
Sep 9, 2026 · 1h 8m
AQR investor challenges portfolios built on hidden equity risk
E427: AQR's Peter Hecht on AI, Market Bubbles & How the Best Investors Build Portfolios
Peter Hecht argues that resilient portfolios require understanding their underlying exposures, testing signals rigorously, and diversifying beyond familiar stock-and-bond mixes.
- 1Systematic investing seeks small, repeatable edges across many positions rather than relying on a few high-conviction bets.
- 2AI can accelerate research, but investment judgment still depends on economic reasoning, source quality, and disciplined signal testing.
- 3Trend following and portable alpha can diversify portfolios whose apparent balance conceals substantial equity and inflation risk.
Don't miss
Hecht explains why investors may be carrying hidden equity exposure through private assets and why traditional stock-bond diversification failed during inflationary shocks such as 2022.
The brief
Peter Hecht of AQR describes systematic investing as a search for small, repeatable edges across thousands of positions, with risk measured through underlying exposures rather than holding counts.
The conversation tests familiar factors such as value, size, and momentum against data problems, behavioral explanations, and the practical need to improve a client’s existing portfolio.
AI can help researchers analyze information and weight signals, but Hecht insists that primary-source discipline, economic theses, and judgment remain essential to deciding what deserves inclusion.
Hecht’s sharpest warning is that portfolios often contain more equity risk than investors realize, including through private assets, making scenario analysis and macro diversification indispensable.
Trend following offers a contrasting form of diversification because it can respond to persistent market moves without requiring the negatively correlated, insurance-like hedges that may carry negative expected returns.
What was said on this episode
51 statements · 23 positive · 19 negative · 1 mixed · 8 neutral
Beating a benchmark requires accepting tracking error
“if you want to beat the benchmark, if you want to try to add value, have to take tracking error”
Listen at 1:20
Passive investing’s effect on active managers depends on which managers go passive
“whether they're being more passive or not is good for active managers sort of depends”
Listen at 2:40
The rise of passive investing may have made markets more inelastic
“the rise of passive have made markets more inelastic”
Listen at 3:44
Greater market inelasticity from passive flows could increase stock-market volatility
“which would lead to more volatility”
Listen at 4:23
Diversification can improve returns for a given level of portfolio risk
“through that diversification, through that portfolio construction, we can actually achieve an attractive return for a given level of risk”
Listen at 5:28
Alpha depends on the chosen risk model
“alpha really depends on the risk model you use”
Listen at 7:14
After controlling for beta, AQR found no small-cap effect
“We found there was no small cap effect”
Listen at 11:38
Size was not viewed as a positively rewarded risk factor
“we didn't think it was a rewarded risk factor with a positive return associated with it”
Listen at 11:55
Holding many small positions can limit losses from individual mistaken trades
“you can really minimize the damage to the overall portfolio”
Listen at 14:51
AQR uses document embeddings rather than prompt-based large-language-model analysis.
“When we use large language models, we're doing the latter. We're not doing the prompt based approach.”
Listen at 17:58
The difficult part of financial AI is making text representations predict returns
“The hardest part is, can you find whether that numerical representation of the data can predict future returns?”
Listen at 18:13
Prompt-based investment analysis using language models is inefficient.
“prompt-based is very inefficient”
Listen at 18:31
Forecasting returns from language-model representations requires substantial skill and subjective judgment.
“That is very difficult to do, requires a ton of skill, and it's highly subjective.”
Listen at 18:46
Widespread use of Claude or ChatGPT will not make investment managers converge.
“And that's why I don't think there'll be convergence.”
Listen at 18:53
Shared use of language models will not make investment managers converge
“that's why I don't think there'll be convergence”
Listen at 18:53
Identifying return-predictive signals distinguishes strong systematic managers.
“What separates a good systematic manager from that systematic manager? It's not obvious which signals predict returns and how they predict returns.”
Listen at 18:59
Machine learning is a material component of AQR’s investment process
“AI machine learning. It is definitely a material component to our process”
Listen at 19:33
AQR’s non-machine-learning signals add alpha beyond machine-learning signals
“The non-machine learning based signals have alpha against the portfolio that's only doing machine learning”
Listen at 20:04
Investors should combine machine-learning and non-machine-learning signals.
“you should do both”
Listen at 20:20
Machine learning in finance is limited by the small amount of available data.
“ML will always have some limitation because we actually have small data in finance”
Listen at 20:32
Financial data limits how much AI can comprise AQR’s process
“there will be sort of a cap on how much it will be part of our process”
Listen at 21:07
AQR seeks intellectually curious researchers.
“We want people who are intellectually curious”
Listen at 21:57
Investors should seek another manager if their asset manager is not innovating
“If your asset manager is not innovating, it's not a sufficient condition. If they're not innovating, it's a necessary condition. If they're not innovating, you need to be looking for another manager.”
Listen at 23:47
AQR requires out-of-sample validation for investment signals
“we're looking for out of sample validation”
Listen at 25:45
AQR uses machine learning to adjust signal weights based on live out-of-sample performance
“we've actually systematized it where we actually use machine learning techniques that will sort of downweight and upweight signals based on whether they're working and out of sample live trading”
Listen at 27:46
AQR can reach its target return with roughly a 55% stock-selection hit rate
“if we're right about 55% of the time, we've hit our target return”
Listen at 29:16
AQR seeks to minimize trading impact on its alpha ideas
“we are always trying to minimize the impact on our alpha ideas”
Listen at 33:27
AQR invests over monthly-to-yearly horizons rather than intraday horizons
“We're thinking about the next month, the next quarter, the next year, not the next tick or the next five minutes.”
Listen at 33:41
Multi-strategy funds diversify exposure across strategies believed attractive long term
“If you think they're all good over the long run, get the benefits of diversification in the single line item that does a multi-strategy approach.”
Listen at 34:12
Trend following has the strongest evidence for protective performance in challenging markets
“the only hedge fund sub-strategy that has sort of bona fide evidence for that type of a result is trend following”
Listen at 35:24
Investors should not concentrate portfolios in one signal or theme
“You shouldn't put all your eggs in one signal or one theme like momentum”
Listen at 36:22
Trend following exploits short-run market underreaction to news
“Trend following is taking advantage of the fact that market participants tend to underreact to news in the short run.”
Listen at 37:05
Trend following has delivered attractive long-run returns
“that type of strategy has delivered attractive returns over the long run”
Listen at 37:46
Trend following offers diversification, competitive returns, and protective properties
“trend following, that's the sort of behavioral basis behind it. And that's why people include it in their portfolio for diversifying competitive return that has those protective properties.”
Listen at 39:23
Put options have negative expected returns over the long run
“put options have negative expected returns”
Listen at 40:58
Equity factors explain about 90% of risk in most portfolios
“most people's portfolios, 90% of the risk can be explained by some type of an equity factor”
Listen at 41:58
Scenario analysis can identify portfolio vulnerabilities and improve resilience
“when we do asset allocation, we do those types of exercises to give people a sense of how bad it could get. And are there ways to make it more resilient?”
Listen at 45:33
Inflation shocks tend to increase stock-bond comovement
“when inflation news becomes a dominant force in the macro economy, stocks and bonds tend to move together”
Listen at 46:47
Multi-strategy and trend-following approaches can perform in growth or inflation shocks
“Multi-strat or trend following can do well in either a growth shock or inflation shock.”
Listen at 47:13
An ideal long-only portfolio combines long-run positive-return assets with risk balancing
“our sort of ideal portfolio would try to get the most out of all of the sort of asset classes that offer a positive average return over the long run and we would sort of risk balance them”
Listen at 48:11
A diversified leveraged portfolio can have higher base-case returns than a 60/40 portfolio
“it will have a higher total return base case than 60-40 because my portfolio is more diversified than just holding stocks and bonds”
Listen at 49:42
Long-horizon investors should measure volatility over their investment horizon
“If you're a long horizon investor and your horizon's 10 year, you should measure 10 year volatility.”
Listen at 50:22
Compounding makes long-horizon return distributions highly skewed
“your distribution gets highly skewed”
Listen at 52:34
Volatility lowers the median long-horizon compounded return through volatility drag
“that volatility of that asset actually lowers the actual base case 50th percentile return”
Listen at 53:01
Higher expected returns do not guarantee long-run outperformance
“just because something has a higher expected return with higher volatility doesn't mean it will eventually outperform over the long run”
Listen at 54:14
Multi-strategy hedge funds and trend following can help address inflation and tail risk
“bringing in those multi-strategy hedge funds and bringing in trend following are ways that I continue to preach can help out if you're worried about inflation or you're worried about tail risk”
Listen at 55:43
Equity valuations are currently stretched
“valuations are stretched”
Listen at 56:55
Portable alpha combines S&P 500 returns with hedge-fund excess cash returns minus frictions
“you would get the total return of the S&P 500 plus the excess of cash return of your hedge fund alpha strategy minus any frictions”
Listen at 59:02
Managed-futures trend following gained roughly 30–40% in 2022
“in 2022, the industry was up, let's call it 30% or 40%”
Listen at 1:01:35
Portable alpha addresses limitations created by long-only active management
“portable alpha, someone came to me and said, so what problems does it solve? One, long only active management. is inconsistent because of the long only constraint.”
Listen at 1:02:55
Portable alpha’s main remaining risk is active-management risk
“the main risk you face with portable alpha is the risk of active management”
Listen at 1:05:08
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.