Risk Models Ch. 3: Risk Parity — Balancing Risk Instead of Capital

Investment Background

A standard 60% stocks / 40% bonds portfolio looks diversified. Sixty percent in one asset class, forty in another — by dollars, that is meaningful diversification. By risk, it is nearly a pure equity bet.

Stocks are roughly three to four times more volatile than investment-grade bonds. Do the arithmetic on risk contribution rather than capital contribution, and the picture inverts: the 60% equity allocation typically contributes 85-90% of the portfolio's total risk (its variance in returns), while the 40% bond allocation contributes only 10-15%. The "diversified" 60/40 portfolio behaves, in a crisis, almost exactly like an all-equity portfolio — because equities are the thing actually driving the outcome.

Risk Parity, popularized institutionally by Ray Dalio's Bridgewater All Weather approach, starts from a different question entirely. Instead of asking "what fraction of my capital goes to each asset," it asks "what fraction of my total risk comes from each asset, and how do I make those fractions equal?"

This chapter covers the logic of risk parity, how it is actually constructed, and — critically — the leverage requirement that makes it inappropriate to replicate directly in most individual retirement accounts.

The Wall Street Translation

The Core Problem With Dollar-Weighting

Capital allocation and risk allocation are different things, and conflating them is the central error risk parity is built to fix.

Allocation view	60/40 portfolio
By capital (dollars)	60% equities, 40% bonds — looks balanced
By risk contribution	~85-90% equities, ~10-15% bonds — is not balanced at all

The consequence shows up exactly when it hurts most. In an equity bear market, a 60/40 portfolio falls substantially, because the 40% "diversification" was never large enough, in risk terms, to offset the 60% that actually drives volatility. The portfolio was diversified on paper and concentrated in practice.

How Risk Parity Actually Balances Risk

The mechanical goal: size each asset class's position so that each contributes roughly equal risk to the total portfolio, rather than equal capital.

Because bonds are less volatile than stocks, achieving equal risk contribution from bonds requires holding more bonds, in dollar terms, than a capital-weighted approach would — often substantially more. Applied naively, this would leave a large fraction of the portfolio in low-return bonds, dragging down overall expected return.

This is why real-world risk parity strategies use leverage. The construction, roughly:

1. Determine target risk contribution for each asset class (commonly: roughly equal across equities, bonds, and sometimes commodities/inflation-linked assets).
2. Weight each asset class inversely to its volatility — more capital to low-volatility bonds, less to high-volatility equities, so that each contributes similar risk.
3. Because this produces a lower-return, lower-volatility mix than an all-equity portfolio, apply leverage to the whole mix to bring total expected volatility back up to a target level (often, to roughly match a traditional stock-heavy portfolio's volatility) while retaining the more balanced risk composition.

Leverage is not incidental to risk parity — it is structurally required to make the strategy competitive on return with a simpler, unlevered equity-heavy portfolio, because reducing equity's dominant risk share necessarily reduces the mix's raw expected return unless leverage restores it.

The Diversification Payoff — and the Correlation Risk

When it works, the appeal is real: a more balanced risk portfolio tends to lose less in equity-driven crashes, because bonds (which often behave differently than stocks in a growth scare, though not always in an inflation scare) carry a real weight in the risk budget rather than a token capital allocation.

But the entire strategy rests on one assumption: that stocks and bonds maintain a low or negative correlation. This assumption held for most of the last four decades. It is not a law of markets.

Period	Stock-bond correlation behavior
Most of 1980s-2010s	Generally negative or low — bonds rallied when stocks fell, delivering genuine diversification
2022	Positive — both stocks and bonds fell together as inflation and rate hikes hurt both simultaneously

2022 is the case study every risk parity investor now has to reckon with. A leveraged risk parity portfolio, built on the assumption that bonds hedge equities, took losses on both legs at once — and because the strategy uses leverage, those simultaneous losses were amplified rather than cushioned. This is the single largest risk in the risk parity approach: it depends on a correlation relationship that can and does break, precisely in the inflationary regime where investors need the hedge most.

Executable Trading Rules

1. Do not evaluate your portfolio's diversification by dollar allocation. Evaluate it by risk contribution. A 60/40 or 70/30 split is not "moderately risky" in the way its labels suggest — check what fraction of the portfolio's volatility each piece actually drives.

2. If you hold bonds specifically for their diversifying effect against equities, understand that this effect is regime-dependent, not guaranteed. It works well against growth scares (recession fears drive both a stock selloff and a flight to bonds) and can fail against inflation scares (rate hikes hurt both simultaneously).

3. Do not attempt to replicate institutional risk parity with retail leverage (margin, leveraged ETFs) without understanding that leverage amplifies the correlation-breakdown risk described above. The 2022 case is the exact failure mode: leverage turns a bad year into a much worse one when the hedge fails.

4. Treat "equal risk contribution" as a useful lens for evaluating any portfolio, even one you never intend to leverage. Simply knowing that your stock allocation dominates your risk far more than its dollar weight suggests changes how you interpret "diversified."

Relevance to a Retirement Portfolio

Risk parity, as institutions run it, is not something we recommend replicating directly. It requires leverage, active rebalancing, and sophisticated risk modeling that are appropriate for an institutional balance sheet and inappropriate for most individual retirement accounts — the 2022 correlation breakdown is exactly the kind of risk that a levered retail investor is least equipped to absorb.

What genuinely transfers to your retirement portfolio is the diagnostic lens, not the leveraged implementation:

Risk parity's lesson	How it applies without leverage
Dollar allocation misrepresents true risk exposure	Check your actual risk contribution, not just your stock/bond split, when assessing how "aggressive" your portfolio really is
Bonds' diversification benefit is conditional, not guaranteed	Do not assume your bond allocation will cushion every downturn — it cushions growth scares reliably, inflation scares unreliably
True diversification means genuinely different risk drivers, not just different asset labels	Favor a core that spans equities, bonds, and geographies with genuinely different return drivers over one that is diversified in name only
Leverage amplifies both the strategy and its failure mode	This is the strongest argument against using margin or leveraged ETFs to "improve" your retirement account's risk-adjusted return

Our standard position holds here too: this chapter does not recommend leveraging your retirement portfolio to chase risk parity's institutional returns. It recommends understanding that your risk, not your capital, is what actually determines your outcome in a crisis — and using that understanding to size your equity and bond allocations with open eyes, rather than assuming a round-number split like 60/40 is automatically balanced simply because the numbers look moderate.

A low-cost, globally diversified core remains the foundation. The risk parity lens is valuable for understanding what that core actually exposes you to — not for building a leveraged replica of it.

Chapter 4 goes deeper into the mechanics behind this chapter's central tool: correlation matrices — what they actually measure, why they are unstable exactly when you need them most, and how to use them without being misled by them.