A payment that may not come: what it is worth today

14 min · 2 drills

In the textbook: Ch. 2.1 Verzinsungsmodelle

What is a promise worth today if it may never be kept?

Some payments are promised only on a condition: a pension if the person is alive, a bonus if the shop is still open, a buy-back if the machine still runs. Such a payment is worth less today than a certain one: on top of discounting comes the chance that it does not come at all. This is how insurers price every life insurance policy.

A city promises a café 20000 €:

  • the money comes in 5 years, if the café is still open then;
  • in each year a café like this closes with probability 10 %.

What is the promise worth today, at 4 % a year?

Discounting the 20000 € alone does not give the answer: the café may close before then, and then the city pays nothing.

So the calculation takes two steps:

  1. The café has to last 5 years in a row, each year with probability 0.9. The chance of that is 0.95=0.59049, so on average the city pays 20000⋅0.59049=11809.80 €.
  2. That money arrives in 5 years, so it is discounted: 11809.80/1.045≈9706.79 €.

Today the promise is worth about 9706.79 €, less than half of 20000 €. The chance 0.95 that the receiver is still there when the payment is due is called the survival probability.

The basics

The calculation for the café rested on discounting, which has already come up in the chapter. Even for a payment that may not come, two rules from the article “A top-up after a rate cut and unequal deposits” apply. The rate r is written as a fraction, and the factor is q=1+r:

  • you move t years back by dividing: the amount Kt in t years is worth K0=Kt/qt today: this is discounting. For example, 5000 € in 3 years at 2 % are worth 5000/1.023=4711.61 € today;
  • you may only compare amounts moved to the same day: a price today is compared with the payout discounted to today.

The answer to the question from the start:

  1. If the promise were certain, it would be worth 20000/1.045=16438.54 € today.
  2. In each year the café stays open with probability 1−0.1=0.9, for all 5 years with 0.95=0.59049. Why the probabilities are multiplied, the next section shows with 1000 shops.
  3. Fair value today: 16438.54⋅0.59049=9706.79 €, 6731.75 € less than a certain promise.

What a payment that may not come is worth

For the café the chance of closing in a year was given and the price of the promise was asked. The exam likes the reverse: the price is known and the chance of dropping out in a year has to be found. Here is such a task about a shop:

Exam-style task

Ein Getränkehändler zahlt dem Betreiber eines Dorfladens mit 200 m² Verkaufsfläche heute 148304 GE. Dafür bekommt er in 10 Jahren Ware im Wert von 280000 GE zurück, aber nur, wenn der Laden dann noch geöffnet hat. Der Händler rechnet mit 4.75% Zinsen pro Jahr, hält den Betrag für netto fair und geht davon aus, dass ein solcher Laden in jedem Jahr gleich wahrscheinlich schließt. Mit welcher Schließungswahrscheinlichkeit pro Jahr (in Prozent) rechnet er? Runden Sie auf zwei Nachkommastellen.

  1. a)1.15
  2. b)2.16
  3. c)1.58
  4. d)1.70
  5. e)1.10

Two things in the setup give the task away:

  • the payment only comes if something still exists after some years: the shop is open, the insured person is alive, the machine runs;
  • there is a fair amount today, “on average” or “net”.

That amount is either asked for or given, and then the question is the probability of dropping out in a year. The amounts are in GE, the exam's money units.

Two ideas about probability are easiest to see on 1000 identical shops.

YearOpen: each year ⋅ 0.976If 24 closed every year
010001000
1976976
2953952
4907904
8823808
16678616

The middle column computes 1000⋅0.976t, where t is the year, the right one 1000−24 t. The fewer shops are left, the fewer close each year, so the middle column falls more slowly.

The probability of lasting all the years. A shop closes in a given year with probability 2.4 %. Of 1000 shops, about 1000⋅0.976=976 are open after year 1. In year 2, 2.4 % of these 976 close again, leaving 976⋅0.976=1000⋅0.9762≈953. Every year the ones left are multiplied by 0.976, the way money is multiplied by q. The years do not affect each other, so the probabilities of the years are multiplied: the share 0.97616 lasts all 16 years. Call the probability of dropping out in a year w; you last t years with probability (1−w)t.

The fair amount: the average payout. It is fair to pay today as much as makes the money balance on average over many contracts. Of 1000 shops, the share 0.97616 gets the payout after 16 years and the rest get nothing. Each shop gets, on average, the payout times this share. This average payout arrives in t years and is worth qt times less today. Payout Kt, fair amount today P:

P=Kt⋅(1−w)tqt
Price of 148304 GE today and its value after 10 years012345678910148304 GEpriceafter 10 yearsyearsrate = 4.75 %

Price P=148304 GE today, payout Kt=280000 GE in t=10 years, rate 4.75 %. Here the formula is read backwards: Kt⋅(1−w)t=P⋅qt. On the right, 1.047510=1.5905243 and 148304⋅1.5905243=235881.12 GE: this is the average payout in 10 years that the wholesaler takes to be fair.

The average payout is 280000⋅(1−w)10, so (1−w)10=235881.12/280000=0.842433. This is the probability that the shop stays open all 10 years. The 200 m² of sales area play no part in the calculation.

1−w=0.8424331/10=0.983, w=0.017=1.70 % a year. Option (d). Check: 280000⋅0.98310/1.5905243=148304 GE.

Ask yourself: what would it mean if 148304⋅1.047510 came out larger than 280000?

A foundation promises a mountain farm 60000 € in 10 years if it is still being farmed then. Such a farm is abandoned in each year, independently of the other years, with probability 1.5 %. The rate is 4 % a year.

What lump sum today is fair?

Discounting the 60000 alone is not enough: the farm may be abandoned by then, and nothing is paid.

Payout of 60000 € after 10 years and its value today01234567891060000 €payouttodayyearsrate = 4 %

1.0410=1.4802443, 60000/1.4802443=40533.85 €.

The farm lasts one year with probability 1−0.015=0.985, all 10 years with probability 0.98510=0.8597304.

40533.85⋅0.8597304=34848.18 €.

Check: on average the foundation pays out 60000⋅0.98510=51583.83 €. The fair amount, moved 10 years forward: 34848.18⋅1.4802443=51583.82 €. The one-cent difference comes from rounding 34848.18.

A leasing company promises to buy back a machine for 50000 € in 5 years if it still runs then. In each year, independently of the other years, the machine breaks down for good with probability 1 %. Rate 3 %.

Discounting the buy-back alone is not enough: the machine may have broken down by then.

  1. Discount: 1.035=1.159274, 50000/1.159274=43130.44 €.
  2. Survival probability: 0.995=0.950990.
  3. What amount is fair today?

Your turn2 of 2

A shipping company promises a port 80000 € in 14 years if its ferry is still running then. In each year, independently of the other years, the ferry goes out of service for good with probability 2 %. What lump sum today is fair at 3.5 % a year? Round to whole euros.

The probability is sometimes given per mille, in thousandths: 6 ‰ is 0.006. The task may call it by the letter q. This is how an insurer's row is printed in a task:

q ist die Sterbewahrscheinlichkeit pro Jahr in ‰

EintrittsalterZinsqLaufzeitEinmalprämie
362.25 %610 Jahre113000 GE

Here q=6 is the probability of dying in a year, 6 ‰. In the formula for the fair amount P above this is w=0.006. This q is not the interest factor: the factor in this row is 1.0225.

Survival probability traps

  • Adding the risks instead of multiplying. 1−10w=0.842433 gives w=1.58 % instead of 1.70 %, option (c) of the exam task. Check: of the 976 shops left, 2.4 % of 976 close in year 2, not of 1000.
  • Carrying forward over the wrong term. 1.04759 instead of 1.047510 gives 2.16 %, option (b). Check: all 10 years lie between the price and the payout.
  • Forgetting the root. 1−0.842433=15.76 % is the probability of closing in some year out of the 10. Check: the question asks for the probability per year, which is much smaller.
  • Taking the probability of closing. In the farm task, multiplying by 1−0.98510: 5685.67 instead of 34848.18 €. Check: the payment comes if the farm is still worked, that is the larger share.
  • Forgetting the probability. Only discounting: 40533.85 € in the farm task. Check: the answer must be smaller than the guaranteed amount.
  • Reading per mille as percent. 6 ‰ as 0.06: over 10 years 0.9410=0.539 instead of 0.99410=0.942.

On the sheet: Kt=K0(1+r)t for discounting; survival probability is not on the sheet.

Formula sheet and what to memorise

Of the whole calculation the sheet gives only discounting. The survival probability and the way back to the probability per year you put together yourself, with the steps in the overview:

  • Payment with survival probability

    The payout Kt in t years only comes if something lasts all that time; in a year it drops out with probability w. If the price P is given and w is asked for, read the formula backwards.
    P=Kt⋅(1−w)tqt

    Trap: taking 1−t⋅w instead of (1−w)t.

On the formula sheet (rate r, term t; we write q=1+r):

  • compound interest Kt=K0(1+r)t: read backwards, K0=Kt/qt, it discounts the payout to today.

Memorise yourself:

  • you last t years with probability (1−w)t, not 1−t⋅w;
  • if the price P is given, (1−w)t=P⋅qt/Kt, then the t-th root.

So far the unknown was an amount or the probability per year. In the article “Term and interest rate: when the unknown is in the exponent” it is the exponent itself. Tasks on payments that may not come are in the “Compound interest” trainer.

Now practise it

A payment that may not come: what it is worth today

Trainer · Elementary financial mathematics

Practise
Next: Term and interest rate: when the unknown is in the exponent
0/2

Was this article helpful?