CIA.DCAT

This reading has technically been removed from the syllabus effective Fall 2020 but please read IMPORTANT NOTE

Dynamic Capital Adequacy Testing is a stress-testing technique. It is a critical tool for OSFI in assessing the financial condition of an insurer. Poor DCAT results will cause OSFI to intervene.   Forum

Study Tips

Update for 2020.Spring: Appendix B is included in the syllabus. (The information below regarding the inclusion/exclusion of Appendix B will be removed after the 2020.Spring exam.)

For 2019.Fall, the CAS clarified that Appendix B is not on the syllabus.

CAS response: (June 6, 2019)

• Both appendices are not required study material and the syllabus is correct.

This change has been reversed! (See Syllabus - 6C - 2020 - Spring)

• Appendix B is included. (Note that Appendix A on Life Insurance is still excluded.)

The source text itself has also been updated but the changes are mostly minor. The only significant change is the definition of plausible adverse scenario that appears on page 10. This is discussed in the wiki article below.

The old exam questions on DCAT are quite interesting. They often present you with an analysis of a company's MCT ratio projected into the future for various adverse scenarios like significant adverse development or a catastrophe. They then ask you to identify ripple effects from these adverse scenarios and possible management actions that could mitigate the impact.

DCAT is a relatively easy topic. If you do the old exam problems, you should be fine.

Estimated study time: 1-2 days (not including subsequent review time)

Pop Quiz    :-o

Consider this important question: Based on a DCAT analysis involving base & adverse scenarios, when can the AA report that an insurer is in good financial condition?

Answer:

for the BASE scenario: must have MCT > 150%   (all years in forecast period)

for the BASE scenario & all PAS: must have A > L   (all years in forecast period)

(Note that PAS = Plausible Adverse Scenarios, A = Assets, L = Liabilities)

Now, here's the Pop Quiz:

(a) Given the following, can the AA report that the insurer is in good financial condition?

Scenario	Component	2018	2019
Base Scenario	MCT Ratio	195%	170%
	Capital (A-L)	300m	250m
Adverse Scenario: Earthquake	MCT Ratio	195%	145%
	Capital (A-L)	300m	225m

(b) Given the following, can the AA report that the insurer is in good financial condition?

Scenario	Component	2018	2019
Base Scenario	MCT Ratio	195%	170%
	Capital (A-L)	300m	250m
Adverse Scenario: High Inflation	MCT Ratio	195%	145%
	Capital (A-L)	300m	-10m

Note: This pop quiz is very similar to E 2016.Fall #19b

BattleTable

Based on past exams, the main things you need to know (in rough order of importance) are:

• adverse scenario: examples, ripple effects management actions
• evaluating the financial condition of an insurer
• how to select a PAS (Plausible Adverse Scenario)
• steps in the DCAT process (especially the base scenario)
• definition of reverse stress-testing
• definition of social inflation

Top Questions ← Questions you absolutely need to know!

Questions held out from most recent exam: #20, 21. (Skip these now to have a fresh exam to practice on later. For links to these questions, see Exam Summaries.)

reference	part (a)	part (b)	part (c)	part (d)
E (2019.Spring #19)	DCAT process: - base scenario	financial condition: - evaluate	adverse scenario: (c) management actions (d) ripple effects	(e) CIA.Subseq
E (2018.Fall #20)	base scenario: - identify	base scenario: - vs business plan	adverse scenario: (c) example (d) management actions	(e) CIA.Mat (f) financial condition: evaluate
E (2018.Spring #19)	define: 4 - Plausible Adverse Scenario	adverse scenario: - propose & justify 2	adverse scenario: - ripple effects 2	financial condition: - evaluate
E (2018.Spring #24)	SCENARIO: - identify AA errors 3
E (2017.Fall #22)	financial condition: - evaluate	define: - social inflation	adverse scenario: - off-B/S risk
E (2017.Fall #23)	define: 4 - Plausible Adverse Scenario	adverse scenario: - premium volume
E (2017.Spring #21)	PAS: - how to select
E (2016.Fall #19)	DCAT process: - base scenario	financial condition: - evaluate	adverse scenario: - frequency & severity	adverse scenario: - policy liabilities misestimation
E (2016.Spring #21)	base scenario: - identify	adverse scenario: - example	adverse scenario: - impact on MCT ratio	adverse scenario: - management actions
E (2015.Fall #20)	PAS: - how to select 5	adverse scenario: - ripple effects	adverse scenario: - management actions	financial condition: - comment (see OSFI.MCT)
E (2015.Spring #29)	financial condition: - evaluate	adverse scenario: - various	definition & examples: - social inflation
E (2014.Fall #26)	define: - reverse stress-testing
E (2014.Spring #25)	DCAT filing: - actuary's actions	DCAT base scenario: - cash injection	DCAT PAS: - cash injection
E (2014.Spring #27)	define: - reverse stress-testing
E (2014.Spring #28)	financial condition: 1 - evaluate	adverse scenario: - OSFI intervention	adverse scenario: - premium risk	adverse scenario: - frequency & severity
E (2013.Fall #29)	DCAT process: - base scenario	financial condition: - evaluate	adverse scenario: - premium risk	adverse scenario: - policy liabilities risk
E (2012.Fall #29)	financial condition: - evaluate

¹ This problem relies on an invalid assumption.

² This is a GREAT PROBLEM. Make sure you study it thoroughly. It comes up in Risk Categories section of this wiki article at the end of mini BattleQuiz 6

³ This is a good Bloom's Taxonomy problem for learning the DCAT process. It's listed under mini BattleQuiz 7 under old exam problems, but it relates closely to the DCAT "method" discussed here.

⁴ The definition of plausible adverse scenario in the examiner's may be outdated. The 2017 version of the DCAT reading and updated CSOPs provide a slightly different definition. This is discussed in the wiki article below.

⁵ See this forum thread for an explanation of how capital available is adjusted.

In Plain English!

Overall Comments

The DCAT reading is one of my favourites because there is less memorizing and more analysis. They often involve interpreting data related to an adverse scenario. Sometimes you also have to recalculate the MCT ratio for the adverse scenario.

• Super-Secret Hint: The MCT ratio for the adverse scenario is usually lower than for the base scenario!

This paper is quite long at roughly 50 pages, but you can ignore a big section in the middle (Appendix A: pages 21-33) because it relates to life insurance. The most important definition in this paper is for Plausible Adverse Scenario or PAS for short. The pop quiz above used earthquake and high inflation as examples of a PAS.

The new (2017) version of the source text defines an adverse scenario as a scenario...

• with an adverse outcome on the insurer's financial condition
• developed by stress-testing assumptions (risk factors) used in forecasting the business plan

They then go on the define a plausible adverse scenario as a particular type of adverse scenario.

Plausible Adverse Scenario: an adverse scenario that is credible and has a non-trivial probability of occurring

But what do adverse, credible and non-trivial mean in practice:

credible	based on a reasonably predictable stochastic model
adverse	> 95th percentile on the loss distribution
non-trivial probability	< 99th percentile on the loss distribution

Putting these terms together means that a PAS should lie between the 95^th and 99^th on the loss distribution from a credible model. Note that the definition of PAS was asked on 2 exams prior to 2019 and the answers in the examiner's reports were based on the older version of the source text. Now, I don't know how strict the grading will be going forward. The key idea, of course, is that the PAS lies between the 95th and 99th percentile of the loss distribution, but the updated version seems to make a bigger deal of having a credible stochastic model for these percentiles.

In any case, you can see from the BattleTable above that a very common type of exam question follows this pattern:

• identify a PAS
• identify ripple effects
• identify management actions

The information you need to answer this type of question is in Appendix B - Risk Categories. (Discussed later in this wiki article.) As a reminder, the official position of the CAS is that appendix B is not (and never was?) on the syllabus. This doesn't make sense to me because most old exam problems do in fact refer to it. My advice is to study appendix B.

Sections 1,2: Intro & Method

Here, we answer 2 basic questions:

Question 1: why do we do stress-testing [Hint: risk-complement-Cap-Liq]

risk: risk identification & control (stress-testing helps identify and control risk)

complement: stress-testing complements other risk management tools such as Best's Capital Adequacy Ratio as described in BCAR.Cdn2018 and BCAR.Cat (may also involve simulation of shocks like earthquakes)

Cap: stress-testing supports capital management

Liq: stress-testing improves liquidity management

This question has been asked on several recent exams: (2016.Fall #20b), (2013.Fall #13a), (2012.Fall #28a) In the BattleCards, these exam references are listed with the OSFI.Stress reading. (There is significant overlap between CIA.DCAT and OSFI.Stress)

mini BattleQuiz 1 You must be logged in or this will not work.

Question 2: how do we do stress-testing

1. review: REVIEW 3 years operations & financial position @ end of each year
2. BASE: create the BASE scenario - a set of assumptions on risk factors that are CONSISTENT with business plan (if plan is realistic & consistent) OVER forecast period. See (2016.Fall #19a), (2013.Fall #29a)
3. risk: RISK CATEGORIES - select relevant risk categories from F-PIP-RIGOR using sensitivity-testing (F-PIP-RIGOR is a memory trick for remembering P&C insurer risk categories, and is discussed later in this article)
4. PAS: select plausible adverse scenarios from the risk categories identified in step 3 (make sure to also consider ripple effects – discussed in the next section)
5. report: REPORT scenarios with highest surplus sensitivity & those where surplus < (supervisery capital required)
6. Action(mgmt): identify management action to facilitate risk mitigation & contingency planning
7. Action(regulator): identify possible regulatory action for falling below regulatory capital required (& SIGN opinion)

If you think about these steps, they really are just common sense:

(Steps 1,2): Before doing anything, you have to understand the state of your company. This is why you review operations and create a BASE scenario.

(Steps 3,4,5): Then you look at company-specific risks, create a few adverse scenarios (with ripple effects), and write a report

(Steps 6,7): Finally, management uses the report to take action to mitigate the risks. (If the analysis shows bad results, you may need to consider what to do in the case of regulatory action.)

And, as stated above, A.M. Best also provides a different stress-testing methodology in their paper on Catastrophe Analysis. See BCAR.Cat.

mini BattleQuiz 2 You must be logged in or this will not work.

Miscellaneous Items

Other things from Sections 1,2 are concepts relating to the forecast period, base scenario, and plausible adverse scenarios. These are covered in the BattleCards. This next mini BattleQuiz has exam questions that have been asked more than once. There are 2 terms related to Plausible Adverse Scenario that deserve special mention:

ripple effect	an event that occurs WHEN an adverse scenario triggers a change in 1 or more independent assumptions	Example: a ripple effect of an earthquake may be loss of reinsurance
management action	actions that management may take to mitigate adverse ripple effects	Example: management may seek alternate reinsurance arrangements

mini BattleQuiz 3 You must be logged in or this will not work.

---

Section 3: Modeling

Purpose

The purpose of any model is to portray, on a simplified level, a process occurring in the real world. A DCAT model must reproduce these 3 key elements of the Financial Statements:

1. B/S(A, L, R/E): Balance Sheet items - Assets, Liabilities, Retained Earnings
2. I/S (revenue, expenses): Income Statement items - revenue, expenses
3. regulatory measures of capital adequacy: MCT ratio

Model Validation

Once you have a model that outputs these elements, you need to verify that the outputs make sense. This is model validation.

mini Pop Quiz: Given Cash on Hand for the base scenario of models 1 & 2, which model is more likely to be valid?

Model	Latest Year	Projection Year 1	Projection Year 2	Projection Year 3
Model 1 - Base Scenario (Cash)	20,000	20,800	21,600	22,400
Model 2 - Base Scenario (Cash)	20,000	18,225	23,125	15,800

Answer: Model 1. You can see that the cash in Model 1 is smoothly increasing whereas in Model 2, it is erratic. So, the first thing to look for when validating a model is whether the base scenario exhibits continuity of results from year to year. (This, and other validation items, are contained in the BattleCards.)

This example is somewhat similar to E (2016.Spring #21)

mini BattleQuiz 4 You must be logged in or this will not work.

Other Modeling Topics

Section 3 in the DCAT paper also covers the topics listed below, but they aren't likely exam questions so we'll skip over them for now. If you want to, you can cover them in Custom Battles. (accessible from the Main Page)

• stochastic versus deterministic models
• how to model ripple effects
• considerations in model segmentation.

---

Section 4: Reporting

Alice the Actuary stayed up all night writing her DCAT report.

Question: why did she bother??!!

• Because her mean boss told her to.
• Also, because that's how the actuary communicates the state of the company (as well as other information) to the Board of Directors.
• The "other information" could include risks the company faces as well as ways to reduce that risk.

Readers of the report may also include management and regulators, but they would get a more detailed statistical report than the Board of Directors.

mini BattleQuiz 5 You must be logged in or this will not work.

---

Appendix B: Risk Categories

• Recall that Appendix B has been removed from the syllabus for 2019.Fall. We verified this with the CAS, but their answer was too brief and I don't trust their response. If you trust them, you can skip this section. (I'm going to wait a couple of more exam cycles before I remove Appendix B from the wiki.)

What is F-PIP-RIGOR?

F-PIP-RIGOR is my memory trick for remembering the 9 P&C risk categories from Appendix B:

F	-	P	I	P	-	R	I	G	O	R
Frequency & Severity	-	Policy Liabilities	Inflation	Premiums	-	Reinsurance & Counter-parties	Investment risk	Government risk	Off-B/S risk	Related Company risk

Each of these categories can be used to construct various adverse scenarios. But not all categories will be relevant in all situations. You must use judgment to assess which are most appropriate. The DCAT paper discusses each in detail. In the next section, I'm going to explain how I summarized that discussion.

Example 1: The first 'F'

F = Frequency & Severity Risk. There are 3 important items related to each risk category:

causes: a change in frequency & severity can be caused by...

• catastrophe
• social inflation

ripple effects: these may include...

• post-event inflation
• loss of reinsurance

management action: to mitigate the risk, management may decide to...

• raise rates
• review reinsurance

This is the pattern for learning each of these categories.

Example 2: The 'I' in RIGOR

I = Investment Risk

causes: a change in investment risk can be caused by...

• significant changes in the yield curve
• decrease in return on equity investments,...

ripple effects: these may include...

• forced sale or liquidation
• significant +/- change in cash flows

management actions: to mitigate the risk, management may decide to...

• sell assets
• change investment strategy

Tricks for Memorizing F-PIP-RIGOR

It's a big job to memorize causes, ripple effects, and mgmt actions for all 9 risk categories, but there are a few tricks:

Trick 1: Instead of memorizing the ripple effects for each risk category individually, just memorize some of the common ripple effects. These are:

higher LR (due to higher losses or operating costs)

loss of reinsurance

post-event inflation

forced sale or liquidation

mix shift

Policyholder actions

regulatory action

• If you're then asked for a ripple effect for a particular adverse scenario, you can take an educated guess and pick a couple of things from this list that seem most appropriate. (2013.Fall #19d)

Trick 2: Do the same thing for management actions. Here are some common management actions:

tighten U/W

raise rates

review reinsurance

sell assets

review mix (geography, limit,...)

• If you're then asked for management actions for a particular adverse scenario, take an educated guess from this list! (2013.Fall #19c)

mini BattleQuiz 6 You must be logged in or this will not work.

2018.Spring Q24 (A Good Bloom's Q)

This question concerns the how of stress-testing that we discussed above in terms of a 7-step method. It is a long question for only 3 points. At 3.25 minutes-per-point, you have roughly 10 minutes to do it, but I think it may take you longer than that. The first step is understanding what you have to do. This information is given in the first paragraph:

• identify 4 incorrect actions by the AA
• describe the correct actions

Make sure you do both of the above items. The second step in this problem is making sense of the given information, and you have 2 choices: You can either highlight the relevant information as you read it. (You brought a highlighter to the exam, right?) Or you can make bullet points for each paragraph. I made bullet point notes as shown below. The text in red & green font was added after I made the paragraph notes. Red indicates an incorrect AA action. Green indicates the corrected action. (Note that I used abbreviations to save time.)

para 2:

- opns begin 2012

- current yr-end: 12/31/2017

- AA analysis: July 2018

- bus. plan uses results through 12/31/2017

para 3:

- AA reviews opns in 2017 (1yr) → should review ≥ 3 yrs

- prepares base scen (uses bus. plan)

- forecast period: 12/31/17 to 12/31/19 (2 yrs) → forecast period should be ≥ 3 yrs

- reduces prem. growth assumptions

- MCT > 150%

para 4:

- review risk categories

- test change in risk factors to get surplus < 0

- adjust risk factors to get 95-99% prob. for adv. scen.

- discard reins. risk & off-B/S risk

para 5:

- rank scen. by MCT (low to high)

para 6:

- select 3 scen. for more analysis → all with MCT < 150% should be further analyzed

- consider: interactions, ripple, macro-eco, mgmt actions → should also consider regulator action due to low MCT

para 7:

- AA writes DCAT rpt with all sections

- explains bus. plan deviation in base scen. (lower prem. growth)

- concludes fin. cond. not satis. (cat event MCAT < 150%)

I've noted 4 items above but there are more. See if you can find them. Check your answers with the examiner's report.

Final Words

The best way to learn this paper is to work on the old exam problems at the same time you're learning the BattleCards. The final mini BattleQuiz for this article just contains a bunch of old exam problems.

mini BattleQuiz 7 You must be logged in or this will not work.

BattleCodes

Memorize:

• conditions that must be satisfied by the DCAT analysis for Alice the Actuary to report that an insurer is in good financial condition.
• 4 purposes of stress-testing: rsk.compl.C.L
• 7 steps in stress-testing: rBrPr.Act(MR)
• defns plausible adverse scenario, reverse stress-testing, ripple effects
• The 9 P&C risk categories F-PIP-RIGOR
• common examples of ripple effects (7) & management actions (5)

Conceptual:

• Explain how the 7 steps of the stress-testing methodology are just common sense.
  • How is this paper, CIA.DCAT, similar to the related paper OSFI.Stress? In what ways are these papers different? (Note that I haven't answered that question in this article. It's something for you to think about after you've covered OSFI.Stress.)
  • What other papers are related to CIA.DCAT? How about OSFI.MCT, the #1 ranked paper on the syllabus.
  • A Bloom's Taxonomy Question: What extra detail does a DCAT analysis provide that the MCT ratio alone does not?

• See also BCAR.Cat:
  • It's a U.S.-specific reading, but it's interesting to compare and contrast the BCAR stress-test methodology (which has 5 steps) to the 7 steps of the DCAT methodology.
  • (Note that the 5 steps of the BCAR methodology has been asked more than once on past exams, so make sure you learn it when you get to that paper!)

Calculational:

• Many DCAT problems from past exams are more like case studies than calculation problems. (2016.Fall #19)
  • You may be presented with MCT ratios and other balance sheet metrics for a base scenario & several adverse scenarios.
  • A typical question is whether the insurer is in good financial condition, and to identify possible ripple effects & appropriate management actions.
• Sometimes a DCAT problem is combined with an MCT problem. (2017.Spring #21), (2016.Spring #21), (2015.Fall #20).

• Expect about 2 pts from this paper on the exam. This may seem low, but it's important because DCAT often underlies a high-value MCT problem.

Full BattleQuiz You must be logged in or this will not work.

  Forum

POP QUIZ ANSWERS

(a) YES. All conditions are satisfied. (The MCT ratio of 145% for the adverse scenario in 2019 doesn't matter.) (b) NO. The adverse scenario has A-L = -10m < 0 in 2019, which is not allowed.

(It always seemed a bit STRANGE to me that the MCT ratio for the adverse scenario isn't relevant to the appointed actuary's decision! But whatever...)