Indian investors put ₹32,490 crore into SIPs in September 2026 alone, across roughly 10.9 crore SIP accounts (AMFI industry data, September 2026). That is the market every MFD, bank RM and paraplanner now works in. Yet when I ask a room of CFP students in Ahmedabad to work out on paper what return a five-year SIP actually earned, most of them reach for a phone within ten seconds.
The phone gives you a number, but it gives you no way to judge whether that number is sensible, and it is no help at all when a client slides a statement across the table and asks why the app says 13% when the fund’s factsheet says 19%. This article works through three pieces of SIP arithmetic I think every advisor should be able to reproduce with a pen: XIRR versus CAGR, step-up SIP maths, and an honest reading of rupee cost averaging. Every figure below is a labelled example you can rebuild in a spreadsheet. The only market statistics are AMFI’s, cited at the end.
What CAGR measures, and where it stops working
CAGR answers one narrow question. If a single sum of money went in on one date and came out on another, what constant annual growth rate connects the two?
CAGR = (Ending value ÷ Beginning value)^(1 ÷ years) − 1
It is the right tool for a lump sum, and it is what a factsheet reports for point-to-point scheme returns (I have written separately about reading those numbers in How to Read a Mutual Fund Factsheet). A SIP fits it badly, because a SIP is not one cash flow. It is sixty or a hundred and twenty separate investments, each with its own holding period.
Watch what happens when someone applies CAGR to a SIP anyway. Take ₹10,000 a month for 5 years into a fund that grows at exactly 1% a month (about 12.7% a year compounded). The standard SIP future value formula, with instalments at the start of each month, is:
FV = P × [((1 + i)^n − 1) ÷ i] × (1 + i)
where P is the instalment, i the monthly rate, n the number of instalments. Plugging in P = ₹10,000, i = 0.01, n = 60:
- 1.01^60 = 1.8167
- (1.8167 − 1) ÷ 0.01 = 81.67
- 81.67 × 1.01 = 82.49
- FV = 10,000 × 82.49 = ₹8,24,864
Total invested: 60 × ₹10,000 = ₹6,00,000. Now the naive calculation a client (or a careless advisor) does:
- 8,24,864 ÷ 6,00,000 = 1.3748
- 1.3748^(1/5) = 1.066
- “CAGR” = 6.6% a year
The fund delivered 12.7% a year by construction, and the arithmetic has just reported 6.6%. Nothing went wrong with the fund. The error is in treating the whole ₹6,00,000 as though all of it sat invested for five years, when the last instalment was invested for one month. In an even SIP the money is deployed, on average, for only about half the tenure. Spread a five-year journey’s gains over five years when the money actually worked for two and a half, and the rate comes out roughly halved.
What XIRR measures
XIRR is the internal rate of return calculated on the actual dates and amounts of every cash flow. In plain words, it finds the single annual rate which, applied to each instalment for exactly as long as that instalment stayed invested, reproduces today’s value. Excel’s XIRR works this out iteratively on a 365-day year. It needs at least one inflow and one outflow, and it begins its search from a default guess of 10% (Microsoft’s XIRR documentation, linked in Sources).
In a sheet, the client’s whole SIP becomes three columns: dates, cash flows (instalments as negatives, current value as a final positive), and one formula:
=XIRR(values, dates)
Run the five-year example above through it and XIRR returns about 12.7%, exactly the rate the fund delivered. It is immune to the deployment-timing illusion that broke CAGR.
A second example, small enough to check by hand. Say an investor puts ₹10,000 in on 1 January 2025 and another ₹10,000 on 1 July 2025, and the holding is worth ₹22,000 on 1 January 2026. The gain is ₹2,000 on ₹20,000, which looks like 10%. But the first tranche worked for twelve months and the second for only six, so the annual rate has to be higher than 10%. XIRR on those three rows gives about 13.4%. The naive 10% answer quietly penalised the investor for the advisor’s own staggered deployment.
Two mistakes I see repeatedly from students and new MFDs. First, mixing sign conventions: every purchase must be negative and the closing value positive, or Excel returns an error or nonsense. Second, comparing a portfolio XIRR with a factsheet CAGR and declaring underperformance. Those two numbers answer different questions. A factsheet’s point-to-point CAGR is the scheme’s journey between two dates; the XIRR is this investor’s journey, weighted by when their money arrived. The honest comparison is to feed the same dates and amounts into the benchmark index and compute the benchmark’s XIRR. (The five checks planners run on funds, rolling returns among them, are in How CFPs Evaluate Mutual Funds.)
A paper shortcut for checking an XIRR
You will not run Newton-Raphson iterations at a client’s dining table, and you do not need to. For an even, ongoing SIP the money is deployed on average for about half the tenure, so the true annualised return sits a little below double the naive CAGR.
Check that on our example. The naive CAGR came to 6.6%; double it and you get 13.2%, against a true answer of 12.7%. Close enough to catch a lie. If an app or a pamphlet claims a SIP XIRR of 24% and your doubling says the absolute gain supports only about 14%, something in the picture is wrong, usually the period being quoted or a few cash flows being left out.
The same logic runs in reverse, and that direction comes up more often in practice. A client shows you an absolute gain and wants to know whether it is any good. ₹6,00,000 grown to ₹8,24,864 over five years of SIP is a 37.5% absolute gain. Divide by roughly half the tenure, 2.5 years, and you get about 15% simple per year of actual deployment, which lands near 13% once you allow for compounding. Under a minute on paper, and you are in the neighbourhood of the true answer.
The step-up SIP, worked out fully
A step-up SIP raises the instalment every year, usually by a fixed percentage, to track a rising income. It is an easy thing to pitch, which is why the arithmetic is worth doing before you pitch it.
The clean way to compute it on paper is in yearly blocks. Each year is an ordinary 12-month SIP at that year’s instalment. Find each block’s value at its own year-end, then compound it forward to the horizon. The example below uses a starting instalment of ₹10,000, stepped up 10% every year, a fund assumed to grow 1% a month (12.68% a year) and a 20-year horizon.
Step 1, the 12-month block factor at 1% a month, instalments at month-start:
- ((1.01^12 − 1) ÷ 0.01) × 1.01 = 12.68 × 1.01 = 12.81
Step 2, year one’s block: ₹10,000 × 12.81 = ₹1,28,093 at the end of year one. That sum then compounds for 19 more years at 12.68%: 1,28,093 × 1.1268^19 = 1,28,093 × 9.67 ≈ ₹12.38 lakh. So ₹1,20,000 of year-one instalments ends up worth about ₹12.4 lakh.
Step 3, the final year, for contrast: by year 20 the instalment has grown to 10,000 × 1.1^19 = ₹61,159 a month. That block is worth 61,159 × 12.81 ≈ ₹7,83,400 at the horizon, with no time left to compound. The year-20 instalment is over six times the size of year one’s, and its block still ends up worth less than two-thirds of what year one’s ₹1,20,000 became.
Step 4, sum all 20 blocks (a spreadsheet confirms the hand method):
| 20 years, 12.68% a year assumed (example) | Flat ₹10,000 SIP | 10% step-up SIP |
|---|---|---|
| Total invested | ₹24,00,000 | ₹68,73,000 |
| Corpus at year 20 | ≈ ₹99.9 lakh | ≈ ₹1.99 crore |
| Corpus ÷ invested | 4.16× | 2.89× |
The last row is the one most step-up illustrations leave out. The corpus roughly doubles, and the pitch usually stops right there. But the investor also put in 2.86 times the money, and each invested rupee actually multiplied less, 2.89 times against 4.16 times, because the later and larger instalments had less time to compound. When an MFD in my classroom presents step-up as a return enhancer, I stop them. It is a savings enhancer, and a genuinely excellent one. It automates the discipline of investing your increments before lifestyle absorbs them, and for a lot of salaried clients that is the difference between a corpus that meets the retirement number and one that falls short (the corpus arithmetic itself is worked out in The Retirement Corpus Maths Planners Actually Use). That is what belongs in front of the client: the savings behaviour is theirs to control, the rate of return is not.
What rupee cost averaging is worth
AMFI’s investor-education material describes a SIP as investing a fixed amount at regular intervals, from as little as ₹500 a month, so that one invests “in a disciplined manner without worrying about market volatility.” The mechanism behind that claim is rupee cost averaging: a fixed rupee amount buys more units when the NAV is low and fewer when it is high.
The arithmetic is real. Take ₹10,000 a month for four months at NAVs of ₹100, ₹80, ₹125 and ₹100.
- Units bought: 100 + 125 + 80 + 100 = 405 units for ₹40,000
- Average cost per unit: 40,000 ÷ 405 = ₹98.77
- Average of the four NAVs: ₹101.25
The average cost comes in below the average price, and it always will: a fixed-rupee buyer pays the harmonic mean of the prices, which never exceeds the arithmetic mean. Here the NAV started and ended at ₹100, yet 405 units at ₹100 are worth ₹40,500, a gain of ₹500 out of volatility alone.
Beating the average price is not the same as beating the market, though, and averaging stops helping once prices trend upward. Run the same ₹40,000 through a steadily rising market, NAVs of ₹100, ₹110, ₹120 and ₹130:
- Units: 100 + 90.9 + 83.3 + 76.9 = 351.2 units
- Value at ₹130: ≈ ₹45,652
- A lump sum of ₹40,000 on day one: 400 units, worth ₹52,000
The SIP trails by about ₹6,350 on a ₹40,000 outlay, because in a rising market every delayed rupee bought fewer units. The same gap turns up over long horizons. In our five-year example, ₹6,00,000 invested as a lump sum at the same 12.68% grows to about ₹10,90,000, against ₹8,24,864 via SIP. The SIP did nothing wrong here. Money that arrives later simply compounds for less time.
So my view, stated plainly for clients and students alike: rupee cost averaging is not the reason to do a SIP. The reason is that salary arrives monthly, so a SIP matches the investment to the cash flow while removing the monthly temptation to time the entry. If a client is already sitting on a large idle sum, phasing it in over years is a behavioural comfort with an expected cost, not a free lunch, and the conversation worth having is how much that comfort is worth to them. Averaging is a pleasant side effect in choppy markets and a small drag in trending ones.
What to check in your own practice
None of this arithmetic needs a certification behind it. If all you want is to verify your own portfolio’s XIRR, you do not need any course, mine included. One afternoon with Excel’s XIRR documentation and your consolidated account statement will do the job.
If you want to put it to work, start with one client’s statement. Build the three columns, dates, instalments as negatives and the current value as a positive, compute the XIRR yourself, and then run the same dates and amounts against the benchmark before you say a word about performance. Keep the doubling shortcut in your head for the conversations where there is no spreadsheet open. And go back to any step-up illustration you have already shared: if it compares corpus with corpus but stays quiet about the extra contributions, add the total-invested row, the way the table above does.
What the paper fluency buys a working advisor is the ability to catch a wrong number while the client is still in the room, and to make a step-up or lump-sum recommendation for reasons the client can check for themselves. For those building a career in advice, this is the level of working arithmetic the CFP curriculum drills until it is reflex, and it is covered in depth in the programme we teach at House of Financial Planners.
Figures from AMFI and Microsoft’s XIRR documentation as cited below. All worked numbers are illustrative examples at assumed rates, not forecasts.
Frequently asked questions
Is XIRR the same as CAGR?
No. CAGR connects one starting value and one ending value with a constant annual rate, so it is only meaningful for a single lump sum. XIRR handles many cash flows on their actual dates and finds the one annual rate consistent with all of them. For a single investment with no additions or withdrawals, XIRR and CAGR give the same answer.
Why is my SIP XIRR lower than the fund’s 5-year return on the factsheet?
The factsheet shows a point-to-point return between two dates, while your XIRR weights each instalment by when it arrived. If recent years were weaker than earlier ones, your later instalments earned less, pulling XIRR below the headline figure even in the same fund. Compare like with like by computing the benchmark’s XIRR using your own dates and amounts.
How do I calculate XIRR for my SIP in Excel?
List every instalment date in one column and every amount in the next, entering purchases as negative numbers. Add a final row with today’s date and the current value as a positive number, then use =XIRR(values, dates). Per Microsoft’s documentation the function needs at least one positive and one negative value and computes on a 365-day year.
Does a step-up SIP give higher returns than a normal SIP?
It gives a bigger corpus, not a higher rate of return. Both SIPs earn whatever the fund earns. In the worked example above, a 10% annual step-up over 20 years at an assumed 12.68% roughly doubled the corpus, but the investor also contributed 2.86 times the money. Step-up is a savings discipline tool, and it should be presented as one.
Does rupee cost averaging guarantee a profit?
No. Averaging guarantees your cost per unit equals the harmonic mean of purchase prices, which is always at or below the average price, but your outcome still depends on the price at the end. A SIP through a market that finishes lower than where you bought shows a loss regardless of how well the averaging worked.
What is a good XIRR for a SIP?
There is no universal number. It depends on the asset class and on what the benchmark did over the same dates. The sound method is to compute the benchmark’s XIRR with your exact cash flow dates and compare. Judging a SIP against a round number like 12% or 15% without that context rewards and punishes the wrong things.
Sources
- AMFI, industry data including monthly SIP contribution of ₹32,490 crore and SIP accounts outstanding for September 2026: https://www.amfiindia.com/mutual-fund
- AMFI (Mutual Funds Sahi Hai investor education), definition of SIP and minimum instalment of ₹500: https://www.mutualfundssahihai.com/en/what-systematic-investment-plan-sip
- Microsoft, XIRR function documentation (syntax, 365-day basis, default guess, input requirements): https://support.microsoft.com/en-us/office/xirr-function-de1242ec-6477-445b-b11b-a303ad9adc9d
