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

This framework characterizes inflation as a near-endogenous, inevitably persistent economic phenomenon (rooted in the time value of money); the inflation triangle model decomposes it into three components — the output gap, inflation expectations, and imported factors (oil prices/exchange rates); at the forecasting layer it uses a two-layer structure of “output gap to determine the annual direction + carry-over effects and other within-year factors to determine the intra-year distribution”; at the tracking layer it assembles a high-frequency indicator system from non-food items (WTI crude / tertiary-hospital consultation fees / traditional Chinese medicine materials) and food items (the vegetable-basket price index / Shandong wholesale vegetable prices / Qianhai agricultural-product wholesale prices, with pork as the key item); and it notes that the low elasticity of rent in the CPI basket (cumulative increase of approximately 70–80% over 15 years, far below actual market rents) may cause the CPI to systematically underestimate inflation, which in turn leads to an underestimated DCF discount rate and persistent asset-pricing bias.

The Framework As It Stands

This section is compiled from research drafts: the original framework’s structure, terminology, and key formulations are preserved, including editorial bridging and external factual annotations; charts are drawn by the compiler following the original framework’s structure.

The Endogeneity of Inflation: Why It Must Persist Long-Term

Inflation is a near-endogenous, inevitably present economic phenomenon. Its root lies in the time cost of money — consuming now versus consuming later requires interest as compensation; as long as money has time value, the economy will necessarily produce inflation. Inflation only varies in degree; it must exist in the long run. The starting point for analyzing inflation is “it will inevitably be present,” not “will there be any.” This framework diverges from the Wage-Price Spiral at the following fork: this entry focuses on decomposing the sources of inflation and the tracking tools; that entry focuses on the divergent conditions for inflation persistence.

The Inflation Triangle Model

Inflation is jointly determined by three factors:

Output gap (actual output vs. potential output)
Inflation expectations
Imported factors (oil prices, exchange rates, etc.)

The Two-Layer Forecasting Framework

Layer One: Output gap determines the annual direction

Actual output < potential output → economic growth heading upward, inflation pressure limited
Actual output > potential output → endogenous inflation gradually rising
From historical experience, the output gap can roughly determine the annual level and direction of inflation

Layer Two: Carry-over effects and other within-year factors determine the intra-year distribution Once the annual direction is established, carry-over effects and other within-year variables determine how inflation is distributed at the monthly level (timing of peaks and troughs).

High-Frequency Tracking Indicator System (as of course reference point 2019-01)

Non-food items

Fuel: WTI crude oil prices
Medical services: tertiary-hospital consultation fees, Chengdu traditional Chinese medicine material prices (All non-food items have corresponding high-frequency tracking systems)

Food items

Ministry of Agriculture vegetable-basket price index
Shandong wholesale vegetable price index
Qianhai agricultural-product wholesale price index
Key individual item: pork — the core indicator driving inflation within the food category; whenever an inflation cycle hits, macro researchers are compelled to become agriculture, forestry, animal husbandry, and fishery analysts

The Mechanism of CPI Underestimating Inflation and Spillover to Asset Pricing

Rent CPI deviation (data as of course reference point 2019-01):

Rent CPI rises approximately 2–3% per year, with low elasticity and stable growth
Cumulative increase over the past 15 years approximately 70–80%
In reality, rent increases in first- and second-tier cities (and even third- and fourth-tier cities) far exceed the above figures
From the rent perspective alone, the CPI may to some extent underestimate inflation

Spillover to asset pricing: if the discount rate is underestimated (the economy’s money depreciates faster than the model assumes), a DCF calculation might yield a company value of 100 but the share price hovers around 50 for an extended period; beyond “market inefficiency,” another explanation is that the applicable discount rate is too low — an efficient market will automatically incorporate the rate of monetary depreciation, causing a persistent divergence between DCF valuations and market prices.

Temporal caveat: rent CPI of 2–3% per year, 15-year cumulative 70–80%, the DCF valuation 100 vs. market price 50 example, WTI crude oil and other specific high-frequency indicator names are all empirical values or illustrative examples from the time of the course (2019-01); to judge the current inflation level, intra-year distribution, and whether CPI is underestimating, one must use the current output gap, carry-over effects, high-frequency indicators, and the latest data sources.

Compiler’s Perspective

Coordinates: Observation Indicators & Signals · Qi · What It Is

Interface Layer

The exclusive incremental insight of this entry lies in the stacking of two levels of “instrument limitation.” First level: relying solely on CPI to read inflation is equivalent to sensing the true speed of monetary depreciation through a low-elasticity rent index — rent CPI accumulated 70–80% over 15 years, while actual market rents surpassed that long ago. This deviation is not a statistical error; it is a systematic underestimation built into the index construction itself. Second level: the old approach, when confronted with the divergence between a DCF valuation of 100 and a long-term market price of 50, would offer “market inefficiency” as the explanation while bypassing another diagnostic path: the discount rate input (the inflation assumption) was already too low to begin with. Only with both layers of instrument limitation stacked together does the complete asset-pricing bias mechanism emerge. This is the exclusive incremental contribution of this entry that distinguishes it from general inflation discussions — deriving a systematic asset-pricing bias directly from the limitations of the tracking tools.

The specific error of the old approach: after monthly CPI data is released, directly extrapolating trends from month-over-month and year-over-year figures without decomposing the output gap (annual direction) and carry-over effects (intra-year distribution) into two layers. This causes misreading of months with strong carry-over effects (e.g., food prices around the Spring Festival) as trend inflation, or misreading weak carry-over months as evidence that inflation has been eliminated. The correct sequence is first to use the output gap to set the annual tone, then use carry-over effects and high-frequency indicators to read monthly disturbances.

Pork’s key position on the food side means that inflation research, when a pig cycle erupts, must descend to agricultural indicators such as breeding farm capacity and the number of sows in production — this is core homework that cannot be outsourced to food-sector analysts, yet most macro research frameworks have not built a dedicated tracking mechanism for this in normal times, only beginning to track it reactively after inflationary pressure has already built up, forfeiting the leading advantage.

In The Four Fiscal Accounts and Broad Fiscal Aggregates: Three-Step Infrastructure Estimation, the pull of broad fiscal policy on infrastructure is the upstream driver of the output-gap component in this entry’s inflation triangle — broad fiscal expansion → rise in infrastructure growth rate → narrowing of the output gap → rising endogenous inflation pressure; the two frameworks converge at the output gap.

Sources

Compiled research draft · collected 2026-07

Compiled draft z-0102 · collected 2026-07

Z-Finance

Explorer