7FNCE026W GARCH(1,1) Volatility Assessment: Model Answer Extract for Westminster

For information and reference only – not to be submitted as your own work.

This extract shows one way to answer the 7FNCE026W Financial Risk Management stock volatility assessment at the University of Westminster. It uses an unnamed FTSE-100 constituent and illustrative parameters, so every figure below is illustrative and must be replaced by your own estimates from your own price data. If you want the planning logic behind each section, read our guide on how to write a GARCH(1,1) volatility assessment. Always check your own brief and module handbook for the word limit, layout and referencing rules.

Planning table

Section Suggested words What markers look for
Executive summary about 150 Purpose, method, headline forecast and the hedging decision
Rationale and model risk of GARCH(1,1) about 450 Critical, not descriptive; limits named and linked to the data
Data and estimation about 400 Log returns, ARCH test, parameter table, persistence check
Volatility forecast about 350 Correct recursion, horizon, annualisation
Hedging recommendations about 450 Instrument choice tied to the forecast and to a stated exposure
Conclusion about 150 Answer to the briefing question and one honest limitation

Executive summary (extract)

This article models the daily volatility of an illustrative FTSE-100 share using a GARCH(1,1) specification and turns the forecast into a hedging recommendation for a £1,000,000 holding. Estimated persistence is high (α + β = 0.98), so volatility shocks fade slowly. The one-day-ahead forecast is 1.62% (25.6% annualised), above the long-run level of 22.4%, and the ten-day horizon volatility is 5.06%. On that basis a partial hedge using protective puts is recommended, with the model’s limits stated openly.

1. Rationale and model risk of GARCH(1,1)

Share returns show volatility clustering: large moves tend to follow large moves. Engle’s ARCH model captured this by letting the conditional variance depend on past squared shocks (Engle, 1982), and Bollerslev generalised it by adding lagged conditional variance, which gave the GARCH family (Bollerslev, 1986). In the GARCH(1,1) form the conditional variance is:

σ²t = ω + α·ε²t−1 + β·σ²t−1, with ω > 0, α ≥ 0, β ≥ 0 and α + β < 1 for covariance stationarity.

The rationale is parsimony. Three parameters describe a mean-reverting variance process, and Hansen and Lunde (2005) found that, for exchange-rate data, little beat a GARCH(1,1) out of sample. Engle (2001) explains why the model suits risk measurement: the forecast updates mechanically each day and reverts to a long-run variance of ω / (1 − α − β).

Model risk is the stronger half of the question. Four points deserve space in a Level 7 answer:

  • Symmetry. A negative and a positive shock of equal size raise variance equally, yet equity volatility usually rises more after bad news. Asymmetric models such as GJR-GARCH and EGARCH address this (Glosten, Jagannathan and Runkle, 1993; Nelson, 1991).
  • Distribution. Maximum likelihood with normal errors understates tail risk, so a Student-t error assumption is a common check.
  • Parameter instability. Estimates drift across regimes, and a window that includes a crisis can overstate persistence.
  • Horizon. Forecasts converge to the long-run level, so long-horizon forecasts say little about current conditions.

The learning outcomes also ask you to defend the chosen methodology. A fair defence is that GARCH(1,1) is transparent, easy to validate and appropriate for a single-stock, daily-horizon question, provided the alternatives above are acknowledged.

2. Data and estimation

The sample is daily adjusted closing prices for the chosen company, downloaded from the library’s FAME database, with the sample period and number of observations stated in a table. Prices are converted to continuously compounded returns, rt = ln(Pt / Pt−1). Before estimating, the squared returns are tested for ARCH effects with an ARCH-LM test; a significant result supports using a conditional variance model. Parameters are then estimated by maximum likelihood.

Parameter Illustrative estimate Interpretation
ω 0.000004 Constant term of the variance equation
α 0.08 Reaction to yesterday’s squared shock
β 0.90 Memory of yesterday’s variance
α + β 0.98 High persistence; stationary because it is below 1
Long-run variance ω / (1 − α − β) 0.0002 Daily volatility of 1.41%, or 22.4% annualised
Half-life of a shock 34.3 days ln(0.5) / ln(0.98)

Annualisation uses 252 trading days: 1.414% × √252 = 22.4%. Diagnostics follow the estimates: standardised residuals and their squares should show no remaining autocorrelation, and the parameter constraints must hold. If α + β is at or above one, the model is not stationary and the forecast would not revert, so the specification would need changing.

Need a second pair of eyes on your estimation output? Our finance assignment help page explains what support is available.

Buy Non Plagiarized & Properly Structured Assignment Solution

3. Volatility forecast

Suppose the last observed return is −2.0% and the model’s conditional variance for that day was 0.00025 (a volatility of 1.58%). The one-step-ahead variance is:

σ²T+1 = 0.000004 + 0.08 × (−0.02)² + 0.90 × 0.00025 = 0.000004 + 0.000032 + 0.000225 = 0.000261.

That is a daily volatility of 1.62%, or 25.6% annualised. For longer horizons the forecast decays towards the long-run variance V = 0.0002 through σ²T+h = V + (α + β)h−1(σ²T+1 − V).

Horizon Forecast variance Daily volatility
1 day 0.000261 1.62%
10 days 0.000251 1.58%
20 days 0.000242 1.55%
Long run 0.000200 1.41%

The ten-day cumulative variance is the sum of the ten daily forecasts, which equals 10V + (σ²T+1 − V)(1 − 0.9810) / (1 − 0.98) = 0.002558. The ten-day volatility is therefore √0.002558 = 5.06%, slightly above the 4.47% that constant long-run volatility would imply (1.414% × √10). On a £1,000,000 position, a ten-day 95% value at risk under normality is 1.645 × 5.06% × £1,000,000 = about £83,200. The normality caveat from section 1 applies, so the figure is a guide rather than a limit.

4. Hedging recommendations

The forecast says current volatility sits above its long-run average and will decay slowly, with a half-life of about 34 days. That supports protection over a one-to-three-month window rather than a permanent hedge. Hull (2023) sets out how market risk is measured and managed, and the options below follow that logic.

Option How it works Fit with the forecast Main drawback
Protective put Buy puts on the share or the FTSE 100 to set a price floor Good: caps loss during an elevated-volatility window Premium is higher when implied volatility is high
Collar Buy a put, sell a call to offset the premium Good if the investor accepts capped gains Gives up upside
Index futures short Sell FTSE 100 futures against the holding Cheap, but removes market exposure only Basis risk and stock-specific risk remain
Reduce position Sell part of the holding Simple and transparent Transaction costs and tax consequences

The recommendation is a collar covering about half the position for the next three months, reviewed each time the model is re-estimated. A put-only hedge is preferable if the investor values upside, provided the premium is justified by comparing implied volatility with the 25.6% GARCH forecast: if the market charges more than the model expects, the hedge is expensive, and a partial reduction may be cheaper. Hedge effectiveness should be monitored, and the model re-estimated after any volatility regime change.

5. Conclusion

The GARCH(1,1) model gives a transparent and testable forecast of this share’s volatility, and the illustrative estimates point to elevated but mean-reverting risk. Its symmetry, normality assumption and parameter instability mean the forecast should be one input to the hedging decision, supported by implied volatility and an asymmetric specification as robustness checks. Your own answer should state the chosen company, the sample dates and the actual estimates.

If you need structured support with the written report, our risk management assignment help and statistics assignment help pages show the options. Related Westminster MSc finance samples include the 7FNCE015W commercial lending appraisal sample. More briefs are listed in our University of Westminster assignment help section.

Please Write Fresh Non Plagiarized Assignment on this Topic

References

Bollerslev, T. (1986) ‘Generalized autoregressive conditional heteroskedasticity’, Journal of Econometrics, 31(3), pp. 307–327.

Engle, R.F. (1982) ‘Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation’, Econometrica, 50(4), pp. 987–1007.

Engle, R. (2001) ‘GARCH 101: the use of ARCH/GARCH models in applied econometrics’, Journal of Economic Perspectives, 15(4), pp. 157–168.

Glosten, L.R., Jagannathan, R. and Runkle, D.E. (1993) ‘On the relation between the expected value and the volatility of the nominal excess return on stocks’, The Journal of Finance, 48(5), pp. 1779–1801.

Hansen, P.R. and Lunde, A. (2005) ‘A forecast comparison of volatility models: does anything beat a GARCH(1,1)?’, Journal of Applied Econometrics, 20(7), pp. 873–889.

Hull, J.C. (2023) Risk management and financial institutions. 6th edn. Hoboken, NJ: Wiley.

Nelson, D.B. (1991) ‘Conditional heteroskedasticity in asset returns: a new approach’, Econometrica, 59(2), pp. 347–370.

Prepared by
Students Assignment Help UK Editorial Team

Students Assignment Help UK Editorial Team

This model answer was prepared by our editorial team in response to a real assignment brief we received. It shows one way to approach the task and is meant as a study guide, not for submission.

do you want plagiarism free & researched assignment solution!

UPTO 35% DISCOUNT

Get Your Assignment Completed At Lower Prices

Plagiarism Free Solutions
100% Original Work
24*7 Online Assistance
Native PhD Experts
Hire a Writer Now

Facing Issues with Assignments? Talk to Our Experts Now!Download Our App Now!

Have Questions About Our Services?
Download Our App!

Get the App Today!

QRcode