Through-Cycle Multiples and Peak-Trough Analysis

Introduction

Caterpillar's trailing EV/EBITDA has ranged from 7.65 to 26.0 over the past 13 years. At the trough of the 2015-2016 mining downturn, with EBITDA compressed and the stock beaten down, the trailing multiple expanded above 20x. By 2023, with EBITDA at a record $16.14 billion (up 39% year-over-year), the trailing multiple compressed to roughly 11-13x. The company looked "cheapest" when earnings were highest and "most expensive" when earnings were lowest. This is not a Caterpillar-specific anomaly; it is the defining valuation characteristic of every cyclical industrial company, and understanding why it happens, what it means, and how to use it is one of the most important analytical skills in the sector.

Peter Lynch warned investors about this exact trap with cyclical stocks: buying on low P/E ratios at earnings peaks can be dangerous because those peak earnings are about to decline, while high P/E ratios at troughs often signal the best buying opportunity because earnings are about to recover. In M&A, the same principle applies: a sell-side banker positioning a company at 10x trailing EBITDA during a peak must help buyers understand that the "cheap" trailing multiple actually represents a full mid-cycle valuation once you normalize earnings, while a buyer acquiring at 18x trailing EBITDA during a trough may actually be getting a bargain on normalized economics.

The mechanics are straightforward but often misunderstood. A cyclical company's enterprise value (the numerator of EV/EBITDA) reflects the market's assessment of the company's long-term earning power across many future cycles, while trailing EBITDA (the denominator) reflects only the most recent twelve months of actual performance. Because enterprise value is forward-looking and relatively stable (varying perhaps 20-35% peak to trough), while trailing EBITDA is backward-looking and volatile (varying 40-70% peak to trough due to operating leverage), the ratio between them swings dramatically.

Inverse Multiple Trap

The counterintuitive phenomenon where cyclical companies display their lowest trailing valuation multiples at the peak of the earnings cycle and their highest trailing multiples at the trough. Caterpillar's EV/EBITDA median over 13 years is approximately 13x, but the trailing multiple swings from below 8x at earnings peaks to above 20x at troughs. An investor who screens for "cheap" industrials using low trailing multiples will systematically identify companies at earnings peaks, buying at the worst possible moment. In M&A, a buyer who offers 10x trailing EBITDA for a company at peak earnings may actually be paying 14x+ on a mid-cycle basis, which is a full price that the trailing multiple disguises.

Consider the math with simplified numbers. A capital goods manufacturer has mid-cycle EBITDA of $200 million. At peak, operating leverage pushes EBITDA to $320 million (60% above mid-cycle). At trough, EBITDA collapses to $100 million (50% below mid-cycle). If the enterprise value ranges from $2.4 billion at peak to $1.8 billion at trough (a 25% decline, far less than the 69% EBITDA swing):

The mid-cycle EV/EBITDA column reveals the truth that trailing multiples obscure: the company is cheapest at the trough (9.0x mid-cycle) and most expensive at the peak (12.0x mid-cycle), the exact opposite of what trailing multiples suggest.

The through-cycle approach pairs mid-cycle EBITDA (the normalized earnings base calculated using historical averaging, margin regression, or capacity utilization adjustment) with a through-cycle multiple (the EV/EBITDA that the market has historically applied to normalized earnings for comparable companies). The product is the through-cycle enterprise value.

The through-cycle multiple should reflect the company's quality tier. Using Caterpillar's median of 13.0x as a benchmark for standard cyclical OEMs, the tiers adjust up or down:

European cyclical industrials (Siemens, ABB, Atlas Copco, Sandvik) typically trade at similar through-cycle multiples to US peers, though market-specific factors (investor base composition, listing liquidity, currency) can create 1-2 turn differentials. Atlas Copco, for example, commands a persistent premium to US diversified industrials peers because of its compressor aftermarket model and disciplined capital allocation.

While through-cycle multiples provide a point estimate of value, peak-trough analysis provides the valuation range by explicitly mapping how earnings and enterprise value behave at each extreme. This analysis is particularly valuable for PE sponsors evaluating LBO candidates (who need to understand downside risk for debt capacity sizing) and for boards of directors evaluating acquisition offers (who need to know whether the offered price is fair across the full cycle, not just at the current earnings level).

The peak-trough framework involves three scenarios:

Peak scenario. Model the company's EBITDA at the peak of the cycle (using historical peak margins applied to current or near-term revenue). Apply a compressed trailing multiple (reflecting the market's anticipation of future earnings decline). The peak enterprise value represents the theoretical maximum if the company is sold at the perfect moment.

Mid-cycle scenario. Model mid-cycle EBITDA using the normalization methods. Apply the through-cycle multiple. This is the central value estimate and the most defensible in M&A negotiations.

Trough scenario. Model trough EBITDA (using historical trough margins, typically 30-50% below mid-cycle due to operating leverage). Apply an expanded trailing multiple (reflecting the market's anticipation of future earnings recovery). The trough enterprise value represents the theoretical minimum and is critical for LBO downside testing.

The peak-trough framework also helps bankers advise on M&A timing. The best time to sell is when trailing earnings are at or near peak levels and the CIM can frame the earnings as sustainable (using secular growth arguments and leading indicator data that support continued momentum). The best time to buy is at the trough, when the trailing multiple looks scary but the mid-cycle multiple reveals genuine value.