Regime-Dependent Indicator Selection
Track the indicators that matter for the economy's current regime
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 95%
Economic indicators do not have constant value across every phase of the cycle. The framework begins with a stable baseline, such as unemployment and payrolls, then identifies the dominant current risk. When recession probability is low, weekly unemployment claims may add little actionable information. If growth deteriorates and recession becomes plausible, a sharp rise in claims becomes a key confirmation signal. During an unusual inflation episode, CPI and PCE components may deserve greater attention, especially when a measurement issue such as lagged rents affects interpretation. The method therefore treats indicators as conditional tools rather than a fixed ranking. Analysts define activation conditions in advance, increase the frequency or weight of relevant measures when those conditions occur, and stop giving them disproportionate attention once the regime changes.
Origin
Bill McBride explained that the most useful economic measure changes with the situation and used weekly unemployment claims as a conditional recession indicator. Extracted from Masters in Business.
Core principles
- 01Indicator importance changes over time
- 02Match each metric to a specific economic question
- 03Escalate recession indicators when slowdown risk rises
- 04Avoid monitoring every release with equal intensity
- 05Use conditional signals rather than permanent predictions
How to run it
- 1
Define the active risk
State the primary question the monitoring system must answer, such as whether recession risk is rising or inflation is becoming persistent. Keep the question specific enough to guide indicator selection.
Pro tip Write the risk as a falsifiable sentence rather than a broad concern.
Watch out Trying to monitor every possible risk equally recreates information overload.
- 2
Maintain baseline indicators
Track a small group of measures that remain broadly informative, such as unemployment and payrolls. Use them to establish the economy's normal state.
Pro tip Keep the baseline dashboard stable so changes remain comparable over time.
Watch out A baseline indicator may be useful without being decisive for the active risk.
- 3
Choose conditional indicators
Select metrics whose mechanisms directly address the active question. For recession monitoring, weekly unemployment claims become valuable when labor-market deterioration is plausible.
Pro tip Document why each conditional indicator should move if the risk is materializing.
Watch out Historical correlation alone is weaker than a clear causal or operational mechanism.
- 4
Set activation thresholds
Specify the conditions that will increase an indicator's importance, such as a sustained sharp rise rather than a single noisy release. Decide what would count as confirmation or disconfirmation.
Pro tip Use persistence and rate of change together where possible.
Watch out Vague thresholds invite hindsight bias.
- 5
Rotate attention with the regime
Increase attention to activated indicators and reduce it when their associated risk fades. Replace them when a new economic question becomes dominant.
Pro tip Review the active indicator set on a fixed monthly or quarterly schedule.
Watch out Do not retain a once-useful signal merely because it previously made a correct call.
In the wild
Payroll growth is weakening, housing activity has declined, and recession risk is now credible. The investor activates weekly unemployment claims as a high-frequency confirmation indicator and watches for a sustained sharp increase rather than reacting to one report.
→ Monitoring intensity rises only when claims can materially change the recession assessment.
Employment remains resilient and recession indicators are stable, but uncertainty about central-bank policy is high. The investor reduces the weight given to weekly claims and focuses on CPI, PCE, and the composition of inflation until the policy outlook becomes clearer.
→ The dashboard directs attention toward the data most relevant to current portfolio risk.
Common mistakes
Following every release equally
Equal attention creates noise and obscures the few indicators capable of changing the current decision. Monitoring should be selective and conditional.
Using a signal outside its regime
An indicator can be valuable during a specific risk phase and unimportant at other times. Applying it continuously can generate false urgency.
Changing indicators after the fact
Selecting whichever metric supports a preferred conclusion creates confirmation bias. Activation rules should be documented before the data arrive.
Is it for you?
Best for
It is best for investors who need a focused monitoring system that adapts as inflation, growth, and recession risks change.
Not ideal for
It is not ideal for analysts who must maintain comprehensive databases or cover every release professionally.
From the transcript
“What's important over time changes.”
“That's important when you really do think that there's a possibility of a recession, if that really starts climbing sharply, that's probably your key indicator.…”
“investors should realize they don't need to follow every data release, every news report, every economic announcement that comes out”
From the episode
At the Money: What Data Matters and What Doesn't