Paper: arXiv 2307.01779

Authors: Giuseppe Cavaliere, Thomas Mikosch, Anders Rahbek, Frederik Vilandt

Abstract

Engle and Russell (1998, Econometrica, 66:1127–1162) apply results from the GARCH literature to prove consistency and asymptotic normality of the (exponential) QMLE for the generalized autoregressive conditional duration (ACD) model, the so-called ACD(1,1), under the assumption of strict stationarity and ergodicity. The GARCH results, however, do not account for the fact that the number of durations over a given observation period is random. Thus, in contrast with Engle and Russell (1998), we show that strict stationarity and ergodicity alone are not sufficient for consistency and asymptotic normality, and provide additional sufficient conditions to account for the random number of durations. In particular, we argue that the durations need to satisfy the stronger requirement that they have finite mean.

Complexity vs Empirical Score

  • Math Complexity: 9.0/10
  • Empirical Rigor: 2.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper is densely packed with advanced asymptotic theory, stochastic processes, and proofs (e.g., central limit theorems for random sample sizes, functional convergence to Brownian motion), earning a high math score. It lacks any empirical implementation, backtesting, or data analysis, focusing solely on theoretical asymptotic results, resulting in low empirical rigor.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Asymptotic Analysis of ACD Model"] --> B{"Key Assumption<br>Strict Stationarity & Ergodicity?"}
  B -->|"Engle & Russell 1998"| C["Apply GARCH Results<br>Assume Finite Mean"]
  B -->|"This Paper"| D["Identify Flaw:<br>Random N of Durations"]
  D --> E["Derive Sufficient<br>Conditions: Finite Mean Required"]
  E --> F["Prove Asymptotic<br>Properties<br>Consistency & Normality"]
  C --> G["Outcome:<br>Validates QMLE in ACD Model"]
  F --> G
  G --> H["Key Contribution:<br>Generalized Theory for ACD"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes additional sufficient conditions for consistency and asymptotic normality of the QMLE in the ACD(1,1) model, specifically requiring that durations have a finite unconditional mean.

Source sentence
Thus, in contrast with Engle and Russell (1998), we show that strict stationarity and ergodicity alone are not sufficient for consistency and asymptotic normality, and provide additional sufficient conditions to account for the random number of durations. In particular, we argue that the durations need to satisfy the stronger requirement that they have finite mean.

Assumes 4

  • The ACD process is strictly stationary and ergodic. model
    Source sentence
    Engle and Russell (1998, Econometrica, 66:1127--1162) apply results from the GARCH literature to prove consistency and asymptotic normality of the (exponential) QMLE for the generalized autoregressive conditional duration (ACD) model, the so-called ACD(1,1), under the assumption of strict stationarity and ergodicity.
  • Durations have a finite unconditional expectation (α₀+β₀ < 1). statistical
    Source sentence
    In this note, by using a new Lemma which extends Lee and Hansen (1994) to the case of a random number of observations, we show that under the additional more restrictive condition α0 + β0 < 1 (or, equivalently, that the durations have finite unconditional expectation) the QMLE is consistent and asymptotically normal at the standard √ T rate. Theorem 2.1 For the ACD model (1.1)-(1.2) with true parameter value denoted
  • The i.i.d. innovations εᵢ have support (0,∞), pdf bounded away from zero on compact subsets, E[εᵢ]=1 and finite second moment. statistical
    Source sentence
    Theorem 2.1 For the ACD model (1.1)-(1.2) with true parameter value denoted by θ0, if: (i) {εi} is an i.i.d. sequence of r.v.s with support (0, ∞), pdf fε (·) bounded away from zero on compact subsets of (0, ∞), E[εi] = 1 and E[ε2 i ] < ∞, (ii) α0 + β0 < 1, then with θ0 an interior point, the maximizer of Ln(T) (θ) in (1.3) will be consistent and asymptotically normal at the standard √ T-rate,
  • The random number of durations n(T) satisfies n(T)/T → 1/μ almost surely, with μ = E[xᵢ] < ∞. statistical
    Source sentence
    The key additional condition needed is that the random number of durations n(T) satisfies n(T)/T a.s. →1/µ , with µ = E (xi) . Here µ = E (xi) < ∞, and ΩI = E[ζi], ΩS = E[ξiξ′ i] are given by (1.7) and (1.6) respectively.

Data

No dataset named in the text.

Costs

Not applicable The paper reports no trading or execution result.

Finds 2

  • Under the additional condition that durations have a finite mean (α₀+β₀ < 1), the QMLE is consistent and asymptotically normal at the standard √T rate. theoretical
    Source sentence
    In this note, by using a new Lemma which extends Lee and Hansen (1994) to the case of a random number of observations, we show that under the additional more restrictive condition α0 + β0 < 1 (or, equivalently, that the durations have finite unconditional expectation) the QMLE is consistent and asymptotically normal at the standard √ T rate. Theorem 2.1 For the ACD model (1.1)-(1.2) with true parameter value denoted
  • The random sampling effect is handled by showing that n(T)/T → 1/μ a.s. ensures the deterministic LLN and CLT extend to the random‑n case. theoretical
    Source sentence
    The key additional condition needed is that the random number of durations n(T) satisfies n(T)/T a.s. →1/µ , with µ = E (xi) . (1.8) This in turn (as n(T) →∞a.s.) is sufficient for the deterministic n LLN in (1.7) to imply that its random n analogue (1.5) holds.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 8 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.