I have a nerdy love for both spreadsheets and finding the best deal. Auto insurance providers do not value loyalty, so I shop our coverage every six or twelve months. Once we reached the point where we had enough savings to cover any deductible level if needed, I wanted to figure out the deductible with the greatest expected value for the premium. This post describes the process I used to shop for the optimal collision and comprehensive deductibles using a breakeven analysis.
Choosing a Provider
I chose the provider almost solely based on price. Insurance is merely a transfer of risk product, and therefore I valued price over other "quality" factors. As long as the provider could be expected to pay up if I needed to make a claim, they were an equal competitor in my quote evaluations.
When comparing quotes, it was important to select baseline coverage values to compare across providers. The chart below shows a reasonable set of coverage selections for which I could get comparable quotes from an array of insurers. These are recent real quotes for covering our pair of Honda Fits in Memphis, TN. I gathered these figures by visiting the websites of each provider individually and providing consistent application information. (Pro Tip: Use a temporary email address when running the quotes to avoid spam.)
We currently insure through Geico for obvious reasons. We are certainly not tied to them permanently, however. The hour it takes to shop quotes again could provide a substantial return.
To make a good deductible selection, I needed to know how many more years would pass before we would need to file a claim on one of our cars. Since I could not predict the future, I used the best available average data for my calculations.
The Insurance Information Institute reported that passenger auto owners made a collision claim 6.05 times out of every 100 car years in the United States in 2015. This value can be translated to 1 claim every 16.5 years (6.05/100). Similarly, insured made comprehensive claims only 2.73 times out of every 100 car years (or once every 36.6 years). Our actual expected claim frequencies undoubtedly vary widely by factors such as driver skill and location, but these averages provided reasonable assumptions for calculating a breakeven analysis.
|Coverage||Claim Every X Years|
The deductible values of the policy had a significant impact on the premium. Assuming that I had enough set aside to cover any deductible, how do I know the optimal deductible to purchase?
Since I decided above that we would purchase from Geico, I ran their quote at all deductible levels. Fortunately, Geico breaks out their total premium into each of the different coverage types. The chart below compares the cost of comprehensive and collision coverage at various deductibles for each of our two cars.
Raising the deductible lowers the cost of the policy because I am transferring less risk to the provider. I assumed that we will file claims at roughly the rates described above. For each deductible level, I calculated how many claim-free years must pass for the premium savings to surpass the extra deductible risk.
For example, raising the deductible from $100 to $150 for collision coverage on our 2013 Fit lowered the annual premium from $290.30 to $269.68. It would take 2.42 years of no claims for the extra risk associated with this higher deductible to pay off in premium savings. Since I assumed that we will only file such a claim every 16.5 years, I opted to take on that extra risk.
The figures in bold are the lowest premiums where the years to breakeven in savings is less than the expected years of claim-free driving; thus, they are the selections I chose.
The best way to improve this process would be to get more accurate claim frequency data. While we obviously can never know how many years will pass between claims, it would be good to be able to get averages based on factors such as location, age, and car value. As far as I know, though, such data is not publicly available (at least for free).
Feel free to grab a copy of this Google Sheet to run your numbers. Let me know if it helped, or how I can improve my analysis.