Why is trend growth often harder to forecast than business cycles, even though trends are about long-term averages?

This is a genuinely counterintuitive point in the curriculum, and understanding it sharpens your entire approach to long-horizon CME.

The Intuitive (But Wrong) View:

On the surface, trends seem easier to forecast than cycles. Cycles involve short-term turning points, inflection moments, and timing decisions — notoriously difficult tasks. Trends are long-term averages that smooth out short-term noise. So surely long averages are more stable and predictable?

Why That Logic Fails:

The trend-is-easier argument only holds if trend growth rates are CONSTANT. If the trend is a fixed 2.5% real per year, then yes — you just use the historical average and you're done. But trend growth rates are not constant. They change — and when they change is precisely when forecasting matters most for investment outcomes.

Two Categories of Trend Changes:

1. Forecastable Changes (Demographics, Capital, Education):

These drivers evolve slowly and are directly observable. Japan's demographic collapse was clear by the 1990s — anyone projecting Japanese trend growth for 2010-2030 could incorporate it with high confidence. Similarly, the rise of China's working-age population through 2015 was mechanically predictable from fertility data 20 years prior.

An analyst using 40 years of Japanese data to set long-run equity CMEs would be using a trend that includes the high-growth pre-1990 era — systematically overestimating forward returns. Demographic analysis would have flagged this years before the data confirmed it.

2. Unforecastable Changes (Shocks):

Truly exogenous shocks cannot be specifically predicted. Worse, even after they occur, their effect on the trend is difficult to identify until the change is "well-established and retrospectively revealed in the data." This creates a multi-year blind spot where your trend assumption is wrong and you cannot yet prove it.

Example — Crestfield Institutional Research:

In 2010, Crestfield's analyst wondered whether the 2008-09 financial crisis had permanently impaired US trend growth or whether the economy would revert to the pre-crisis 3% real trend.

By 2015 — six years post-crisis — GDP had recovered but the gap vs. the pre-crisis trend line was widening rather than closing. By 2018, it was clear trend growth had downshifted to roughly 2%.

But in 2010-2012, the data was too noisy to prove the regime change. An analyst waiting for statistical confirmation would have used a 3% trend for 3-4 years while the true trend was 2% — producing systematically biased CMEs through a critical recovery period.

The Cyclical Contrast:

Ironically, cycles are often easier to assess in real time because you have more tools:

• Leading indicators (yield curve, ISM, claims)
• Coincident indicators (employment, production)
• Policy signals (central bank guidance)
• Market-based signals (credit spreads, breakeven inflation)

Cycles are noisy but observable. Trend changes are invisible until they're obvious in hindsight.

Practical CME Implications:

1. Don't just compute a historical mean and call it your trend estimate
2. Decompose the trend into forecastable drivers (demographics, productivity, capital) and estimate each
3. After major shocks, actively hypothesize whether the trend has changed, rather than passively waiting for the data to "prove" it
4. Use wider ranges for long-horizon forecasts than for short-horizon ones — the opposite of what intuition suggests

Explore more trend forecasting in our CFA Level III question bank.