analysis of financial management levers for stock selection · “market eye q” analysis of...

27
“Market Eye Q” Analysis of financial management levers for stock selection This presentation has been prepared for the Actuaries Institute 2015 Actuaries Summit. The Institute Council wishes it to be understood that opinions put forward herein are not necessarily those of the Institute and the Council is not responsible for those opinions. Cary Helenius, Kevin Gomes & Graham Taylor

Upload: phamdat

Post on 03-Aug-2018

217 views

Category:

Documents


0 download

TRANSCRIPT

“Market Eye Q”

Analysis of financial management levers for

stock selection

This presentation has been prepared for the Actuaries Institute 2015 Actuaries

Summit.

The Institute Council wishes it to be understood that opinions put forward herein

are not necessarily those of the Institute and the Council is not responsible for

those opinions.

Cary Helenius, Kevin Gomes &

Graham Taylor

2

Phase 1

Analysis of financial management levers

for assessment of company

management

Case Studies

Phase 2

Analysis of financial management levers for stock selection

Outline of Presentation:

Total Shareholder Return (TSR) is affected by a number of

factors (“Ps” & “Qs”):

Total Shareholder Return*

Q's

P's

Quantitative – Financials under

company influence

Price (short term) Perception Prospects

3

* “Total Shareholder Return” = Share price change and dividend yields

Management can drive a company’s TSR over future

periods by appropriately managing company financials…

Total Shareholder Return

Q's

P's

Quantitative – Financials under company influence

Price (short term) Perception

Prospects

4

We have explored the relationship between a company’s

reported financials and Relative TSR*…

= 81,000 data points

6 monthly data points over past 12 years

225 companies analysed

We have only included ratios which were clear and objective

measures, derived from the publicly released results

…and found a strong relationship for certain ratios * “Relative TSR” = TSR relative to the market average

5

6

Taken in combination, the financial ratios were used to

model the Relative TSR

Company Predicted Relative TSR Rank

Predicted Relative TSR Quartile

1. aaa 17 1st

2. bbb 205 4th

3. ccc 115 3rd

… … …

… … …

225. xyz 81 2nd

…with companies ranked and grouped into quartiles

according to their predicted Relative TSR level

Companies in the top quartile represent those

with optimal mix of financial parameters

across P&L, Cashflow and Balance Sheet

a

7

Price related ratios were excluded for purposes of assessing

company management performance … Financial Ratio P&L Cashflow Balance

Sheet Share Price

ROE

Change in ROE

ROFE

NPAT growth

ROFE growth

Net debt to equity

Change in net debt to equity

Net debt to total funds employed

Net interest cover

Pay out ratio

Net cash flow to NPAT

EPS growth

Revenue growth

Dividend Yield

P/E

…but were included in

our stock selection

model as they provide

useful information on

valuation levels

Price related ratios excluded from our analysis of company management on the basis that

share price is not within management control

8

Phase 1

Analysis of financial management levers

for assessment of company

management

Case Studies

Phase 2

Analysis of financial management levers for stock selection

Outline of Presentation:

Case Study 1: Fine-tuning corporate strategy

Increasing

Inc

rea

sin

g

Year on year change in ROFE (increasing )

9

The company had used dividend and EPS growth as its primary share price drivers

The heat maps demonstrated that the company was not in the optimal area for pay-out ratio and yield, and also showed that EPS growth was a poor measure

The data highlighted that Return on Funds Employed

was deteriorating and it appeared the market was more

focused on this measure

The company was able to incorporate the findings in the development of their corporate strategy

Case Study 2: Setting executive remuneration

Inc

rea

sin

g

Increasing

10

The company identified a financial ratio which it wanted to use in its executive remuneration

We determined the ratio (ROCE) had a relationship with Relative TSR and determine the companies ranking on this measure within an industry peer group

We were able to align the executive bonus targets

with the Relative TSR findings

The company was able to demonstrate to proxy advisers and investors that the bonus setting mechanism was aligned to driving shareholder returns

-40%

-20%

0%

20%

40%

60%

80%

-15%

0%

15%

30%

scenario A scenario B scenario C

average RTSR 0 (LHS) positive RTSR occurences (RHS) level of debt funding (RHS)

level of debt funding

aver

age

RTS

R

0%

ratio +ve &

-ve

Capital mix for aquisitions

Case Study 3: Capital funding mix for acquisitions

A B C

11 11

We determined that there was a Relative TSR relationship for companies that increased their assets by more than 15%

The scenario analysis demonstrated that certain

mix of debt and equity funding had a high

frequency of positive Relative TSR outcomes (A)

whereas an alternative strategy (C) had a high

frequency of negative Relative TSR outcomes

The company was able to optimise its funding composition of a potential acquisition (B) , within its internal constraints to minimise the potential negative Relative TSR aspects and maximise the positive Relative TSR aspects

Case Study 4: Capital management initiatives

12

We determined that there was a Relative TSR relationship for companies changing their dividend policy and introducing share buy backs

The scenario analysis demonstrated that certain

combination of dividend and share buy back

had a high frequency of negative Relative TSR

outcomes (A,B and D) whereas an alternative

strategy had a high frequency of positive Relative

TSR outcomes (C)

The company was able to discuss the alternative strategies at the board with supporting equity market analysis of the likely investor response

A B C D

13

Phase 1

Analysis of financial management levers

for assessment of company

management

Case Studies

Phase 2

Analysis of financial management levers for stock selection

Outline of Presentation:

We have constructed a model of relative share price to

company financials….

• GLM (normal error term)

• Explanatory variables: Company financials

• Response: Relative change in share price plus dividends*

Financials Relative change in TSR eg

* Relative (share price t+1 + dividends t to t+1) / share price t - 1

Release period of financials (by mid Sep 13)

14

A previous “Phase 1” finding…

RO

E

(in

cre

as

ing -

-->

)

Change in ROE (increasing --->)

ROE vs change in ROE

High ROE and high

positive change is

associated with

outperformance

Low ROE and

reduction in ROE is

associated with

underperformance High Relative TSR

Low Relative TSR

15

For “Phase 2”, we’ve introduced dividend yield and PE

(trailing) to our model of Relative TSR…

Dividend

yield

PE (trailing)

Actual relative TSR <= June 2010

Dividend

yield

PE (trailing)

Exposure of observations <= June 2010

Low PE associated with outperformance

and low PE & high Dividend Yield associated with

highest outperformance

Relative

TSR

-5% to 0%

0% to 5%

5% to

10%

10% to

15%

Increasing

Inc

rea

sin

g

Increasing

Inc

rea

sin

g

16

The contribution of dividend yield and PE to Relative TSR

was observed to change post June 2010 …

17

Relative

TSR

-5% to 0%

0% to 5%

5% to

10%

10% to

15%

Dividend

yield

PE (trailing)

Actual relative TSR > June 2010

Increasing

Dividend

yield

PE (trailing)

Exposure of observations > June 2010

Inc

rea

sin

g

Broadly similar relationships as

for period prior to Jun 2010 but

notable differences

Inc

rea

sin

g

Increasing

17

18

Dividend

yield

PE (trailing)

Exposure of observations > June 2010

Inc

rea

sin

g

Increasing 0%

5%

10%

15%

20%

pre Jun 2010 post Jun 2010

Distribution of observations

Dividend yield Increasing

18

There was also a notable migration to higher dividend yields

post June 2010 …

Comparing heatmaps pre and post Jun 2010 highlights the

differences …

19

Dividend

yield

PE (trailing)

Actual relative TSR <= June 2010

Relative

TSR

-5% to 0%

0% to 5%

5% to

10%

10% to

15%

Dividend

yield

PE (trailing)

Actual relative TSR > June 2010

Inc

rea

sin

g

Low PE and low Dividend Yield now associated with under-

performance

Mid-level PE and above average

Dividend Yield now associated with outperformance

Increasing Increasing

Inc

rea

sin

g

19

Absence of systematic bias in actual vs predicted Relative

TSR validates the overall model…

-3%

-2%

-1%

0%

1%

2%

3%

Ju

n 2

00

3

De

c 2

003

Ju

n 2

00

4

De

c 2

004

Ju

n 2

00

5

De

c 2

005

Ju

n 2

00

6

De

c 2

006

Ju

n 2

00

7

De

c 2

007

Ju

n 2

00

8

De

c 2

008

Ju

n 2

00

9

De

c 2

009

Ju

n 2

01

0

De

c 2

010

Ju

n 2

01

1

De

c 2

011

Ju

n 2

01

2

De

c 2

012

Ju

n 2

01

3

De

c 2

013

Ju

n 2

01

4

Re

lative

TSR

Actual vs predicted Relative TSR

Actual

Predicted

Financials at Jun

2014 predict Relative

TSR at Jun 2015

Model is robust

over time -

including

predicted

future period

Date of financials

20

“Phase 1” model (excl. dividend yield and PE) successfully

groups companies into quartiles according to Relative TSR…

-15%

-10%

-5%

0%

5%

10%

15%

20%

Ju

n-0

2

Ju

n-0

3

Ju

n-0

4

Ju

n-0

5

Ju

n-0

6

Ju

n-0

7

Ju

n-0

8

Ju

n-0

9

Ju

n-1

0

Ju

n-1

1

Ju

n-1

2

Ju

n-1

3

Actual Relative TSR by predicted

quartile

1st quartile

2nd quartile

3rd quartile

4th quartile

All years 1st qtle

All years 2nd qtle

All years 3rd qtle

All years 4th qtle

21

-15%

-10%

-5%

0%

5%

10%

15%

20%Ju

n-0

2

Ju

n-0

3

Ju

n-0

4

Ju

n-0

5

Ju

n-0

6

Ju

n-0

7

Ju

n-0

8

Ju

n-0

9

Ju

n-1

0

Ju

n-1

1

Ju

n-1

2

Ju

n-1

3

Actual Relative TSR by predicted

quartile

1st quartile

2nd quartile

3rd quartile

4th quartile

All years 1st qtle

All years 2nd qtle

All years 3rd qtle

All years 4th qtle

22

“Phase 2” model (incl. dividend yield and PE) improves the

performance…

“Phase 2” model (incl. dividend yield and PE) is clearly

superior at predicting top and bottom quartile performers…

-15%

-10%

-5%

0%

5%

10%

15%

20%Ju

n-0

2

Ju

n-0

3

Ju

n-0

4

Ju

n-0

5

Ju

n-0

6

Ju

n-0

7

Ju

n-0

8

Ju

n-0

9

Ju

n-1

0

Ju

n-1

1

Ju

n-1

2

Ju

n-1

3

Actual Relative TSR by predicted quartile

1st quartile - Model

excl DivY/PE

1st quartile - Model

incl DivY/PE

4th quartile -

Model excl DivY/PE

4th quartile -

Model incl DivY/PE

23

The suitability of our “Phase 2” model for fund management

purposes was back-tested on a “real-time” basis…

Jun 13

Jun 13 company

financials released

by mid-Sep 13

Companies ranked into

predicted quartiles by

model – post the market

response to financials

Sep 13 Dec 13 Jan 14

Real-time calculation

of RTSR over following

4 months

30 Jun 13 balance

date

24

Back testing the “Phase 2” model on a real-time basis

shows that it has consistently outperformed the market…

Period:

pre GFC (2003

to 2007)

GFC affected

(2008 to 2009)

post GFC (2010

to 2014)

Market

outperformance

pa (predicted

top quartile)

+3.7% +6.3% +0.5%

25

…with potential for further improvement as the model is

refined

Early testing suggests

that model performance

in post GFC period can be

significantly improved

through use of “price

momentum” filters

26

In summary, our research has some important implications…

• We found a strong relationship between certain financial ratios derived from a company’s reported financials and prospective shareholder return

• These financial ratios can be combined to form a multivariate model which predicts a company’s TSR relative to the rest of the market

• A “Phase 1” model (excl. price related measures) was initially developed which was able to successfully group companies according to relative TSR

• A “Phase 2” model (incl. price related measures) was subsequently developed and found to offer superior performance, with enhanced ability to predict top and bottom quartile performers

27

…with a range of potential applications

• Our “Phase 1” model provides an objective framework for assessing management

performance, setting executive remuneration rewards and evaluating strategic

alternatives – all based on factors which are within management control and

independent of share price

• Our “Phase 2” model can be used as a stock selection tool which is based entirely

on a company’s reported financials and current share price i.e. does not rely on

forecasts.