We often asked this question, what is Long-term investing?
In a simple term it is easier to answer this but
academically It doesn't have a single, fixed definition; it length varies by
context, like taxes, regulations, or academic studies and most of academic
research consistently shows it means holding assets for 5 years or more, often
10-20+ years, to capture growth while riding out market ups and downs. In this article I will try explaining inferences
of a few popular research papers published to answer this very important
question.
Papers from CFA Institute and SSRN highlight how this
horizon reduces risks through patterns in returns, making stocks more appealing
over time. Let me break this down simply, paper by paper, with clear
explanations of their findings.
CFA Institute Research Foundation Brief (2024)
This paper, "Investment Horizon, Serial Correlation,
and Better (Retirement) Portfolios" by David Blanchett and Jeremy
Stempien, dives deep into US stock and bond data from 1872 to 2023, plus
international markets. Think of it like this: Imagine returns as a bumpy road.
Short trips (1 year) feel chaotic with big drops scaring you off. But long
drives (20 years) smooth out because the bumps don't all hit at once; they
zigzag predictably.
Simple Explanation of Serial Correlation: Returns
aren't random; they show "serial correlation," meaning today's gain
or loss slightly predicts tomorrow's in a negative way (high today, often lower
next). This acts like a shock absorber. For a super-cautious investor (high risk
aversion coefficient; risk aversion coefficient is a mathematical term
to explain risk averseness of investor for mean variance optimization of
portfolio), a 1-year horizon suggests just 20% in stocks and the rest in safe
bonds. While stretching it to 20 years, and it jumps to 50% stocks because the
zigzag cuts long-term volatility relative to bonds.
Key Inference: Don't use one-year math for long goals
like retirement. Standard mean-variance models (basic risk-return balancing)
fail over decades; they undervalue stocks. The paper's fix: Factor in serial
patterns. Bootstrapping (shuffling returns randomly) proves it's not luck; real
history shows 10-20 year horizons favor 40-70% equities, even for the conservative
investor (high risk averseness coefficient). Additionally this paper also explains
commodities shine for inflation protection after 10 years; their correlation
with inflation hits 0.62, versus near-zero short-term.
SSRN Paper: Determinants of Investment Horizon
(2014/2017)
Titled "Long-Term Investing: What Determines Investment
Horizon?", this study catalogs 12 factors shaping why investors hold long
versus trade short. It says, no hard number like "exactly 10 years"
can be termed long term but it is about mindset and setup. Long-term means low
trading, focusing on company fundamentals (earnings growth) over daily prices.
Simple Breakdown: Horizon is not random. It is driven
by:
- Investor
Traits: Pension funds hold investments for longer.
- Environment: Rules
like mutual fund lock-ins and taxes.
- Decisions: Choosing
funds over hot stocks.
Key Inference: Long-term wins because it ignores noise.
Short horizons chase momentum, inflating bubbles. Long ones align with
intrinsic value, cutting costs (fees from trading). Data shows institutions
with long horizons outperform by 1-2% annually via compounding. But barriers
like career risk (managers fired for underperformance) shorten it artificially.
ScienceDirect: Long Horizon Predictability (2019)
"Long Horizon Predictability: An Asset Allocation
Perspective" tests if future returns can be forecasted over many years,
aiding portfolio tweaks. Simple idea: Markets aren't efficient forever;
patterns emerge slowly.
Core Finding: Predictability boosts welfare (your
"certainty-equivalent return," like guaranteed pay after risk). R²
(forecast accuracy) rises with horizon, but gains aren't linear; they peaks at
5-10 years, then plateau. Why? Noise drowns signals short-term; long-term,
valuations (like P/E ratios) predict 3-7% annual excess returns.
Key Inference: Multi-period models beat one-shot plans.
A risk-averse investors gets 0.5-1% higher lifelong returns by tilting to
predicted winners (mostly value stocks) over 10 years. But do not overbet; noise
always persists.
LSE/Vayanos: Long-Horizon in Non-CAPM World (2022)
Dimitri Vayanos' "Long-Horizon Investing in a Non-CAPM
World" challenges basic models (CAPM assumes one risk measure). Real
markets have anomalies like value (cheap stocks win) and momentum, varying by
time.
Core Finding: Short-term, momentum rules (buy
risers). Long-term (5+ years), value dominates as prices revert. Predictability
ties to economic variables (GDP growth-return forecasts better over decades).
Inference: Horizon matters for factors. A 1-year horizon
need to evaluate100+ strategies which shrinks to 10 reliable ones at 10 years.
Portfolios ignoring this miss 2-4% alpha. So, Blend short term (momentum) and
long term (value) bets.
Broader Patterns Across Papers
Common Thread among all these research papers: 5 years
is entry-level long-term but true power kicks at 10 years; it effects halve
risk, predictability doubles utility.
This Inferences converge:
- Equities
from 30% (short term perspective) to 60%+ (long term perspective).
- Behavior
beats math; design horizons via SIPs/lock-ins.
- Huge
alpha gap for disciplined advisors.
Links for the mentioned research papers-
https://www.sciencedirect.com/science/article/abs/pii/S0377221719303777
https://personal.lse.ac.uk/vayanos/WPapers/LHINCW.pdf
