High-Frequency Data: Becoming One with the Running Economy

High-frequency data tracking is an observational dimension in macro research that runs parallel to official statistical data. It provides continuous perception of the running economy through six categories of daily/weekly-frequency indicators outside the National Bureau of Statistics framework — real estate transactions and land, power generation coal consumption, key industrial goods prices, operating rates, export container freight indices, and agricultural & sideline product prices. Three daily-frequency core data points serve as the rhythmic anchor (30-city property sales, six major power groups’ coal consumption, and rebar prices), and through intensive training the analyst’s perception becomes “one with” the actual running economy.

The Framework As It Stands

This section is organized from compiled research notes: the original framework’s structure, terminology, and key formulations are preserved, including editorial bridging and supplementary external facts; diagrams are drawn by the compiler according to the original text’s structure.

Why look at high-frequency data (the limitations of statistical data and how they complement each other)

Even if statistical data has fast-or-slow biases, as long as those biases are continuous and stable (analogous to a watch that consistently runs ten minutes slow or fast — you can still tell the time), they remain usable for macro judgment. Macro analysts simultaneously access large amounts of high-frequency data outside the NBS framework, which can be cross-verified against statistical data to reduce the risk of misjudgment from any single source.

Six categories of high-frequency indicators (what to look at)

High-frequency indicators broadly fall into six categories: real estate related / power generation and consumption / key industrial goods prices / operating rates / export container freight indices / agricultural & sideline product prices.

• Category 1 · Real estate related: 30-city property transaction volume (daily/monthly transaction floor area), tracking property sales demand. Upstream of property sales is construction starts, and upstream of starts is land acquisition — land acquisition reflects developers’ expectations for the future, so land transaction volume and land premium rates are key indicators for judging the future trajectory of real estate.

• Category 2 · Power generation and consumption: The six major power generation groups publish daily power generation coal consumption (coal consumed for thermal power generation). Both industrial and residential sector activity requires electricity, so coal consumption is a key derived indicator for observing the running economy.

• Category 3 · Key industrial goods prices: Rebar prices, cement prices, coal prices. Price indicators never lie — they concentrate the supply-demand conditions of upstream industrial goods and expectations for the future (derived directly from transaction outcomes, not subjective judgment). When market participants anticipate an economic downturn, rebar prices will fall first and can serve as a leading signal for industrial demand and economic expectations. Cement prices have regional pricing characteristics (East China / South China / Northwest China each have independent prices), providing evidence of regional growth — for example, a price rise for East China cement implies that East China infrastructure construction may have already started.

• Category 4 · Operating rates: Blast furnace operating rates, coking enterprise operating rates, and semi-steel tire operating rates represent the actual operation of key industrial sectors and allow simultaneous tracking of enterprise business conditions.

• Category 5 · Export container freight indices: CCFI (China Containerized Freight Index) and SCFI (Shanghai Containerized Freight Index) — their year-on-year changes can broadly serve as a coincident or slightly leading indicator of export-sector conditions.

• Category 6 · Agricultural & sideline product prices: Prices for pork, vegetables, grains, and similar items are high-frequency data for tracking inflation.

Three must-watch daily-frequency data points + becoming one with the running economy (how to use them)

Three daily-frequency data points to watch every day: 30-city property sales, six major groups’ power generation coal consumption, and rebar prices — representing property demand, economic activity intensity, and supply-demand expectations for industrial goods, respectively.

Only when high-frequency tracking becomes a habit does economic growth appear continuous in the analyst’s view and forecasting develop “feel”; those who pay attention only at low frequency (monthly/annually) are prone to sudden shocks. Macro forecasting requires using high-frequency data to take the pulse of the economy; through intensive training, perception becomes synchronized with actual economic activity — the Leon sniper technique: move with the target until your pace matches, then pull the trigger — achieving the state of “becoming one.”

Statistical data has lags/biases, but stable biases remain usable + non-statistical high-frequency data cross-verifies
        ↓ Classify high-frequency indicators into six categories
Six categories: property transactions & land / power gen. coal consumption / industrial goods prices / operating rates / container freight indices / agricultural & sideline product prices
  ├ Property transactions = demand; land acquisition & premium rate = developer expectations
  ├ Power gen. coal consumption = economic activity intensity (derived indicator)
  ├ Industrial goods prices "never lie": rebar senses downturns first / cement regional pricing reveals regional growth
  ├ Operating rates (blast furnace / coking / semi-steel tire) = enterprise production conditions
  ├ CCFI/SCFI = export conditions (coincident / slightly leading)
  └ Agricultural & sideline product prices = inflation
        ↓ Select daily-frequency core from the six categories
Three must-watch: 30-city property sales + six major groups' power gen. coal consumption + rebar price
        ↓ Intensive training
Perception syncs with actual economic activity ("becoming one") (Leon-sniper-style pace alignment) → forecasting develops "feel"

Timing caveat (course recorded in early 2019): The six indicator categories, the three must-watch daily-frequency data points, CCFI/SCFI, and the “six major power groups” coal consumption data all reflect indicator practices and examples current at the time of the course (early 2019); some definitions have since been adjusted (e.g., subsequent changes to the six major power plants’ coal consumption data framing). Judging “current conditions” requires drawing on current-period high-frequency data and labeling it accordingly — these figures must not be cited as the current state of affairs.

Compiler’s Perspective

Coordinates: Category = Observation Indicators & Signals / axis_h = Qi (Instruments) / axis_v = Its Place in the Whole

Bridging Layer

The usage error for this framework comes in almost only one pattern: treating high-frequency tracking as a “publication-day ritual” — reviewing data in a concentrated burst when monthly figures are released, then ignoring it mid-month. The result is that one only notices GDP growth falling after it happens, while rebar prices had already been declining continuously three to four weeks earlier, and 30-city transaction volumes had already shown dual year-on-year and month-on-month declines two weeks prior. Analysts who observe only at monthly low frequency are naturally one beat behind at turning points — the lag is not a failure of computational ability but a failure to establish a daily-frequency rhythm.

The three must-watch daily-frequency data points do not carry equal weight: rebar (a price indicator) is a leading signal, power generation coal consumption (a derived indicator) is a coincident signal, and 30-city transaction volume (terminal demand) is a signal that is only complete when combined with land premium rates. Simply averaging the three and ignoring their temporal sequence means missing a roughly one-to-two-week window of leading signals around turning points — this is the structural trap unique to this entry, and it does not overlap with the discussion of the general six-category framework.

Another common misreading of the six categories is to interpret “price indicators never lie” as “price indicators are the only trustworthy ones.” The original intent of the framework is that prices are derived directly from transaction outcomes rather than subjective statistics — but they only reflect supply, demand, and expectations, not transaction volume; a leading decline in rebar prices only shows that market participants anticipate a downturn, not its magnitude. Magnitude still requires combining blast furnace operating rates and sales volume data for a comprehensive judgment.

Proprietary increment: Cement prices’ regional pricing characteristics (East China / South China / Northwest China each have independent prices) enable them to provide regional signals earlier than a national average price — East China cement rising alone while Northwest China is flat implies infrastructure construction has started within that region rather than a nationwide demand recovery. Analysts who read only the national cement average without examining regional breakdowns will systematically misread regional policy-driven pushes as weak signals.

See Also

• The Nature of Macro Research and the Sense of Position — high-frequency tracking is the operational foothold for “sense of position” training: synchronizing daily-frequency rhythm with the running economy is the path from abstract position-sense to something perceptible

• Export Cycle Tracking: From Large to Small on Dual-Cycle Coordinates — CCFI/SCFI serve in this entry as high-frequency freight indicators; in the export entry they serve as leading verification of export-sector conditions — the two roles are different

• Property Indicator Chain Tracking: The Three Sales Drivers and Investment Transmission — 30-city transaction volume and land premium rates are the daily-frequency core in this entry and the starting point of the driver chain in the property entry; the two should be read together

• Finding High Probability Through Empirical Regularities: Six Case Studies in Practice — the perceptual patterns accumulated through high-frequency training are the data foundation for empirical regularity induction

• The Four Fiscal Accounts and Broad Fiscal Aggregates: Three-Step Infrastructure Estimation

• The Inflation Triangle Model: Two-Layer Forecasting and High-Frequency Tracking

Sources

• Compiled research notes · collected 2026-07

External source: External course (collected with identity removed), Section 4.2 “High-Frequency Data: Becoming One with the Running Economy”; time point January 2019; course name and instructor have been stripped per the identity-removal method, framework content preserved in full.