The experiment

Hemscott is a website dedicated to providing financial information. In 2003/2004, Tom Stevenson was Hemcott's head of research and in that capacity he wrote about the systematic approach to selecting shares. He regularly reviewed systems such as Jim Slater's price/earnings to growth (PEG) ratio, Michael O'Higgins high-yield approach and David Dreman's contrarian low price/earnings ratio (PER) method. Using these systems of selecting shares and using Hemcott's database, he drew up four shortlists from the FTSE Small Cap index. These were the following:

• The five with the lowest PERs

• The five producing the highest yields

• The five with the lowest price/tangible asset book value ratios

• The five with the lowest price/cash ratios

Ten months later at the end of June 2004, Tom Stevenson noted that although some shares produced amazing results in each category, there were also heavy losers. Overall, each section achieved net gains of 16%, 29%, 23% and 46% compared with an 8% gain for the market as a whole in the same period.

Again using Hemscott's database, he produced a list of the top five shares in each category as at the end of June 2004. These were:

• The lowest PERs - Chaucer Holdings, Cox Insurance, Goshawk Insurance, Molins and Jarvis

• The highest yields - Beattie (James), Molins, Hardy Underwriting and Ultraframe;

• The cheapest price/tangible asset ratio - Goshawk Insurance, Lavendon, Molins, Tops Estates and UNITE Group;

• The lowest price/cash flow - Hitachi capital, Ashtead, Lavendon, Danka and Beazley.

As can be seen, there were some duplications, so the original list of 20 companies came down to 16:

Ashtead Group, Beattie (James), Beazley, Chaucer Holdings, Cox Insurance, Danka, Goshawk Insurance, Hardy Underwriting, Hitachi Capital, Jarvis, Lavendon, Molins, Regent Inns, Tops Estates, Ultraframe and UNITE Group.

Based on the previous experiences described by Tom Stevenson for the 2003 selected companies (range: gain of 213% to loss of 96%), it seemed likely that the companies listed for 2004 would do either spectacularly well or excruciatingly bad. The question was: Could analysis of the latest published accounts distinguish between the two extremes?

On 1 July 2004, the latest published accounts for each of the 16 companies were examined. Because of timing constraints (analyse the accounts and write an article based upon the analysis), none of the accounts was reviewed in detail, which would be the case if investing real money had been envisaged. Instead, the exercise was simply an experiment to see if the accounts would give obvious clues without carrying out time-consuming reviews.

Three companies out of the sixteen were eliminated because they failed one or more of the tests described in Chapter 2. These were: Cox Insurance, Jarvis and Molins. Another four companies were eliminated because the latest accounts showed negative earnings. More time spent on these companies might have revealed reasons why future earnings could be better, but that was outside the scope of the rules laid down. Negative earnings took out Ashtead Group, Danka, Goshawk Insurance and UNITE Group.

The remaining nine companies were analysed and were placed in the following order:

(1) Tops Estates, (2) Chaucer Holdings, (3) Hitachi Capital, (4) Hardy Underwriting, (5) Beattie (James), (6) Beazley, (7) Ultraframe, (8) Lavendon, (9) Regent Inns

The top four shares were selected and backed against the whole 16 shares, based on £4000 being invested in each of the selected four shares compared with £1000 being invested in each of the 16 shares; the test to run over the 10 months (to be consistent with Tom Stevenson's selected timeframe) ended on 30 April 2005. The article was written on 1 July 2004 and was published in the October 2004 edition of'Financial Management' (Journal of Chartered Institute of Management Accountants).

The results of this test were published in the June 2005 edition of 'Financial Management' (selected companies in bold, rejected on test in italics):

Opening share Closing share Percentage of price, 1 July 2004 price 30 April capital returneda

Opening share Closing share Percentage of price, 1 July 2004 price 30 April capital returneda

(pence)

2005 (pence)

Jarvis

78.00

10.25

13.1

Danka

63.00

16.50

26.2

Ultraframe

127.50

51.50

46.6

Beazley

96.50

88.00

91.5

Molins

175.00

162.50

92.9

Beattie (James)

126.50

116.50

94.9

Hardy Underwriting

237.50

212.50

100.0

Goshhawk Insurance

40.25

42.50

105.6

Chaucer Insurance

48.00

56.00

118.2

Cox Insurance

71.50

92.00

129.4

Hitachi Capital

197.50

256.50

131.4

Tops Estates

332.00

442.50

134.5

UNITE Group

197.50

285.00

145.6

Lavendon

125.00

184.50

149.4

Regent Inns

43.50

80.50

185.1

Ashtead Group

27.50

86.75

315.5

a Percentage takes into account dividends paid in the period (excluded if ex-dividend).

The four selected shares showed a profit of £3363.67 in the period on the investment of £16 000 (£2842 after transaction costs) compared to the profit of £2796.44 for the whole market (£2283 after transaction costs), giving a gross percentage profit of 21.0% and 17.5%, respectively, compared to the FTSE 100 index increasing 7.6% in the same period.

a Percentage takes into account dividends paid in the period (excluded if ex-dividend).

The four selected shares showed a profit of £3363.67 in the period on the investment of £16 000 (£2842 after transaction costs) compared to the profit of £2796.44 for the whole market (£2283 after transaction costs), giving a gross percentage profit of 21.0% and 17.5%, respectively, compared to the FTSE 100 index increasing 7.6% in the same period.

However, the methodology used for this experiment was flawed in that it used only one valuation method, that being valuation based on earnings. To evaluate insurance companies, the methodology should include an assessment of the risks associated with underwritten policies that were still live. In addition, property investment companies such as Tops Estates and UNITE Group are primarily valued on net asset values (NAVs) and not earnings (see UNITE Group's accounts in Chapter 3). In the latter case, it was the potential for increased property values in the future rather than the lack of earnings that should have been the key feature of any valuation.

Six months later, the overall 16 shares were performing much better than the four that had been selected. What this tells you is that reviewing a set of accounts once, making a decision and forgetting about it is not a rational strategy. Once a share is purchased, that company's accounts must be reviewed every 6 months.

This review should assess whether or not the methodology used for a particular company was the right one and should also look for clues in the latest accounts to see if the company is progressing or not. The old adage that 'you should buy into a good company and stay with it' is valid only if the company is continuing to perform. If a review of subsequent accounts cannot detect a problem, then by all means stay with a share even if the price is going down, but if the accounts reveal a reason why the share price is falling, then selling might be the best strategy.

Although the method of selecting companies had successfully eliminated the four worst-performing companies, it had also failed to spot the four companies that were going to show the greatest improvement. The reason why these winners were missed was a combination of using only one valuation methodology (as detailed above) and not looking at the potential for future earnings to improve. But further analysis revealed that the investment rule that states companies failing the tests shown in Chapter 2 (and later on in this chapter) should be avoided was largely validated. However, although it is not impossible that there will be exceptions to this rule, given the probabilities such companies may best be avoided.

The fate of the bottom four and top four companies is shown below:

Share price (pence)

1 July 2DD4

3D April 2DD5

31 March 2DD6

5 April 2DD7

Bottom four

Jarvis

78.GG

1G.25

G.2Ga

G.18a

Danka

63.GG

16.5G

2G.GG

16.GG

Ultraframe

127.5G

51.5G

38.75

b

Beazley

96.5G

88.GG

118.GG

161.5G

a Based on restructuring where investors received one new ordinary share for every 400 old ordinary shares.

b Taken over by Latium Holdings Limited; price not known.

a Based on restructuring where investors received one new ordinary share for every 400 old ordinary shares.

b Taken over by Latium Holdings Limited; price not known.

Share price (pence)

Top four Ashtead Group Regent Inns Lavendon UNITE Group

1 July 2004 30 April 2005 31 March 2006 5 April 2007

27.5G 43.5G 125.GG 197.5G

86.75 8G.5G 184.5G 285.GG

223.GG 97.5G 28G.GG 45G.GG

153.75 113.75 416.GG 535.GG

0 0

Post a comment