Research · NEPSE ALERT

The Headline Myth: Why NEPSE Doesn't Follow the News

25 years of shocks, budgets and liquidity, put to the test

Written 11 October 2026 (25 Asoj 2083 B.S.)

By Ananta Dhungana · data 2000-01-03 to 2026-10-16 (6,191 trading sessions)

Every Nepali investor knows the ritual. NEPSE drops forty points, and by evening there's a reason: a political rumour, an NRB circular, something said in Singha Durbar. NEPSE jumps the next day and there's a reason for that too, often the same news read the other way. We've watched this long enough that it started to bother us. If the news really drove the market, you'd think someone would get rich trading it. Hardly anyone does.

Robert Prechter makes the case bluntly in the first chapters of The Socionomic Theory of Finance (free to read on Elliott Wave International). He takes the most dramatic American events of the last sixty years, a presidential assassination, the 2003 blackout, 9/11, Hurricane Katrina, and shows that knowing them a day early would have lost you money. Commentators, he argues, fit the news to the price after the fact. We wanted to know whether that holds for a small, retail-driven, rumour-heavy market like ours. So we ran the same kind of test on NEPSE.

The test

We picked 16 events that nobody argues about: shocks and turning points whose dates are on the record. For each we wrote down what the usual "news moves markets" logic would predict. An earthquake or a lockdown should push NEPSE down. A new constitution or a clean election should lift it. Then we looked at what NEPSE actually did on the next session, a week later, a month later and three months later. The starting point is always the last close before the news broke. NEPSE was often shut for days afterwards, and pretending otherwise would flatter the results.

One more number matters, and it's the one most market commentary skips: the base rate. Over any random month since 2000, NEPSE rose 51% of the time. A theory that is right half the time isn't a theory. It's a coin.

What happened after the news

On the very next session the news logic looks decent: right 11 times out of 16. People do react, at least for a day. After that it falls apart. One week, one month, three months: 8, 8 and 8 out of 16. That is the base rate, give or take a coin toss.

A few charts are worth a second look.

The Gorkha earthquake. If ever there was a reason for stocks to fall, this was it. NEPSE stayed shut for 31 days. When it reopened it slid -10.2% in a week. Anyone who sold that week sold within a stone's throw of the low: three months after the quake NEPSE was +23.7% higher than before it. The country was still rebuilding. The market had already moved on.

The COVID lockdown. Same story, louder. The exchange closed for 51 days, the economy stopped, remittances were in doubt. NEPSE went on to rise +14.5% in the month after it reopened and +25.9% in three months, the start of the biggest rally in its history. Note what came before: NEPSE had already fallen -10.9% in the twenty sessions before the lockdown. The fear was largely in the price before the news arrived.

The constitution. The "good news" case. Years of political limbo ended. A month later NEPSE was -9.3%. The border blockade started three days later, and you could argue that was the cause. But that's exactly the game Prechter describes: when the expected reaction doesn't show up, find another headline that fits.

The House dissolution, December 2020. A constitutional crisis. NEPSE did drop -4.8% on the next session, and that's the one-day reaction we keep seeing. A month later it was +13.9% higher. The rally that started in the middle of a pandemic ran straight through a political crisis.

The royal massacre. This one went the "expected" way: -17.4% over three months. Look at the line before the event, though. NEPSE was already falling, -6.3% in the preceding twenty sessions. That pattern repeats: in 10 of 16 events the market was already moving in the direction the news "should" have caused. The news fit the trend. It didn't start it.

The Gen Z protests are the most recent test. NEPSE went -1.3% the next session and -4.0% a month later, after a government fell and Parliament burned. A crisis that historic should, by the headline logic, have done far more damage. The 2024 floods are similar: -0.9% the next day and +7.7% a month later. NEPSE had already fallen -10.6% before the rain.

The RSP landslide, March 2026. If any result should have set off a lasting "stability rally", it was this one. A party only a few years old took 182 of 275 seats, two short of two-thirds, and the old parties collapsed around it. NEPSE did celebrate: +6.0% on the first session after the vote, the biggest one-day reaction in this whole study. Then the cheering wore off. A month later the gain was down to +1.8%, and three months on it was +1.6%. That's the pattern in its cleanest form: a real first-day jolt that the market quietly gives back.

Prachanda becomes Prime Minister, 2008. The earlier "big mandate". The Maoists won the Constituent Assembly vote and Prachanda took office in August 2008. NEPSE had already climbed +10.4% in the twenty sessions before. Afterwards it fell -8.2% in a month and -35.6% in three, the start of the long post-2008 bear market. The mandate made headlines either way. It didn't set the index's direction.

Budget day

Budget day in Nepal is a festival of predictions. Since the constitution fixed it on 15 Jestha, we have ten of them to check, and they line up as cleanly as any event can.

NEPSE closed lower the session after 7 of the 10 budgets. That's a real tendency, and probably "buy the rumour, sell the speech". Only 4 of those moves were bigger than twice an ordinary day's swing (0.54%), though. A month later NEPSE was up after 4 budgets and down after the rest. The speech that everyone reads line by line barely leaves a mark beyond the first day or two.

Our own news bot, put on trial

A confession here. This site now reads every finance headline our news bot collects, and the AI writes up what the news "means" for the market. It would be dishonest to publish this research and not test that too. So every day's news gets a tone (positive, negative or mixed), and the table below checks it against what NEPSE did on the next session, a week later and a month later — news can take time to play out, so one day alone proves little.

So far the clear calls matched the next session 11 out of 11 times, but a week later only 5 out of 8, and the month-later results are still pending. Most days read as "mixed". So we are not claiming any accuracy yet: 11 days is nothing, and a call that is right tomorrow but wrong a week later was not much of a call. There's also a catch that proves Prechter's point rather than refuting it. A large share of Nepali business news is the market reporting on itself: "NEPSE falls 30 points", "banking sector under pressure". A negative day produces negative headlines, and markets tend to drift a little in the same direction the next day. The news isn't steering. It's reading the speedometer. This table updates every day, and we'll come back to it when there are a few hundred days in it.

The market makes the news

Prechter's bigger idea is worth stating plainly: news doesn't make the market, the market makes the news. In socionomics, the crowd's mood comes first. It shows up in prices, then in the headlines, and often later in events themselves. A rising mood produces rallies, confident budgets and hopeful elections. A souring mood produces sell-offs, angry streets and gloomy front pages. The news is the last place the mood shows up, not the first.

Our data lets us check one piece of that. If headlines mostly describe what the market already did, a day's news tone should match the move that came before it at least as well as the move that comes after. So far the tone matched the earlier move 9 times out of 11, and the next move 11 out of 11. The event study points the same way. In 10 of 16 big events NEPSE was already heading in the direction the news "should" have caused: down before the lockdown, down before the floods, up before Prachanda took office. The headline arrived to explain a move that was already underway.

That changes how you read the evening bulletin. A wave of bad news after a long fall is usually a description of the fall, not the start of a new one. Often it's a sign the fall is getting old.

"But liquidity moves NEPSE"

This is the objection we hear most, and it deserves a straight test because it isn't really news, it's money. When the CD ratio is near its cap and interbank rates spike, brokers' margin lines dry up and NEPSE suffers. When banks are flush and interbank money is nearly free, NEPSE runs. Every trader in Nepal has said some version of that.

We pulled ten years of monthly NRB data as republished by ShareSansar, 106 months from 2073-4 to 2083-5 B.S.: the interbank rate, the 91-day T-bill rate, and loan growth minus deposit growth. NRB doesn't publish the CD ratio itself as a series, so we use that last one as a proxy for its direction: when loans outgrow deposits, the ratio is climbing and liquidity is tightening.

Here the traders have a point, and we'll say so plainly. In cheap-money months NEPSE averaged +2.34% and rose 61% of the time. When interbank money cost more than 3%, the average month was negative and fewer than four months in ten were up. Over a decade that's not noise.

Now try to trade it. A change in the interbank rate in a given month tells you almost nothing about NEPSE that month, the month before or the three after. Every one of those correlations sits within ±0.15, which is statistically close to zero. 2080 B.S. shows why. Interbank fell from about 7% at the start of the year to near 3% by Ashad, and the money got cheap. NEPSE kept falling for another four months anyway, to around 1,850 in Kartik, before it finally turned. If you had bought the day liquidity eased, you'd have sat through a drop of more than 11% first.

So liquidity is the weather, not a switch. It tells you what kind of season you're in. It doesn't tell you which day the season changes, and very often the market turns first and the cheap money shows up later to "explain" it.

Hydropower, the loan-heavy case. The other argument you hear: hydro companies are built on bank loans, so cheaper money means a lower interest bill, better profits and higher share prices. If that's true anywhere, it should show up in the Hydropower index, so we ran the same months on it.

In cheap-money months the Hydropower index averaged +2.58% a month against +2.34% for NEPSE, and in tight months -1.17% against -1.18%. So hydro does swing harder with the cost of money, as the loan argument says. The timing problem is the same, though: a rate cut this month says little about hydro next month. Cheaper loans take quarters to reach a company's interest bill, while the share price moves on the mood of the day.

Cheap money, no new high: why 3,227 still stands

There's an obvious problem with the cheap-money table, and plenty of readers will spot it. NEPSE's all-time high is 3,227.11, touched on 19 August 2021 (3 Bhadra 2078). For most of the time since 2023, interbank money has been cheap. NEPSE still hasn't gone back. If liquidity drove the market, it should have.

Start with the top. The crunch everyone remembers came after it. The first month our data shows interbank at 6.56% (Falgun 2078), NEPSE was already -17.3% below its peak. By the time the rate hit its high of 8.5% (Bhadra 2079), NEPSE was -40.8% off. The market turned first. The liquidity headlines arrived slowly, one circular and one rate hike at a time, to explain a fall that was already well underway. That's the market making the news again.

Now the cheap money. Since Asoj 2080 the interbank rate has averaged 2.8%, 36 months of it. NEPSE did rise +40.1% from there. Its best month-end was still -7.0% short of the top, and today it sits -19.0% below it. Cheap money helped. It didn't bring a new high.

The difference is where the money went. In the year around the 2021 top, loans were growing up to 15.3 points faster than deposits: banks lent hard, and a good share of it went into shares on margin. Since Asoj 2080, loans have grown on average 6.8 points slower than deposits. The banks are full, but borrowers aren't borrowing. Money sitting in a bank at 3% isn't the same as money chasing shares. Behind that is mood again. After a mania like 2021, a crowd takes years to get its appetite back, and in Elliott wave terms a correction of that size simply takes time. A rate cut doesn't shorten it.

What we take from all this

We're not saying news is irrelevant to your life or your portfolio. A company's results, a bonus share, a merger: those change what you own. What the evidence says is narrower and more useful. Big, dramatic, everyone-is-talking-about-it news is a poor guide to where the index goes next. The first-day reaction is real, and it usually fades. The trend that was in place before the news is a better guide than the news.

So in practice, keep these in mind. Don't sell because the evening bulletin was scary. Check what the market was already doing for the past month. Treat liquidity as the backdrop and not the trigger. And when someone gives you a confident reason for today's move, ask them what they said yesterday. It's the same exercise Elliott wave analysts have argued for all along: the pattern of the market, which is the pattern of the crowd's mood, tells you more than the headlines the crowd is reacting to.

Read next: Seasonality — what NEPSE usually does each month · Today's market summary and news read