[Audio] %[break]% Good afternoon, My name is Conrad Valk and I'm glad to talk you through my presentation of Assignment 1. 

Scene 1

Kantitatief Integriteitsmanagement
Module 4 Assignment 1

[Audio] I bring to the floor the following causal claim High sickness absence is a predictor for low compliant behaviour Its not a claim I have seen before in any literature or in practice its just a causal relation that "I" imagine would logically exist I would include this indicator in the Compliance Mananagement Dashboard if I were a compliance professional which I'm not by the way. 

Scene 2

A high sickness absence rate [the cause or predictor] is a predictor for  low compliant behaviour [the effect or outcome].

[Audio] The negative sign below at the arrow tells us that we expect a negative association between the sickness absence rate and the level of employee compliance Higher sickness is a proxy for lower compliance behaviour.. 

Scene 3

[Audio] We have a continuos predictor variable (the SAR) this variable can be 1% or 2.43% or any value in between The outcome variable is discrete
it can have a value of either 1, 2, 3, 4 or 5 the outcome value is "1" if registered non compliant events range between zero and 10 The outcome is "2" if it ranges between 11 and 15
And so forth with the graph to the right we show how we expect the distribution of the association will be. It is a desktop picture. Only after we have collected and plotted the data the true distribution will be manifest A regression factor tells us something about the strength of the association between predictor and outcome the regression is based on real historic data. 

Scene 4

How to measure the effect and the
predictor and the expected distribution pattern?

For the predictor (the sick absense) we will use the company wide sick absense rate (further: SAR) calculated by the HR department as the continuous variable of measurement  of choice. The definition of this sick absense rate % is as follows: Total aggregated number of man days clocked by  employees in a month (excluding employees on vacation leave) devided by the total number of normative man days minus 1 times 100%
 For the effect (being employee compliance with company policies) we use the compliance score that is determined by ranges of numbers of non compliant events accounted for in one month from the central registry of the compliance department as the discrete variable of measurement. Compliance score derived from intervals of non compliant events: 5: High compliant  = between 0-10 non compliant events registered a month 4: High-to-Average comliant  = between 11-15 non compliant events a month 3: Average compliant = between 16 - 30 non compliance events a month  2: Average-to-Low compliant = 31- 35 non compliance events a month 1: Low compliant = 36-50 non compliant events a month

A graph with blue lines and red writing

Description automatically generated

[Audio] A predicted lower compliance score would affect Enterprise Risk Management. To the effect that the intrinsic value for operational and fraud risk could be adjusted upwards The conclusion might be that Internal Audit will perform more substantive system testing and in general there might be a he need for more intense managerial review, monitoring and surveillance Higher compliance as predicted by the model migh on the other hand merit a more laissez faire regime The outcome of Sickness Absense Rate will have its own set of predictors for instance: unsafe or unhealthy working space, dissatisfaction with salary level. The factors driving a high Sickness Absense Rate is a separate study with its own associations outside the scope of this presentation.. 

Scene 5

Relevance of the causal claim 
for compliance

The causal claim is that SAR functions as a proxy for compliant/non-compliant behaviour. The claim is that  aggregated SAR goes up  up the level of intrinsic compliance risk goes up for the company. Not adressed here are the deeper lying association of the sickness absense rate and one or more root causes driving this rate. Is the SAR driven by physical or by mental illnessses (demotivation, anxieties, burn out).  Especially in case of mental illnesses management should investigate and address the deeper lying factors causing the increase in SAR.

[Audio] A quantitative model is only as good as for how it is used and what you use it for ===== - is the purpose clear?
- is the model used as intented?
- is the logic sound?
- are design, theory and underlying logic sound?
- is the data quality high?
- is the model appropriately documented?. 

Scene 6

The first step is to gather enough statistical data by way of historical back testing on the SAR outcome and compliance score (compliant behaviour) on a statistically representative  population of data points. Quantify the regression between SAR and compliance score with a sufficiently high confidence level to determine how strong the association is between SAR and compliant behaviour. A confidence level of <0,5 is considered a good benchmark for a statistical sound model.
 Once a strong correlation is determined based on a sufficient population of observations then the distribution is  considered ”statistically  reliable” and we can use the  SAR as a proxy for intrinsic compliance risk (compliant behaviour score). Perform a solid model risk assessment

[Audio] Thanks for taking the time to listen to my presentation. I will be happy to answer your questions next time I'm there live.. 

Scene 7

%[break]%

Good afternoon, 
My name is Conrad Valk and I'm glad to talk you through my presentation of Assignment 1

avatar

I bring to the floor the following causal claim

 
High sickness absence is a predictor for  low compliant behaviour

 
 Its not a claim I have seen before  in any literature or in practice

 
its just a causal relation that "I" imagine would logically exist

 
I would include this indicator in the  Compliance Mananagement Dashboard if I were a compliance professional 

which I'm not by the way

The negative sign below at the arrow tells us that we expect a negative association between the sickness absence rate and the level of employee compliance
 
Higher sickness is a proxy for lower compliance behaviour.

We have a continuos predictor variable (the SAR) 
 
 this variable can be 1% or 2.43%  or any value in between
 


The outcome variable is discrete
it can have a value of either 1, 2, 3, 4  or 5 

the outcome value is "1" if registered non compliant events range between zero and  10 
The outcome is "2" if it ranges between 11 and 15
And so forth

 
 
 
with the graph to the right we show how we expect the distribution of the association will be.    
It is a desktop picture. Only after we have collected and plotted the data the true distribution will be manifest

 
A regression factor tells us something about the strength of the association between predictor and outcome
 
the regression is based on real historic data

A predicted lower compliance score would affect Enterprise Risk Management.
 
To the effect that the intrinsic value for operational and fraud risk could be adjusted upwards

 
The conclusion might be that  Internal Audit  will perform more substantive system testing and in general there might be a he need for more intense managerial review, monitoring and surveillance
 

Higher compliance  as predicted by the model migh on the other hand merit a more laissez faire regime

 
 
The outcome of Sickness  Absense Rate will have  its own set of predictors  for instance: unsafe or unhealthy working space, dissatisfaction with salary level.  
The factors driving a high Sickness  Absense Rate is a separate study with its own associations outside the scope of this presentation.

A quantitative model is only as good as for how it is used and what you use it for =====
 
- is the purpose clear?
-  is the model used as intented?
- is the logic sound?
- are design, theory and underlying logic sound?
- is the data quality high?
- is the model appropriately documented?

Thanks for taking the time to listen  to my presentation.
 
 I will be happy to answer your questions next time I'm there live.

Kantitatief Integriteitsmanagement Module 4 Assignment 1

Scene 1 (0s)

Scene 2 (9s)

Scene 3 (33s)

Scene 4 (48s)

Scene 5 (1m 41s)

Scene 6 (2m 33s)

Scene 7 (3m 1s)