Paper: arXiv 2309.02970

Authors: Christian Oliver Ewald, Kevin Kamm

Abstract

We study the effect of stochastic feeding costs on animal-based commodities with particular focus on aquaculture. More specifically, we use soybean futures to infer on the stochastic behaviour of salmon feed, which we assume to follow a Schwartz-2-factor model. We compare the decision of harvesting salmon using a decision rule assuming either deterministic or stochastic feeding costs, i.e. including feeding cost risk. We identify cases, where accounting for stochastic feeding costs leads to significant improvements as well as cases where deterministic feeding costs are a good enough proxy. Nevertheless, in all of these cases, the newly derived rules show superior performance, while the additional computational costs are negligible. From a methodological point of view, we demonstrate how to use Deep-Neural-Networks to infer on the decision boundary that determines harvesting or continuation, improving on more classical regression-based and curve-fitting methods. To achieve this we use a deep classifier, which not only improves on previous results but also scales well for higher dimensional problems, and in addition mitigates effects due to model uncertainty, which we identify in this article. effects due to model uncertainty, which we identify in this article.

Complexity vs Empirical Score

  • Math Complexity: 8.0/10
  • Empirical Rigor: 7.5/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced stochastic calculus (SDEs for Schwartz-2-factor models) and deep learning for optimal stopping, resulting in high mathematical complexity. It demonstrates strong empirical rigor through the use of real market data (soybean futures), extensive calibration algorithms (Kalman filter and nested minimization), and includes code and data availability for reproducible backtesting.

Research Flowchart

  flowchart TD
  A["Research Goal: <br>Assess impact of stochastic feeding costs <br>on aquaculture valuation & decisions"] --> B["Model Inputs: <br>Stochastic Soybean Futures & Salmon Feed <br>via Schwartz 2-Factor Model"]
  B --> C["Methodology: <br>Deep Neural Network Classifier"]
  C --> D{"Decision Rule Comparison"}
  D --> E["Scenario 1: <br>Deterministic Feeding Costs"]
  D --> F["Scenario 2: <br>Stochastic Feeding Costs"]
  E & F --> G["Computational Process: <br>Inference of Harvest/Continue <br>Decision Boundary"]
  G --> H["Key Findings & Outcomes"]
  H --> I["Stochastic rules outperform <br>deterministic proxies in key cases"]
  H --> J["DNNs mitigate model uncertainty <br>& scale efficiently"]
  H --> K["Feeding cost risk integration <br>adds significant value <br>with negligible computational cost"]

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 a deep‑neural‑network classifier to infer the optimal harvesting decision boundary under stochastic feeding costs, comparing it to deterministic cost rules.

Source sentence
From a methodological point of view, we demonstrate how to use Deep-Neural-Networks to infer on the decision boundary that determines harvesting or continuation, improving on more classical regression-based and curve-fitting methods.

Assumes 6

  • Salmon feed price follows a Schwartz‑2‑factor commodity model. model
    Source sentence
    we assume to follow a Schwartz-2-factor model.
  • Salmon spot price (S1) and soybean spot price (S2) are independent. model
    Source sentence
    we assume independence between S1t and S2t
  • A diagonal covariance matrix is used for the Kalman filter implementation. statistical
    Source sentence
    We chose in our implementation a diagonal covariance matrix D.
  • The calibration period is limited to 01/01/2018–01/01/2022, avoiding regime‑switching models. data
    Source sentence
    we only consider the slice from 01/01/2018 to 01/01/2022, taking only the most recent price developments into account making a regime-switching model unnecessary.
Show 2 more
  • Feeding cost volatility determines whether stochastic modeling yields a benefit. other
    Source sentence
    If the volatility of the feed price is high enough, then it is beneficial to use stochastic feeding costs, otherwise deterministic feeding costs can lead to an almost equal performance.
  • Computational effort of the stochastic rule is negligible compared to deterministic rule. execution
    Source sentence
    the additional computational costs are negligible.

Data

  • Soybean futures and salmon spot price data · fishpool.eu (forward price history) market-prices public
    salmon and soybean commodities
    Source sentence
    We studied the market data of both commodities and found different results for their correlation.
  • Code and data sets for numerical experiments · GitHub repository https://github.com/kevinkamm/AquacultureSt synthetic public
    Source sentence
    The Matlab/ Python code and data sets to produce the numerical experiments are available at https://github.com/kevinkamm/AquacultureStochasticFeeding.

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • Stochastic feeding‑cost rules outperform deterministic ones when feed price volatility is high, while the gap is small for low volatility. performance
    Source sentence
    If the volatility of the feed price is high enough, then it is beneficial to use stochastic feeding costs, otherwise deterministic feeding costs can lead to an almost equal performance.
  • Including stochastic feeding costs never impairs performance and adds only marginal computational effort. performance
    Source sentence
    In none of the cases, however, does the inclusion of stochastic feeding costs lead to impaired performance, and as the computational effort only marginally increases when using our proposed methodology...
  • Model uncertainty is large: Kalman filter and nested‑minimization calibrations yield significantly different parameters, affecting real‑option valuation. theoretical
    Source sentence
    We found that there is a lot of model uncertainty, highlighted by the significantly different parameters obtained from Kalman filtering and the technique by Cortazar and Schwartz (2003).
  • Deep classifiers provide a relative improvement metric (RIτ, RIf) over least‑squares Monte Carlo stopping rules. performance
    Source sentence
    In Table 5, we show two metrics for the relative improvement of using stochastic feeding costs compared to deterministic feeding costs. For this we use the pathwise comparison and the one using Deep Classifiers, denoted by RIτ
  • Simulation of 100 000 paths for LSMC with stochastic feeding costs takes roughly 5 seconds on a CPU. performance
    Source sentence
    we used M = 100000 simulations in the LSMC, which takes roughly 5 seconds on the CPU in the case of stochastic feeding costs.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 24 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 1 paper in the archive

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