Annotated exemplar · AI SL

Should I lease or buy my first car?

Financial mathematics and exponential models · about 12 pages · suggested mark 17/20

Exemplar written by IB Math Revision for teaching — do not submit or copy (academic misconduct). The data and the student voice are illustrative. Schools check coursework with similarity software such as Turnitin.

An AI SL exploration comparing leasing, buying new with a loan and buying used, using an exponential depreciation model fitted to 30 listings, a repayment formula derived from a geometric series, and a sensitivity analysis on interest rates, resale value and repairs.

How this exemplar would be marked

CriterionMarkWhy
A · Presentation4/4Coherent and well organised: the comparison is defined before the calculations, and the conclusion answers the aim with its conditions.
B · Mathematical communication3/4Mostly appropriate and consistent, but GDC input notation appears in the text and R² for transformed data is not fully explained.
C · Personal engagement3/3The student's own decision drives every choice; family expertise and the student's own data collection are used to test assumptions.
D · Reflection3/3Critical: predictions revisited, three key assumptions varied and the conclusion shown to be robust, specific limitations with their direction of effect.
E · Use of mathematics4/6Relevant AI SL mathematics used correctly with understanding (exponential regression interpreted and its assumption tested, repayment formula derived). Held at 4 because the break-even values come from graph intersections with no discussion of accuracy, and the uncertainty in the fitted depreciation rate is never quantified or carried into the sensitivity analysis.
Total17/20A strong AI SL exploration; tidying the notation and adding residual analysis would push it to 18–19.

Marks are our judgement of this teaching exemplar against the current criteria, explained criterion by criterion. They are not IB moderation results.

Excerpts with examiner annotations

Free sections are shown below with comments; the rest is in the full exemplar, available in the protected viewer with the IA package or a Pro plan.

1. Introduction

In a year I will start university in a town with poor public transport, and my parents have said they will help with a car only if I can show them the cheapest sensible way to get one. Every car advert I have looked at shows a low monthly price, but the adverts compare different things — a lease, a loan, a deposit, a “balloon payment” — so I could not tell which option was actually cheapest.

The aim of this exploration is to find which of three ways of having the same model of car for four years — leasing it new, buying it new with a loan, or buying a three-year-old one with savings — costs least overall, and how sensitive that answer is to the interest rate and the car's future value.

I chose four years because that is the length of my course. I expected leasing to be the most expensive (because the car is never yours) and the used car to be the cheapest, but I did not know how big the differences would be, or whether they would survive realistic changes in interest rates.

C A real decision with real stakes (the parents' condition) — the student's reason is authentic and specific.

A A precise, answerable aim that already signals the sensitivity analysis.

2. The three options and my assumptions

I chose a popular small hatchback because it is the car my older cousin drives and it is common on used-car websites. The new price is €19,500. The three options are:

Lease (personal contract hire): an initial payment of €2,000 then 47 monthly payments of €255, for up to 10,000 km a year, after which the car is returned. These figures come from a leasing website's quote in October (Appendix B).

Buy new with a loan: a 10% deposit (€1,950) and a 48-month loan for the remaining €17,550 at a nominal annual interest rate of 6.9%, compounded monthly (the rate I was quoted by a bank's online calculator). At the end I own the car and could sell it.

Buy a three-year-old car with savings: pay for it outright, keep it for four years, and sell it when it is seven years old.

To compare the options fairly I measure each one's net cost over four years: everything paid, minus the value of the car at the end. I also include the interest my family's savings would have earned (3% per year, compounded annually) on any money taken out of savings, because that money would otherwise have grown. I assume the same mileage, insurance and fuel for all three options, so I leave those out; the difference in repair costs is considered in Section 8.

A The comparison is defined clearly before any calculation — the reader knows exactly what “cheapest” means.

D Including the opportunity cost of savings shows thoughtful, meaningful reflection on what a fair comparison needs.

B Sources of every figure are given (in appendices); rates are stated as nominal with the compounding period — precise financial language.

3. How fast does the car lose value?

Two of the options depend on what the car will be worth in the future, so I needed a model for its value. I recorded the asking price and age of every listing of this model (same engine, same trim) on a large used-car website on one day, which gave 30 cars aged 1 to 9 years (Table 1, Appendix A). I define \(t\) = age of the car in years and \(P\) = asking price in euros.

The website listed 34 cars when I searched. I removed four before analysing the data: two were listed as “accident damaged”, one was a different engine size mislabelled as my model, and one had covered over 200,000 km, far more than any other car of its age. I decided to remove these because my question is about the value of a normal car of this model, and each of them would have pulled the model towards prices that do not apply to the car I would buy. I kept one 6-year-old car that looked cheap but had no stated reason, because removing data just because it does not fit would bias my model.

Table 1: Summary of the 30 listings (all listings are in Appendix A).
Age (years)Number of listingsMean asking price
14€17,425
24€14,338
34€12,525
44€10,788
53€9,283
63€7,600
73€6,450
83€5,500
92€4,775

0246810age of car, t (years)02500500075001000012500150001750020000asking price, P (€)Used prices of the same modellistings (jittered)exponential: P ≈ 20,308 × 0.85ᵗlinear
Figure 1: Asking prices of 30 listings of my chosen model, with linear and exponential models.

The scatter graph shows prices falling quickly at first and then more slowly, which suggested an exponential model: a car loses a similar percentage of its value each year, not a similar amount. I compared this with a linear model. The linear model has a strong correlation (\(r \approx −0.973\)), but it predicts a price of about €-1,299 for a 12-year-old car, which is impossible, and its residuals are positive at both ends and negative in the middle.

To fit the exponential model I used my GDC's exponential regression, which fits a straight line to \(\ln P\) against \(t\) and then converts back. This gave \(P \approx 20,308 \times 0.850^{t}\) with \(R^2 \approx 0.981\) for the transformed data. The value 0.850 means the car keeps about 85% of its value each year, i.e. it loses about 15% a year. The model predicts a new price (\(t = 0\)) of about €20,308, which is close to the actual €19,500; I think it is a little higher because asking prices are usually above the price people finally pay.

From the model: a 3-year-old car costs about €12,482, a 4-year-old one is worth about €10,613, and a 7-year-old one about €6,523.

E Appropriate choice of model justified from the context (percentage loss) and from residuals; the regression is interpreted, not just reported. The linearisation is explained in words.

D Checks the model against a known value (the new price) and explains the difference — meaningful reflection.

B Mostly good, but R² is quoted for the transformed data without saying clearly why that matters when comparing with the linear model's r.

4. Is the percentage loss really constant?

In the full exemplar (about 1.5 pages). Tests the constant-percentage assumption with year-on-year losses and residuals. Open in the protected viewer

5. The loan

In the full exemplar (about 2 pages). Finance solver, then the repayment formula derived from a geometric series, and an amortisation table. Open in the protected viewer

6. Comparing the total costs

In the full exemplar (about 2 pages). Net four-year cost of each option, with and without lost interest on savings. Open in the protected viewer

7. What does each option cost per month?

In the full exemplar (about 1 page). Turns net costs into true monthly costs and compares them with the adverts. Open in the protected viewer

8. How robust is the answer?

In the full exemplar (about 2.5 pages). Break-even interest rate, resale-value sensitivity and an estimate of repair costs. Open in the protected viewer

9. Conclusion and reflection

The aim was to find the cheapest of three ways of having the same car for four years. With the figures I found, buying a three-year-old car with savings costs about €7,525 (about €9,925 with extra repairs), buying new with a loan about €11,715 and leasing about €13,985. Leasing is the most expensive option unless loan rates rise above about 12.5% or the car loses value much faster than the market suggests.

The most important thing I learned is that monthly payments are misleading: the lease has the lowest monthly payment of the two new-car options, but because the car is returned it is the most expensive overall. The exponential model was central to the answer, because both buying options depend on what the car is worth later.

My conclusion is limited by the data. The listings were asking prices on one day from one website, and there were only two 9-year-old cars; a model fitted to actual sale prices might show faster depreciation. I also ignored inflation: €1 in four years is worth less than €1 now, which slightly favours options where the money is paid later (the lease and the loan). A fuller model would discount future payments, which I would like to learn how to do. Finally, the cheapest option is not automatically the best one for me: a new car comes with a warranty and less risk, which may be worth paying for — but now I can tell my parents exactly how much.

A Answers the aim directly with the three figures and the conditions under which the answer changes.

D Specific limitations (asking prices, few old cars, inflation) with the direction of their effect — critical rather than generic.

Bibliography and appendices

In the full exemplar (about 1 page). Appendices A–C. Open in the protected viewer

What would push it higher?

  • Replace GDC input notation with mathematical notation and define every symbol (PV, i, n, R).
  • Discount future payments and resale values to present value so that the timing of the money is compared fairly.
  • Estimate how much the depreciation rate could vary (e.g., by fitting to half the data) and carry that into the sensitivity analysis.

Read the full exemplar All exemplars