Hierarchical DFM factor paths
Data through Sep 2026 · Updated Oct 2, 2026
13 factors · history from 2000 · COVID window blanked (line breaks) · forecast 12 months
These are latent statistical factors from the hierarchical mixed-frequency DFM used on the Ohio economy page—not plotted series. Soft titles are interpretation aids from the largest absolute loadings in the current estimate.
- Sign is arbitrary (weakly normalized); higher is not automatically “better.”
- Scale is standardized (near mean zero in sample), not percent growth or rate points.
- MSA and block factors are residuals after higher layers (national, then Ohio, then local/sector).
- Labels can shift when loadings re-estimate; technical ids (
national1, …) stay stable. - Plots start in 2000. The hard/soft COVID window is missing on the chart (line breaks, not connected through the gap) so 2020–21 spikes do not dominate the scale.
National factors
Back to topExtracted from the national panel first; all series may load on these.
National activity / labor common factor
Technical id national1 · latest Sep 2026: +1.00 · central path Sep 2027: -0.06
Largest |loadings|: +0.91 PAYEMS; +0.70 PCEPI; +0.69 OHNA; −0.69 UNRATE; +0.69 OHPHCI.
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
National second factor (prices–activity mix)
Technical id national2 · latest Sep 2026: +2.41 · central path Sep 2027: -0.06
Largest |loadings|: +0.53 UNRATE; +0.52 CPILFESL; +0.48 Unemployment Rate, Dayton OH (MSA); −0.45 INDPRO; +0.45 PCEPILFE.
Second national factor is a residual mix — less clean than national1.
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Ohio & regional factors
Back to topStatewide residual factors (ohio*) then sector block residuals (labor / prices / housing / activity / consumption), all after national factors are removed.
Ohio residual factor 1
Technical id ohio1 · latest Sep 2026: +0.12 · central path Sep 2027: -0.02
Largest |loadings|: −0.65 OHPHCI; +0.60 OHUR; +0.60 COLU139UR; +0.58 Unemployment Rate, Akron OH (MSA); +0.57 CINC139UR.
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Ohio residual factor 2
Technical id ohio2 · latest Sep 2026: -0.42 · central path Sep 2027: +0.01
Largest |loadings|: +0.78 OHRQGSP; +0.68 Total wages and salaries, Ohio (BEA, SAAR); +0.51 OHNA; −0.48 Realtor.com median listing price, Cleveland (…; +0.47 Ohio state and local tax collections (Census ….
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Labor block residual
Technical id labor1 · latest Sep 2026: -0.92 · central path Sep 2027: -0.03
Largest |loadings|: +0.86 OHLF; +0.77 LBSSA39; +0.73 LASMT391741000000006; +0.66 COLU139LF; +0.30 All Employees: Total Nonfarm, Canton OH (MSA).
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Prices block residual
Technical id prices1 · latest Sep 2026: +0.15 · central path Sep 2027: +0.05
Largest |loadings|: +0.68 CUUR0200SA0L2; +0.63 CUUR0200SA0; +0.55 APUS2007471A; +0.55 CUUR0200SAH2; +0.55 CUUR0200SEHF02.
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Housing-construction block residual
Technical id housing1 · latest Sep 2026: +0.90 · central path Sep 2027: -0.07
Largest |loadings|: +0.88 OHBPPRIVSA; +0.79 COLU139BPPRIVSA; +0.75 OHBPPRIV; +0.62 OHBP1FHSA; +0.46 COLU139BP1FHSA.
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Activity / income block residual
Technical id activity1 · latest Sep 2026: -0.42 · central path Sep 2027: +0.04
Largest |loadings|: +0.63 BROH39M647NCEN; +0.55 QCEW total wages, Ohio (all ownerships); +0.50 Retail Electricity Sales: Industrial in Ohio …; +0.39 Retail Electricity Sales: All Sectors in Ohio…; −0.37 COLU139MFGN.
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Consumption / fiscal block residual (sales-tax base, retail sales, OBM receipts)
Technical id consumption1 · latest Sep 2026: +1.11 · central path Sep 2027: -0.02
Largest |loadings|: +1.15 Implied taxable sales (S-1 county sales tax b…; +1.06 Implied taxable sales (S-1), sum of Akron/Day…; +0.90 Implied taxable sales (S-1), Cleveland MSA (O…; +0.89 Implied taxable sales (S-1), Cincinnati MSA (…; +0.84 Implied taxable sales (S-1), Columbus MSA (Oh….
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Cleveland, OH (MSA) residual
Back to topMSA residual after national and Ohio factors — local deviation, not a standalone “economy index.”
Cleveland residual
Technical id cleveland1 · latest Sep 2026: +0.60 · central path Sep 2027: +0.01
Largest |loadings|: +0.94 CEXRSA; +0.40 Zillow Home Value Index: Cleveland, OH metro; −0.24 Realtor.com median listing price, Cleveland (…; +0.23 Realtor.com active listings, Cleveland (CBSA …; +0.19 Realtor.com new listings, Cleveland (CBSA 17410).
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Columbus, OH (MSA) residual
Back to topMSA residual after national and Ohio factors — local deviation, not a standalone “economy index.”
Columbus residual
Technical id columbus1 · latest Sep 2026: -0.00 · central path Sep 2027: +0.12
Largest |loadings|: +0.81 COLU139NA; +0.30 COLU139MFGN; −0.19 ATNHPIUS18140Q; +0.10 COLU139UR; +0.09 MEDLISPRI18140.
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Cincinnati, OH-KY-IN (MSA) residual
Back to topMSA residual after national and Ohio factors — local deviation, not a standalone “economy index.”
Cincinnati residual
Technical id cincinnati1 · latest Sep 2026: +0.12 · central path Sep 2027: -0.10
Largest |loadings|: +0.73 CINC139NA; +0.29 SMU39171404200000001SA; +0.19 CINC139UR; +0.18 CINC139MFG; +0.18 CINC139LFN.
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.
Other Ohio metros (pooled) residual
Back to topMSA residual after national and Ohio factors — local deviation, not a standalone “economy index.”
Other Ohio metros residual (pooled: Akron, Dayton, Toledo, Youngstown, Canton)
Technical id othermetros1 · latest Sep 2026: +0.24 · central path Sep 2027: +0.08
Largest |loadings|: +0.83 Zillow Home Value Index, Toledo OH metro; +0.82 Zillow Home Value Index, Canton OH metro; +0.81 Zillow Home Value Index, Akron OH metro; +0.78 Zillow Home Value Index, Youngstown OH metro; +0.45 FHFA All-Transactions HPI, Akron OH (MSA).
Standardized latent factor (mean near zero in sample). History from 2000; COVID months blanked with a broken line. Bands: 68% and faint 95% Monte Carlo paths from the sparse factor VARX. Sign and scale are statistical identification choices, not percent growth or unemployment points.