Second-Order Thinking

Thinking Tools4.3 Crafting Solutions for the Future StateAlso known as Second-Level Thinking, And Then What?

Set out as second-level thinking by Howard Marks (The Most Important Thing, 2011); the habit of tracing an act's later effects goes back to Frédéric Bastiat (That Which Is Seen, and That Which Is Not Seen, 1850) and Henry Hazlitt (Economics in One Lesson, 1946).

Follows an action past its first effect to the reactions and consequences that follow, including how competitors, customers, and regulators will adjust.

A 5% price cut followed past its first effect to where it settles

The top branch is what the plan counted; the lower branch asks who responds, then what follows from each response. Select a box to read it, and select the link under the tree to set the plan's figure beside the result. Illustrative.

Links between branches:

Select a box to read it, or the link to compare the plan with the result.Cut list price 5%. The move: a distributor of office supplies, $60M in sales at a 40% gross margin, cuts list price 5% to win share.First order: what the plan counted. The effect everyone in the room could see.Orders up 20%: margin up $1.2M. At a 5% lower price each order earns less, but 20% more orders more than make up for it: gross margin rises from $24.0M to $25.2M.Second order: who responds. Everyone the move touches gets a turn, judged by what they did the last time something like this happened.Two rivals match within weeks. Both matched the last cut in this market within five weeks.The order gain falls to 5%. Third order: once prices are level again, few customers have a reason to switch.Margin down $2.0M a year. Where it settles: 5% more orders at a 5% lower price leaves gross margin near $22.1M, against $24.0M before the cut.Customers learn to wait for cuts. Buyers at large accounts see that list price is open to change.Orders bunch around promotions. Third order: demand swings from month to month, and the warehouse staffs for the peaks.Reps sell on price, not service. A cut announced from the top tells the sales team what wins deals.Big accounts ask for more off. Third order: the next round of contract renewals opens with a request for a further discount.The plan beside where it settles. First-order view: margin up $1.2M. Once rivals match, margin down $2.0M. The difference, $3.2M a year, is what the plan left out.

An example

The commercial director of an office supplies distributor proposed cutting list prices 5 percent to win share. His model showed orders rising 20 percent and gross margin rising $1.2 million on sales of $60 million. The numbers were right as far as they went.

The finance director asked one question of each party the cut would touch: and then what? The two largest rivals had matched the last cut in that market within five weeks, so the 20 percent gain would shrink to about 5 percent once prices were level again. Large accounts would learn that list price moved and start waiting for promotions. The sales team would hear that price wins deals and open the next renewals with a discount. With the rivals' response alone, the cut left gross margin about $2.0 million lower than before.

The team kept list prices and offered free next-day delivery to mid-size accounts instead, a move the rivals could not copy without new trucks. Orders from those accounts rose 8 percent within two quarters.

Why it works

The first effect of a move is the one everyone can see, so it is already in the plan and often in the rivals' expectations too. The result depends on what comes after: other parties adjust, and their adjustments change the conditions the first effect assumed. A price cut assumes prices stay apart; a match removes that assumption.

The tool works because it forces a turn for each party, judged by what that party has done before. That turns vague worry about consequences into a short list of specific responses, each with a number. It also shows the asymmetry that makes many obvious moves bad: the gain arrives now and shows in this quarter's results, while the cost arrives later and lands on someone else's report. Moves that hold up are often ones whose later effects work in their favor.

The thinking behind it

Howard Marks Howard Marks, cofounder of Oaktree Capital Management and known for his memos to investors, opens The Most Important Thing (2011) with second-level thinking. First-level thinking asks what will happen; second-level thinking also asks what others expect to happen and what follows from the gap between the two. His subject is investing, where a forecast everyone shares is already in the price, but the habit carries to any move that others will respond to. The book is a practitioner's account.

Frédéric Bastiat Frédéric Bastiat, a French economist and pamphleteer, made the point in That Which Is Seen, and That Which Is Not Seen (1850). In its parable of the broken window, the glazier's new work is the effect everyone sees; the effect not seen is whatever the window's owner would have bought with the money instead. A good economist, he argued, counts both, and second-order thinking applies the same rule to a decision's later effects.

Henry Hazlitt Henry Hazlitt, an American economic journalist, restated Bastiat's point in Economics in One Lesson (1946): the art of economics lies in looking at the longer as well as the immediate effects of a policy, and at its consequences for all groups, not only one. The book is a popular argument for free markets. Its one lesson supplies the tool's two questions: what happens later, and to whom else.

How to run it

  1. State the move and its first effect. Write both with a number: cut list price 5 percent, orders up 20 percent.
  2. List everyone the move touches. Name the specific rivals, customer groups, staff, suppliers, and regulators.
  3. Ask how each will respond. Use what they did the last time something similar happened.
  4. Ask the same of each response. A rival's match changes what customers do; follow each response one step further.
  5. Stop when the effects settle or blur. Two or three steps are usually enough; past that, estimates turn into guesses.
  6. Set the end state beside the first effect. Put a number on where it settles and compare it with the plan's figure.
  7. Look for a move whose later effects help. One rivals cannot copy quickly often beats the obvious one.

Where it goes wrong

The chain never ends. Every move has some later downside, and the team stops deciding. Spot it when the third and fourth steps are guesses; stop at the step where effects blur and ask whether going further would change the choice.

The responses are chosen to fit. Every party reacts in the way that makes the favored move look good. Spot it when no response hurts; have one person argue the most damaging plausible response, based on what that party has done before.

Timing gets ignored. The gain shows this quarter and the cost shows next year, so the plan gets credit before its costs appear. Report the plan against the settled figure, not the first one.

When to use something else

  • Game Theory. When a few named rivals choose in response to each other and their payoffs can be laid out side by side.
  • Causal Loop Diagram. When the responses feed back into the original cause, so the effects run in loops.
  • Futures Wheel. When the question is the wide spread of consequences from a trend or event, not one decision.

Where it sits in SOLVE eX

Go to the source

  • The Most Important Thing: Uncommon Sense for the Thoughtful Investor, Columbia University Press, 2011
  • That Which Is Seen, and That Which Is Not Seen, 1850
  • Economics in One Lesson, 1946

Back to the catalog