Criticism of the Chancy Islands Model

Several criticisms, concerns, and objections have been raised about the Chancy Islands model:

1. Were the initial conditions and parameters of the model deliberately set to produce the outcome you wanted?

Though the model is relatively simple, there are a dozen parameters that determine how it plays out. These parameters could be set to a wide variety of values that would lead to diverse results.

To make the model a useful and understandable tool for exploring wealth stratification caused by chance expenses, we chose particular values that are objectively reasonable and self-consistent, and that were also illustrative. These chosen settings eliminate a whole slew of complicating factors that would obscure the main points we are trying to make, and they accentuate certain developments. Specifically, the parameters were chosen to meet these five criteria:

  1. Identical households: Every household on the Chancy Islands is identical, with the same initial wealth (except on Skewed Island), the same expenses, and the same income. There is no discrimination or other barriers that prevent any household from prospering. There is no financial exploitation or fraud that would pit one household against another. There is no wealth stratification at the beginning (except on Skewed Island), and no intention or effort to create wealth stratification. This ensures that only luck is responsible for the effects we see.
  2. Comfortably prosperous society: Every household on the Chancy Islands (except on Skewed Island) begins the study period with a fully-paid-for house (worth $200,000) and a significant amount of savings ($5,000).[1] Also, on average, every Chancy Islands household has exactly enough income to meet their expenses. If the Chancy Islands were, instead, a highly constrained society (overall poor and struggling), competition for limited resources could be fierce and wealth stratification might be expected. By making life on the Chancy Islands prosperous overall, there is no inherent reason for societal wealth stratification to develop or persist (as Parity Island demonstrates).
  3. Values similar to a real society: Household income is comparable to the median value in the United States and expenses are comparable to the household average in the U.S. Bankruptcy provisions, interest rates, and tax rates on the Chancy Islands are somewhat reflective of (and in most cases, more favorable than) those in the United States (see the box below). These settings ensure that the spreadsheet model has some relevance to our own society and that there are some lessons that might be applicable to our society — it is not just a silly computer game or an arbitrary garbage-in/garbage-out exercise.
  4. Initial wealth set just above the level necessary to obtain loans: The initial wealth amount is set so that any household that encounters a large expense or catastrophic loss will need to immediately seek a home equity loan — and hence be subject to interest payments which will pull the household’s finances down. This setting is not unreasonable — in American society, half of households have a home mortgage and many households live on the edge of default. Many poorer households in rental housing live even more tenuously. This paramenter setting enables the model to clearly demonstrate how random costs can quickly and easily skew a society’s wealth distribution.
  5. Routine expenses are the same every year: Real households, when faced with severe adversity, would likely cut back their routine expenses to the bare minimum — though most real households find it difficult to reduce their expenses very much because cutting big expenses like housing or transportation often requires moving and cutting back on healthcare (like not taking expensive medications for diabetes) can lead to bad long-term consequences. On the other hand, real households, when faced with severe adversity are often affected for more than a single year (recovery often takes a while).
    The model assumes all Islanders behave exactly the same way every year and their routine expenses are identical. The model also assumes that every year their luck is completely random, so a bad year is often followed by a good year. Though unrealistic, these assumptions have offsetting impacts, make the model simple, and enable us to see more clearly how random costs can skew a society’s wealth distribution.

Ways that Chancy Islanders are better off than people in the United States

  • Every Chancy Islander is a fully educated and capable adult. In contrast, all Americans are dependent infants, toddlers, and children for many years and then, at the end of life, many are disabled or feeble elders. Even in the prime of life, most Americans have one or more physical or emotional limitations.
  • Chancy Islanders have no obligations to anyone outside of their completely capable spouse/partner. In contrast, most Americans have substantial responsibilities for taking care of children, aging parents, ailing friends, etc.
  • In the beginning, every Chancy Islander is decidedly upper-middle class with a house fully paid off and wealth higher than most Americans.
  • In the beginning, Chancy Islanders have no debt — no mortgages, no car loans, no credit card debt, and no student loans. In contrast, most Americans have some debt and many are swimming in oceans of debt.
  • Chancy Islanders never get laid off or fired from their jobs. They are only prevented from working when they are sick or suffering from a natural disaster. In contrast, Americans are subject to the vagaries of the fluctuating job market, especially when recessions batter the economy.
  • Chancy Islanders never lose their savings when the stock market plummets, like Americans do.
  • Everyone in the Chancy Islands model begins at the same level, so no one feels inferior and no one is treated badly. In contrast, many Americans are ignored or devalued for being poor, female, non-white, gay, foreign-born, or disabled.

2. The Chancy Islands model is a really poor economic simulation of a society.

The Chancy Islands model is not intended to actually simulate a real society. It is a very simple model that ignores a great number of crucial aspects of a real society including birth, death, childhood, old age, disability, etc. Regular household consumption (routine expenses) remains the same even when households grow extraordinarily rich or wretchedly poor, which is not at all reflective of real societies. It does not include important structures that we have built in our real societies to cope with bad luck, such as extended family ties and responsibilities, college scholarships, insurance policies, and government safety net programs (like Unemployment Insurance, Social Security, Social Security Disability Insurance (SSDI), VA Benefits, Medicare, and Medicaid). Also, the mitigation measures in the model (loans, bankruptcy, and taxes) are implemented in a very rudimentary way, unlike how they typically operate in real societies.

The model was deliberately designed to be simple so that a few features could be spotlighted and the results of those features could be seen clearly without being tangled in the myriad strands of reality that would typically obscure them. A real society is so complicated and economically messy that it is difficult to pick out a clear signal from the noise surrounding it. In a simple, controlled model, we can isolate just a few features and see what happens when we alter them.

The Chancy Islands model should never be considered realistic — it is by design simple and focused to highlight just a few ideas. The model should never be twisted to explain or demonstrate anything beyond its simple features. The numerical results should be viewed, at best, as rough indicators of the direction and magnitude that would likely occur in a real system.

3. But doesn’t that mean that we really can’t learn much from this model?

It means we should be very careful in drawing any large or solid conclusions about what would happen in the real world based merely on what happens in the Chancy Islands model. For example, a ten percent wealth tax would likely have many more effects in a real society than it does on Combo Island. Some of those effects might be helpful in mitigating wealth stratification, but others might be destructive or counter-productive.

Still, the model does suggest some important ideas:

4. In real societies, insurance ensures that households do not have to pay the full cost of major losses — these costs are spread over many years (through monthly premiums) and across thousands of policyholders. Why is there no insurance on the Chancy Islands?

When it is available, insurance does a good job of spreading risk and shielding people from the full cost of illness, accidents, and natural disasters. We chose not to include it here for three reasons:

  1. A comprehensive insurance program — one that covered everyone in a society — would provide similar protection (at similar costs) as the benevolent bankruptcy system does in the Chancy Island model: Insurance mitigates severe costs much like the Chancy Island bankruptcy-and-equity-restoration program does in helping those hurt by injury or disaster, and the cost of comprehensive insurance (that covers everyone) is equivalent to everyone paying for bankruptcies.
  2. Insurance is often not available or is prohibitively expensive: Insurance is often not available for some of the largest risks that people face. For example, in the United States, flood insurance is often not offered in flood zones, earthquake insurance is often not offered in earthquake zones, and — before the Affordable Care Act — health insurance was often not available to those with pre-existing health conditions. When insurance is available, the cost is often prohibitive for everyone but the wealthiest,[2] so most people do not buy insurance unless the government requires it or subsidizes it, which then makes it similar to a government social program.
  3. Adding insurance would complicate the model, making it difficult to understand: The Chancy Islands model is very simple, which makes it much easier to see how each component affects the results. Typical insurance policies are complex, with some conditions/events not covered or only partially covered if they exceed a deductible amount. Insurance premiums change annually and depend partly on the insurance company’s payouts and investment returns. If insurance were available, but voluntary, then there would be another source of randomness — those who chose insurance versus those who chose not to — intersecting with the randomness of illness, accidents, and weather disasters. Adding insurance would muddle the basic economic mechanisms that we are trying to highlight without providing any unique insight.

5. By leaving out so many aspects of a real society, isn’t this model unrealistic and limited?

The Chancy Islands model does not include any of the personal factors that might cause economic stratification because the purpose is to explore the impersonal and random factors. And though the model is quite simple, it does include elemental versions of all the main impersonal levers that affect wealth distribution:

The details and stringency of all the real-world impersonal financial levers differ, but at an elementary level they have the same basic effects as the simple representations in the Chancy Islands model. So we can use the model to see the how these basic elements affect wealth distribution and gain some understanding of how real-world systems may operate.

6. Why did you use exponential distributions for the cost of Unexpected Extra Expenses and Catastrophic Losses? Why did you round down the calculated amounts to zero — doesn’t that risk skewing the results?

To mirror what occurs in the real world, Unexpected Extra Expenses and Catastrophic Losses should strike households randomly. In addition, most households should have no costs or very small costs each year (few households experience catastrophes most years), a few households should have very high costs (those subjected to the full brunt of the worst disasters), and a somewhat larger number should have moderate costs (those on the edge of major disasters or who experience minor disasters/illness/ accidents).

It is easy to generate that kind of cost distribution by feeding a random number into an exponential equation with a negative exponent. In the Chancy Islands model, the results are truncated to the next lower integer, which makes it easier to see the results and it avoids possible accumulated round-off errors. This also means the smallest generated results are truncated to zero (which in 50 years add up to very little) to better reflect the reality that most households have zero costs from distant disasters and from illnesses and accidents that happen to others. Any other method of generating costs that produces a similar distribution should also yield similar results to the Chancy Island model.

7. The Chancy Islands model has no productivity growth and no investment sector. Doesn’t this make it unrealistic and skew the results?

For the sake of simplicity, the Chancy Islands model assumes a steady-state society. Physically, the model assumes that the environment provides an endless supply of natural resources to the Islanders in the form of plants and animals that can be harvested, trees that can be lumbered, and minerals that can be mined, but also regularly takes resources away from them in the form of destructive natural disasters, illnesses, and accidents. This provision of resources to, and taking away from, the Chancy Islanders is assumed to be balanced over time to make the model simpler. It is monetized in the form of income and costs spread evenly across the whole society, reflecting the equal talent and effort of every Chancy Islander.

Productivity growth would make it possible for the Chancy Islanders to reap more physical goods from the environment and lose fewer resources to destruction over time. If this growth were also distributed equally, it would mean that all Chancy Islanders would be better off over time, but the financial distribution among them would remain the same. Since all Chancy Islanders are equally talented and exert equal amounts of effort, there is no reason why productivity growth would be distributed in any way other than equally.

8. Are you saying that talent and effort have no effect on wealth inequality? Does hard work account for nothing?

No. Clearly those with abundant talent who work hard are much more likely to become wealthy in the United States and most real societies. But there are many talented and hard-working people who do not become rich (and some who are dreadfully poor). There are also wealthy people who have little talent and have never worked hard. There is clearly some correlation between effort/talent and wealth, but there are other factors that also significantly impact the outcome (wealth inheritance being the most obvious one).

The question explored here is whether random luck — in the form of illnesses, accidents, and severe weather — is also an important factor. This analysis, using parameters that are reflective of a real society, finds that luck can produce inequality similar to what we find in reality. It appears that luck is an important enough factor in wealth inequality that it cannot be ignored.

Back to TopUp Arrow
Previous Page Arrow Previous | Next Next Page Arrow

1. ^

In contrast, the Federal Reserve Board in 2017 found that 44 percent of American households surveyed could not cover a $400 emergency expense.

2. ^

For example, the insurance industry touts annuities as an effective alternative to Social Security; but typical annuity policies are difficult to understand, have high fees, and are troublesome to alter or cancel once purchased.