βeta Solutions Quarterly Style Analysis – Q1 2025

27four Style Analysis Quarterly Note Q1 | March 2025

1

Methodology italization while factor weighting implies applying the standardised style 27four Beta Solutions conducts style-based analysis on the top 100 shares

score (proxy) as the initial weighting. The results of both market capitaliza-

(by market capitalization) listed on the JSE. The research conducted in-

tion and factor weighting are presented. Importantly, both market capital-

volves style back-testing on 27four’s proprietary JSE dataset where shares

ization and factor weighted portfolios are identical in terms of their share

are sorted on several style proxies. Long-only style benchmark returns are

constituents, the only difference being the initial weight assigned at each

calculated on a buy-and-hold basis applying market capitalization and

portfolio sort date.

factor weighting. Market capitalization weighting implies assigning each share an initial in-portfolio weight based on their sorting date market cap-

The proxy for each style is described on the table below:

Style

Proxy

Style Factor Name/Detail

Inverse Style Factor

Style Factor Premium

Momentum

Historical cumulative return measured over the previous 252 trading days excluding the most recent 21 days of return

Winner = top 30 shares based on momentum rank

Loser = bottom 30 shares based on momentum rank

Winner minus Loser

Value

Linear combination of earnings yield (EY) and cashflow per share (CFPS)

Value = top 30 shares based on z-score combination of EY and CFPS

Growth = bottom 30 shares based on z-score combination of EY and CFPS

Value minus Growth

Low Volatility

Linear combination of idiosyncratic risk and market beta measured over the previous 252 trading days

Low Vol = bottom 30 shares based on z-score combination of standard deviation and market beta

High Vol = top 30 shares based on z-score combination of standard deviation and market beta

Low Vol minus High Vol

Quality

Linear combination of return on assets (ROA), return on equity (ROE)and debt to equity ratio (D/E)

Quality = top 30 shares based on z-score combination of ROA, ROE less DE

Junk = bottom 30 shares based on z-score combination of ROA, ROE less DE

Quality minus Junk

Size

Natural logarithm of market capitalization on portfolio sort date

Small = bottom 30 shares based on natural log of market capitalization

Big = top 30 shares based on natural log of market capitalization

Small minus Big

Liquidity

Average value traded measured over the previous 252 trading days relative to market capitalization

Illiquid = bottom 30 shares based on value traded scaled by market capitalization

Liquid = top 30 shares based on value traded scaled by market capitalization

Illiquid minus liquid

Dividend Yield

One year trailing total dividend (inclusive of special dividends) scaled by the current share price

High Yield = top 30 dividend paying shares based on natural log of dividend yield

Low Yield = bottom 30 dividing paying shares based on natural log of dividend yield

High Yield minus Low Yield

Currency Risk

Currency beta measured over the previous 252 trading days

Rand hedge = top 30 shares based on currency beta

Rand tracker = bottom 30 shares based on currency beta

Rand hedge minus rand tracker

Style Performance at 31 March 2025 Market capitalisation weighting Time Period

Winner

Loser

Value

Growth

Low Vol

High Vol

Quality

Junk

1 Month

1.07%

-0.53%

2.12%

0.30%

1.27%

8.15%

2.50%

-4.47%

1 Month Vol.

19.78%

18.46%

17.06%

16.17%

11.79%

23.52%

20.29%

16.87%

3 Month

17.93%

7.45%

3.18%

6.11%

8.84%

8.90%

11.22%

0.03%

3 Month Vol.

18.78%

16.00%

13.74%

15.08%

17.12%

19.68%

20.32%

15.69%

6 Month

16.58%

-4.52%

-3.49%

0.95%

-2.91%

4.81%

7.89%

-8.01%

6 Month Vol.

15.64%

16.21%

13.47%

14.64%

15.07%

18.57%

17.45%

13.58%

1 Year Return

45.28%

6.05%

20.30%

13.63%

3.17%

22.55%

28.92%

-0.22%

1 Year Vol.

15.17%

16.68%

16.94%

15.66%

14.98%

18.73%

17.37%

13.02%

3 Year Return

15.12%

11.66%

1.01%

14.19%

5.23%

14.25%

13.76%

6.92%

3 Year Vol.

19.47%

23.22%

20.96%

20.92%

14.82%

24.82%

23.17%

15.33%

5 Year Return

19.48%

17.65%

18.45%

17.31%

9.38%

27.34%

16.59%

18.94%

5 Year Vol.

20.64%

25.67%

22.05%

23.31%

16.70%

27.58%

23.43%

19.90%

10 Year Return

8.69%

4.35%

5.89%

11.43%

4.31%

11.68%

10.62%

3.15%

10 Year Vol.

20.37%

25.75%

21.81%

22.71%

17.01%

27.80%

21.58%

20.21%

Small

Big

Illiquid

Liquid

High Yield

Low Yield

Rand

Rand

Hedge

Tracker

1 Month

-1.27%

-0.49%

-1.67%

-0.25%

0.83%

-1.15%

-0.55%

3.87%

1 Month Vol.

9.73%

15.17%

14.45%

16.24%

14.10%

17.64%

13.37%

18.79%

3 Month

-4.47%

7.67%

7.05%

-3.14%

2.08%

13.72%

9.54%

7.89%

3 Month Vol.

10.92%

13.75%

14.82%

14.31%

12.65%

18.22%

14.64%

14.85%

6 Month

1.63%

0.24%

-1.13%

-8.28%

-5.00%

6.67%

2.11%

3.27%

6 Month Vol.

9.94%

13.18%

14.28%

14.02%

12.51%

16.15%

14.16%

14.72%

1 Year Return

31.18%

8.77%

4.13%

14.15%

3.80%

13.59%

5.01%

23.73%

1 Year Vol.

11.43%

13.94%

15.66%

17.19%

14.76%

15.81%

15.70%

16.63%

3 Year Return

10.77%

10.11%

11.56%

-0.14%

6.43%

17.46%

7.68%

13.79%

3 Year Vol.

12.24%

17.69%

18.33%

19.12%

19.63%

20.39%

16.53%

23.74%

5 Year Return

25.76%

16.62%

16.47%

22.04%

20.93%

13.64%

11.85%

34.76%

5 Year Vol.

15.42%

18.47%

19.07%

21.98%

21.73%

21.81%

17.80%

25.44%

10 Year Return

7.02%

7.51%

7.71%

6.71%

9.14%

6.62%

7.39%

11.40%

10 Year Vol.

16.12%

18.34%

19.00%

21.94%

21.99%

20.92%

18.20%

25.31%

3

Style Performance at 31 March 2025 Factor weighting Time Period

Winner

Loser

Value

Growth

Low Vol

High Vol

Quality

Junk

1 Month

5.33%

1.80%

1.22%

3.51%

1.06%

9.11%

5.50%

1.21%

1 Month Vol.

12.57%

17.66%

14.12%

14.52%

9.22%

24.51%

14.20%

13.73%

3 Month

3.17%

6.27%

0.97%

4.09%

0.05%

13.82%

4.91%

1.42%

3 Month Vol.

12.38%

14.70%

12.90%

11.90%

10.47%

19.87%

12.50%

12.33%

6 Month

6.99%

-2.88%

-1.05%

3.79%

-2.78%

8.48%

5.15%

-2.65%

6 Month Vol.

11.45%

16.06%

12.55%

12.02%

8.95%

20.18%

11.53%

11.79%

1 Year Return

40.39%

17.71%

25.87%

22.22%

15.15%

23.00%

28.93%

10.50%

1 Year Vol.

12.32%

17.68%

14.82%

12.90%

10.16%

21.40%

13.14%

12.98%

3 Year Return

14.94%

12.78%

5.26%

5.00%

5.19%

6.35%

10.83%

5.75%

3 Year Vol.

14.35%

17.88%

18.53%

12.75%

10.53%

22.02%

17.18%

13.66%

5 Year Return

27.45%

27.19%

29.49%

17.00%

10.59%

29.81%

22.19%

20.22%

5 Year Vol.

18.42%

22.41%

20.86%

15.23%

11.69%

24.39%

18.16%

18.61%

10 Year Return

12.71%

7.12%

12.93%

2.93%

1.47%

14.73%

13.02%

3.75%

10 Year Vol.

19.54%

22.54%

20.76%

16.50%

13.62%

24.26%

17.76%

18.89%

Small

Big

Illiquid

Liquid

High Yield

Low Yield

1 Month

-0.76%

2.75%

0.16%

1.77%

0.90%

1 Month Vol.

10.52%

14.67%

14.05%

15.12%

3 Month

-4.27%

11.24%

2.42%

3 Month Vol.

11.76%

11.89%

6 Month

2.01%

6 Month Vol.

Rand

Rand

Hedge

Tracker

5.43%

1.21%

8.84%

10.17%

15.33%

10.19%

21.43%

-2.24%

1.30%

8.45%

2.46%

12.38%

13.70%

13.88%

11.51%

12.89%

11.28%

18.08%

4.52%

-1.70%

-4.14%

-3.10%

2.86%

0.24%

5.25%

10.40%

11.89%

12.17%

13.56%

10.56%

12.57%

10.04%

18.66%

1 Year Return

32.97%

19.09%

11.75%

28.78%

24.22%

19.12%

13.43%

30.48%

1 Year Vol.

11.84%

12.76%

12.52%

16.40%

12.59%

13.26%

9.91%

20.35%

3 Year Return

10.42%

8.37%

7.51%

6.87%

9.92%

12.12%

9.45%

7.98%

3 Year Vol.

12.10%

16.50%

14.09%

17.15%

15.16%

14.37%

11.52%

21.56%

5 Year Return

24.11%

21.67%

21.08%

26.30%

26.17%

21.62%

17.61%

34.32%

5 Year Vol.

14.85%

17.89%

15.62%

21.21%

19.47%

16.70%

13.53%

25.29%

10 Year Return

5.69%

9.17%

7.66%

9.04%

9.40%

10.69%

10.09%

10.56%

10 Year Vol.

15.49%

18.66%

15.55%

21.12%

20.28%

17.24%

15.91%

24.73%

4

Momentum

Quality

Momentum’s performance over Q1 shows the diver-

Where low vol underperformed, quality outperformed.

gence of outcomes associated with weighting meth-

As noted above, quality as an investment style con-

odologies. Market cap weighting dominated factor

siders shares that have both high quality earnings and

weighting over the quarter, with the former achieving

strong balance sheets. The increased global volatili-

the highest returns across styles of 18%, compared to

ty drove SA investors into safer companies, as quality

the 3.2% for the latter. The discrepancy was even more

outperformed junk over Q1 2025. The quality premium

stark when comparing the momentum premium (win-

was significantly stronger when applying market cap

ner minus losers), with market cap weighting produc-

weighting, with quality outperforming junk by 11.2%

ing an excess return of 10.5% compared to -3.1% for

(factor weighting produced a premium of 3.5%).

factor weighting.

Size Value and Growth

The good fortune of SA inc. small and mid-caps expe-

Value and growth produced consistent results across

rienced over 2024 was reversed in Q1, with large caps

weighting methodologies with growth outperforming

clearly outperforming. Under both weighting meth-

value over Q1 2025, with market cap weighting produc-

odologies, small and mid-caps posted the worst per-

ing a growth premium of 2.9% while factor weighting

formance across styles, dropping by -4.5% and -4.3%

returned 3.1%. The result is consistent with the market

respectively (market cap and factor weighting). Con-

cap weighted momentum results for the quarter, given

versely, large caps performed positively, posting re-

growth’s naturally high correlation with momentum.

turns of 7.7% and 11.24% assuming market cap and factor weighting, resulting in large cap premiums of 12.1%

Low Volatility

and 15.5%.

Unlike global style performance for the quarter, low volatility shares dramatically underperformed their

Liquidity

high volatility counterparts in Q1 2025. Factor weight-

Typically, the illiquidity premium is expected to be

ing produced the starkest results, with high vol shares

highly correlated with the size effect. Interestingly, this

outperforming low vol by 13.8%. More importantly, on a

did not manifest over Q1 2025, with illiquid counters

long-only basis, factor weighted high vol produced the

outperforming their liquid peers by 10.2% and 4.7% on

highest return across styles. In contrast to the positive

a market cap and factor weighted basis respectively.

performance of global low vol, the results seem to be largely driven by the composition of the JSE relative to

Dividend Yield

the MSCI world. Almost every MSCI world style port-

Given that dividend yield (DY) is used as proxy for value,

folio holds some degree of the tech heavy NASDAQ

academic literature typically indicates a positive rela-

which was hardest hit due to the increased volatility as-

tionship between the two styles. This did manifest on

sociated with Trump’s return to office and swath of ex-

the JSE with high DY shares following value in Q1, un-

ecutive orders. Conversely, the JSE was largely helped

derperforming low DY shares by 11.6% and 7.2% assum-

by precious mineral counters, especially the gold min-

ing market cap and factor weighting.

ers which are notoriously volatile while simultaneously being viewed as a safe-haven in market turmoil scenarios.

5

Rand Hedge Rand hedge produced inconsistent performance across weighting methodologies. Over the past quarter, market cap weighted rand hedge shares outperformed rand trackers marginally by 1.7%. Conversely, factor weighting resulted in rand tracker shares dominating rand hedge by 9.9%. The result is logical, especially considering the precious metal surge over the quarter. Gold counters typically maintain above average Rand betas (high Rand beta implies rand hedge) as revenues are largely driven by commodity prices priced in USD yet maintain ZAR based expenses. Shares with the highest ZAR beta’s typically have ‘purer’ offshore exposure, with both income and expenses being denominated in hard currency. As mentioned, market cap weighting naturally upweights larger shares while factor weighting is determined based on the respective style proxy, therefore, in this case, precious metals were effectively upweighted in the market cap weighted Rand hedge portfolio and downweighted when applying factor weighting.

Recent Excess Style Performance Instead of showing cumulative performance, we now present full year and last quarter performance of factors in excess of the JSE Capped SWIX (J433). The figures that follow describe excess style performance on a market cap and factor weighted basis.

6

The results show: •

Market cap weighted momentum and quality convincingly outperformed the Capped SWIX over the quarter

•

Factor weighting negatively impacted momentum and quality, with both underperforming the Capped SWIX over the quarter

•

Under factor weighting, large cap and rand trackers posted the best returns while small and mid-caps were the biggest laggard for the quarter under both weighting methodologies.

7

Style Correlation Winner

Winner

100.0%

Loser

Loser

Value

50.8% 100.0%

Value

Growth

Low Vol

High Vol

50.0%

73.3%

63.4%

55.2%

78.3%

61.1%

100.0%

54.2%

63.8%

Growth

100.0%

Low Vol

Quality

Junk

Small

76.2%

77.7%

68.3%

81.6%

72.3%

83.6%

54.4%

60.8%

56.8%

Low Yield

Rand Hedge

Rand Tracker

43.1%

81.3%

74.0%

65.6%

50.6%

73.4%

59.5%

81.8%

52.7%

40.8%

69.2%

Big

Illiquid

Liquid

High Yield

65.6%

77.1%

40.3%

74.5%

74.5%

75.5%

39.1%

81.4%

49.6%

58.2%

88.8%

60.5%

95.9%

75.7%

76.3%

77.9%

85.1%

79.1%

79.1%

42.9%

84.8%

51.7%

80.7%

67.5%

80.6%

100.0%

52.3%

67.4%

72.1%

74.5%

64.6%

57.6%

70.5%

64.6%

66.4%

63.3%

67.7%

79.9%

77.5%

69.3%

82.5%

41.0%

80.3%

46.5%

85.0%

71.4%

79.7%

100.0%

71.3%

70.1%

84.8%

42.5%

82.8%

51.9%

77.8%

64.5%

82.0%

77.1%

79.5%

42.7%

85.1%

53.2%

78.0%

66.9%

82.2%

100.0%

62.3%

41.2%

75.9%

48.1%

70.7%

60.7%

71.3%

43.1%

81.1%

50.2%

84.1%

73.9%

80.7%

100.0%

37.9%

91.4%

42.7%

39.1%

49.3%

54.5%

79.4%

62.8%

87.2%

100.0%

44.5%

36.8%

63.7%

100.0%

78.1%

73.9%

High Vol

100.0%

Quality Junk

100.0%

Small Big

100.0%

Illiquid Liquid

100.0%

High Yield Low Yield Rand Hedge

100.0%

Rand Tracker

47.8% 100.0%

Top 10 tickers Rank

Momentum

Value

Growth

Low Volatility

Quality

Small

Illiquid

Yield

Rand Hedge

1

WE BUY CARS

CAPITEC

MULTICHOICE

MULTICHOICE

MR PRICE

SUN INTERNATIONAL

KAROOOOO

THUNGELA

BATS

2

PREMIER

THUNGELA

CLICKS

NEPI

THUNGELA

OMNIA

POWERFLEET

KUMBA

ABinBEV

3

PICK N PAY

ABSA

OUTSURANCE

LIGHTHOUSE

SUN INTERNATIONAL

REUNERT

PREMIER

EXXARO

RICHEMONT

4

BARLOWORLD

ANGLO PLATINUM

BHP

ABinBEV

AVI

SOUTHERN SUN

SANTAM

CORONATION

SHAFTESBURY CAPITAL

5

KAROOOOO

EXXARO

PICK N PAY

SIRIUS

CORONATION

HOSKENS

HOSKENS

SUN INTERNATIONAL

SIRIUS

6

OUTSURANCE

MOTUS

AVI

FORTRESS

BYTES

EQUITES

SUPERMARKET INC

REDEFINE

BYTES

7

DISCOVERY

KUMBA

DIS-CHEM

AFRICAN RAINBOW CAPITAL

CLICKS

WBHO

PRIMARY HEALTH PROPERTIES

HYPROP

MONDI

8

SOUTHERN SUN

NEDBANK

SHOPRITE

PSG

PICK N PAY

MAS P.L.C.

WBHO

AFRICAN RAINBOW MINERALS

KAROOOOO

9

PAN AFRICAN RESOURCES

FORTRESS

MTN

BID CORP

WOOLWORTHS

DATATEC

AFRICAN RAINBOW CAPITAL

EQUITES

BID CORP

10

AFRICAN RAINBOW CAPITAL

SASOL

MR PRICE

SHAFTESBURY CAPITAL

KAROOOOO

ITALTILE

ITALTILE

GROWTHPOINT

MAS P.L.C.

The top 10 tickers per style category are described above with each tickers rank being determined by its respective styles z-score. The top 10 momentum maintained We Buy Cars at the top with the rest of the top 10 being a fair mix of small and large cap counters across industries. 3 of the big 4 banks made it into the top 10 value counters as well as both coal miners (Thungela and Exxaro). The top 10 growth counters were once again made up of largely consumer discretionary and consumer staples while BHP entered at number 4. Mr Price once again tops the quality counters in Q1 2025 as well as the other retailers (Clicks, Pick n’ Pay and Woolies). Thungela replaced by Exxaro as the top DY share, followed by Kumba and Exxaro, while the large RIETS also featured. The top 10 rand hedge counters are highly consistent given the relationship with the USDZAR exchange rate with all having either offshore hard currency earnings and asset bases (usual suspects, namely BAT, AB InBev & Richemont) while BidCorp came in at number 9.

Datascience Corner Back-tests provide an educated glimpse of historical style related investment opportunities. Consistent with theoretical framework of asset pricing, we view styles as risk-factors that explain returns. Therefore, by implication, style returns are risky and move stochastically across the performance spectrum. This means that historical analysis provides minimal insight about the future performance of styles, especially over the short-term. To this end, 27four Beta Solutions applies data science and econometric modelling to provide short to mid-term views on future style performance. The models applied are vector auto-regression (“VAR”) and long and short-term memory (“LSTM”) deep learning neural networks. Both are well established in literature, with the former a staple in economics and finance studies while the latter is a more recent and highly popular machine learning algorithm. The benefit of both techniques is that they allow for the estimation of a ‘system’ where all variables in question are modelled to be inter-related. Once the system is modelled, recent data is fed into the models to estimate future performance using the style in question as well as the other styles present on the JSE.

9

A summary of the results is presented graphically above. The figure shows the next quarters forecast for style performance on the JSE where the LSTM predicts relative outperformance for momentum, small and mid-caps as well as illiquid shares while having a negative outlook for the benchmark. Conversely, VAR is predicting superior performance for momentum, rand hedge and growth as well as positive performance for the benchmark. The results of the forecast analysis are synthesised in the table below where the prior quarters forecast is evaluated in terms of accuracy (green means materialised, yellow means partially materialised and red means did not materialise) as well as presenting in the last two columns the VAR and LSTM’s next quarter forecast.

Style

Prior Quarter VAR Forecast

Prior Quarter LSTM Forecast

Next Quarter VAR Forecast

Next Quarter LSTM Forecast

Momentum

Neutral

Positive

Positive

Positive

Growth

Neutral

Positive

Positive

Positive

Quality

Positive

Positive

Positive

Neutral

Value

Positive

Neutral

Neutral

Positive

Low Volatility

Neutral

Positive

Positive

Negative

Size

Negative

Positive

Negative

Positive

Dividend Yield

Neutral

Neutral

Neutral

Positive

Illiquidity

Positive

Neutral

Positive

Positive

Rand Hedge

Positive

Neutral

Positive

Negative

JSE

Positive

Neutral

Positive

Negative

10

11