ECON0006 Chap.3 Eigenvalues and Spectral Methods
Eigenvalues and Spectral Methods
Eigenvalues and Spectral Methods focuses on directions preserved by a linear transformation. An eigenvector remains on its original line, while its eigenvalue records scaling and possible reversal. Enough independent eigendirections can simplify repeated transformations, matrix powers and matrix functions.
The chapter also connects sign patterns to curvature reasoning while warning that one eigenpair does not establish a full spectral basis.
What this chapter covers
- 01
Preserved directions
- 02
Eigenvalue scaling
- 03
Direct substitution checks
- 04
Independent eigenvectors
- 05
Spectral representation
- 06
Matrix powers
- 07
Hessian interpretation
Worked example · free
Audit a claimed eigenvector
- 2Compare the scale implied by every nonzero component.
- 1Observe that two and three are not one common scalar.
- 1Reject the proposed eigenvector unless a zero component changes the comparison.
Key terms
- Eigenvector
- A nonzero vector whose direction is preserved by a transformation.
- Eigenvalue
- The scalar applied along a corresponding eigendirection.
- Spectral representation
- A decomposition that uses eigenstructure to describe a linear transformation.
Eigenvalues and Spectral Methods FAQ
Can an eigenvalue be negative?
Yes. A negative eigenvalue keeps the vector on the same line while reversing its orientation and scaling its magnitude. Direction preservation concerns the line, not the arrow's sign.
Why are eigenvectors useful for repeated change?
Each eigendirection is scaled without being rotated away. Decomposing a state into independent eigendirections can make repeated transformations easier to understand and calculate because each component then evolves by its own scalar factor.
Does one eigenvector prove diagonalisation?
No. A full spectral basis needs enough independent eigenvectors for the space being represented. One valid pair proves only that preserved direction and its scale.
What is the quickest check of a claimed eigenpair?
Multiply the matrix by the proposed nonzero vector and compare the result with the claimed scalar multiple component by component. Every nonzero component must imply the same scale, and zero components must remain compatible with the relation. This direct substitution catches false pairs before any characteristic-polynomial work and keeps the algebra tied to the preserved-direction definition.
Exam move
Start every exercise with the definition: the transformed nonzero vector must be one scalar multiple of the original. Verify proposed pairs by direct multiplication before using a longer method. Sketch preserved and non-preserved directions to make the geometry visible. Keep eigenvalue, eigenvector, spectral representation and definiteness in separate columns because they answer different questions.
When the topic is linked to a Hessian or dynamic system, state what the sign or magnitude means in that setting rather than treating eigenvalues as context-free labels. Use direct multiplication as the first check whenever a vector and candidate eigenvalue are supplied. Write the original and transformed components in aligned columns, then ask whether every nonzero component gives the same scale factor.
A zero component requires care because it cannot be used as a ratio, but its transformed component must still agree with the proposed relation. Rejecting a false pair this way is faster and clearer than constructing a characteristic equation. Sketch the original vector and its image to connect the calculation with the geometric statement that an eigendirection remains on the same line.
When eigenpairs must be found, keep a stable sequence: form the characteristic condition, solve for candidate scalar values, substitute each value into the homogeneous system and find a nonzero direction. Verify every resulting pair in the original matrix. Record repeated scalar values separately from the number of independent eigenvectors they provide.
A repeated eigenvalue does not guarantee enough directions for a full spectral basis. If diagonalisation is requested, count independent eigenvectors and show how the basis and diagonal entries correspond rather than assuming the characteristic roots alone complete the argument. Build a translation table for applications.
In repeated transformations, magnitude describes expansion or contraction along an eigendirection and a negative sign also reverses orientation. In local dynamic systems, these quantities help classify nearby movement under the stated timing convention. In Hessian analysis, signs contribute to curvature and must be interpreted with the relevant symmetry and definiteness conditions.
Do not transfer a discrete-time stability rule into continuous time without checking the model, and do not call a negative eigenvalue economically bad without identifying what transformation or quadratic form it describes. For revision, pair symbolic work with a numerical verification.
After finding a decomposition, choose a simple vector, express it in the eigenvector basis and apply the transformation repeatedly to see which component dominates. Keep exact arithmetic long enough to distinguish a genuine dependency from a rounding artefact. End each exam response by naming the preserved direction, its scalar factor and the implication in context.
If a full basis does not exist, state that limitation explicitly and use only the valid spectral information rather than forcing an unsupported diagonal representation.
Working through Eigenvalues and Spectral Methods in ECON0006? Sia is AskSia’s AI Economics tutor — ask any ECON0006 Eigenvalues and Spectral Methods question and get a clear, step-by-step explanation grounded in how ECON0006 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.