Skip to content

Growth-Trend Timing and 60-40 Variations: Lethargic Asset Allocation (LAA): summary and key takeaways

The paper applies growth-trend timing, which flags a bear market only when both US unemployment and the S&P 500 trend badly, to a 60-40 style portfolio. A search on an early sample yields LAA, which switches between two near-identical four-asset portfolios. Over the full sample LAA earned more than 60-40 with about half the maximum drawdown.

Keller, W.J. · 2019 · SSRN Electronic JournalSummary updated 20 November 2026
Published2019Working paper
Asset focusMulti-assetTAA
TypeBothTactical Allocation
Reading time4 minthis summary

Growth-Trend Timing and 60-40 Variations: Lethargic Asset Allocation (LAA) at a glance

The question
Can growth-trend timing, combined with an equal-weight portfolio chosen in-sample, beat a static 60-40 portfolio with far fewer trades than usual trend-following tactical strategies?
Dataset & period
US ETF proxies (SPY, IWD, QQQ, GLD, IEF, SHY and others) plus US unemployment data, monthly, Feb 1949 - Oct 2019 (nearly 72 years). In-sample Feb 1949 - June 1981, out-of-sample June 1981 - Oct 2019, plus the most recent 10 and 20 years. A 0.1% one-sided transaction fee is applied.
Method
Growth-trend timing signals risk-off only when the unemployment rate is trending up and the S&P 500 is below its own moving average. The author starts from 60-40 (SPY-IEF), then runs a stepwise in-sample search that adds risky assets one at a time to maximise the Ulcer Performance Index. The risky portfolio found is QQQ, IWD, GLD and IEF. The cash portfolio swaps QQQ for SHY, chosen as the best in-sample bond. The timing signal moves between these two portfolios monthly.

Key findings of Growth-Trend Timing and 60-40 Variations: Lethargic Asset Allocation (LAA)

  • Static 60-40 over the full sample returned 9.5% a year with a maximum monthly drawdown of 29.5%.
  • Switching SPY to IEF with growth-trend timing lifted the full-sample return to 13.2%, but the maximum drawdown stayed about the same, falling only from 29.5% to 29.4%.
  • The final LAA portfolios hold 25% each: QQQ, IWD, GLD and IEF in the risky state, with SHY replacing QQQ in the defensive state.
  • LAA returned 10.5% over the full sample with a maximum drawdown of 15.0%, roughly half the 60-40 drawdown.
  • The timing signal sat in cash 13% of months from Feb 1949 to Oct 2019 and under 2% over the last 10 years, so the strategy trades rarely, about once in three years on average (25 switching months in total).
  • Over the most recent 10 years LAA lagged 60-40: 9.5% versus 10.2% in return, and 5.13 versus 6.21 in Ulcer Performance Index.

Robustness, as reported in the paper

  • Adding a one-month publication lag for the unemployment data moved full-sample return and drawdown from 10.5% and 15.0% to 10.4% and 15.0%, and the in-sample search picked the same portfolios.
  • Raising the transaction fee tenfold changed full-sample return only from 10.5% to 10.2%.
  • With a faster trend filter on unemployment, full-sample return stayed above 10% and drawdown below 16% for all slower filters tested. Very fast filters pushed drawdown to 18% and 20%.
  • Making both trend filters much faster, plus the lag, gave a full-sample return of 9.9% and a drawdown of 20%.

Caveats and limits

  • Asset selection and the cash portfolio were optimised on the same history the paper reports, and the author notes the split design is meant to limit data snooping, not remove it.
  • Long-history ETF returns are proxies built from index and Fama-French data and adjusted by regression, so pre-ETF numbers are constructed, not traded (inferred from the data notes).
  • The out-of-sample window is dominated by a long run of falling yields and strong tech returns, which helped QQQ in the risky portfolio (inferred).
  • Over the most recent 10 years the strategy trailed 60-40, and the signal rests on a single US unemployment indicator.

Items marked (inferred) are our reading of the design, not statements by the authors.

Why it matters

The paper shows a low-turnover way to add crash protection to a diversified portfolio, with a risk-off state that changes only one of four holdings. It also shows how sensitive such results are to the choice of in-sample window, assets and cash universe. Practitioners comparing tactical overlays can use the cash-fraction and transaction-cost figures as a benchmark against heavier trend-following rules.

Questions about this paper

How often does growth-trend timing move to the defensive state?

From Feb 1949 to Oct 2019 it signalled risk-off in 13% of months. Over the last 10 years the figure was under 2%.

What does LAA hold?

In the risky state it holds QQQ, IWD, GLD and IEF at 25% each. In the defensive state SHY takes the place of QQQ.

How much do transaction costs matter for LAA?

Very little in the paper's test. Raising the fee tenfold lowered the full-sample return from 10.5% to 10.2%.

Where can I read the original paper?

The original is at its publisher: https://doi.org/10.2139/ssrn.3498092. We link to the paper and do not host the PDF.

Follow new quant research

Download MacroFinance market data as CSV or Excel, or pull it by API, and test the ideas in these papers yourself. New summaries are added as we review them.

Create a free account
Summaries are written by MacroFinance from the published papers, in our own words, for information only. They are not investment advice and may omit detail. Papers are not hosted here: each page links to the original at its publisher.