Total Public Construction Spending: Total Construction in the United States (MPCPXXXXS) | FRED | St. Louis Fed
- Rows
- 291
- Columns
- 2
- Data range
- 2002-02-01 to 2026-04-01
Graph and download economic data for Total Public Construction Spending: Total Construction in the United States (MPCPXXXXS) from Feb 2002 to Apr 2026 about public, expenditures, construction, and USA.
AI analysis
Dataset Analysis: Total Public Construction Spending (MPCPXXXXS)
1. What This Dataset Represents
This series, sourced from the Federal Reserve Bank of St. Louis's FRED database (via fredgraph.csv?id=MPCPXXXXS), tracks total public construction spending in the United States. FRED is a highly authoritative, actively maintained source for macroeconomic time series, and the fact that this pull extends to April 2026 suggests either forward-looking seasonally-adjusted projections or that the dataset was retrieved with a future vintage/forecast horizon built into FRED's release schedule — worth verifying against FRED's actual publication cadence, since real observations rarely extend that far past the current date. With monthly observations from February 2002 onward (291 rows), this gives roughly 24 years of history, which is a strong length for time-series correlation, seasonality decomposition, and lead/lag analysis against other economic indicators.
2. Data Quality Observations
The dataset is clean on the surface: zero null cells across both columns, and the join key (observationdate) has 291 distinct values matching the 291 total rows — indicating no duplicate timestamps and a well-formed, continuous monthly index. However, the duplicate-row count itself is marked as pending ("Phase D — recomputed against Parquet"), so full deduplication confirmation is outstanding and should be re-checked once that computation completes. Similarly, the sample data rows are not yet available (0 rows shown), which limits deeper manual spot-checking of formatting or encoding issues at this stage — this is a gap to close before final sign-off on quality.
3. Key Columns and Distributions
The core value column, MPCPXXXXS, appears to already be transformed (likely a percent-change or growth-rate series rather than raw dollar spending, given a range of -6 to 7.4 and a mean near zero). Its distribution is roughly symmetric (skewness = 0.09), with mean (0.32) sitting close to median (0.3), suggesting minimal skew-driven distortion. The spread is moderate (σ = 1.69), with an interquartile range from -0.7 to 1.3, and only 4 outliers detected — a small, manageable fraction of 291 observations (~1.4%). With 69 distinct values across 291 rows, there's meaningful repetition, consistent with a bounded percentage-style metric rather than continuous raw spending figures.
4. Recommended Join Key
observationdate is the clear and only sensible join key — it's complete (0 nulls), fully distinct (291/291), and monthly-grained, making it straightforward to align with other FRED series or any dataset reportable at monthly frequency (resampling may be needed for weekly/quarterly sources).
5. Suggested Pairings for Correlation Analysis
Strong candidates include: Private Construction Spending (PRRESCONS or TLPRVCONS) to compare public/private investment cycles; interest rate series (e.g., 10-Year Treasury Yield, Federal Funds Rate) to test financing-cost sensitivity; PPI for construction materials to explore cost-pass-through effects; government spending/deficit data for fiscal-policy correlation; and employment data in construction (CES2000000001) to link spending to labor demand. Given the monthly cadence and long history, this series also pairs well with leading indicators like housing starts/permits for lead-lag causality testing.
Columns
- observation_date (date)
- MPCPXXXXS (decimal)