{
 "bundle_of": "C3",
 "schema_version": "0.43",
 "paper": {
  "id": "entrylib:2026.000011",
  "version": 1,
  "title": "Multi-step estimators and shrinkage effect in time series models",
  "licence": "CC-BY-4.0",
  "status": "active",
  "date": "2026-11-16"
 },
 "claims": [
  {
   "about": {
    "comparator": "D2",
    "conditions": [
     "correctly specified model"
    ],
    "object": "parameter estimates",
    "relation": "yields parameter estimates identical to",
    "subject": "D7"
   },
   "assumptions": [
    "C19",
    "C21"
   ],
   "depends_on": [
    "C7",
    "C13"
   ],
   "id": "C3",
   "limitations": [
    "L3"
   ],
   "novelty": "no_prior_found",
   "provenance": "author_stated",
   "relations": [
    {
     "id_type": "doi",
     "locator": "p. 77",
     "note": "Clements and Hendry (1998) formulate the multi-step predictive log-likelihood and note diagonal covariance under exact assumptions.",
     "reference": "B5",
     "type": "extends"
    }
   ],
   "role": "principal",
   "scope": {
    "conditions": [
     "correctly specified linear state space models with Gaussian errors"
    ],
    "excludes": [
     "misspecified models where |\\hat{A}| \\neq 1"
    ]
   },
   "statement": "Maximising the [[D7:General Predictive Likelihood]] yields parameter estimates identical to those of the conventional one-step-ahead estimator [[D2:MSE_1]] under correct model specification.",
   "support": [
    {
     "evidence": "E2",
     "kind": "formal_proof"
    }
   ],
   "type": "formal",
   "verification": "unchecked",
   "verifiability": "open",
   "verifiability_flags": []
  },
  {
   "about": {
    "comparator": null,
    "conditions": [
     "D1",
     "C19"
    ],
    "object": "cross-horizon forecast error covariance",
    "relation": "has non-zero cross-horizon covariance formula",
    "subject": "D1"
   },
   "assumptions": [
    "C19"
   ],
   "id": "C7",
   "novelty": "no_prior_found",
   "provenance": "author_stated",
   "relations": [
    {
     "id_type": "doi",
     "locator": "p. 11",
     "note": "Martinez (2017) assumes zero correlation between multi-step forecast errors when assumptions hold.",
     "reference": "B8",
     "type": "contradicts"
    }
   ],
   "role": "supporting",
   "statement": "In an [[D1:SSOE state space model]] with independent homoscedastic errors, the covariance $\\sigma_{i,j}$ between $i$-step-ahead and $j$-step-ahead forecast errors is given by $$\\sigma_{i,j} = \\begin{cases} \\sigma_1^2 c_{|i-j|} & \\text{when } i=1 \\text{ or } j=1 \\\\ \\sigma_1^2 \\left(c_{|i-j|} + \\sum_{l=1}^{i-1} c_{|l-j|} c_{|i-l|}\\right) & \\text{otherwise} \\end{cases}$$ so multi-step forecast errors have non-zero cross-horizon covariance whenever $c_j \\ne 0$.",
   "support": [
    {
     "evidence": "E1",
     "kind": "formal_proof"
    }
   ],
   "type": "formal",
   "verification": "unchecked",
   "verifiability": "open",
   "verifiability_flags": []
  },
  {
   "about": {
    "comparator": null,
    "conditions": [
     "correctly specified model"
    ],
    "object": "determinant of matrix \\hat{A}",
    "relation": "decomposes log GV and satisfies |\\hat{A}| = 1",
    "subject": "D8"
   },
   "assumptions": [
    "C19",
    "C21"
   ],
   "id": "C13",
   "novelty": "no_prior_found",
   "provenance": "author_stated",
   "role": "supporting",
   "statement": "The logarithm of the [[D8:Generalised Variance]] decomposes as $\\log GV = \\log s_1^2 + \\log |\\hat{A}|$, where $|\\hat{A}| = 1$ under correct model specification.",
   "support": [
    {
     "evidence": "E2",
     "kind": "formal_proof"
    }
   ],
   "type": "formal",
   "verification": "unchecked",
   "verifiability": "open",
   "verifiability_flags": []
  },
  {
   "assumption_category": "statistical",
   "check": "unchecked",
   "id": "C19",
   "provenance": "author_stated",
   "role": "supporting",
   "statement": "The one-step-ahead innovation errors $\\epsilon_t$ of the state space model are independent and identically distributed with zero mean and constant variance $\\sigma_1^2$.",
   "support": [
    {
     "kind": "assumption"
    }
   ],
   "type": "formal",
   "verifiability": "not_applicable",
   "verifiability_flags": []
  },
  {
   "assumption_category": "statistical",
   "check": "unchecked",
   "id": "C21",
   "provenance": "author_stated",
   "role": "supporting",
   "statement": "The one-step-ahead innovation errors $\\epsilon_t$ follow a Normal distribution $\\mathcal{N}(0, \\sigma_1^2)$.",
   "support": [
    {
     "kind": "assumption"
    }
   ],
   "type": "formal",
   "verifiability": "not_applicable",
   "verifiability_flags": []
  }
 ],
 "definitions": [
  {
   "definition": "An additive-error linear state space model of the form $$\\begin{aligned} y_t &= w' v_{t-1} + \\epsilon_t \\\\ v_t &= F v_{t-1} + g \\epsilon_t \\end{aligned}$$ where $y_t$ is the observed value at time $t$, $v_t$ is the state vector, $F$ is the transition matrix, $g$ is the persistence vector, $w$ is the measurement vector, and $\\epsilon_t \\sim \\text{i.i.d. } \\mathcal{N}(0, \\sigma_1^2)$, with multi-step coefficients $c_j = w' F^{j-1} g$.",
   "id": "D1",
   "kind": "theory",
   "source": "B1",
   "term": "SSOE state space model"
  },
  {
   "definition": "Conventional one-step-ahead mean squared error estimator minimising $$MSE_1 = \\frac{1}{T-1} \\sum_{t=1}^{T-1} e_{t+1|t}^2$$ where $e_{t+1|t} = y_{t+1} - \\hat{\\mu}_{t+1|t}$ is the one-step-ahead in-sample forecast error over $T$ observations.",
   "id": "D2",
   "kind": "method",
   "term": "MSE_1"
  },
  {
   "definition": "Geometric Trace Mean Squared Error estimator introduced in this paper, minimising the sum of logarithmic multi-step mean squared errors across horizons $1$ to $h$: $$GTMSE = \\sum_{j=1}^{h} \\log \\left( \\frac{1}{T-h} \\sum_{t=1}^{T-h} e_{t+j|t}^2 \\right)$$",
   "id": "D6",
   "kind": "method",
   "term": "GTMSE"
  },
  {
   "definition": "General Predictive Likelihood for consecutive observations $Y_t = (y_{t+1}, \\dots, y_{t+h})'$ under Gaussian errors, whose concentrated log-likelihood is $$l_{GPL_h}(\\theta, \\hat{\\Sigma} \\mid Y) = -\\frac{T-h}{2} \\left( h \\log(2\\pi e) + \\log|\\hat{\\Sigma}| \\right)$$ where $\\hat{\\Sigma} = \\frac{1}{T-h} \\sum_{t=1}^{T-h} E_t E_t'$ is the sample multi-step forecast error covariance matrix.",
   "id": "D7",
   "kind": "method",
   "term": "General Predictive Likelihood"
  },
  {
   "definition": "Determinant of the sample multi-step forecast error covariance matrix, $$GV = |\\hat{\\Sigma}|$$ whose minimisation is equivalent to maximising the [[D7:General Predictive Likelihood]].",
   "id": "D8",
   "kind": "method",
   "term": "Generalised Variance"
  }
 ],
 "evidence": [
  {
   "availability": {
    "level": "public"
   },
   "description": "Analytical derivation in the single source of error (SSOE) state space framework showing that multi-step loss functions minimise multi-step forecast error variances and covariances, shrinking dynamic state-space parameters toward zero (and regression parameters not at all because $g=0$), along with specific derivations for ARIMA(1,1,1) and pure additive ETS models (Sections 2, 3, and Appendices A–B).",
   "foundation": "Kolmogorov probability axioms and linear algebra",
   "id": "E1",
   "profile": "formal",
   "steps": [
    {
     "by": [
      "D1"
     ],
     "id": "S1",
     "rule": "definition",
     "statement": "\\eta_{t+h|t} = y_{t+h} - \\mu_{t+h|t} = \\sum_{j=1}^{h-1} c_j \\epsilon_{t+j} + \\epsilon_{t+h}, \\quad c_j = w' F^{j-1} g"
    },
    {
     "by": [
      "S1",
      "C19"
     ],
     "id": "S2",
     "rule": "variance and covariance of linear combinations",
     "statement": "\\sigma_h^2 = \\sigma_1^2 \\left(1 + \\sum_{j=1}^{h-1} c_j^2\\right), \\quad \\sigma_{i,j} = \\sigma_1^2 \\left(c_{|i-j|} + \\sum_{l=1}^{i-1} c_{|l-j|} c_{|i-l|}\\right) \\text{ for } i,j > 1"
    },
    {
     "by": [
      "S2"
     ],
     "id": "S3",
     "rule": "sample substitution and minimisation",
     "statement": "\\text{Minimising } MSE_h, TMSE, \\text{ or } MSCE \\text{ minimises } s_1^2 \\text{ and } \\hat{c}_j^2 \\text{ toward } 0 \\text{ for } h > 1, \\text{ whereas } g=0 \\text{ in static regression implies } c_j = 0 \\text{ and no shrinkage.}"
    }
   ]
  },
  {
   "availability": {
    "level": "public"
   },
   "description": "Analytical derivation of concentrated predictive log-likelihoods for MSE_h, MSCE, and the General Predictive Likelihood (GPL), proving that log GV = log s_1^2 + log |\\hat{A}| and that |\\hat{A}| = 1 under correct specification, and defining GTMSE via diagonal restriction on \\hat{\\Sigma} (Section 4 and Appendices C–E).",
   "foundation": "Kolmogorov probability axioms and multivariate normal likelihood theory",
   "id": "E2",
   "profile": "formal",
   "steps": [
    {
     "by": [
      "D7",
      "C21"
     ],
     "id": "S1",
     "rule": "concentrating multivariate normal log-likelihood",
     "statement": "l_{GPL_h}(\\theta, \\hat{\\Sigma} \\mid Y) = -\\frac{T-h}{2}\\left(h \\log(2\\pi e) + \\log|\\hat{\\Sigma}|\\right), \\quad GV = |\\hat{\\Sigma}|"
    },
    {
     "by": [
      "S1"
     ],
     "id": "S2",
     "rule": "matrix determinant reduction",
     "statement": "\\log GV = \\log s_1^2 + \\log|\\hat{A}|, \\quad |\\hat{A}| = 1 \\text{ under correct model specification}"
    },
    {
     "by": [
      "S1",
      "D6"
     ],
     "id": "S3",
     "rule": "diagonal restriction and logarithmic scaling",
     "statement": "GTMSE \\propto \\sum_{j=1}^{h} \\log(s_j^2) = h \\log(s_1^2) + \\sum_{j=2}^{h} \\log\\left(1 + \\sum_{i=1}^{j-1} \\hat{c}_i^2\\right)"
    }
   ]
  }
 ],
 "limitations": [
  {
   "affects": [
    "C3"
   ],
   "category": "model_specification",
   "id": "L3",
   "impact": "could_reverse",
   "mitigation": "none",
   "statement": "The equivalence between General Predictive Likelihood and one-step-ahead MSE_1 relies on correct model specification so that $|\\hat{A}| = 1$; under model misspecification, off-diagonal covariances do not cancel and GPL estimates diverge from MSE_1."
  }
 ],
 "references": [
  {
   "id": "B1",
   "id_type": "doi",
   "target": "10.1111/j.2517-6161.1985.tb01355.x"
  }
 ]
}
