Financial Management and Business Data Analytics · Data Analysis and Modelling

Time Series Analysis and Forecasting for CMA Inter

Updated 10 October 2026 · Fact-checked

A time series is a set of values recorded at regular intervals, such as monthly sales. You split it into trend, seasonal, cyclical and irregular parts. To forecast, smooth the data with a moving average, fit a trend line, adjust for seasonal indices, and project the result into the next period.

Understand Forecasting and Time Series Analysis

A time series is a list of observations of one variable, taken at equal time gaps and kept in time order. Monthly sales, quarterly profit and yearly production are all time series. The aim is to understand the past pattern so you can forecast the future value.

A time series has four components:
- Trend (T): the long-run movement, upward or downward, over many years.
- Seasonal variation (S): a pattern that repeats within a year, such as higher sales of sweets in the festive quarter.
- Cyclical variation (C): swings around the trend that last more than a year, linked to business cycles. They are not fixed in length.
- Irregular variation (I): random shocks such as floods, strikes or a one-off order.

The components can be combined in two ways. In the additive model, Y = T + S + C + I. In the multiplicative model, Y = T × S × C × I. Use the multiplicative model when seasonal swings grow as the level of sales grows. Most exam questions assume the multiplicative model for seasonal indices.

The moving average method smooths out short-term ups and downs. You replace each value with the average of a fixed number of nearby values. Seasonal and irregular effects get averaged out, so what is left shows the trend. The least squares method goes further. It fits a straight line Y = a + bX through the data, and you read the forecast from the line.

Forecasts are estimates. They assume the past pattern continues. The further ahead you forecast, the less reliable the figure.

Key rules to remember

Additive model: Y = T + S + C + I; Use when seasonal swings stay about the same size.
Multiplicative model: Y = T × S × C × I; Use when seasonal swings grow with the level. Seasonal indices are then ratios or percentages.
Simple moving average (odd period, n = 3): MA = (Y₁ + Y₂ + Y₃) ÷ 3; Place the result against the middle period, here the second one.
Centred moving average (even period, n = 4): Centred MA = (sum of two consecutive 4-period averages) ÷ 2; A 4-period average falls between two periods, so average two consecutive ones to align with a real period.
Moving average forecast: Forecast for next period = average of the last n actual values; The simplest forecast. It lags behind if there is a strong trend.
Least squares trend line: Y = a + bX, where b = (nΣXY − ΣX·ΣY) ÷ (nΣX² − (ΣX)²) and a = (ΣY − bΣX) ÷ n; X is time, coded 0, 1, 2, ... or around zero.
Trend line with ΣX = 0: a = ΣY ÷ n and b = ΣXY ÷ ΣX²; Code time so the middle period is 0. Easy when n is odd (…, −2, −1, 0, 1, 2, …).
Seasonal index (average method): Seasonal index = (average of that season ÷ grand average of all values) × 100; The indices over a full year average 100.
Seasonal index (ratio to moving average): Ratio = (Actual Y ÷ centred moving average) × 100; Average the ratios for each season across years, then adjust so the indices average 100.

How to solve Forecasting and Time Series Analysis questions

Use this order for any time series question. It works for moving averages, trend fitting and seasonal forecasting.

1Read what is asked: smoothed values, a trend equation, seasonal indices or a forecast. Note the period length (3-year, 4-quarter) and the model (additive or multiplicative).
2Write the data in a table in time order. Add columns for the working you need, such as total, average, X and XY.
3For a moving average, add up each group of n values and divide by n. Place each result against the middle period. For an even n, centre it by averaging two consecutive averages.
4For a trend line, code X so ΣX = 0 where possible. Find ΣY, ΣXY and ΣX². Then calculate a and b and write the equation Y = a + bX.
5For seasonal indices, find the average of each season and the grand average. Divide, multiply by 100, and check the indices add up to 100 × number of seasons.
6Substitute the future X into the trend equation. If seasonality is asked for, multiply the trend value by the seasonal index ÷ 100 (multiplicative model).
7State the forecast with units and one line of interpretation, such as the direction of the trend. Mention that it assumes past patterns continue.

Quickest way: Code time around zero and use the shortcut

When to use it: Use when you must fit a straight-line trend or forecast under time pressure, especially with an odd number of years.

Mark the middle year as X = 0 and number the others −1, −2, … and 1, 2, … so ΣX = 0.
Find a = ΣY ÷ n directly. This is just the average of Y.
Find b = ΣXY ÷ ΣX². Only two sums are needed.
Forecast by putting the next X value (one more than the last coded value) into Y = a + bX.
For moving averages, keep a running total: drop the oldest value, add the newest, then divide. This avoids re-adding every group.

Common mistakes in Forecasting and Time Series Analysis

Placing a moving average against the first or last period instead of the middle period.
Students write each average next to the group's starting row because it is easier to fill the table.
Fix: Always put an odd-period average against the centre of its group. For a 3-year average of years 1-3, write it against year 2.
Forgetting to centre a 4-period moving average.
The average of four values falls between two periods, and students ignore this.
Fix: Average two consecutive 4-period averages to get a centred value that lines up with an actual quarter.
Coding X wrongly for an even number of years, or using the wrong X for the forecast year.
Students mix up the origin and continue the coding with the wrong step when forecasting.
Fix: Write the X values in a column beside each year. For an even n, use steps of 0.5 (for four years: −1.5, −0.5, 0.5, 1.5) so ΣX = 0. The next value after 1.5 is then 2.5. Alternatively, use X = 0, 1, 2, … with the full formula, because ΣX is not 0 in that case.
Using the additive method for a multiplicative question, or the reverse.
Students skip the model named in the question.
Fix: Circle the model in the question first. Additive: subtract or add seasonal values. Multiplicative: multiply or divide by index ÷ 100.
Seasonal indices whose total is not 100 × number of seasons (for example, not 400 for four quarters).
Averaging ratios can leave a small error, and students do not check it.
Fix: Add the indices. If the total is not 100 × number of seasons, multiply each by (100 × number of seasons ÷ total).
Calling a forecast certain or giving no interpretation.
Students stop after the number.
Fix: Add one sentence: the forecast assumes the trend and seasonal pattern continue, and actual sales can differ because of cyclical and irregular factors.

Worked examples

Example 1

A firm's sales (₹ lakh) for seven years are: 30, 36, 33, 39, 42, 45, 48. (a) Calculate the 3-year moving averages. (b) Forecast sales for year 8 using the latest 3-year moving average.

Show the solution

Year 1-3 total = 30 + 36 + 33 = 99. Average = 99 ÷ 3 = 33, placed against year 2.
Year 2-4 total = 36 + 33 + 39 = 108. Average = 108 ÷ 3 = 36, placed against year 3.
Year 3-5 total = 33 + 39 + 42 = 114. Average = 114 ÷ 3 = 38, placed against year 4.
Year 4-6 total = 39 + 42 + 45 = 126. Average = 126 ÷ 3 = 42, placed against year 5.
Year 5-7 total = 42 + 45 + 48 = 135. Average = 135 ÷ 3 = 45, placed against year 6.
Forecast for year 8 = average of the latest three actual years (5, 6 and 7) = 135 ÷ 3 = 45.

Answer: 3-year moving averages for years 2 to 6: 33, 36, 38, 42, 45 (₹ lakh). Forecast sales for year 8 = ₹45 lakh. The averages rise steadily, so the trend is upward. Because the method lags, actual sales may be higher.

Example 2

Annual sales (₹ lakh) of a company for five years are: 20, 24, 26, 30, 35. Fit a straight-line trend Y = a + bX by the least squares method, taking the middle year as origin, and forecast sales for the sixth year.

Show the solution

Code X for the five years as −2, −1, 0, 1, 2. Then ΣX = 0.
ΣY = 20 + 24 + 26 + 30 + 35 = 135.
XY values: (−2 × 20) = −40; (−1 × 24) = −24; (0 × 26) = 0; (1 × 30) = 30; (2 × 35) = 70. ΣXY = −40 − 24 + 0 + 30 + 70 = 36.
X² values: 4, 1, 0, 1, 4. ΣX² = 10.
a = ΣY ÷ n = 135 ÷ 5 = 27.
b = ΣXY ÷ ΣX² = 36 ÷ 10 = 3.6.
Trend line: Y = 27 + 3.6X.
The sixth year has X = 3. Y = 27 + 3.6 × 3 = 27 + 10.8 = 37.8.

Answer: Trend equation: Y = 27 + 3.6X (X in years, origin at the third year). Sales grow by about ₹3.6 lakh a year. Forecast for the sixth year = ₹37.8 lakh, assuming the trend continues.

Exam tips

In the 15 compulsory MCQs (2 marks each, no negative marking), expect quick checks: name the component from a description (for example, festive-season rise is seasonal, a multi-year boom and slump is cyclical), or compute one moving average. Always attempt every MCQ.
For written answers, draw the table first. Step marks usually go for correct setup, the formula and each calculation, not just the final number.
State the formula before you substitute. Show ΣY, ΣXY and ΣX² clearly so the examiner can award marks even if you slip on one figure.
Always add one line of interpretation, such as the direction and size of the trend or which season is strongest, and a note that the forecast is an estimate.
Practise questions on the difference between trend, seasonal and cyclical components. Theory parts often ask for definitions with a business example.

Practice questions from Data Analysis and Modelling

Forecasting and Time Series Analysis in other exams

The same ground in other exams, if you are preparing for more than one or want another angle on it.

Forecasting and Time Series Analysis: frequently asked questions

What are the components of a time series?

There are four: trend, seasonal variation, cyclical variation and irregular variation. Trend is the long-run direction. Seasonal repeats within a year. Cyclical lasts more than a year and is not fixed in length. Irregular is random, such as a strike or a natural disaster.

How is a moving average used to forecast?

A moving average replaces each value with the average of the last few values, which smooths short-term ups and downs. The simplest forecast for the next period is the average of the latest n actual values. It works best when the series has no strong trend.

What is the difference between seasonal and cyclical variation?

Seasonal variation repeats within one year and has a fixed, predictable length, such as quarterly or monthly patterns. Cyclical variation spans more than a year and has no fixed length, as it follows business conditions. Seasonal effects can be measured by indices; cyclical effects are harder to predict.

How do I forecast sales using a time series?

Fit a trend line, usually by least squares, as Y = a + bX. Put the future time value into the equation to get the trend forecast. If the data has seasons, multiply by the seasonal index ÷ 100 under the multiplicative model. Then state that the forecast assumes past patterns continue.

Why code X so that ΣX = 0?

It makes the calculation short. The formulas reduce to a = ΣY ÷ n and b = ΣXY ÷ ΣX². This saves time and reduces arithmetic errors in the exam.