Skip to main content
Financial Innovation

Table 5 Results for components of state-level economic conditions

From: Do U.S. economic conditions at the state level predict the realized volatility of oil-price returns? A quantile machine-learning approach

Benchmark/rival model

h = 1

h = 2

h = 4

 

Expectations

HAR-RV vs. HAR-RV-states/q = 0.95

0.0218

0.1292

0.1412

HAR-RV vs. HAR-RV-states/q = 0.75

0.0205

0.0654

0.0833

HAR-RV vs. HAR-RV-states/q = 0.5

0.0154

0.0390

0.0733

HAR-RV vs. HAR-RV-states/q = 0.25

0.0030

0.0186

0.0732

HAR-RV vs. HAR-RV-states/q = 0.05

0.0052

0.0341

0.0749

HAR-RV-US vs. HAR-RV-states/q = 0.95

0.0211

0.1279

0.1356

HAR-RV-US vs. HAR-RV-states/q = 0.75

0.0202

0.0654

0.0832

HAR-RV-US vs. HAR-RV-states/q = 0.5

0.0153

0.0389

0.0732

HAR-RV-US vs. HAR-RV-states/q = 0.25

0.0030

0.0184

0.0729

HAR-RV-US vs. HAR-RV-states/q = 0.05

0.0002

0.0326

0.0746

 

Financials

HAR-RV vs. HAR-RV-states/q = 0.95

0.0416

0.1628

0.2245

HAR-RV vs. HAR-RV-states/q = 0.75

0.0490

0.0833

0.1458

HAR-RV vs. HAR-RV-states/q = 0.5

0.0271

0.0580

0.1083

HAR-RV vs. HAR-RV-states/q = 0.25

0.0017

0.0577

0.1209

HAR-RV vs. HAR-RV-states/q = 0.05

0.0267

0.1026

0.1075

HAR-RV-US vs. HAR-RV-states/q = 0.95

0.0410

0.1616

0.2195

HAR-RV-US vs. HAR-RV-states/q = 0.75

0.0487

0.0833

0.1456

HAR-RV-US vs. HAR-RV-states/q = 0.5

0.0270

0.0579

0.1081

HAR-RV-US vs. HAR-RV-states/q = 0.25

0.0017

0.0575

0.1206

HAR-RV-US vs. HAR-RV-states/q = 0.05

0.0218

0.1012

0.1072

 

Households

HAR-RV vs. HAR-RV-states/q = 0.95

0.0000

0.1586

0.1788

HAR-RV vs. HAR-RV-states/q = 0.75

0.0128

0.0844

0.1204

HAR-RV vs. HAR-RV-states/q = 0.5

0.0208

0.0434

0.1010

HAR-RV vs. HAR-RV-states/q = 0.25

0.0386

0.0457

0.0947

HAR-RV vs. HAR-RV-states/q = 0.05

0.0349

0.0556

0.1232

HAR-RV-US vs. HAR-RV-states/q = 0.95

− 0.0007

0.1574

0.1734

HAR-RV-US vs. HAR-RV-states/q = 0.75

0.0126

0.0844

0.1203

HAR-RV-US vs. HAR-RV-states/q = 0.5

0.0206

0.0433

0.1009

HAR-RV-US vs. HAR-RV-states/q = 0.25

0.0386

0.0455

0.0944

HAR-RV-US vs. HAR-RV-states/q = 0.05

0.0300

0.0541

0.1229

 

Labour market

HAR-RV vs. HAR-RV-states/q = 0.95

0.0000

0.1580

0.2217

HAR-RV vs. HAR-RV-states/q = 0.75

0.0000

0.0801

0.1318

HAR-RV vs. HAR-RV-states/q = 0.5

0.0279

0.0254

0.1002

HAR-RV vs. HAR-RV-states/q = 0.25

0.0364

0.0277

0.1013

HAR-RV vs. HAR-RV-states/q = 0.05

0.0307

0.0457

0.1295

HAR-RV-US vs. HAR-RV-states/q = 0.95

− 0.0007

0.1568

0.2166

HAR-RV-US vs. HAR-RV-states/q = 0.75

− 0.0003

0.0801

0.1317

HAR-RV-US vs. HAR-RV-states/q = 0.5

0.0278

0.0252

0.1001

HAR-RV-US vs. HAR-RV-states/q = 0.25

0.0364

0.0275

0.1010

HAR-RV-US vs. HAR-RV-states/q = 0.05

0.0258

0.0442

0.1292

 

Mobility

HAR-RV vs. HAR-RV-states/q = 0.95

0.0060

0.1409

0.2098

HAR-RV vs. HAR-RV-states/q = 0.75

0.0253

0.0660

0.1128

HAR-RV vs. HAR-RV-states/q = 0.5

0.0295

0.0249

0.0784

HAR-RV vs. HAR-RV-states/q = 0.25

0.0012

0.0427

0.0671

HAR-RV vs. HAR-RV-states/q = 0.05

0.0001

0.0947

0.1317

HAR-RV-US vs. HAR-RV-states/q = 0.95

0.0053

0.1396

0.2047

HAR-RV-US vs. HAR-RV-states/q = 0.75

0.0251

0.0659

0.1126

HAR-RV-US vs. HAR-RV-states/q = 0.5

0.0294

0.0247

0.0783

HAR-RV-US vs. HAR-RV-states/q = 0.25

0.0012

0.0425

0.0668

HAR-RV-US vs. HAR-RV-states/q = 0.05

− 0.0049

0.0933

0.1314

 

Real activity

HAR-RV vs. HAR-RV-states/q = 0.95

0.0280

0.1327

0.2037

HAR-RV vs. HAR-RV-states/q = 0.75

0.0219

0.0684

0.1026

HAR-RV vs. HAR-RV-states/q = 0.5

0.0193

0.0353

0.0892

HAR-RV vs. HAR-RV-states/q = 0.25

0.0000

0.0436

0.0829

HAR-RV vs. HAR-RV-states/q = 0.05

0.0000

0.0601

0.1284

HAR-RV-US vs. HAR-RV-states/q = 0.95

0.0273

0.1314

0.1985

HAR-RV-US vs. HAR-RV-states/q = 0.75

0.0216

0.0684

0.1025

HAR-RV-US vs. HAR-RV-states/q = 0.5

0.0192

0.0352

0.0891

HAR-RV-US vs. HAR-RV-states/q = 0.25

0.0000

0.0434

0.0826

HAR-RV-US vs. HAR-RV-states/q = 0.05

− 0.0050

0.0587

0.1281

  1. The relative-performance statistic, RP, statistic is computed as \(RP = 1 - \sum _{t=1}^T \rho _\alpha \left( e_{t, R} \right) / \sum _{t=1}^T \rho _\alpha \left( e_{t, B} \right)\), where \(e_t\) denotes the model prediction errors. The benchmark (B) model is the first model given in the first column of the table, and the rival (R) model is the second model given in that column. The HAR-RV-states model includes the state-level components in the vector of potential predictors. The benchmark model is estimated by the quantile-regression technique, while the HAR-RV-states model is estimated by the quantile Lasso technique. The intercept and the classic HAR-RV terms are not penalized. The penalty parameter is determined by tenfold cross-validation. A positive RP statistic shows that the rival model outperforms the benchmark model. The parameter h denotes the forecast horizon. The parameter q denotes the quantile being analyzed. The dependent variable is the natural log of the realized volatility of oil-price returns
Back to article page