Introduction

The Federal Reserve Bank of San Francisco publishes two indexes drawn from its Beige Book surveys of 12th District businesses: a Business Sentiment Index reflecting overall economic conditions, and an Inflation Gauge capturing price expectations. Historically these indexes have moved together — when businesses felt pessimistic, inflation expectations were also subdued, and vice versa. This co-movement makes intuitive sense: in a typical demand-driven inflationary episode, strong economic conditions and rising prices go hand in hand.

This paper uses those two indexes, together with realized West Region CPI data from FRED, to examine whether the current divergence between sentiment and inflation expectations in the 12th District resembles the 2022 inflation episode — and what it might tell us about the nature of inflation when it doesn’t. The exercise is primarily descriptive: an attempt to show how publicly available Fed data can help distinguish between demand-pull and cost-push inflationary dynamics in real time, using the Western United States as a case study.

Data

This analysis draws on two sources. The SF Fed 12th District Business Sentiment data provides two indexes — a Business Sentiment Index and an Inflation Gauge Index — both derived from Beige Book surveys and released on Beige Book publication dates. Macroeconomic context comes from FRED: West Region CPI (CUUR0400SA0), West Region unemployment (CWSTUR), and the federal funds rate (FEDFUNDS).

A few data management decisions are worth noting.

Geographic proxy. The 12th District covers nine western states but there is no direct CPI series for the district boundary. West Region CPI from FRED is the closest available approximation and was used throughout as a proxy for realized inflation in the district.

Sample period. Data was filtered to begin in January 2014, which corresponds to the earliest reliable observations in the SF Fed sentiment dataset. Including earlier observations would have introduced structural gaps.

CPI transformation. Raw CPI levels were converted to year-over-year percent change to capture inflation dynamics rather than price levels. This is the most natural comparison to the SF Fed indexes, which are also expressed in terms of directional change relative to a baseline.

Merging on nearest date. Because the Beige Book releases on an irregular schedule rather than fixed monthly dates, SF Fed sentiment observations were matched to the nearest available CPI observation using a rolling nearest-date join via the data.table package. A standard calendar merge would have introduced unnecessary missingness.

Scaling for regression. Prior to the distributed lag regression, CPI YoY was standardized using scale() to bring it onto a comparable unit scale with the SF Fed indexes. Without this step the raw CPI values — which average around 5% over the sample period — would have distorted the intercept and made the coefficients difficult to interpret.

Lag structure. Three and six month lags of the Inflation Gauge were tested based on the assumption that survey-based expectations would take at least one quarter to show up in measured prices. The six month lag proved statistically significant while the three month lag did not, suggesting the transmission from expectations to realized inflation operates on a longer horizon than a single quarter.

Results

Business Sentiment, 2014–2026

Business sentiment declined sharply during the initial COVID-19 outbreak in 2020, recovered through 2021, and fell again during the Federal Reserve’s 2022 tightening cycle. A third negative episode emerges in early 2025, coinciding with the imposition of broad tariffs on imports.

Inflation Expectations vs. Business Sentiment

Prior to 2025, the Business Sentiment and Inflation Gauge indexes tracked closely, rising and falling together across business cycles. The current episode is visually distinct: the Inflation Gauge has moved sharply positive while Sentiment has remained negative, a decoupling not previously observed in the data.

Inflation Expectations as a Leading Indicator

The Inflation Gauge tends to lead realized West Region CPI by several months, most visibly around the 2021–2022 inflation surge where expectations peaked before measured prices followed. This leading relationship motivates the regression analysis below.

Two Episodes of Divergence

The divergence between the SF Fed Inflation Gauge and Business Sentiment Index reached 2.98 in April 2025 — the largest value in the dataset and 54% above the peak divergence recorded during the 2022 inflation episode. Eight of the ten largest divergences in the full dataset occur in 2025–2026.

The 2022 inflation episode fit the historical pattern well. Business sentiment was elevated, inflation expectations were elevated, and realized West Region CPI followed both upward. The divergence between the two indexes was modest and temporary — both were pointing in the same direction, consistent with a demand-driven episode where strong economic conditions and rising prices reinforced each other.

The 2025 episode looks different. Inflation expectations have returned to levels not seen since the 2021-2022 surge, but business sentiment has remained persistently negative. Businesses across the 12th District appear to expect prices to rise while simultaneously feeling pessimistic about overall economic conditions. That combination is historically unusual in this dataset and is more consistent with cost-push inflation — prices rising not because consumers are spending freely but because tariffs and supply disruptions are pushing input costs higher — than with the demand-driven dynamics that characterized 2022.

This distinction matters because the two mechanisms can produce different outcomes and may respond differently to monetary policy. The current divergence doesn’t tell us definitively which dynamic is at work, but it is a useful signal that the current episode may not follow the same path as 2022.

Regression: Does the Inflation Gauge Predict Realized CPI?

## 
## Call:
## lm(formula = west_cpi_yoy_scaled ~ inflation_gauge_lag3 + inflation_gauge_lag6, 
##     data = merged)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.5051 -0.6581 -0.2643  0.5937  2.5358 
## 
## Coefficients:
##                      Estimate Std. Error t value Pr(>|t|)   
## (Intercept)           0.02529    0.10180   0.248  0.80441   
## inflation_gauge_lag3 -0.03939    0.12823  -0.307  0.75944   
## inflation_gauge_lag6  0.37940    0.12830   2.957  0.00404 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.9417 on 83 degrees of freedom
##   (13 observations deleted due to missingness)
## Multiple R-squared:  0.1341, Adjusted R-squared:  0.1133 
## F-statistic: 6.429 on 2 and 83 DF,  p-value: 0.002536

A one standard deviation increase in the SF Fed 12th District Inflation Gauge predicts a 0.38 standard deviation increase in West Region CPI YoY approximately six months later (p = 0.004). The three-month lag is not significant (p = 0.76), suggesting the predictive relationship operates on a longer horizon. The model explains 13% of the variation in realized CPI — modest, but consistent with the gauge capturing one meaningful signal among many drivers of regional inflation.

Discussion

The SF Fed 12th District data offers a useful lens for distinguishing between inflationary episodes in real time. When the Inflation Gauge and Business Sentiment Index move together — as they did in 2022 — it is broadly consistent with demand-pull dynamics. When they diverge sharply, as they have in 2025, it may signal something different: inflation expectations becoming embedded even as underlying economic confidence weakens, more consistent with cost-push pressures.

The leading relationship between the Inflation Gauge and realized CPI adds a further dimension. If the historical pattern holds in the current environment, elevated expectations today would suggest upward pressure on West Region prices through late 2025 and into 2026. However, that relationship was estimated largely during demand-driven episodes, and it is an open question whether it holds with the same force when the source of inflation pressure is different.

Conclusion

The SF Fed 12th District data offers a useful lens for distinguishing between inflationary episodes in real time. The divergence reached its largest recorded value in April 2025 — 54% above the 2022 peak — and has remained historically elevated through mid-2026. Whether realized West Region CPI ultimately confirms the Inflation Gauge’s signal, or whether the divergence reflects uncertainty rather than embedded price pressure, remains to be seen. The next several index releases will be informative.

This analysis is necessarily preliminary and descriptive. Its primary contribution is to illustrate how two publicly available SF Fed indexes, combined with FRED macroeconomic data, can be used to characterize the nature of an inflationary episode in the 12th District in real time and to document that the current episode appears structurally distinct from the most recent comparable period.

Data Sources