top of page
Writer's pictureCharles Miller

Abraham Wald & why he matters today

This is not investment advice. Do your own research before taking a position in any of the securities mentioned in the post below.


For those who have never heard the story, Abraham Wald was a Hungarian-Jewish statistician who reviewed air bombers that returned to base during the war. He also used detail maths to address a difficult war-time problem...


You don’t want your planes to get shot down by enemy fighters, so you reinforce them with armour. But armour makes the plane heavier, and heavier planes are less maneuverable and use more fuel. Somewhere between too little & too much armour is the easiest way to save lives and/or win the war.


He noticed that you cannot conclude the cause of failure for planes that do not return to base, therein lies the concept of survivorship bias. So, analysing the damaged parts of a returning plan was the wrong place to start.


In fact, the parts of the plan which were untouched suggested that these were the areas of the plan which were the weakest and where armour/reinforcement was required. The mathematics was more precise than this, however, his formulaes concluded that armour was being missued. With the new analysis in hand, crews reinforced the bombers' cockpit, engines, and tail armor. The result was fewer fatalities and greater success of bombing missions.


How does this apply to markets & investing?


The adages I take from Abraham Wald is

  • The best performing companies are the ones that reinforce the right parts of their business model (noting that it may not be the same for each company)

  • Sometimes the immediate conclusion (e.g. reinforce the broken parts) is not the best solution

  • Survivorship bias is another bias that we should not forget about

Examples of reinforcing the 'important areas' of the business

Admittedly, I want to delve into this in a number of other posts, however, examples of companies that need to 'reinforce' key: parts of the downturn

  • Overlevered finance (e.g. Resimac and Liberty): Appreciating these are all very different companies... the key question is whether that these growing non-bank lenders manage their gearing and financing sources in a downturn

    • Note these business have quite different business models for example RMC = RMBS issuer and LMI insurer, LFG = part non-conforming commercial loan originator

  • Small M&A led company: There are many examples, however, companies which are heavily dependent upon "bolt-on" acquisitions for growth. Old examples included Slater & Gordon or ABC Learning

  • Effectiveness vs. Efficiency: Although I like efficient companies, that do do buy-backs & pay big dividends, I also admit that companies with limited growth profile (e.g. telstra) are also no better either. Companies need to grow or die.

When obvious truths are not obvious

I won't dwell on this point too much because the application is very wide and there are way better writers in this space such Thinking Fast Thinking Slow or Reflexivity however below are some 'current affair' questions that you may think are easy but they may be complicated to answer

  • Omnicron is less deadly vs. RSA fatality is simpy lower given smaller population?

  • If markets slow, will asset fallr or is further economic stimulus drive things higher than they otherwise would

  • If Amazon UK blocked Visa; is their business model under threat?

Survivorship bias is one of many biases which can impact your investing

  • Finally - on the topic of survivorship bias, it is important to remember that this bias along with mny others are possible sources of error in how you invest. A great list of biases is set out here

So in conclusion, Abraham Wald taught us a great deal about mathematics & analysis. Some takeaways I think are worth remembering today are:

  1. Make sure you/your investments are armouring the right part of their business. Too much armour is a bad thing, but so is too little

  2. Immediate conclusions whilst easy, are not always the best ones

  3. Biases are always impacting our judgement so make sure you identify these biases

9 views0 comments

Recent Posts

See All

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
Post: Blog2_Post
bottom of page