Pith. sign in

REVIEW 5 major objections 4 minor 9 references

Reproducing kernel Hilbert space methods for modelling the discount curve

T0 review · 5 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper shows that polynomial-exponential kernel regression over fully consistent kernels yields discount curves that satisfy the no-arbitrage HJM drift condition, and that reducing such fits to a low-dimensional affine model gives a…

desk verdict The paper's main theorem fails its own definition: the proposed kernel slices do not vanish at the origin, so the claimed no-arbitrage consistency is unsupported. read the letter →

arxiv 2506.03342 v1 pith:HAS6GOO2 submitted 2025-06-03 q-fin.MF

Andreas Celary , Paul Krühner , Zehra Eksi This is my paper
classification q-fin.MF MSC 91G3046E2260H15
keywords HJMbonddiscountreproducingkernelHilbertspacetermstructuremodelsfinite-dimensionalrealizationsregressionno-arbitrageyieldcurve
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper proposes a way to fit discount curves, bond prices as a function of maturity, that are automatically consistent with no-arbitrage, using reproducing kernel Hilbert spaces whose kernels generate only admissible curves. It characterises a family of 'fully consistent' polynomial-exponential kernels, proves that every curve produced by kernel regression over these kernels satisfies the HJM drift condition, and tests the idea on a year of US Treasury data. In a second step the fitted high-dimensional curves are reduced to a low-dimensional affine model; the no-arbitrage condition fixes the drift of the driving process, leaving only the diffusion to be estimated from the data. The paper claims this yields a fully calibrated stochastic model whose simulated term structures look reasonable over the observed maturity range.

What carries the argument

The central object is the 'fully consistent' kernel: a kernel $k$ is fully consistent if any finite collection of slices $k_{y_i}(x)=k(y_i,x)$ is contained in a finite-dimensional, derivative-invariant function space. This property is equivalent to every slice having the quasi-exponential form $\varphi((1_{d+1} - e^{xM})e_0)$, which is exactly the admissible curve shape forced by the HJM drift condition under the linearity assumption. The kernels of Proposition 3.8, a nonnegatively-coefficient polynomial times the exponential kernel $e^{\beta xy - \alpha(x+y)}$, are shown to be fully consistent, and their RKHS is described explicitly as a weighted Taylor-series space. These kernels carry the argument because they make the infinite-dimensional curve fitting problem reduce to finite-dimensional ridge regression whose output is automatically admissible.

What would settle it

Compute the quadratic drift test on the extracted factor paths: for the reduced model, $Z_{t+1} - Z_t$ should match $(D + \langle\lambda, Z_t\rangle 1) Z_t \Delta t$ plus a martingale increment. A regression of $\Delta Z_t$ on the linear and quadratic terms $Z_t$ and $\langle\lambda, Z_t\rangle Z_t$ that rejects the theoretical coefficients, or residuals with tenor-dependent variance, would refute the claim that the fitted model is the consistent stochastic model.

Watch

Extended reading notes

Core claim

Under the linearity assumption the admissible discount curves are exactly the quasi-exponentials, of the form $1 - \langle e^{xM} e_0, Z \rangle$. The paper shows that kernels of the form $k(x,y) = p((\sqrt{\beta}x - \alpha/\sqrt{\beta})(\sqrt{\beta}y - \alpha/\sqrt{\beta})) e^{\beta xy - \alpha(x+y)}$ with $p$ a polynomial with nonnegative coefficients, and finite sums of such kernels, have the property that every kernel slice is such an admissible curve. Consequently, applying the Representer Theorem to these kernels produces, day by day, fitted curves that already satisfy the no-arbitrage drift condition, and the coefficients of the fit can be read as the realisation of the stochastic factor process $Z_t$. The calibrated two-step model therefore delivers both an arbitrage-free interpolation of market prices and a simulation scheme for future term structures.

Load-bearing premise

The paper assumes the fitted coefficient process $C_t$ is one realisation of the stochastic process $Z_t$ and therefore satisfies the quadratic no-arbitrage drift condition (11), but it never verifies this drift on the observed time series; if the fitted paths do not follow that drift, the 'calibrated consistent stochastic model' and its simulations are unsupported.

Editorial extensions

If this is right

  • Every fitted day's discount curve from the kernel regression is an admissible curve, so no-arbitrage is built into the interpolation rather than checked after fitting.
  • The no-arbitrage condition supplies the drift of the factor process analytically, so calibration reduces to estimating the diffusion matrix from the fitted coefficient paths.
  • Reduced models with around 20 exponential factors reproduce the full model and observed market prices to a negligible error on the training range, while being 4–6 times faster than a naive exponential regression.
  • Simulated paths of the factors and of bond prices over 252 days look reasonable for tenors up to about 25 years, enabling scenario simulation for arbitrary time frames within the learned maturity range.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A direct test of the paper's main empirical premise: regress the increments of the fitted coefficient paths on the quadratic terms $(\lambda_i + \langle\lambda, Z\rangle Z_i)\Delta t$ and check whether the coefficients match the theoretical drift; the paper does not perform this check.
  • The parameter set $\Theta = \{\alpha/\beta > y_{\max}\}$ proposed for bounded extrapolation suggests a regularisation scheme: restricting kernel parameters to $\Theta$ should improve long-tenor extrapolation at a possible in-sample cost, which is testable on the same data.
  • The time-inhomogeneous extension of the drift theorem indicates a route to maturity-dependent kernels; whether fully consistent time-inhomogeneous kernels exist is open.
  • If the affine model class is correct, the eigenvalues $\lambda_i$ estimated from different sub-samples should be stable; instability across sub-samples would indicate that the linearity assumption is too rigid for this dataset.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 4 minor

Summary. The paper studies the bond discount H_t(x)=1-P(t,t+x) in the HJM framework under finite-dimensional affine realizations. It derives drift conditions under NAFLVR, defines 'fully consistent' kernels whose slices should generate admissible no-arbitrage discount curves, and claims in Proposition 3.8 that a family of polynomial-exponential kernels has this property. The second half of the paper uses kernel ridge regression with these kernels on US Treasury data, then reduces the fitted high-dimensional curve to a low-dimensional affine model with a quadratic drift, estimates a constant diffusion matrix from the fitted coefficient paths, and simulates the term structure over 252 days. The central theoretical claim is that the polynomial-exponential kernels are fully consistent and that the subsequent two-step procedure yields a fully calibrated, arbitrage-free stochastic discount model.

Significance. If Proposition 3.8 were correct, the paper would offer a valuable bridge between RKHS curve fitting and no-arbitrage term structure modelling: a tractable regression basis with a guaranteed consistency property, followed by a reduction to a low-dimensional diffusion. The empirical pipeline is clearly described, uses real CRSP Treasury data, reports cross-validated parameter choices, presents sensitivity analysis, and explicitly acknowledges in the conclusion that global existence of the SDE is open. However, the main consistency theorem is false as stated, and the dynamic interpretation of the fitted coefficients is asserted rather than verified. These are load-bearing issues, not presentation defects.

major comments (5)
  1. [Section 3, Proposition 3.8 (with Proposition 3.4)] For p≡1 the kernel in Proposition 3.8 is k(x,y)=e^{βxy-α(x+y)}, and the slice k_y(x)=e^{-αy}e^{(βy-α)x} satisfies k_y(0)=e^{-αy}>0. However, by Proposition 3.4 every element of U has the form φ((I_{d+1}-e^{xM})e0), which vanishes at x=0 because e^{0M}=I. Hence k_y∉U, and the kernel is not fully consistent by the paper's own criterion. The proof of Proposition 3.8 only establishes that k_y is a polynomial-exponential function; it does not verify the boundary condition built into U. Since the regression basis in Section 4.1 is selected on the strength of this theorem, the no-arbitrage admissibility of the fitted curves is unsupported.
  2. [Section 4.2, full model construction] The text asserts that the fitted coefficient process C_t is 'one realisation of the stochastic process Z_t' and therefore that the full model satisfies the quadratic drift of Corollary 2.9, Eq. (11). This is not verified: the day-by-day regressions do not estimate or test the drift relation, the soft constraint h(0)=1 means the fitted curves may violate H_t(0)=0, and setting coordinates to zero for tenors that are absent on a given day is not shown to be compatible with the drift of any Z_t. Consequently the simulated paths in Figures 13-15 are not demonstrated to be paths of a no-arbitrage discount model.
  3. [Section 4.2, Proposition 4.5] Proposition 4.5 is the existence result for the reduced-model minimization, but its proof begins 'Without loss of generality, we may assume β=1, α=0 and d=1' and no reduction from general d to d=1 is given. Because E_d(k) is a union of d-dimensional subspaces rather than a fixed vector space, the weak-closedness argument for d=1 does not carry over automatically. The theoretical support for the numerical model-reduction step is therefore incomplete.
  4. [Section 4.2, reduced-model specification and Proposition 4.6] The reduced model is defined by eigenvalues λ_i in Eq. (29), but the manuscript never states how these exponents are selected. The matrices K'_t and K'' in Eq. (34) depend on λ_i, so the minimization problem leading to Proposition 4.6 is not fully specified without an exponent-selection rule. A reader cannot reproduce the reported reduced-model fits, and the claim that the model is 'fully parameterized' is not supported.
  5. [Section 5 and Conclusion] The conclusion states that the model can be calibrated 'for simulation purposes for arbitrary time frames', while Section 5 concedes that the global existence of the SDE is not clear because of the quadratic drift. The drift (11) is quadratic in Z_t, so finite-time explosion is a genuine concern, and the 252-day simulation horizon does not establish long-time behavior. This discrepancy should be resolved before the simulation claim is made.
minor comments (4)
  1. [Section 4.1, Proposition 4.2] The index set is written as I1 = {1≥i≥M}, which should be {i∈{1,...,M}: w_i=∞}; the convention λ/∞:=0 is also informal.
  2. [Section 2.5, Proposition 2.5] The process is written as (\tilde P(t,T))_{0≥t≥T}; the intended index range is 0≤t≤T.
  3. [Throughout] Cross-references are inconsistent with the labels of the cited results: for example, 'Theorem 2.8' for Corollary 2.8, 'Theorem 3.2' for Definition 3.2, 'Theorem 3.7' for Lemma 3.7, and 'Theorem 2.9' for Corollary 2.9.
  4. [Section 4.1 and Figure 7] The soft constraint for h(0)=1 is introduced, but the realized deviations h_t(0)-1 across the sample are not reported; given the high sensitivity of yields to small maturity-zero errors, this quantity would help assess how close the fitted curves are to the bond-price boundary condition.

Circularity Check

2 steps flagged · score 6.0 of 10

Prop 3.8's 'fully consistent' kernel is made admissible by construction: the proof equates full consistency with the polynomial-exponential slice form (46) the kernel was built to have, while Prop 3.4 requires k_y(0)=0 and in fact k_y(0)=e^{−αy}p(−αy+α²/β)≠0; Section 4.2 then renames the fitted coefficients C_t as 'one realisation' of the no-arbitrage process Z_t.

  1. self definitional [Prop. 3.8 (Sec. 3.1) and proof in App. B, 'Proof of Theorem 3.8', Eqs. (46)-(47); criterion in Prop. 3.4]
    "It therefore remains to be shown that k is fully consistent in the sense of Theorem 3.2. For this to hold, we must have that k(·, y) ∈ U for all y ∈ R_+. By the Jordan normal form, any element g ∈ U can be written as a sum of products of polynomials with exponentials. Therefore, observe that k must be of the form (46)... We may set q_i(y) := e^{−αy}b_i(y) and λ(y) := √βy − α which shows that k is of the desired form and thus indeed fully consistent."

    Under Prop. 3.4, full consistency means k_y ∈ U, and every element of U = {φ((1−e^{xM})e_0)} vanishes at x = 0 because (1−e^{0·M})e_0 = 0. The proof substitutes the weaker property 'k_y is a polynomial-exponential of the form (46)' for U-membership; but that slice form is exactly what the Lemma 3.7 construction (h(t) := e^{−α²/β}p(t)e^t) was chosen to produce, so the conclusion is packed into the ansatz. Against the paper's own Prop. 3.4 the claim fails: k_y(0) = e^{−αy}p(−αy + α²/β), which for p ≡ 1 equals e^{−αy} ≠ 0, hence k_y ∉ U. The missing boundary condition (H(0)=0, i.e. h(0)=1) is later reintroduced only as a 'soft constraint' in the regression (Sec. 4.1).

  2. fitted input called prediction [Sec. 4.2, 'Second step optimisation': identification of fitted C_t with Z_t; drift claim]
    "Indeed, with this specification, we obtain a consistent M-dimensional model H := 1 − Ĥ of the type specified in Theorem 2.9, where M ≤ T max_{t≥0} M_t, K ≡ g and C_t is one realisation of the stochastic process Z_t, that is C_t = Z_t(ω_0) for some ω_0 ∈ Ω."

    The coefficients C_t and the exponents come from the two-step fit to market data; the paper then declares them to be 'one realisation of the stochastic process Z_t' of Theorem 2.9, so the quadratic no-arbitrage drift (11) is attributed to the fitted series 'due to the no-arbitrage quadratic drift condition' without any check that the fitted paths satisfy it. The consistency of the 'full' and 'reduced' stochastic models — the paper's central numerical claim — thus rests on renaming the fitted output as the theoretical process. The covariance used for simulation is extracted from the same fitted paths, and the simulated paths are validated against them in-sample; the static fit itself only imposes h(0)=1 as a soft constraint, so the asserted Theorem 2.9 realisation is not verified.

full rationale

No self-citation chain is load-bearing: the HJM discount-drift theory is credited to [Fil23] (no author overlap), the RKHS machinery to [PR16]/[Aro50], and the numerics are benchmarked against CRSP Treasury data with reported errors, so most of the pipeline is externally supported. The circularity is concentrated in the central theoretical claim. Prop. 3.4 defines a fully consistent kernel by k_y ∈ U, and U forces k_y(0)=0; the proof of Prop. 3.8, however, concludes full consistency directly from the polynomial-exponential slice form (46) — the very property the kernel was constructed to have — and the paper's own equations give k_y(0)=e^{−αy}≠0 for p≡1, so the slice is not in U. Hence the advertised result (kernels 'admissible under no-arbitrage') holds in the proof only because admissibility is identified with the ansatz; the distinguishing boundary condition re-enters only as a soft constraint in the numerics, where the paper notes deviations at tenor 0 cause large yield errors. Downstream, Sec. 4.2 identifies the fitted coefficients C_t with the no-arbitrage process Z_t of Theorem 2.9 and 'obtains' the quadratic drift (11) without testing it on the fitted paths; the covariance used for simulation is bootstrapped from the same paths, and the simulated paths are compared back to them over the in-sample 252-day window. The admitted limitation that 'the global existence of the SDE for consistent discount models is not entirely clear' is a correctness risk rather than circularity. Overall: the central admissibility claim reduces to the quasi-exponential construction (partial circularity), while the RKHS characterisation (Lemma 3.7), the externally sourced drift theory, and the empirical fitting retain independent content; hence score 6 rather than 8-10.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The theoretical core relies on standard domain assumptions: the linearity assumption (LA) in Section 2.5, the NAFLVR no-arbitrage notion from CKT16 and Fil23, and the Hilbert space conditions (H1) and (H2). The empirical stochastic model adds an unverified assertion that fitted coefficients follow the quadratic drift (11), and the numerical results rely on cross-validated tuning parameters whose optimization is only partially described.

free parameters (6)
  • kernel parameter alpha = 0.2
    Chosen by cross-validation in Section 4.1; shifts the exponential kernel and controls decay.
  • kernel parameter beta = 0.04
    Chosen by cross-validation; controls exponential growth and is highly sensitive in the authors' sensitivity analysis.
  • ridge parameter lambda = 0.001
    Chosen by cross-validation; penalizes the RKHS norm in the regression.
  • reduced model dimension d = explored 1 to 30, 20 or more chosen visually
    No formal selection criterion; the paper reports fitted curves become visually indistinguishable around d=20.
  • reduced model exponents lambda_i = not reported
    The optimization over exponents is not described in the numerical procedure, so the reduced model calibration depends on an unspecified selection.
  • diffusion matrix sigma = empirical covariance of fitted coefficients
    Used as a constant diffusion in the simulation step without checking whether the fitted paths satisfy the quadratic no-arbitrage drift.
assumptions (5)
  • domain assumption Discount curves lie in a finite-dimensional affine subspace, Assumption (LA) in Section 2.5
    This forces admissible curves to be quasi-exponentials; without it the kernel consistency construction does not apply.
  • domain assumption No asymptotic free lunch with vanishing risk (NAFLVR) is the no-arbitrage notion
    Adopted from CKT16 and Fil23; the HJM drift condition in Proposition 2.2 is proved under this notion.
  • standard math The Hilbert space H satisfies shift semigroup and continuous evaluation, conditions (H1) and (H2)
    Needed for the Musiela SDE formulation and the drift condition; standard in HJM theory.
  • ad hoc to paper The set of diagonalisable matrices is dense, so restricting the reduced model to diagonal M is justified
    Invoked in Section 4.2; density does not by itself justify the approximation quality for a finite sample.
  • ad hoc to paper Fitted coefficient sequence C_t is a realization of the process Z_t satisfying the quadratic drift (11)
    Used to claim the full model is a consistent stochastic model; this is asserted rather than verified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Reproducing kernel Hilbert space methods for modelling the discount curve." pith.science (2026). https://pith.science/paper/HAS6GOO2

@misc{pith2026250603342,
  author       = {Pith},
  title        = {Pith review of: Reproducing kernel Hilbert space methods for modelling the discount curve},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HAS6GOO2}},
  note         = {Machine review of arXiv:2506.03342}
}
read the original abstract

We consider the theory of bond discounts, defined as the difference between the terminal payoff of the contract and its current price. Working in the setting of finite-dimensional realizations in the HJM framework, under suitable notions of no-arbitrage, the admissible discount curves take the form of polynomial, exponential functions. We introduce reproducing kernels that are admissible under no-arbitrage as a tractable regression basis for the estimation problem in calibrating the model to market data. We provide a thorough numerical analysis using real-world treasury data.

Figures

Figures reproduced from arXiv: 2506.03342 by the authors.

Figure 1
Figure 1. Bond prices and implied yield on 31st of December, 2021 The contracts offered on the market are coupon bonds with prices quoted on each trading day. Any coupon bond can be written as a linear combination of zero￾coupon bonds with different times to maturity multiplied with respective coupons. That is, for any coupon bond P, we have P = X N i=1 Cih(xi), where {x1, . . . , xN } is a collection of tenors, Ci denotes th… view at source ↗
Figure 2
Figure 2. Cashflow matrix extracted from coupon bond on the 31st of December, 2021 4.1. First step optimisation. Given a dataset with zero-bond prices and cashflow matrices, our aim is to fit a zero-coupon bond curve h, which will minimize the pricing error. To this end, we shall fix an appropriate curve space from which to draw our curve. Since we are interested in models fulfilling no-arbitrage, we will use a fully consiste… view at source ↗
Figure 3
Figure 3. Observed data against fitted data on the 31st of De￾cember, 2021 [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗
Figures from the paper (13 more)
Figure 4
Figure 4. Figure 4: Implied zero-coupon price curve and yield curve on the 31st of December, 2021 We also perform a sensitivity analysis, where we take the parameter set for the best fit and calibrate the model several times while varying the parameters in a range of 20% of the original v…
Figure 5
Figure 5. Figure 5: root mean square error of fitted yields across all con￾tracts on each trading day in the observed time window [PITH_FULL_IMAGE:figures/full_fig_p017_5.png]
Figure 6
Figure 6. Figure 6: Sensitivity of the model with respect to change in pa￾rameters. In each plot, one of the parameters is kept fixed. 4.2. Second step optimisation. We consider the results we obtained in the first step optimisation. For each observation day, we fitted a bond-price curve …
Figure 7
Figure 7. Figure 7: Curves produced by reduced models with full model (blue) on the 31st of December, 2021 see that the curves become indistinguishable to the eye starting around dimension d = 20. In particular, the shape of the curve seems to be captured very well in the minimisation wit…
Figure 8
Figure 8. Figure 8: Price fits on reduced models on the 31st of December, 2021 reduced model to observe how well the real market bond prices are approximated using our model. The prices in data set are plotted in a scatter plot against the model implied prices, along with the identity lin…
Figure 9
Figure 9. Figure 9: Relative pricing errors of reduced models (truncated at 100%) on the 31st of December, 2021 mean square errors of fitted yields across all contracts on a given trading as a time series evolution on all days in our dataset for the reduced models. We again observe [PITH…
Figure 10
Figure 10. Figure 10: Root mean square errors of fitted yields on each trad￾ing day in the observed time window that the higher-dimensional reduced models produce a very good fit while the lower￾dimensional models have some errors. Again, these occur mainly in the long-term contracts. The …
Figure 11
Figure 11. Figure 11: Comparison of the RMSE of the best performing mod￾els to the full model (blue) To conclude our analysis, we consider the stochastic driving factors inherent to our model. Indeed, we have already observed a high amount of redundancy in going from the full model to the …
Figure 12
Figure 12. Figure 12: Estimated covariance matrices of extracted stochastic coefficients. to recognise which factors are negligible. It seems some of the correlated factors perform the task of a control factor for the higher exponents in the model. When left out, this leads to the high err…
Figure 13
Figure 13. Figure 13: Extracted paths vs. simulated paths of stochastic processes calibrated using the covariance data normalised by their starting values. model, it is now easy to simulate the entire term structure for arbitrary time frames. However, due to the observed bad extrapolation …
Figure 14
Figure 14. Figure 14: Simulation of time series of bond prices and yields for contracts with several different tenors. We observe that for longer tenors, the time series becomes more volatile. This is mainly due to fall in performance of the model for long tenors. Indeed, we notice a deter…
Figure 15
Figure 15. Figure 15: Simulation of bond price curves for every observation point in the time frame. bounded for any fixed y ≥ 0 if and only if βxy − α(x + y) < 0. Assuming x, y > 0, we may write α  1 x + 1 y  > β. As x → ∞, we may rewrite this as α β > y. Since the inequality has to hol…
Figure 16
Figure 16. Figure 16 [PITH_FULL_IMAGE:figures/full_fig_p028_16.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

9 extracted references · 9 canonical work pages

  1. [1]

    Theory of Reproducing Kernels

    [Aro50] Nachman Aronszajn. “Theory of Reproducing Kernels”. In:Transac- tions of the American Mathematical Society68.3 (1950), pp. 337–404. url: http://www.jstor.org/stable/1990404 (visited on 04/01/2023). [Bj¨ o04] Tomas Bj¨ ork. “On the Geometry of Interest Rate Models”. In:Paris- Princeton Lectures on Mathematical Finance

    work page arXiv 1950

  2. [2]

    Function Analysis, Sobolev Spaces and Partial Differen- tial Equations

    [Bre10] Haim Brezis. “Function Analysis, Sobolev Spaces and Partial Differen- tial Equations”. Springer-Verlag, Jan. 2010.isbn: 978-0-387-70913-0. [Car61] Frank W. Carroll. “A Polynomial in Each Variable Separately is a Poly- nomial”. In:The American Mathematical Monthly68.1 (1961), pp. 42– 42.issn: 00029890, 19300972.url: http : / / www . jstor . org / s...

    work page 1961

  3. [704]

    Real Analysis: Modern Techniques and Their Ap- plications

    eprint: https : / / onlinelibrary. wiley. com / doi / pdf / 10 . 1111 / jofi . 42 REFERENCES 12488.url: https : / / onlinelibrary. wiley. com / doi / abs / 10 . 1111 / jofi . 12488. [Fol13] Gerald B. Folland. “Real Analysis: Modern Techniques and Their Ap- plications”. Pure and Applied Mathematics: A Wiley Series of Texts, Monographs and Tracts. Wiley, 20...

    work page 2022

  4. [1760]

    Term-Structure Models: A Graduate Course

    Springer, 2001.url: https://infoscience.epfl.ch/handle/20.500.14299/49680. [Fil09] Damir Filipovi´ c. “Term-Structure Models: A Graduate Course”. Springer Finance. Springer Berlin Heidelberg, 2009.isbn: 9783540680154.url: https://books.google.at/books?id=KqcSh6CavaAC. [Fil23] Damir Filipovi´ c. “Discount Models”. In:Finance and Stochastics. Swiss Finance ...

    work page 2023

  5. [1992]

    Markov Processes: Charac- terization and Convergence

    [EK09] Stewart N. Ethier and Thomas G. Kurtz. “Markov Processes: Charac- terization and Convergence”. Wiley Series in Probability and Statistics. Wiley, 2009.isbn: 9780470317327.url: https://books.google.at/books? id=zvE9RFouKoMC. [Fil00a] Damir Filipovi´ c. “Consistency problems for HJM interest rate models”. en. Diss. Mathematische Wissenschaften ETH Z¨...

    work page 2009

  6. [2000]

    Invariant manifolds for weak solutions to stochas- tic equations

    [Fil00b] Damir Filipovi´ c. “Invariant manifolds for weak solutions to stochas- tic equations”. In:Probability Theory and Related Fields118 (2000), pp. 323–341.url: https://api.semanticscholar.org/CorpusID:6968887. [Fil01] Damir Filipovi´ c. “Consistency Problems for Heath-Jarrow-Morton In- terest Rate Models”. Lecture Notes in Mathematics

    work page 2000

  7. [2003]

    by Ren´ e A

    Ed. by Ren´ e A. Car- mona et al. Springer Berlin Heidelberg, 2004, pp. 133–215.url: https: //doi.org/10.1007/978-3-540-44468-8

    work page doi:10.1007/978-3-540-44468-8 2004

  8. [2016]

    Fitting Dynamically Consistent For- ward Rate Curves: Algorithm and Comparison

    [WJ24] David Wu and Robert Jarrow. “Fitting Dynamically Consistent For- ward Rate Curves: Algorithm and Comparison”. In:International Jour- nal of Theoretical and Applied Finance(2024).url: https://doi.org/ 10.1142/S0219024924500213

    work page doi:10.1142/s0219024924500213 2024

Show all 9 references
  1. [2882]

    Bond Pricing and the Term Structure of Interest Rates: A New Methodology for Contin- gent Claims Valuation

    [HJM92] David Heath, Robert Jarrow, and Andrew Morton. “Bond Pricing and the Term Structure of Interest Rates: A New Methodology for Contin- gent Claims Valuation”. In:Econometrica60.1 (1992), pp. 77–105.url: https://ideas.repec.org/a/ecm/emetrp/v60y1992i1p77-105.html. [JS87] ...

    1992

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.