Sector Rotation & Relative Strength

Sector rotation is the movement of capital from one group of businesses to another. When a large institution decides that banks will earn more over the next four quarters than metal producers will, it does not simply buy banks — it usually funds that purchase by selling something else. The market is close to a closed system in the short run, so leadership is relative by construction. Somebody has to lag.

The reason it shows up as a *sector* pattern rather than a stock-by-stock pattern is that companies inside a sector share the same input costs, the same customers and the same regulator. If the price of coking coal doubles, it does not damage one steel maker — it damages all of them, in roughly the same direction and at roughly the same time. So the market treats them as a block and repositions the block.

The practical consequence is uncomfortable: you can be completely right about the market direction and still make nothing, because you were in the wrong block. Two investors can both be long the Indian market for the same twelve months and end up with completely different outcomes purely on sector allocation.

Three things push money between sectors, and it is worth separating them because they operate on different clocks.

The first is the **rate of change in earnings**. Markets pay for acceleration, not for level. A sector growing profits at a steady, respectable pace can badly underperform a sector whose profits are terrible but improving. This catches people out constantly: the sector with the worst headline numbers is often the one being bought, because the second derivative has turned.

The second is the **cost of money**. When the Reserve Bank of India is cutting rates, businesses that borrow to grow — lenders, real estate developers, vehicle buyers — get a direct benefit, and businesses whose value sits far in the future get re-rated because future cash flows are discounted less harshly. When rates rise, that runs in reverse.

The third is **positioning**. If every large fund is already overweight a sector, there is very little marginal money left to buy it, however good the story is. Crowded trades stop working not because the story broke, but because everyone who was going to act on it already has.

Relative strength here means something very specific and very simple: divide one price series by another. Take the closing level of a sectoral index — say the NIFTY Bank index — and divide it by the closing level of the NIFTY 50 on the same date. Do that for every date. Plot the result. That line is the relative strength ratio.

The number itself is meaningless. If the ratio reads 2.1 it tells you nothing except that the sector index happens to be quoted at a higher level than the broad index. What matters is the **direction the ratio is moving**. A ratio that is climbing means the sector is gaining ground on the market. A ratio that is falling means it is losing ground. That is all it says, and that narrowness is exactly why it is trustworthy.

This is a different thing from the Relative Strength Index, despite the unfortunate name overlap. The Relative Strength Index compares a security to its own past. A relative strength ratio compares one security to a different one. They answer different questions.

Here is the part that most people get wrong on the first pass. A rising relative strength ratio does **not** mean the sector is going up. It means the sector is doing better than the index. Those are different statements, and in a falling market they can point in opposite directions.

If the NIFTY 50 drops and a sector drops less, the ratio rises. The sector lost you money. It is still, by this measure, the leader. That is not a flaw in the tool — it is the entire point of it. Leadership in a drawdown is where the next up-leg usually starts, because it identifies the group that people were unwilling to sell.

So always read the ratio together with the direction of the market itself. The two together give you four meaningfully different situations.

The numbers below are **invented for teaching**. They are not a reading of any real index on any real date, and they are not a forecast. They exist only so you can follow the arithmetic once and then do it with live closing values yourself.

Take a hypothetical sectoral index and the NIFTY 50, sampled every four weeks over a twelve-week window.

Read down the last column. The ratio climbs at every single step. Now read the middle column: between week 8 and week 12 the NIFTY 50 actually **fell**, from 25,000 to 24,500. The ratio kept rising anyway, because the sector rose while the market slipped.

Over the full window the sector gained 8% and the broad index lost 2%. The ratio moved from 0.800 to 0.882 — roughly 10% of relative gain, which is exactly the 8% and the 2% combining. That decomposition is the useful habit: every relative strength move is a market move plus a sector-versus-market move, and you want to know how much of your result came from each.

If you were only watching the sector index in isolation you would have seen a decent uptrend. If you were only watching the NIFTY you would have seen a wobble. The ratio is the only view that shows you the sector was quietly taking share the whole time.

One ratio tells you about one sector. To see rotation you need the whole board. The practical version is a ranking table: list every sectoral index published by the NSE (National Stock Exchange), compute its plain percentage return over 1 week, 1 month and 3 months, and rank the columns independently.

Why three timeframes rather than one? Because each answers a different question. The 3-month column tells you who has been leading. The 1-month column tells you who is leading now. The 1-week column is mostly noise on its own — but it is where a change of leadership shows up first, and noise that keeps pointing the same way stops being noise.

The interesting information is in the **disagreements** between columns.

These two states feel similar on a screen and are completely different in what they demand from you.

Early rotation is a sector that has started to outperform but has not yet been recognised. The ratio has turned up from a long base. The financial press is not writing about it. Most people who own it are people who have owned it through the pain. It offers the best entry and carries the highest chance of being wrong, because turns fail constantly.

Confirmed leadership is a sector where the ratio has been rising for months, the story is well understood and the flows are visible. The chance of being right about the direction is far higher. The chance of being early is zero, and the amount already priced in is large.

Neither state is 'better'. They are different trades with different risk, and the mistake is treating them as the same trade.

The single most repeated retail mistake in rotation is buying the sector that tops the 3-month or 1-year return table. It is a completely understandable mistake. That table is what gets published, screenshotted and shared. By the time a sector is famous for outperforming, the outperformance is in the price.

There is a mechanical reason this hurts more than it looks like it should. A sector that has run hard has usually also re-rated — investors are paying a higher multiple of the same earnings than they were before. So you are buying a higher price *and* a higher valuation *and* a more crowded position. When the rotation ends, all three unwind together.

The better use of the leaders table is as a source of questions, not conclusions. Ask why that sector led. Then ask whether the reason is still in force or whether it has already fully played out. Sometimes the honest answer is that it is still early. Often it is not.

Rotation analysis is a sizing input, not an entry signal. It tells you which way the wind is blowing across the market; it does not tell you where to stand or when to step out.

In practice it gets used for two things. It is one input into how large or small an existing exposure should be — a gradual tilt informed by the direction of the ratio, never a switch flipped on a single reading. And it surfaces concentration you did not know you had. A book of eight stocks spread across banks, an NBFC (a non-banking financial company — a lender that is not a bank) and a housing finance company is not eight positions. It is one interest-rate bet held eight times, and it will move as one.

The honest framing: relative strength changes how much, rarely whether. Your entry, your invalidation level and your maximum loss per position are set by the individual chart and your own rules, not by a sector ranking.

This does not need software. A spreadsheet updated once a week on a weekend is enough, and doing it by hand is what builds the intuition.

The tool has hard limits and they are worth stating plainly, because most of the damage done with rotation analysis comes from asking it questions it cannot answer.

It is entirely backward-looking. The ratio is built from prices that have already printed. A rising ratio is a description of what happened, and it will keep describing the past right up until the day the leadership changes. There is no mechanism inside it that anticipates anything.

It says nothing about any individual company. A sector index is a weighted average; inside it will be a company with a stretched balance sheet and a company with net cash, and the ratio treats them identically. Sector leadership is a starting filter, never a substitute for reading a company's accounts.

It does not know about valuation. A sector can lead all the way to a multiple that later takes years to grow into. And it breaks down entirely in a market where correlations go to one — in a genuine panic, everything falls together, ratios flatten and the tool has nothing to say for a while. That is normal, and forcing a reading out of it during those stretches is how you manufacture a false signal.