How to Write a GARCH(1,1) Stock Volatility Assessment for Westminster 7FNCE026W
A strong GARCH(1,1) volatility article starts with the model’s logic and its weaknesses, not with the software output: spend about a fifth of the words on why GARCH suits share returns and where it fails, a third on clean estimation and forecasting, and the rest on hedging advice that follows directly from your forecast. If you are working on the 7FNCE026W Financial Risk Management stock volatility assessment at Westminster, this guide shows how to plan it. Stuck on the estimation itself? Our finance assignment help page explains the support available.
What the assessment is really asking
The brief gives you four jobs: critically assess GARCH(1,1) and its model risk, choose a FTSE-100 company, build and forecast with the model, and give hedging recommendations. It is an article-style piece of about 2,000 words with a cover page, executive summary and contents page, and it carries 75% of the module mark, so each section has to be tight.
Notice the marking weights. Quality of discussion carries 25%, critical understanding of the model 20% and accuracy of calculations 15%. In other words, a perfect estimation with a thin discussion will still lose the largest block of marks. Check your own brief for the exact deadline, word limit and layout rules, because they can change between sittings.
A suggested structure and word split
The brief does not prescribe a section split, so the figures below are suggestions only. Adjust them to the shape of your argument.
| Section | Suggested words | What it must show |
|---|---|---|
| Executive summary | 150 | Purpose, method, headline forecast, hedging decision |
| GARCH(1,1) rationale and model risk | 450 | Critical evaluation, not a textbook definition |
| Data, returns and estimation | 400 | Why the company, how returns were built, parameter table |
| Forecast | 350 | Correct recursion, annualisation, horizon |
| Hedging recommendations | 450 | Instrument choice tied to the forecast |
| Conclusion | 150 | Direct answer and one honest limitation |
The executive summary and contents page are normally outside the word count, but the brief decides that, so confirm it before you trim anything.
A step-by-step plan
- Pick the company first. Choose one with a long, clean price history and no major corporate action in your window, because stock splits and demergers distort returns.
- Download daily adjusted closing prices from the library database and record the sample dates.
- Convert prices to log returns and plot them. Volatility clustering should be visible by eye.
- Test for ARCH effects before you estimate. If there is none, say so, because GARCH may not be justified.
- Estimate ω, α and β by maximum likelihood and put the results in a table.
- Check that α + β is below one and that the long-run variance is sensible.
- Forecast one day ahead, then a longer horizon, and annualise correctly.
- Choose a hedge that matches the forecast, then criticise your own model.
Section-by-section guidance
Rationale and model risk
Explain the idea in plain terms: variance today depends on a constant, yesterday’s squared shock and yesterday’s variance. Then spend most of the section on weaknesses. Symmetry is the classic point, since a fall and a rise of the same size have the same effect on forecast variance, although equity volatility usually rises more after bad news. Normal errors, unstable parameters and forecasts that revert to a long-run mean are also fair targets.
Name an alternative, such as an asymmetric GARCH model, and say when you would prefer it. That one sentence usually separates a merit from a distinction.
Data and estimation
Say why you chose the company and justify the sample window. Give the number of observations, the mean and standard deviation of returns, and the result of the ARCH test. Then show the parameter table with standard errors and interpret every number in words. A table that is never discussed earns very little.
Forecast
Show one worked calculation, not only a software chart. A one-step forecast takes the constant, adds α times the last squared return and β times the last variance. For longer horizons the forecast decays towards the long-run variance at a speed set by α + β. Explain what the half-life of a shock means for the length of any hedge.
Annualise with the number of trading days you used, usually 252, and keep the units consistent. Mixing daily variance with annual volatility is the most common arithmetic error.
Hedging recommendations
Tie the advice to a stated exposure, for example a £1,000,000 holding, so the recommendation is concrete. If the forecast is above the long-run level, protective puts or a collar for the coming weeks are logical. If the forecast is below it, a cheaper partial hedge or no hedge may be better. Compare the cost of protection with what your forecast implies, and state clearly what each option gives up.
For a worked illustration of this structure, see our GARCH(1,1) model answer extract, for reference only. It uses illustrative figures and an unnamed company, so build your own from your own data.
What a higher-scoring answer looks like
| Weaker sentence | Stronger sentence |
|---|---|
| GARCH is widely used to model volatility. | GARCH(1,1) suits daily share returns because it captures clustering with three parameters, but its symmetric response understates the effect of bad news. |
| Alpha plus beta is 0.98. | Alpha plus beta of 0.98 implies shocks fade slowly, with a half-life of about 34 days, so a hedge should run for weeks rather than days. |
| Volatility will rise, so hedge. | The one-day forecast sits above the long-run level, which supports a partial collar for three months, reviewed when the model is re-estimated. |
| The model has limitations. | Normal errors understate tail risk, so the value at risk figure is a guide, and a Student-t check is needed. |
If you want a second opinion before you submit, our proofreading service can check structure, tables and Harvard formatting.
Mark-maximising checklist
- The cover page shows the module title, code, explicit title and word count.
- The executive summary states the result and the hedging decision, not just the plan.
- The ARCH test is reported before the model is estimated.
- Every table and figure has a title and is discussed in the text.
- The condition α + β < 1 is checked and stated.
- One forecast calculation is shown step by step.
- The hedge is tied to a stated exposure and horizon.
- At least one model limitation is linked to your own data.
- Referencing is consistent Harvard, with every source listed.
- Font, spacing and file format follow the brief.
Common mistakes
- Describing instead of evaluating. Defining GARCH for 300 words earns little. Evaluate it.
- Estimating without testing. Skipping the ARCH test leaves the choice of model undefended.
- Mixing units. Daily variance, daily volatility and annualised volatility are three different numbers.
- Ignoring stationarity. If α + β is close to or above one, the forecast does not revert and needs comment.
- Generic hedging. Advice that could fit any share ignores the point of the forecast.
- Unlabelled output. Pasting software screenshots without titles or interpretation costs presentation marks.
Common student worries
“I have never used the software.” Choose the tool you can finish with. What matters is that you can report the parameters, explain them and reproduce one forecast by hand.
“My estimates look odd.” Odd results are fine if you explain them. A very high persistence can reflect a crisis in the sample, and saying so shows judgement.
“I do not know which hedge to choose.” Start from the exposure, the forecast and the horizon, then pick the instrument that fits all three and explain what it costs. If you want to know who prepares our material, read about our editorial team. For extra support with the write-up, our risk management assignment help page and the Westminster University assignment help page are good starting points.
FAQs
Which company should I choose?
Choose a FTSE-100 constituent with a long, continuous price history and no major corporate action in your window. A company you can discuss sensibly in the context of the economy also makes the hedging section easier.
How many years of data do I need?
The brief does not fix a period, so justify your own. A few years of daily data usually gives enough observations for stable estimates, but a longer window may include regimes that reduce relevance. State the dates and explain the choice.
Do I use prices or returns?
Use returns. GARCH models the variance of returns, so convert prices to log returns before testing and estimating.
What if alpha plus beta is above one?
The process is not covariance stationary and the forecast will not revert to a long-run level. Report it, discuss possible causes such as a structural break, and consider a different window or specification.
Can I use an asymmetric model as well?
The brief asks for GARCH(1,1), so keep that as the main model. A short comparison with an asymmetric version can strengthen the critical discussion, provided it does not take over the word count.
How specific should the hedging advice be?
Specific enough to act on: name the exposure, the instrument, the horizon and the reason, and state what the hedge costs or gives up. Vague advice to hedge when volatility rises will not score well.
If you are short of time before the deadline, send your brief through our submit assignment page and we will advise on the next step.




