I think that there will be higher growth in the data science field within the insurance industry because many of those jobs simply did not exist before. The advent of more readily available data and cheaper computing power makes advanced predictive modeling possible for smaller and smaller carriers and cost effective for more lines of business. Data scientists are not replacing actuaries so much as filling these new roles. This presents more of an opportunity for young actuaries to expand their skills than a threat to the profession.
The problem with the sexy pattern discovery of data science is that the world is a dynamic place. Recognizing when the pattern stops working is just as important as discovering the pattern in the first place. And often, that occurs due to something outside the data (a change in the building code, for example, or tort reform). There is also the risk that improved predictive analytics transforms insurance from risk transfer to risk financing.
Giving context to the data is the value that an actuary brings. Running models is not helpful unless you talk to UW, Claims, Sales, etc. to understand what is changing. The guy mentioned in the article is probably very proficient in creating models but needs someone to bridge the gap between pure data and the real world. Actuaries are good at that and will continue to provide that service.
We should sit down for a cup of coffee before you make any assumptions about my communication skills! In general, I think it is hard to make a blanket statement comparing the communication skills of the two professions. Backgrounds are usually very similar and the experience of taking exams vs. the experience of studying some flavor of machine learning, deep learning, or reinforcement learning isn’t going to crown a clear winner in the communication sector.
If you’ve interacted mostly with data scientists in an insurance company, it is quite possible you have a small, biased sample. All the top actuaries work at insurance companies. My hypothesis is that the top data scientists do not.
In the coding world, people are highly social and share information at incredible rates. StackOverflow is a prime example. This industry will finally start seeing a much needed long delayed shakeout, insurance is the least efficient industry in the world. Its about time this changes.
In the long term, underwriting positions will be minimized. I think of this as the “Money Ball” effect of insurance. For years, underwriters follow old rules on risk acceptance and application of schedule mods where applicable. The long term impact of data science will be to allow machine learning to select risks and modifications. The difficult challenge for smaller markets will be obtaining the appropriate data to make proper selections. In previous years, a small company could always obtain rates and guidelines from big entities as a barometer for their own rates and guidelines. ISO filled that void in the commercial world. Now the challenge is that those rates and guidelines are not available as easily and ISO has not kept pace. Some vendors have stepped in, but keep their data and processes confidential.
a dart board rating method works for some companies. You look at their dec pages and the household risk and just shake your head. Some companies lower the rate when you add on a youthful driver, because there is another driver in the household. ummm?
The insurance industry is abundant with data. But just due to regulatory reasons or so, these data were not fully utilized. Actuaries don’t have the necessary skills to do modern data analysis. I understand they are changing curriculums, but the knowledge in the new system still isn’t comparable to what a stats/ds master program entails. Actuaries should concede the data sci work to data scientists, and just focus on the traditional rate filing and reserving stuff. If someday insurance companies don’t need to do rate filing (in other words, a no-file system) and don’t need to list reserve on their liability column in the financial stmt, then they can just stop hiring actuaries. A data scientist with a life or p&c underwriter credential is good enough.
Actuaries have given up working in the one area of expertise where they are specifically trained for; retirement pensions.
Longevity, financial market volatility, and human arrogance all combined to support the statement that individuals are the ones that should take retirement risk, an enterprise in which a minimum number are qualified, for and at odds with the benefits of working with populations of Large Numbers.
The proposal of new retirement models should come from actuaries, but who are focused on this today?
I did extremely well as an “actuary” in a large insurance company. I retired at age 50 with more wealth than I will ever need in two life times. Here is how it happened: I combined computer skills (programming, data analysis, modeling (including stochastic) in my actuarial career. I also added great communication and presentation skills in to the mix. The rest was history. I predict that in the near future there will be a new profession that will be a combination of actuary and data scientist. This combination will be much better than actuarial profession or data scientist, especially if you add communication and presentation skills. This has been my experience.
This is a good discussion of the pros and cons of pursuing a career in my profession. Much better than the usual “best jobs” nonsense.
I think that there will be higher growth in the data science field within the insurance industry because many of those jobs simply did not exist before. The advent of more readily available data and cheaper computing power makes advanced predictive modeling possible for smaller and smaller carriers and cost effective for more lines of business. Data scientists are not replacing actuaries so much as filling these new roles. This presents more of an opportunity for young actuaries to expand their skills than a threat to the profession.
The problem with the sexy pattern discovery of data science is that the world is a dynamic place. Recognizing when the pattern stops working is just as important as discovering the pattern in the first place. And often, that occurs due to something outside the data (a change in the building code, for example, or tort reform). There is also the risk that improved predictive analytics transforms insurance from risk transfer to risk financing.
Giving context to the data is the value that an actuary brings. Running models is not helpful unless you talk to UW, Claims, Sales, etc. to understand what is changing. The guy mentioned in the article is probably very proficient in creating models but needs someone to bridge the gap between pure data and the real world. Actuaries are good at that and will continue to provide that service.
We should sit down for a cup of coffee before you make any assumptions about my communication skills! In general, I think it is hard to make a blanket statement comparing the communication skills of the two professions. Backgrounds are usually very similar and the experience of taking exams vs. the experience of studying some flavor of machine learning, deep learning, or reinforcement learning isn’t going to crown a clear winner in the communication sector.
If you’ve interacted mostly with data scientists in an insurance company, it is quite possible you have a small, biased sample. All the top actuaries work at insurance companies. My hypothesis is that the top data scientists do not.
In the coding world, people are highly social and share information at incredible rates. StackOverflow is a prime example. This industry will finally start seeing a much needed long delayed shakeout, insurance is the least efficient industry in the world. Its about time this changes.
In the long term, underwriting positions will be minimized. I think of this as the “Money Ball” effect of insurance. For years, underwriters follow old rules on risk acceptance and application of schedule mods where applicable. The long term impact of data science will be to allow machine learning to select risks and modifications. The difficult challenge for smaller markets will be obtaining the appropriate data to make proper selections. In previous years, a small company could always obtain rates and guidelines from big entities as a barometer for their own rates and guidelines. ISO filled that void in the commercial world. Now the challenge is that those rates and guidelines are not available as easily and ISO has not kept pace. Some vendors have stepped in, but keep their data and processes confidential.
Actuaries don’t have the best communication skills across all professions, but I would rank them in front of most data scientists I know.
a dart board rating method works for some companies. You look at their dec pages and the household risk and just shake your head. Some companies lower the rate when you add on a youthful driver, because there is another driver in the household. ummm?
The insurance industry is abundant with data. But just due to regulatory reasons or so, these data were not fully utilized. Actuaries don’t have the necessary skills to do modern data analysis. I understand they are changing curriculums, but the knowledge in the new system still isn’t comparable to what a stats/ds master program entails. Actuaries should concede the data sci work to data scientists, and just focus on the traditional rate filing and reserving stuff. If someday insurance companies don’t need to do rate filing (in other words, a no-file system) and don’t need to list reserve on their liability column in the financial stmt, then they can just stop hiring actuaries. A data scientist with a life or p&c underwriter credential is good enough.
Actuaries have given up working in the one area of expertise where they are specifically trained for; retirement pensions.
Longevity, financial market volatility, and human arrogance all combined to support the statement that individuals are the ones that should take retirement risk, an enterprise in which a minimum number are qualified, for and at odds with the benefits of working with populations of Large Numbers.
The proposal of new retirement models should come from actuaries, but who are focused on this today?
I did extremely well as an “actuary” in a large insurance company. I retired at age 50 with more wealth than I will ever need in two life times. Here is how it happened: I combined computer skills (programming, data analysis, modeling (including stochastic) in my actuarial career. I also added great communication and presentation skills in to the mix. The rest was history. I predict that in the near future there will be a new profession that will be a combination of actuary and data scientist. This combination will be much better than actuarial profession or data scientist, especially if you add communication and presentation skills. This has been my experience.