Asset Allocation Explained: How to Split Stocks, Bonds and Cash

Most of the attention in investing goes to picking things. Which fund, which sector, which month to buy in. The choice that actually determines the shape of your outcome is duller and gets a fraction of the airtime: how much of your money sits in stocks, how much sits in bonds, and how much sits in cash.

That split is your asset allocation. It is the structural decision in a portfolio. You do not control returns and you have limited control over your income, but you have near-total control over the mix -- and the mix sets both the range of outcomes available to you and the size of the loss you will have to sit through to reach them.

What follows is the mechanics. What an asset class is, why combining them helps, how to choose a mix you can still hold when holding it is unpleasant, and what maintenance the mix needs afterwards.

#What an asset class actually is

An asset class is a group of investments that earn money the same way and therefore respond to the same forces. Three of them do most of the work in an ordinary portfolio.

Stocks, or equities, are ownership stakes in businesses. You are entitled to what is left after the company pays everyone else: staff, suppliers, lenders, tax authorities. That residual claim is why stocks have historically produced the highest long-run returns, and it is also why they are violent in the short run. The leftover is the most uncertain part of any business.

Bonds, or fixed income, are loans. You lend to a government or a company, collect scheduled interest, and get the principal back at maturity. The return is contractual rather than residual, which caps the upside and usually -- not always -- softens the downside.

Cash and cash equivalents are checking accounts, savings accounts, money market funds and very short-term government bills. The nominal value barely moves. The purchasing power does move, because inflation grinds at it quietly.

Other classes exist -- real estate, commodities, inflation-linked government bonds -- but a portfolio built from the first three and held for decades does most of what a complicated one does, with far fewer moving parts.

Note: These categories describe behavior, not labels. Two funds can both be called bond funds and behave nothing alike if one holds two-year government debt and the other holds twenty-five-year corporate debt.

#Why the mix dominates the picks

Consider what actually moves a portfolio on a given day. If 90 percent of your money is in stocks, the portfolio moves when stocks move, and which particular broad stock fund you own barely registers, because broad stock funds move together. The shared factor swamps the differences.

That is the whole argument for taking allocation seriously. Fund selection changes your result at the margin. Allocation changes it structurally.

Worked example: Two people each hold $200,000. Investor A is 100 percent stocks. Investor B is 60 percent stocks and 40 percent bonds. A year arrives in which stocks fall 35 percent and bonds gain 5 percent.

Investor A: $200,000 x 0.65 = $130,000. A loss of $70,000, or 35 percent.

Investor B: the stock sleeve is $120,000 x 0.65 = $78,000. The bond sleeve is $80,000 x 1.05 = $84,000. Total $162,000. A loss of $38,000, or 19 percent.

Now the part people skip. Recovering a loss takes a bigger percentage gain than the percentage that was lost, because the gain works on a smaller base. Investor A needs $200,000 / $130,000 = 1.538, a gain of 53.8 percent, to get back to even. Investor B needs $200,000 / $162,000 = 1.235, a gain of 23.5 percent. Both will probably get there. One of them has a much harder time staying invested while it happens.

#Diversification is about correlation, not count

Owning twenty funds is not diversification if all twenty hold large American companies. You would own one bet, expensively, twenty times over.

The property that matters is correlation: the tendency of two holdings to move together. Correlation runs from +1 (they move in lockstep) through 0 (no relationship) to -1 (they move in opposite directions). Combining two assets reduces the swings of the total only when their correlation is below +1, and the further below, the greater the smoothing.

This is why adding bonds to a stock portfolio changes the character of the portfolio while adding another stock fund mostly does not. It is also why the useful question about a new holding is never "is this good?" but "what does this do that I do not already own?"

Correlations are not fixed. They drift with the economic regime, and they have an unhelpful habit of rising toward +1 during panics, exactly when you wanted them low. Diversification is a reliable long-run structure, not a guarantee about any single week.

#Risk capacity versus risk tolerance

These get used interchangeably and they are not the same thing.

Risk capacity is structural: how much loss your circumstances can absorb without forcing a change of plan. It is set by facts. When you need the money, how stable your income is, whether you have other assets or a pension, whether anyone depends on you, how large your emergency fund is.

Risk tolerance is psychological: how much loss you can watch without selling. It is set by temperament and by experience, and people routinely overestimate it in calm markets.

The binding constraint is whichever is lower. A 28-year-old with a thirty-year horizon has enormous capacity, but if a 30 percent drop would make her liquidate at the bottom, her effective tolerance is what governs, and a mix she will actually hold beats a theoretically optimal one she will abandon.

How to tell them apart: capacity questions have factual answers, and tolerance questions are answered by your own history. The most informative evidence is what you actually did during the last serious market decline, not what you predict you would do during the next one.

#Time horizon is the main input

Horizon, not age, is the primary driver. Horizon means when the money will be spent, and over what period.

A house down payment three years out is a single-date goal. The money has to be there on a specific day, so exposure to a 30 percent drawdown is not a risk you are compensated for taking. Cash and short-term instruments do the job; see saving for a house down payment for the mechanics.

Retirement is different, and it is where the most common error lives. Retiring at 67 is not a horizon of zero. It is the start of a spending stream that might run thirty years. The dollars you plan to spend at 90 have a sixty-year horizon on the day you turn 30 and a twenty-three-year horizon on the day you retire. That is why an all-cash portfolio at retirement is itself a risk decision, not the absence of one.

#Age-based rules of thumb and their limits

The classic shorthands say to hold your age in bonds, or 110 minus your age in stocks, or 120 minus your age for a more aggressive version. A 40-year-old under the "110" rule holds 70 percent stocks.

These are useful for one reason: they produce a defensible starting point in ten seconds, and a defensible starting point held consistently beats an optimal one you never get around to choosing.

They break wherever age stops being a good proxy for horizon or capacity. A person with a substantial pension covering essential spending has far more capacity than the rule implies, because the portfolio is funding discretionary spending only. A person whose income is tied to the same economic cycle as their portfolio has less. Someone intending to retire at 50 has a shorter accumulation phase but a longer decumulation phase, which pulls in two directions at once. And someone with a portfolio far larger than they need has capacity the formula cannot see.

When age and circumstance disagree, circumstance decides.

#What bonds actually do -- and why they fall too

Bonds have three jobs in a portfolio, and only one of them is "make money."

1. Dampen the swings, so the whole thing stays holdable.
2. Provide something to sell during a stock decline, whether to spend or to rebalance into stocks.
3. Fund spending that is close enough that stock volatility is intolerable.

The mechanic people find surprising is that bonds can lose money. Bond prices and market yields move inversely. If you hold a bond paying $30 a year on $1,000 of face value and new bonds start paying $50, nobody will buy yours at $1,000. The price falls until the buyer's total return matches the going rate.

The sensitivity is measured by duration, roughly in years. A quick approximation: price change is about minus duration times the change in yield.

Worked example: You hold $10,000 in a bond fund with average duration of 7 years, yielding 3 percent. Market yields rise by 2 percentage points.

Estimated price change: -7 x 2% = -14%, so about -$1,400. The balance drops to roughly $8,600.

But the fund now reinvests at 5 percent instead of 3 percent. Annual income goes from 3 percent of $10,000, about $300, to 5 percent of $8,600, about $430 -- roughly $130 more a year.

Recovery comes sooner than $1,400 / $130 suggests, because the marked-down balance itself compounds at the higher rate. Growing $8,600 at 5 percent overtakes the $10,000 you would have had growing at 3 percent after about 7.8 years -- a little beyond the 7-year duration.

That is the underappreciated symmetry: a rate rise hurts a bond holder whose horizon is shorter than that crossover and helps one whose horizon is longer, and duration approximates where that crossover falls. Choosing bond duration is really choosing to match your spending timeline.

It also explains why stocks and bonds sometimes fall together. When the shock is a growth scare, capital rotates out of stocks and into bonds, and the two offset. When the shock is a jump in inflation and interest rates -- as in 2022 -- the discount rate applied to every future cash flow rises, and stocks and bonds fall at the same time. A stock-bond mix reduces the frequency and depth of bad years. It does not abolish them.

#Domestic versus international

The United States is a large share of global stock market value, but not all of it. Holding only domestic stocks is a concentrated bet on one country's companies, regulations, currency and demographics being the best place to be for the next several decades.

Arguments for holding non-US stocks: leadership between regions has rotated over long stretches; currency exposure diversifies a portfolio whose liabilities are all in dollars; and it removes the need to be right about which country wins.

Arguments people make against: costs and taxes are marginally higher, currency movements add short-run noise, and large domestic companies already earn revenue abroad. That last one is weaker than it sounds, since a company's share price still tracks its home market's sentiment.

There is no correct number. Investors commonly land somewhere between 20 and 40 percent of their stock allocation in international holdings, and the choice matters far less than being consistent about it.

#Allocation profiles by horizon and capacity

The table below is a starting framework, not a prescription. The final column is an illustrative planning assumption for a severe bear market -- the number to look at and ask whether you could sit still. It is not a forecast, and actual outcomes have been both better and worse.

Read the last two columns together. A 90 percent stock portfolio of $400,000 losing 36 percent is a paper loss of $144,000 in a year. If reading that sentence produces a flinch, that is useful information about the row above it.

#Rebalancing, and why it feels wrong

Left alone, a portfolio drifts. Stocks grow faster than bonds most of the time, so a 60/40 mix quietly becomes 70/30 and then 75/25. Nobody decided to take more risk. It happened by arithmetic, and it typically peaks just before the risk shows up.

Rebalancing is selling some of what grew and buying some of what did not, to restore the target. Two common disciplines:

• Calendar rebalancing. Check on a fixed schedule -- annually is common -- and correct any drift. Simple, predictable, ignores what markets are doing between checks.
• Threshold rebalancing. Correct only when an asset class is more than a set distance from target. An absolute band of 5 percentage points (act when the 60 percent stock sleeve leaves the 55-65 range) or a relative band of 25 percent of the target weight are both used.

Many people combine them: look once or twice a year, act only if a band is breached. That keeps trading and taxes down while still catching real drift.

Worked example: You start at $100,000, split $60,000 stocks and $40,000 bonds. Over a year stocks return 25 percent and bonds return 2 percent.

Stocks: $60,000 x 1.25 = $75,000. Bonds: $40,000 x 1.02 = $40,800. Total: $115,800.

Stock weight: $75,000 / $115,800 = 64.8 percent. That is inside a 5-point band (55-65 percent), so a threshold rule takes no action while a calendar rule does.

If acting: target stock value is $115,800 x 0.60 = $69,480. Sell $75,000 - $69,480 = $5,520 of stocks and add it to bonds, bringing them to $46,320.

Every one of those actions feels wrong at the moment you take it. In year 3 you are buying the thing that just fell 30 percent, using money from the thing that held up. That discomfort is the mechanism working, not a sign that something is broken. The purpose is to keep risk at the level you chose; any return benefit is a secondary effect and is not dependable.

Two practical notes. Rebalancing inside a 401(k) or IRA has no tax consequence, so that is where to do it first. In a taxable account, directing new contributions and dividends toward the underweight asset can restore the target without selling anything.

#Glide paths and target-date funds

A glide path is a pre-set schedule that reduces the stock share as a target date approaches. A target-date fund packages one: you buy the fund with the year in its name and the mix shifts automatically for decades.

The design question is what happens at the target. A "to" glide path reaches its most conservative mix on the retirement date and stops. A "through" glide path keeps reducing stock exposure for years afterward, on the reasoning that the horizon does not end at 67. Two funds with the same year on the label can hold very different stock percentages, which is worth checking rather than assuming.

The trade-off is honest. You get automatic discipline and automatic rebalancing, which are worth a great deal. You give up the ability to reflect anything the fund does not know -- your pension, your spouse's portfolio, your actual retirement date -- and you accept whatever fee layering the structure carries.

#Asset location is a separate decision

Asset allocation is what you own. Asset location is which account holds it. They are independent, and confusing them causes real damage.

The general logic: assets that generate regularly taxed income, such as taxable bond funds, are less painful inside a tax-deferred account like a 401(k). Broad stock index funds are relatively tax-efficient and sit comfortably in a taxable brokerage account. Some people place their highest-expected-growth holdings in a Roth account, since qualified withdrawals come out untaxed -- the Roth versus traditional comparison covers why.

Two cautions. Location only produces meaningful savings once you hold significant balances across more than one account type. And it must never distort the allocation: the correct total mix held in slightly suboptimal locations beats a tax-perfect arrangement that leaves you 85 percent in stocks by accident.

#What the usual advice gets wrong

Treating a risk questionnaire as a measurement. Five questions answered during a calm market produce a number that describes your mood, not your capacity.

Calling bonds "safe" without finishing the sentence. Safe from what? High-quality bonds are relatively safe from default and from large drawdowns. They are not safe from inflation, and cash is least safe of all on that measure over long periods.

Counting holdings instead of examining correlation. Six funds that all track large domestic companies are one position with extra paperwork.

Assuming stocks and bonds always move in opposite directions. They often do. In inflation shocks they do not.

Folding the emergency fund into the allocation. Cash reserved for a job loss is not a conservative sleeve of the portfolio; it is a separate account with a separate job.

Changing the allocation because of news. The trigger for a change is a change in your circumstances -- a new horizon, a new income, a discovery about your own tolerance. Not a headline.

Precision theater. The difference between 62/38 and 60/40 is noise. The difference between 90/10 and 60/40 is the whole outcome. Spend effort proportionally.

#Putting it together

1. Set aside cash for emergencies and for anything you will spend within two to three years. That money is not part of the investment allocation.
2. Write down the horizon for each remaining pot of money.
3. Pick a stock/bond split from the profile table using the shorter of what your capacity and your tolerance allow.
4. Split the stock sleeve between domestic and international and leave that ratio alone.
5. Choose a rebalancing rule -- a date, a band, or both -- and write it down before you need it.
6. Review annually, and when your circumstances change. Not when the market moves.

If you are still building the habit of contributing at all, the allocation matters less than the contribution rate for the first several years; regular investing on a schedule and understanding what compounding is doing will carry more weight early on. The allocation decision becomes the dominant one as the balance grows.