Scott.Lampman 4-2-14
Adapting
- Tis a complex world
- Book aims to understand how any problem really gets solved in complexity
- In Search of Excellence (1980s), many of these firms are failed today
- Failure is ubiquitous
- Biologists have a word for the way in which solutions emerge from failure: evolution.
- Complexity emerges from a simple process: variation and selection – V&S (repeat).
- The problems keep changing, biology works by producing “works for now” solutions
- Biological evolution usually takes small steps, but sometimes takes wild leaps.
- V&S work in the market economy as well.
- Ormerod’s discovery: effective planning is rare in modern economy.
- Tetlock’s studies: expert judgment is marginal gain.
- Lesson: don’t cover up or deny failures, we need them to adapt
- Palchinsky Principles (27):
- Seek out new ideas [Variation]
- Do it on a scale where failure is survivable [Survivability]
- Seek out feedback & learn from mistakes [Selection]
- Variation is difficult due to tendencies: grandiosity (large projects) and uniform standards
- Lesson: decentralize and become comfortable with chaos of local approaches and dissent.
- Kahneman & Tversky: “chasing a loss” – the psychology of risk: a person who has notmade peace with his losses is likely to accept gambles that would otherwise be unacceptable.(34)
- To produce new ideas (A), we must overcome our tendency to fall into step with those around us and to overcome those with a vested interest in the status quo.(39)
- To make failure survivable (B), use both small steps (but risk big ones periodically).
- Distiguishing success from failure (C) is difficult (arrogant leaders ignore/deny).
Conflict or: How Organizations Learn
- Military examples = idealized hierarchy: use a big picture strategy, entire team pulling in same direction, proper flow of information vertically.
- Problem:
- Big picture can lead to self-deluded propaganda (GroupThink)
- A single unified direction is counter to Variation.
- Chain of command can prevent feedback/dissention.
- Vietnam: Johnson/McNamara(bad), Iraq: Rumsfeld (bad), Gates/Petraeus/McMaster.. (good)
- Computers – centralization – incapable of capturing local tacit knowledge
- There is a limit to how much variation/experimentation to undertake.
Creating New Ideas that Matter or: Variation
- The Spitfire example. ROI is simply not a useful way of thinking about new ideas.
- The idea that we can predict which new ideas will flourish, flies in the face of evidence.
- In biology, “Speciation” rarely happens without some form of physical isolation.
- Today, more complex problems often require whole teams of specialized personnel
- USG patent system may not encourage the innovations we really need (103)
- USG grants have also often failed to deliver the full potential of innovation
- NIH expert led research grants – low risk, designed to avoid failure not on success.
- H.Hughes Medical Inst. Grants – high risk, high failure rate, few strings attached.
- Lessons: create as many separate experiments as possible, encourage some long-shots – both are the antithesis of government planning (113)
- The Prize methodology (e.g. Longitude, Gates Foundation) don’t cost a penny untilsuccess is achieved, allows combination of open field, failure toleration and risky ideas with big reward
- But, can governments know the costs, benefits and envision innovation to set prize target?
Finding What Works for the Poor or: Selection
- Development as business of governments – which are poorly placed to spot failures
- It is difficult to distinguish between what works and what doesn’t (PlayPump example)
- “randomists” (control trials RCTs) and the ethics of exclusion
- RCTs are powerful tools and help in Selection
- Cochrane’s “God Complex” – assumes we already know the answer (142)
- “Fundamentally unidentified questions” cannot be answered by experiment (e.g.corruption)
- We need better feedback loops.
- Lubeck, Germany 1158, locally distinctive rules for economic development
Climate Change (CC) or: Changing the Rules for Success
- CC discussion confuses targets with policies.
- Any answer is going to come either because of individual voluntary behavior change or because governments change the rules
- A carbon tax uses the market system – each decision maker has an incentive to reduce carbon
- Merton Rule 2003 – any development must include capacity to generate 10% of its energy req.
- Problem with Merton – installation does not mean use, disallows for local logical adaptation
- Time and again environmental regulation makes the Merton mistake – perverse incentive
- Orgel’s Maxim (OM): if the problem is misstated then evolution is likely to find loopholes
- Orgel’s Maxim: if the rules of the economics are poorly written, loopholes will be found, resulting in perverse outcomes (e.g. biofuels cause deforestation)
- Better rules could turn OM to our advantage – harnessing the ingenious solutions
- Root cause of the loophole problem (like Merton) the letter v.s. the spirit of the law (192)
- Nudge (Thaler/Sunstein): subtle influences could be used to direct thoughtless behavior, while preserving individual rights consciously to choose (think marketing)
- Carbon pricing tries to harness Orgel’s Maxim by focusing on what we think the ultimate goal is: a reduction of the greenhouse gas emitted into the atmosphere, at the lowest possible cost.
- OM tells us the economic evolution, with the playing field tilted by the new rule,‘greenhouse gases are expensive’ will produce entirely unexpected ways to reduce greenhouse gases
- Carbon pricing will work because it takes a global objective – reduce GG emissions – and delegates that objective. Business understands the costs. Governments should not pick and choose specific solutions, they should tilt the playing field to encourage us to make all our decisions with the planet in mind.
Preventing Financial Meltdowns or: Decoupling
- Failure is both necessary and useful
- Certain systems allow little room for trial an error (e.g. nuclear plants)
- A dangerous combination is a system that is both Complex and ‘Tightly Coupled’ (e.g. dominos)
- The art of success is to fail productively (202)
- Complexity means there are many different ways for things to go wrong.
- Tightly Coupled means the unintended consequences proliferate so quickly that it isimpossible to adapt to the failure (e.g. banking system)
- To make complex systems safer, install safety measures (Swiss cheese model ofaccidents)(203)
- Examples: the credit default swaps (CDS) and collateralized debt obligation (CDO)transformed small problems into catastrophes (increasing complexity and tight coupling).
- Examples: Lehman Brothers, AIG, zombie banks, decoupling proposals (e.g. CoCobonds)
- Lesson: Palchinsky’s second principle: make failures survivable – hence, decouple. (226)
- Three different types of error:
- Slips – clumsiness or lack of attention resulting in error
- Violations – involve someone deliberately doing the wrong things
- Mistakes – things you do on purpose, but with unintended consequences, because your mental model of the world is wrong.
- Slips are easier to spot than Violations or Mistakes and so, V&M lead to more “latenterrors”
- Latent Errors lurk unnoticed until the worst possible moment – w/o noticing the risk of disaster
- Make a priority of reliable indicators to anticipate systematic problems – latent errors
- Eliminating latent errors is impossible, so simplify, decouple and encouragewhistleblowers
The Adaptive Organization
- A crucial difference: individuals, unlike populations, can[not] succeed without adapting.
- The guppy population evolved camouflage through trial and error, but no individualguppy did
- Three essential steps to using the principles of adaptiong in business (Palchinskyprinciples):
- Try new things, some will fail (V)
- Make failure survivable – create a space for failure or take small steps (find theright scale in which to experiment)
- Make sure you know when you’ve failed
- Organizations having adapted Palchinsky’s principles of “pluralism” often succeed.
- Whole Foods: decentralized, local autonomy, self-selecting, empowered
- The correct balance between centralized control & decentralization depends on circumstances
- Google (and evolutionary org.), has a 20% time policy – providing space for innovation
- Disruptive innovations don’t appeal to traditional customers. Market niches provideopportunity
- More often, the problem is not technological, but psychological and organizational
- The Virgin Group conglomerate maintains strong separation, it allows for failure inisolation
- The concept of “limited liability” encourages people to experiment (265)
Adapting and You
- Be willing to fail
- Find safe spaces to fail
- Three obstacles to learning from mistakes: denial, compensating for losses; rose tinted glasses
- Denial (cognitive dissonance) prevent us from learning from our failures
- “hedonic editing” (Thaler, Nudge) convincing ourselves that the mistake doesn’t matter(274)
- A different psychological process: reinterpreting our failures as successes (validation)
- We systematically reinterpret our past decision as being better than it really was
- “disciplined pluralism” (how markets work) should be a credo for a successful life
- Combine both leaps and small steps – be willing to risk failure, without it we cannot succeed.