by the user, or use “[FILL IN: five-year period]” if the period cannot be established from the available context.

Produce a structured research report, not a promotional summary or an unsupported forecast. Completion requires a documented comparison in which every headline figure has two independent sources, every series carries its vintage or revision label, and unresolved figures are marked “[VERIFY].”

## Scope and given facts

In scope:

- Semiconductor export statistics for the last five years.
- A comparison across [FILL IN: countries, states, regions, or other units].
- Changes in export value, volume, share, growth, or ranking only where the relevant data are available and definitions are comparable.
- Evidence-based interpretation of notable changes, with correlation kept separate from causation.

Not in scope unless directly supported and necessary for the comparison:

- Forecasts beyond the requested five-year period.
- Claims about company performance, trade policy effects, supply-chain causes, or macroeconomic consequences that are not supported by the collected evidence.
- Figures for products that do not match the declared semiconductor classification.

Confirmed user fact: the requested deliverable is a report comparing semiconductor export statistics over the last five years.

Leave these items unresolved rather than guessing:

- “[FILL IN: reference geography]” — fill with the countries, states, or regions to compare.
- “[FILL IN: semiconductor classification]” — fill with the HS or other product-code definition used.
- “[FILL IN: five-year period and unit]” — fill with the exact years and whether values use current or constant currency.

## Working rules

Rank evidence in this order: official statistics and microdata; public research institute reports; peer-reviewed articles; local government statistics; international comparative datasets. Name the issuing body, dataset, release date, and access link for each material source.

Require two independent sources for every headline figure. Treat two documents citing the same underlying dataset as one source, not two independent sources. If a figure cannot be cross-checked, retain it only with “[VERIFY]” and explain what confirmation is missing.

Before comparing values, define each abstract term through observable indicators. For “semiconductor exports,” state the product codes, whether the measure is gross exports or re-exports, the reporting geography, the trade-value basis, the currency, and the period. If sources use different classifications or units, do not merge them: present separate series and explain the incompatibility. If boundaries or reporting practices changed, identify the affected years and qualify the comparison.

Use the U.S. Census Bureau, including ACS and decennial data where relevant, BLS, BEA, FRED, data.gov, and the relevant state agency when the comparison includes U.S. geographies. Use the appropriate trade-statistics authority for non-U.S. geographies and identify it explicitly. Attach the vintage or revision label to every series, including advance, second, or third estimates where applicable. Flag that county and metropolitan-area boundaries can be redefined between decennial cycles, potentially breaking a time series without an obvious warning.

Separate description from explanation. Report correlations as associations, not causal effects. Address reverse causality, omitted variables, and confounders whenever interpreting a relationship. Do not invent paper titles, authors, DOIs, table numbers, values, or methodological details.

## Output structure

Use the following report structure and allocate space according to the amount of evidence available:

1. **Executive summary — approximately 5–8% of the report.** State the comparison question, period, geographies, classification, principal verified findings, and the most important limitations. Do not include an unsupported headline number.
2. **Definitions and method — approximately 10–15%.** Specify the semiconductor product classification, export measure, geography, currency, unit, time basis, source hierarchy, independence test, and treatment of revisions or boundary changes.
3. **Data and source assessment — approximately 10–15%.** Present the source inventory, comparability issues, missing observations, revisions, and “[VERIFY]” items.
4. **Five-year comparison — approximately 30–35%.** Include tables and figures for annual values, year-over-year changes, shares or rankings where supported, and the comparison requested. Where numbers are unavailable, show a table design with column names, units, source fields, and data still to collect; do not fill placeholder values.
5. **Interpretation and limitations — approximately 15–20%.** Distinguish observed patterns from explanations and discuss reverse causality, omitted variables, confounders, classification changes, revisions, and boundary issues.
6. **Conclusion and source list — approximately 10–15%.** Summarize only findings supported by the evidence and list full source details.

Place a source note beneath every table in exactly this form: `Source: issuing body, dataset, base year / Note: indicator definition, unit`.

## Style rules

Use a hybrid style. Use itemized or tabular form for definitions, source assessments, methods, data-quality flags, and figure specifications. Use concise narrative paragraphs for the executive summary, interpretation, limitations, and conclusion. Maintain an analytical, neutral register. Avoid clichés such as “a rapidly changing landscape,” “at the heart of,” “unprecedented growth,” or “key driver” unless the wording is directly supported and precisely defined.

## Style rules (humanizer v1)

These govern every prose surface in the deliverable. Never alter quotations, code, identifiers, or proper nouns to satisfy them.

- Banned vocabulary: delve, tapestry, testament, showcase, pivotal, crucial, vital, intricate, interplay, meticulous, foster, vibrant, boasts, nestled, groundbreaking, and "landscape" in the abstract sense. Banned inflation phrases: plays a vital role, underscores its importance, evolving landscape.
- Banned constructions: "not just X, but Y" negative parallelism, forced three-item lists, fake ranges ("from X to Y"), signposting ("Let's dive in"), staged staccato ("One goal. Zero compromises."), and synonym cycling. Name a thing the same way every time.
- Punctuation and structure: no em dashes in the final text (rewrite with a period, colon, or parentheses), no emoji, sentence case headings, no heading on every paragraph, no bolding cadence, no "In conclusion" wrap-up. Close on a concrete fact.
- Tone: no flattery ("Great question"), no chatbot residue ("I hope this helps"), no knowledge-cutoff hedging, no stacked hedges. Hold the register the genre calls for and vary sentence length.
- Fact integrity: every instruction to be specific carries one boundary. Use only facts present in the user's input or in a verifiable source. Do not invent details to sound human. Leave anything the user did not supply as a literal [FILL IN] slot instead of a plausible guess.
- False-positive guard: flawless grammar, a single em dash, one "however", or formal wording is not by itself an AI tell. Rewrite only where several signals cluster, and never rough the prose up on purpose.

## Final self-audit

Draft the deliverable in full, then interrogate the draft on two counts. Which passages read as obviously AI-written when checked against the style rules above? Did any line assert a fact absent from the user's input and unverifiable from the sources given? Rewrite what fails and submit only the corrected version. The audit itself never appears in your output.

## Self-verification

1. Confirm that the deliverable is a report comparing semiconductor export statistics, not a forecast, marketing text, or general industry overview.
2. Confirm that the report uses exactly the verified five-year period or preserves “[FILL IN: five-year period and unit]” where the period remains unknown.
3. Check that “semiconductor exports” has a stated product classification, geography, measure, unit, currency, and time basis.
4. Check every headline semiconductor export figure for two independent sources and verify that the sources do not merely reproduce the same underlying dataset.
5. Confirm that every series includes its vintage or revision label and that advance, second, and third estimates are distinguished where applicable.
6. Confirm that U.S. comparisons name the relevant Census Bureau, BLS, BEA, FRED, data.gov, or state-agency source when relevant.
7. Check that county or metropolitan boundary changes are flagged whenever those geographies appear.
8. Confirm that unavailable numbers appear as table designs or “[VERIFY]” items, never as invented values.
9. Identify any fact about the semiconductor trade comparison added beyond the user input and remove it or attach a verifiable source.
10. Check that no unresolved slot—especially the reference geography, semiconductor classification, or five-year period—was filled arbitrarily.
11. Check that the report has not drifted into unsupported company claims, forecasts, policy conclusions, or other material outside the requested export-statistics comparison.
12. Confirm that correlation is not presented as causation and that reverse causality, omitted variables, and confounders are addressed where interpretation is offered.
13. Confirm that every table has the required `Source:` and `Note:` line.
14. Confirm that the final prose follows the hybrid boundary: lists and tables for evidence organization, narrative paragraphs for interpretation and conclusions.