CAS path

MAS-II Time Series

MAS-II time series is about dependent observations over time: stationarity, autocorrelation, ARMA structure, forecast uncertainty, and diagnostics.

Credential side
CAS
Primary intent
MAS-II time series
Best next page
CAS PCPA Guide
Official Source Map

CAS Exam MAS-II

Current MAS-II page and 2026 outline are mapped for format, domain weights, assumed knowledge, tables, and the reading list that drives credibility, mixed models, statistical learning, and time series preparation.

source map reviewed
Last verified 2026-05-141 official source filesNo raw exam or textbook text published
Exam facts

What the official PDFs establish

Appointment length
4.5-hour appointment with a 4-hour exam duration.
Scheduled break
The appointment includes a scheduled 15-minute break plus tutorial/confidentiality/survey time.
Assumed knowledge
Calculus, probability, linear algebra concepts at the regression-prerequisite level, and mastery of MAS-I concepts are assumed.
Weights

Topic and domain coverage

TopicWeightSource
Introduction to Credibility15-25%
Linear Mixed Models10-20%
Statistical Learning40-50%
Time Series with Constant Variance15-25%
Cognitive level: Remember5-10%
Cognitive level: Understand and Apply55-60%
Cognitive level: Analyze and Evaluate35-40%
Cognitive level: Create0-5%
Readings

Chapter and reading intelligence

  • Tse

    Credibility work is assigned from Nonlife Actuarial Models, covering classical, Buhlmann, Buhlmann-Straub, and Bayesian credibility sections in chapters 6-9.

  • West

    Linear Mixed Models: A Practical Guide Using Statistical Software is assigned across all chapters, excluding coding examples, with shrinkage notes called out separately.

  • James et al., Salis, and GLM Monograph

    Statistical learning is anchored to ISLR chapters 2.2, 4.4.2, 8, 10, and 12, Salis chapters 3 and 10, and Chapter 7 of Generalized Linear Models for Insurance Rating.

  • Cowpertwait and Metcalfe

    Time series preparation uses Introductory Time Series with R chapters 1-5 excluding selected sections, plus chapter 6 and sections 7.1-7.3.

Materials

Official files used by the map

  • CAS content outlinecontent-outline

    Primary source for domain weights, exam format, assumed knowledge, and official reading assignments.

Source note: some study materials are private references. ActuaryPath links official sources and uses original explanations instead of republishing paid or copyrighted materials.

Quick Answer

The current MAS-II outline gives time series with constant variance a 15-25% weight. It includes time-series concepts and forecasting tools from the assigned time-series reading.

The central difference from ordinary regression is dependence. Observations close in time may be related, and that changes both model setup and diagnostic checks.

Stationarity

Stationarity means the probabilistic behavior is stable enough over time for the model to be meaningful. In many exam settings, that means constant mean, constant variance, and autocovariance depending on lag rather than calendar time.

A series with a trend, structural break, or changing variance may need transformation, differencing, segmentation, or a different model family before a simple ARMA-style model is reasonable.

Autocorrelation

Autocorrelation measures dependence between observations separated by a lag. The ACF and PACF are diagnostic tools for identifying time dependence and candidate model structure.

A large lag-1 autocorrelation says adjacent observations carry information about each other. That is not the same as ordinary cross-sectional correlation between two different variables.

Autocorrelation at lag k
ρk=γkγ0\rho_k=\frac{\gamma_k}{\gamma_0}

ARMA Model Shape

Autoregressive terms use past values of the series. Moving-average terms use past shocks. ARMA models combine both for a stationary series.

The exam may ask for model recognition, interpretation, or forecast behavior rather than full estimation. Know what each term is doing before reading a fitted output table.

ARMA(1,1) shape
Xt=c+ϕXt1+εt+θεt1X_t=c+\phi X_{t-1}+\varepsilon_t+\theta\varepsilon_{t-1}

Forecasts And Diagnostics

Forecasts should include uncertainty, not only a point estimate. Prediction intervals widen as the forecast horizon increases, especially when future shocks accumulate.

Residual diagnostics ask whether the fitted model removed the time dependence. If residuals still show autocorrelation, the model has not captured the dependence pattern well enough.

Original Practice Drill

A quarterly loss-cost index shows a strong upward trend and positive autocorrelation. Explain why fitting a stationary ARMA model directly to the raw series may be inappropriate, and name two checks or transformations you would consider first.

A complete answer names the trend as a stationarity problem, checks ACF/PACF or residuals, and considers detrending, differencing, or modeling the trend before applying a stationary model.

References and official sources