All Courses are free for limited time.
NISM Professor

Deterministic method

Also written Deterministic approach · Deterministic model

The simplest way to calculate a retirement corpus — one fixed inflation rate and one fixed return throughout, giving a single definite corpus figure but ignoring how the sequence of returns can hurt a portfolio.

In plain language

To work out how big a retirement corpus needs to be, an adviser has to assume some numbers — how fast prices will rise, and how much the investments will earn. The simplest way to do this is to pick one number for each and hold it fixed for the whole plan.

That is the deterministic method. It is easy to compute and easy to explain to a client, because it produces just one number. Its weakness is exactly that simplicity: real markets do not deliver the same return every single year.

How it works

The deterministic method assumes a single, fixed set of inputs — one inflation rate and one rate of return, both held constant through accumulation and retirement. Current expenses are inflated forward to the income needed in the first year of retirement, and that income stream is then discounted back to a single lump sum using the inflation-adjusted rate of return.

Step 1 — inflate expenses to retirement: Income required at retirement = Income today × (1 + inflation)^years to retirement.

Step 2 — discount the retirement income stream to a corpus, using the inflation-adjusted rate of return: ((1 + nominal return) / (1 + inflation)) − 1, applied via the Excel PV function over the retirement period in months.

The method's core limitation: it implicitly assumes the retiree withdraws inflation-adjusted income each year from a portfolio earning a steady real rate of return. Real markets do not do this — returns vary year to year, and the order in which good and bad years arrive matters, not just the average. This is sequence-of-return risk, which the deterministic method cannot see at all, since it holds everything constant. The workbook's alternative, the stochastic method, runs thousands of simulated return paths and reports a Safe Withdrawal Rate with an associated probability of the corpus lasting, rather than one fixed figure.

A worked example

Rani needs a monthly income of Rs 35,000 at today's value, retiring in 25 years, and expects to live to 80 (a 20-year retirement). Her savings will go into a bank deposit expected to yield 8%, and inflation is assumed at 6%.

Step 1: Income needed at retirement = 35,000 × (1.06)^25 = Rs 1,50,215 a month.

Step 2: Inflation-adjusted rate of return = ((1.08)/(1.06)) − 1 = 1.89% (per annum). Using the PV function with this rate (divided by 12 for monthly), 240 months, and a monthly payment of Rs 1,50,215, the corpus required works out to Rs 3,00,48,832.

That single number — just over Rs 3 crore — is the deterministic method's answer. It assumes the corpus earns exactly 8% and inflation runs at exactly 6% every single year for the next 45 years, an assumption the workbook itself flags as unrealistic, which is why it goes on to introduce the stochastic method and the Safe Withdrawal Rate as a more robust alternative.

Why NISM asks about it

Chapter 3, section 3.2.7 (Calculation of Retirement Corpus — The Deterministic Method), works through exactly this Rani example, and section 3.2.8 immediately contrasts it with the stochastic method. Expect a full corpus calculation using the two-step (inflate, then discount) method, and a conceptual question on what the deterministic method fails to capture.

Common exam traps

  • Inflate first, then discount using the inflation-adjusted rate — not a single blended rate applied once. Skipping the two-step sequence produces the wrong figure.
  • The deterministic method's single weakness is sequence-of-return risk, not simply "it uses assumptions" — every method uses assumptions; the deterministic method's specific flaw is that it cannot see the effect of bad returns arriving early.
  • The deterministic method is best for a first-pass, simple estimate; the stochastic method suits comprehensive planning — the workbook explicitly ranks them this way, rather than treating one as strictly superior.

Check yourself

  1. 1.The 1994 back-test of the deterministic method exhausted the corpus in 2016 instead of December 2023. What does the workbook say caused this?

    1. a)The return and inflation assumptions turned out to be far too optimistic
    2. b)Not incorrect assumptions — actual realised returns were close to those assumed; the deterministic method simply does not account for volatility and sequence of return risk
    3. c)The retiree withdrew more than the plan allowed
    4. d)Indian inflation over the period was much higher than the assumed 5%
    Show the answer

    Answer: (b) Not incorrect assumptions — actual realised returns were close to those assumed; the deterministic method simply does not account for volatility and sequence of return risk

    This is the most important finding in the chapter. The above outcome was NOT A FUNCTION OF INCORRECT ASSUMPTIONS, as THE ACTUAL REAL REALIZED ASSET RETURNS WERE CLOSE TO THE INITIAL ASSUMPTIONS. The divergence is attributable to the fact that THE DETERMINISTIC METHOD DOES NOT ACCOUNT FOR VOLATILITY AND SEQUENCE OF RETURN RISK.

    The specification: Rs 1 crore, 50% equity / 50% debt, equity 12%, debt 6%, inflation 5%, withdrawing Rs 46,260 a month rising 5% a year — an initial withdrawal rate of 5.5512%. It failed by 7 years 5 months on a 30-year horizon.

    Why the model could not see it coming: a deterministic model, BY CONSTRUCTION, HAS NO ORDER TO REPRESENT AND THEREFORE CANNOT EXPRESS THIS RISK.

    Hence the prescription — overlay stochastic evaluations on deterministic computations, not replace them. The deterministic figure remains easy to compute and easy to explain, and is best suited for first-pass estimation.

    One comparison seals it: he withdrew at 5.55% when Indian research puts the safe 30-year rate at 3% to 3.5%.

Where this is taught

Free preparation for NISM Series XVII

Related terms

← All terms
Something look wrong? Report it