Investing

How to invest money you won't need for at least five years, written for a US-based reader: what each asset class has done historically, risk and drawdowns, diversification, why low-cost index funds beat most active managers, asset allocation, rebalancing, bonds and stocks from first principles, behavioral traps, and a written investment policy. Savings order and the emergency fund come first in money basics, cash, CDs and I bonds are in banking, paying off high-interest debt beats almost any investment (credit and debt), and account types, tax rules and contribution limits are in retirement and taxes. Base rates and risk thinking in general are in probability and risk.

The short version

Asset classes

Historical US returns 1928–2025, nominal (before inflation), computed from Aswath Damodaran's annual dataset at NYU Stern (updated January 2026). Geometric = compound annual growth rate, what an investor actually earned; arithmetic = simple average of yearly returns, always higher when returns vary.

US consumer prices rose about 3% a year over the same period (BLS CPI-U), so real returns are roughly: cash ~0.4%, Treasuries ~1.5%, stocks ~7%.

Risk, return and drawdowns

Risk in everyday terms: the chance you don't have the money when you need it. The textbook proxies are volatility (standard deviation of returns) and drawdown (fall from a previous peak).

Losses are asymmetric: after a fall of $d$, you need a gain of $\frac{1}{1-d} - 1$ to get back.

Volatility drag: the geometric return is roughly the arithmetic return minus half the variance, $g \approx a - \tfrac{\sigma^2}{2}$. For the S&P 500, $11.86\% - \tfrac{0.194^2}{2} \approx 10.0\%$, which matches the 10.02% actually compounded. Two portfolios with the same average return: the less volatile one ends up richer.

Risk premium: stocks return more than bills because they can lose half their value when you most need the money. There is no reliable way to get the return without the risk. Anything claiming stock-like returns with bond-like risk is mispriced, leveraged, illiquid, hiding risk in the tails, or a fraud.

Diversification and correlation

Correlation ($\rho$, from −1 to +1) measures how two assets move together. Portfolio volatility for two assets with weights $w_1, w_2$:

When $\rho < 1$, the portfolio is less volatile than the weighted average of its parts: the "only free lunch" in investing (the phrase is usually attributed to Harry Markowitz, whose 1952 paper founded portfolio theory).

Worked example. 60% stocks ($\sigma = 19.4\%$), 40% Treasuries ($\sigma = 7.9\%$), correlation 0 (in Damodaran's annual data the 1928–2025 stock/10-year Treasury correlation is about 0.02):

A weighted average would suggest $0.6(19.4) + 0.4(7.9) = 14.8\%$. The gap is the diversification benefit.

Correlations are not constant. From 2000 to 2025 US stocks and Treasuries were negatively correlated (about −0.36 in annual data), which made bonds an excellent hedge. In 2022 both fell together (S&P 500 −18.0%, 10-year Treasury −17.8%) because rising inflation and interest rates hit both. In crashes, correlations between risky assets tend to rise toward 1. Diversification reduces risk; it doesn't remove it.

Market efficiency and active management

Efficient market hypothesis (Fama, 1970): prices reflect available information, so beating the market consistently after costs is very hard. In practice markets are not perfectly efficient, but they are efficient enough that the mispricings left are small, fleeting, and eaten by fees.

The arithmetic of active management (William Sharpe, 1991): before costs, the average actively managed dollar must earn the market return, because active and passive investors together are the market. After costs, the average active dollar must therefore trail the average passive dollar by the difference in fees. This is arithmetic, not a theory about efficiency.

The evidence agrees:

Where active has better odds: some bond, real-estate and small-cap categories in some periods (the same Morningstar report found stronger 10-year success rates in fixed income and real estate). Even there, pick the cheapest active funds and expect wide dispersion.

Blunt conclusion: for your core money, own index funds. Paying someone 1% a year to try to beat the market is a bet that they are in the small minority that will, identified in advance, which the persistence data say you can't do.

Index funds, ETFs and mutual funds

What to check before buying a fund

Fees compound too

The expense ratio is taken from fund assets every year, so it compounds against you exactly as returns compound for you.

Worked example: $100,000 invested for 30 years at a 7% gross annual return.

The 1% fee costs $180,500, 24% of the final value, on a single deposit with no further contributions. Layer a 1% adviser fee on a 0.75% active fund and the drag is 1.75% a year. A good fee-only adviser can be worth it for complex situations or for stopping you from panicking; judge the fee in dollars, not percent.

Asset allocation

Asset allocation (the split between stocks, bonds and cash) is the decision that matters most. Choose it from your ability to take risk (time horizon, job stability, other assets like a pension or Social Security) and your willingness (how you'll actually behave in a crash).

Test your willingness: in 2007–09 a 60/40 portfolio fell roughly a third at its worst and a 100% stock portfolio more than half. If you would sell at −30%, you own too many stocks, whatever the rule of thumb says.

The three-fund portfolio

Popularised on the Bogleheads forum by Taylor Larimore: own the whole world with three index funds.

Variants: a two-fund or one-fund version (a total-world stock fund plus a bond fund, or a single target-date or balanced fund). Home bias is common; somewhere between 20% and 40% of stocks in international funds is a typical Bogleheads range. Simplicity is the feature: fewer decisions, fewer mistakes.

Target-date funds

A target-date fund (e.g. "2060") holds a mix of index funds and shifts from stocks to bonds as the date approaches. Good default for a 401(k) if the fees are low (index-based versions from the big providers charge about 0.1% or less; some plan versions charge far more). Hold it alone, not alongside other funds, or you undo its allocation. Less ideal in taxable accounts (bond income and rebalancing trades are taxed).

Rebalancing

Markets drift your allocation: after a long bull market a 60/40 portfolio can become 75/25, with more risk than you chose. Rebalancing sells what has grown and buys what has shrunk.

Rebalancing is risk control, not a return booster: in a long bull market it lowers returns, and in choppy markets it can add a little. Don't rebalance more than necessary in taxable accounts.

Lump sum vs dollar-cost averaging

Dollar-cost averaging (DCA, strictly "cost averaging" when you have the lump sum already) means investing a windfall in instalments instead of all at once. Investing each paycheque as it arrives is not really DCA; it is just investing when you have the money, which is correct.

Vanguard research (Finlay and Zorn, Cost averaging: Invest now or temporarily hold your cash?, February 2023) compared lump-sum investing (LS) with cost averaging (CA) over one-year horizons, 1976–2022 across several markets:

Markets rise more often than they fall, so money waiting in cash usually misses gains. Lump sum wins on expected value; cost averaging reduces regret if a crash follows. If spreading it out is the only way you'll invest at all, do it over a short, fixed schedule (say 3–6 months), automated in advance, and never "wait for a dip".

Sequence-of-returns risk

When you are adding money, the order of returns doesn't change your ending wealth for a given set of returns. When you are withdrawing, it does: losses early in retirement force you to sell more shares at low prices, and those shares never recover for you.

Worked example: $1,000,000, withdraw $50,000 at the start of each year for 10 years. Same ten annual returns (−20%, −10%, 5%, 10%, 15%, 20%, 12%, 8%, 10%, 14%; geometric average 5.7%), in two orders:

Same average, a $333,000 difference, entirely from the order. Defenses: hold several years of spending in bonds/cash near and early in retirement, use flexible withdrawal rules (spend less after bad years), delay Social Security to raise guaranteed income, and consider a "bond tent" (more bonds around the retirement date, fewer later). Withdrawal rates are in retirement and taxes.

Bonds

A bond is a loan: you pay a price now, receive coupons (interest), and get the face value (par) back at maturity.

Price and yield move in opposite directions

where $D_{\text{mod}} = D_{\text{Mac}}/(1+y)$ is modified duration.

Worked example: a 10-year bond, $100 face, 4% annual coupon, bought at par (yield 4%). Macaulay duration is 8.44 years, so modified duration $= 8.44/1.04 = 8.11$ and convexity $\approx 80.8$.

A 30-year bond with the same coupon would fall to $84.63 (−15.4%) for the same 1-point rise. Rule of thumb: if the money is needed in $N$ years, hold bonds with duration of roughly $N$ or less. Over a holding period about equal to its duration, a bond fund's higher reinvestment yield roughly offsets the initial price loss.

Types of bonds

Individual Treasuries and CDs held to maturity have no price risk if you don't sell. Bond funds never mature, so their price moves with rates forever; their duration stays roughly constant.

Stocks

A share is a fractional claim on a company's future profits, with a vote. Its value today is, in theory, the present value of all future cash returned to owners.

Dividend obsession is a mistake: a dividend is not free money; the share price falls by the amount paid. A "dividend income" portfolio is a less diversified, often less tax-efficient version of the total market. Spend from total return (sell shares as needed) rather than restricting yourself to dividend payers.

Factor tilts (small-cap, value, momentum, quality) have academic backing (Fama and French) but long stretches of underperformance; value lagged growth for most of 2007–2020. Optional, never necessary.

Behavioral traps

The biggest threat to returns is usually the investor, not the market. Morningstar's annual Mind the Gap study compares funds' reported (time-weighted) total returns with the dollar-weighted returns investors actually earned after their buying and selling; in recent editions the shortfall has averaged roughly 1 percentage point a year over ten years, larger in volatile funds. (DALBAR's older "behavior gap" figures are much larger but their method has been criticized as overstating it; Morningstar's approach is more defensible.)

Cognitive biases in general are in cognitive biases.

Accounts and tax placement

Details, limits and the traditional vs Roth choice are in retirement and taxes.

Asset location (which fund goes in which account), once your accounts are big enough for it to matter:

Brokerage safety: SIPC protects customers of a failed broker-dealer up to $500,000 per customer (including $250,000 cash) if securities are missing; it does not protect against market losses. Use a large, low-cost broker, turn on two-factor authentication, and name beneficiaries (TOD) on every account.

What to avoid

Investment policy statement

Write it when calm; read it when markets are not. One page is enough.

INVESTMENT POLICY STATEMENT            date: ________
 
1. Goals
   - Retirement at age __, target portfolio $______
   - Other goals (house, education) and dates: ________
 
2. Time horizon and constraints
   - Money needed within 5 years stays in cash/short bonds
   - Emergency fund: __ months of expenses (separate)
   - Other income: pension/Social Security estimate $___/yr
 
3. Target allocation (whole portfolio, all accounts)
   US stocks ___%   International stocks ___%   Bonds ___%
   Glide path: reduce stocks by __ points every __ years
   from age __
 
4. Funds (index, expense ratio <= 0.10%)
   US stock: ______   Intl stock: ______   Bond: ______
 
5. Contributions
   - __% of gross pay, automated on payday
   - Order: 401(k) to match -> HSA -> IRA -> 401(k) max
     -> taxable
 
6. Rebalancing
   - Check every [birthday]; rebalance if any asset class is
     more than 5 points off target
   - Use new contributions first; sell inside tax-advantaged
     accounts before taxable
 
7. Rules I will follow
   - No selling because of market news or a decline
   - No individual stocks/crypto beyond __% (max 5%)
   - No product I cannot explain in two sentences
   - Harvest tax losses in taxable if a loss exceeds $____
 
8. When to revise this document
   - Marriage, divorce, children, job change, inheritance,
     health change, within 10 years of retirement
   - Never because of market performance alone
 
Signed: ____________

Investing checklist

[ ] Emergency fund and high-interest debt handled first
[ ] Employer 401(k) match captured in full
[ ] Target allocation written down (IPS)
[ ] Broad index funds or one target-date fund, ER <= 0.10%
[ ] No loads, 12b-1 fees, or advisory fee above what it's worth
[ ] Contributions automated
[ ] Beneficiaries / TOD set on every account
[ ] Asset location: bonds in tax-advantaged where possible
[ ] Rebalancing rule set (annual or 5-point bands)
[ ] Speculation capped (single stocks, crypto <= 5%)
[ ] Two-factor authentication on every brokerage login
[ ] Annual review: fees, allocation, contributions, IPS

References

• Aswath Damodaran: Historical returns on stocks, bonds, bills, real estate and gold, 1928–2025 (opens in a new tab): annual US returns dataset (NYU Stern, updated January 2026) used for the return, volatility and correlation figures
• Elroy Dimson, Paul Marsh and Mike Staunton, UBS Global Investment Returns Yearbook (UBS, annual): long-run returns for dozens of countries since 1900
• Morningstar: Active fund manager success rates ticked up in 2026, but passive funds still hold the advantage (Paoli, 6 August 2026) (opens in a new tab): mid-year 2026 Active/Passive Barometer results
• S&P Dow Jones Indices: SPIVA scorecards (opens in a new tab): active vs index performance and persistence reports, US and international
• Sharpe (1991), The Arithmetic of Active Management (opens in a new tab), Financial Analysts Journal 47(1): why the average active dollar must lag after costs
• Fama (1970), Efficient Capital Markets: A Review of Theory and Empirical Work (opens in a new tab), Journal of Finance 25(2): the efficient market hypothesis
• Markowitz (1952), Portfolio Selection (opens in a new tab), Journal of Finance 7(1): diversification and mean-variance portfolios
• Barber and Odean (2000), Trading Is Hazardous to Your Wealth (opens in a new tab), Journal of Finance 55(2): the most active traders earned the least
• Vanguard: Cost averaging: Invest now or temporarily hold your cash? (Finlay and Zorn, 2023) (opens in a new tab): lump sum beat cost averaging 68% of the time
• Bogleheads wiki: Three-fund portfolio (opens in a new tab): Taylor Larimore's three-fund approach and variants
• Bogleheads wiki: Rebalancing (opens in a new tab): calendar, threshold and 5/25 bands
• Investor.gov (SEC): Asset allocation (opens in a new tab): the SEC's plain-language introduction to allocation and diversification
• FINRA Regulatory Notice 09-31: Non-traditional ETFs (opens in a new tab): why daily-reset leveraged and inverse ETFs are unsuitable for buy-and-hold investors
• FINRA BrokerCheck (opens in a new tab): check a broker's or firm's registration and disciplinary history
• SIPC: What SIPC protects (opens in a new tab): $500,000 limit including $250,000 cash, and what isn't covered
• Robert Shiller: Online data (opens in a new tab): long-run S&P prices, earnings and CAPE
• John C. Bogle, The Little Book of Common Sense Investing, 10th anniversary ed. (Wiley, 2017): the case for low-cost index funds from Vanguard's founder
• William Bernstein, The Four Pillars of Investing, 2nd ed. (McGraw Hill, 2023): theory, history, psychology and business of investing
• Ian Ayres and Barry Nalebuff, Lifecycle Investing (Basic Books, 2010): the case for more stock exposure early in life
• Burton Malkiel, A Random Walk Down Wall Street, 50th anniversary ed. (W. W. Norton, 2023): efficient markets and indexing for general readers