-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathDecoding Sector Valuation Dynamics.tex
1845 lines (1631 loc) · 225 KB
/
Decoding Sector Valuation Dynamics.tex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
\documentclass[12pt,a4paper,english]{article}
\usepackage[hidelinks]{hyperref}
\usepackage{csquotes}
\usepackage[]{apacite}
\usepackage{indentfirst}
\usepackage{appendix}
% Margins
\usepackage{geometry}
\geometry{margin=1in}
% Double-spaced font
\usepackage{setspace}
\doublespace
% Header and Footer
\usepackage{fancyhdr}
\fancypagestyle{fancy}{
\fancyhf{}
\setlength{\headsep}{0.5\baselineskip}
\renewcommand{\headrulewidth}{0pt} % Remove header line
\renewcommand{\footrulewidth}{0pt} % Remove footer line
\fancyfoot[C]{\makebox[\textwidth]{\noindent\thepage}}
}
\pagestyle{fancy}
\fancypagestyle{main}{
\fancyhf{}
\setlength{\headsep}{0.5\baselineskip}
\renewcommand{\headrulewidth}{0pt} % Remove header line
\renewcommand{\footrulewidth}{0pt} % Remove footer line
\fancyfoot[C]{\makebox[\textwidth]{\noindent\thepage}}
\lhead{DECODING SECTOR VALUATION DYNAMICS}
}
% Graphics (Tables and Figures)
\usepackage{graphicx}
\usepackage{booktabs} % For better-looking tables
\usepackage{tabularx}
\usepackage{makecell}
\usepackage{array}
\usepackage{adjustbox}
\usepackage{rotating}
\usepackage{float} % For the H specifier
\usepackage[normalem]{ulem}
\useunder{\uline}{\ul}{}
\usepackage{etoolbox}
\BeforeBeginEnvironment{table}{\begin{singlespace}}
\AfterEndEnvironment{table}{\end{singlespace}}
% Footnotes
\usepackage{tablefootnote}
\usepackage[hang, bottom]{footmisc}
\setlength{\belowcaptionskip}{0pt}
% Captions
\usepackage{caption}
\captionsetup{font=normalsize,format=plain}
\DeclareCaptionFormat{myformat}{\raggedright\textbf{#1}\\[2ex]\textit{#3}\vspace*{1ex}}
\captionsetup[table]{format=myformat}
\captionsetup[figure]{format=myformat}
\usepackage{subcaption}
\DeclareCaptionFormat{myformat-subt}{\footnotesize{\textbf{#1} #3}}
\captionsetup[subtable]{format=myformat-subt}
% Title formatting
\usepackage{titlesec}
\titleformat{\section}[block]{\fontsize{16pt}{18pt}\selectfont\bfseries\centering}{\thesection}{1em}{\phantomsection}
\titleformat{\subsection}[block]{\fontsize{14pt}{16pt}\selectfont\bfseries}{\thesubsection}{1em}{\phantomsection}
\titlespacing{\section}{0pt}{0pt}{0pt}
\titlespacing{\subsection}{0pt}{0pt}{0pt}
% Itemize formatting
\usepackage{enumitem}
\setlist[itemize]{itemsep=-2pt, topsep=0pt}
% Begin Doc
\begin{document}
\begin{titlepage}
\thispagestyle{fancy}
\vspace*{\fill}
\begin{center}
\textbf{\LARGE{Decoding Sector Valuation Dynamics:}}\\
\textbf{\Large{Exploring the Nexus of Company Fundamentals, Economic Indicators, and Equity Valuations via Predictive Analytics}}\\
\vspace{0.5cm}
Dane Skalski\\
Honors Thesis\\
Spring 2024\\
\vspace{0.5cm}
Advisor:\\
Mario Reyes\\
Department of Finance and Management Science\\
Carson College of Business\\
Washington State University
\end{center}
\vspace*{\fill}
\end{titlepage}
\clearpage
\raggedright
\setlength{\parindent}{1cm}
\setcounter{secnumdepth}{0}
\pagestyle{main}
\section*{Précis}
This study examines the key drivers of equity valuation across various sectors and industries. Equity valuation is a critical metric that captures a company's perceived worth and influences investment decisions, mergers and acquisitions, and risk management strategies. Understanding the primary factors impacting valuation is essential for market participants to make informed choices.
The research seeks to unveil determinants of equity valuation that may have been obscured by traditional linear regression techniques. With modern ensemble machine learning methods, particularly gradient-boosted decision trees, the study aims to leverage non-linear relationships among a diverse set of variables and fluctuations in stock prices, which serves as a proxy for shifts in company valuations.
The analysis employed both linear regression and gradient-boosted decision tree models to explain the variance in monthly stock returns across various sectors within the Global Industry Classification Standard (GICS). The results uncover substantial variations in the factors influencing equity valuation between sectors. Various commodities, economic factors, and company fundamentals emerged as the primary determinants, with their relative importance contingent upon the sector context. Particularly noteworthy is the consistent superiority of gradient-boosted decision tree models over linear regression methods, underscoring the significance of capturing intricate non-linear patterns in financial data analysis.
The identified drivers provide crucial insights that can enhance investors' decision-making, improve valuation modeling practices, and strengthen risk management strategies. By uncovering these pivotal factors, the study aims to guide market participants in navigating the complexities of financial markets with greater foresight.
Future research could explore the incorporation of additional data sources, such as forward guidance from companies, analyst forecasts, news events, and social media sentiment, to further enhance the predictive power of the models. Additionally, investigating industry-specific, sub-industry-specific, and global market dynamics could furnish valuable insights into the determinants of equity valuation transcending the context of the United States and broad sector approach.
Overall, this study contributes to a more comprehensive understanding of the dynamic forces shaping equity valuations, empowering investors, analysts, and industry stakeholders to make more informed decisions and refine their valuation methodologies.
\clearpage
\tableofcontents
\cleardoublepage
\listoffigures
\cleardoublepage
\listoftables
\cleardoublepage
\pagestyle{main}
\phantomsection\addcontentsline{toc}{section}{Introduction}
\section*{Decoding Sector Valuation Dynamics: \\Exploring the Nexus of Company Fundamentals, Economic Indicators, and Equity Valuations via Predictive Analytics}
The valuation of a company represents a pivotal endeavor in the dynamic landscape of financial markets, serving as a guiding compass for investors and shaping capital allocation decisions. This metric encapsulates a company's perceived worth, reflecting its financial health, growth prospects, and market standing. Understanding the primary drivers of equity valuation is critical for investors to make informed choices regarding stock selection, portfolio management, and risk mitigation strategies. Moreover, in mergers and acquisitions, precise and nuanced company valuations influence negotiations and strategic decisions that can restructure entire industries. As articulated by \citeA[p. 3]{koller2010valuation}, \enquote{value is the defining dimension of measurement in a market economy.}
Expanding upon the foundational work laid by \citeA{ely2007equity}, \citeA{liu2002}, \citeA{khan2016}, and \citeA{deng2012}, this study aims to delve into the determinants of equity valuations across diverse sectors and industries. While these prior studies have shed light on the varying influences of certain variables on company valuation within specific industry and sector contexts, their reliance on conventional linear regression methods may have overlooked the potential impacts of nuanced non-linear relationships and more intricate patterns on valuation dynamics.
Similarly, this research encompasses a broad spectrum of metrics, including commodities, economic indicators, and diverse company fundamentals, seeking to uncover the primary drivers of equity valuation within the Global Industry Classification Standard (GICS) sectors. Employing modern ensemble machine learning techniques, such as gradient-boosted decision trees, harnesses relationships that may have eluded previous linear regression analyses. Thus, this study aims to offer a more comprehensive and holistic understanding of the key factors driving fluctuations in valuation.
Acknowledging the unique characteristics and complexities inherent in various sectors, as emphasized by \citeA{li2021improving}, this research endeavors to offer empirical insights that equip market participants with a deeper comprehension of the factors impacting sector-specific valuations with enhanced accuracy and foresight. By decoding these pivotal drivers, this study also aims to steer risk management and speculative strategies, facilitating the anticipation of potential valuation reactions to shifts in the identified determinants.
Moreover, valuation methodologies like Discounted Cash Flow (DCF) analysis and comparable company analysis heavily hinge on underlying assumptions, profoundly influencing the resulting valuation estimates. Through the identification and prioritization of key drivers that may have been previously overlooked or undervalued, there lies the opportunity to refine and optimize due diligence resource allocation in valuation modeling endeavors.
\section{Methodology}
\subsection{Data Collection}
To construct the dataset, historical daily stock prices, quarterly financial statements (10-Q Filings), and earnings call dates were collected for companies contained within the S\&P 500, 400, and 600 indices. Stock data and 10-Q financials were sourced from \citeauthor{yahoofinance}, while the respective earnings call dates and times were collected through the \citeauthor{stocktwits} API. Various commodity prices, economic indicators, and other market factors were retrieved from \citeauthor{yahoofinance} and the Federal Reserve Economic Data (FRED) database.
The dependent variable is the percentage change in a stock's adjusted closing price from the previous period, representing the change in a company's valuation. Independent variables include the percent change for a diverse set of company fundamentals (e.g. revenue, profitability ratios, cash flows), commodity prices, economic indicators, and other market factors potentially influencing equity valuations, refer to Table \ref{tab:variables} in the appendix for a comprehensive list of variables.
A monthly time frame was analyzed, spanning from January 25, 2014, to February 17, 2024, encompassing 525 weeks (approximately 10 years). Percent changes in company fundamentals and select economic indicators reported quarterly were calculated based on their last release.
\subsection{Data Preprocessing}
The raw data underwent several preprocessing and analysis steps to handle outliers and missing values.
For the dependent variable, monthly stock returns, the top and bottom 15 values for each sector were explored to identify potential extreme outliers. Observations due to extraordinary events were removed (e.g. GameStop's short squeeze in 2021, failed acquisitions, bankruptcy) as well as improper historical data for IPOs. Further outlier treatment involved Winsorizing returns to within 4-6 interquartile ranges (IQRs) from the median on a per-sector basis.
To address potential autocorrelation stemming from seasonality, new features were incorporated with lagged values spanning 3, 6, or 12 months. The selection of specific lags for each GICS sector was guided by autocorrelation function plots depicting the median percentage change of each company fundamental. This approach mitigates the effects of seasonal patterns in an overall sector's financial metrics on the model.
Missing values for independent variables were imputed with 0, representing no change, when there was no economic data release or earnings call that particular month. The independent variables were then standardized. Potential outliers for the independent variables were identified and removed using the DBSCAN clustering algorithm for each GICS sector \cite{ester1996density}.
\subsection{Model Development}
Both a multivariate linear regression model from the Python library cuML and a gradient-boosted decision tree regression model (LightGBM) were trained for selected\footnote{Refer to Table \ref{tab:sectors-industries} in the appendix for selected sectors/industries.} GICS sectors/industries to explain the variance in the percentage change in stock prices based on the percentage change in the independent variables \cite{raschka2020machine,zhang2017gpu}.
A 5-fold cross-validation technique was employed to assess the models during both variable selection and hyperparameter tuning stages, aiming to mitigate potential overfitting of the models and to maximize the generalizability of the findings. These folds were stratified to ensure an equal representation of earnings releases within each fold.
A backward stepwise regression approach was used to address multicollinearity and retain the most important independent variables for each model. Highly correlated variables were identified using correlation matrices, scatter plots, and Variance Inflation Factor (VIF) analysis. The groups of correlated variables identified across the full dataset were:
\begin{table}[H]
\centering
\caption{Correlated Variable Groups}
\label{tab:var-correlates}
\resizebox{\textwidth}{!}{%
\begin{tabular}{l|l|l|l|l|l|l|l|l|l}
\toprule
Net Income & Debt Ratio & Total Revenue & EBIT & Current Ratio & Cash Flow Margin & Tangible BV per Share & Gold & Copper & Wheat \\
ROA & Debt to Equity & Operating Revenue & Interest Coverage & Quick Ratio & CROA & BV per Share & Silver & Aluminum & Corn \\
Basic EPS & Assets to Equity & & & & Cash Flow to Debt & & Platinum & & Soybean \\
ROE & & & & & & & & & \\
Net Profit Margin & & & & & & & & & \\ \bottomrule
\end{tabular}%
}
\end{table}
Subsequently, each variable within its group of correlates underwent thorough testing in each model. The individual variable that contributed to the highest average cross-validated $R^2$ value among its correlates was retained. After mitigating multicollinearity, variables were systematically eliminated from the model to maximize the average cross-validated $R^2$, thereby removing variables that did not enhance nor contribute to the predictive capability of the models.
Finally, cross-validated hyperparameter tuning was conducted on the linear regression model (Elastic Net) and the gradient-boosted decision tree model (LightGBM) in an effort to optimize model performance and stabilize the remaining variables through regularization.
\subsection{Model Rationale}
The methodology employed in this study leverages regression techniques to uncover the drivers influencing equity valuations by explaining the variance in stock returns. The percentage change in stock prices as the dependent variable provides a direct window into how various financial metrics and economic factors impact a firm's market value. Regression analysis allows for quantifying the relationships between these independent variables and shifts in company valuation. By using gradient-boosted decision trees (LightGBM), analysis can flexibly capture complex, non-linear patterns that meaningfully contribute to movement in valuations obscured by traditional linear approaches. The variable selection process retains the most influential drivers while mitigating multicollinearity concerns. Ultimately, uncovering these determinants by explaining stock return variance elucidates the factors driving changes in valuations within each sector's context. This crucial insight allows market participants to refine valuation modeling due diligence practices and inform speculative strategies by anticipating potential reactions to changes in identified drivers. Furthermore, it contributes to strengthening risk management protocols through proactive measures that mitigate potential risks associated with fluctuations in the identified determinants.
\section{Analysis and Results}
\newcommand{\selmet}[2]{
\begin{table}[H]
\centering
\begin{minipage}{\textwidth}
\centering
\caption{#1: Top Drivers}
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}rr|rr|rr@{}}
\toprule
\multicolumn{2}{c|}{{\uline{Commodities}}} & \multicolumn{2}{c|}{{\uline{Economic}}} & \multicolumn{2}{c}{{\uline{Company Fundamentals}}} \\
\multicolumn{1}{c}{\thead{Importance\\ Rank}} & \multicolumn{1}{c|}{\thead{Commodity}} & \multicolumn{1}{c}{\thead{Importance\\ Rank}} & \multicolumn{1}{c|}{\thead{Factor}} & \multicolumn{1}{c}{\thead{Importance\\ Rank}} & \multicolumn{1}{c}{\thead{Metric}} \\
\midrule
#2\\
\bottomrule
\end{tabular}
}
\footnotetext{Note: $\uparrow$ denotes that higher variable values are associated with positive returns (i.e. \enquote{higher is better}), $\downarrow$ denotes that lower variable values are associated with positive returns (i.e. \enquote{lower is better}).}
\footnotetext{Drivers based on the median LightGBM model's\tablefootnote[2]{Median Model: Model with the median $R^2$ across 5-fold cross-validation.} SHAP values. Ambiguous relationships were omitted, see Figure \ref{fig:#1-shap} for the model's SHAP plot. \vspace{-\baselineskip}}
\end{minipage}
\end{table}
}
Linear regression models can be interpreted using the coefficients of the standardized independent variables, while gradient-boosted decision tree models can be interpreted using SHAP (SHapley Additive exPlanations) values. SHAP values offer a unified interpretation of models by quantifying the impact of each independent variable on the model's output \cite{lundberg2017,mitchell2022gputreeshap}. Interpretation is based on the median LightGBM model (model with the median $R^2$ across 5-fold cross-validation). Utilizing the median model provides a balanced perspective on model performance and explanatory power, alleviating potential overfitting concerns.
Considering the monthly time frame used in the regression analysis, commodities and most economic indicators are expected to carry high importance to the models. This is due to the month-to-month variation observed in these variables, which consequently impacts valuations more frequently within the chosen time frame. Company fundamentals and certain economic variables are typically reported on a quarterly basis, potentially diminishing their importance because of their lesser frequency. Therefore, it's essential to keep reporting frequencies in mind while interpreting each primary driver's importance ranking. Variables with less frequent reporting should be weighed accordingly, independent of those exhibiting month-to-month volatility. Refer to Table \ref{tab:variables} for a comprehensive list of variables, noting non-monthly reporting frequencies.
Additionally, some economic indicators are not reported in real-time, lagging current economic conditions. As such, these variables should be interpreted as the market's reaction to these economic data releases and its subsequent anticipation of future economic conditions/trends.
\subsection{Summary Table}
\begingroup
\renewcommand*{\thefootnote}{\alph{footnote}} % Use letters for footnotes
\begin{table}[H]
\centering
\begin{minipage}{\textwidth}
\caption{Summary Table: Identified Top Drivers by Sector}
\label{tab:summary-table}
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrrrrrrrr|r@{}}
\toprule
& \multicolumn{11}{c}{{\ul Sector}} & \multicolumn{1}{|c}{} \\
\multicolumn{1}{c}{{Variable}} & \multicolumn{1}{c}{Energy} & \multicolumn{1}{c}{Materials} & \multicolumn{1}{c}{Industrials} & \multicolumn{1}{c}{\thead{Consumer \\ Discretionary}} & \multicolumn{1}{c}{\thead{Consumer\\ Staples}} & \multicolumn{1}{c}{\thead{Health\\ Care}} & \multicolumn{1}{c}{\thead{Financial\\ Services}} & \multicolumn{1}{c}{\thead{Information\\ Technology}} & \multicolumn{1}{c}{\thead{Communication\\ Services}} & \multicolumn{1}{c}{Utilities} & \multicolumn{1}{c}{\thead{Real\\ Estate}} & \multicolumn{1}{|c}{\thead{Total\\ Count}} \\ \midrule
\multicolumn{1}{c}{{\ul Commodities}} & & & & & & & & & & & & \\
Oil & I & I & & I & I & I & I & I & I & & I & 9 \\
Gold & & & & & & & & & & & & 0 \\
Silver & & & & & & & & & I & & I & 2 \\
Copper & & & I & & & I & & & & & & 2 \\
Aluminum & & & & & & & I & I & I & & & 3 \\
Platinum & & & & I & I & & & I & & & & 3 \\
Palladium & I & I & & & I & & & & & I & & 4 \\
Natural Gas & I & & I & & & & & & & & & 2 \\
Wheat & & & & & I & I & & & & I & & 3 \\
Corn & & & & & & & & & & & & 0 \\
Soybean & & I & & & & & & & & & & 1 \\
Coffee & & & I & & I & & & & & & & 2 \\
Sugar & & & & & & I & & & & & & 1 \\
Cotton & I & I & I & I & & & I & & I & & & 6 \\ \midrule
\multicolumn{1}{c}{{\ul Economic}} & & & & & & & & & & & & \\
Federal Funds Rate & & I & I & I & I & I & & I & & I & I & 8 \\
3 Month Treasury Yield & I & & & & & & & & & I & & 2 \\
10 Year - 2 Year Treasury Yield Spread & I & I & I & & I & & & & & & & 4 \\
BAA - AAA Corporate Bond Yield Spread & & & & & & I & & I & & & & 2 \\
VIX & I & I & I & I & I & I & I & & & I & I & 9 \\
30 Year - 15 Year Mortgage Rate Spread & & & & & & & & & I & & & 1 \\
CPI & I & I & & I & & & I & & I & & I & 6 \\
Unemployment Rate & & & & & & & & & & & I & 1 \\
Federal Budget Surplus or Deficit & & & & I & & & & & & & & 1 \\
Housing Starts & & & I & I & I & & I & I & I & & & 6 \\
Real GDP & & & & & & & & & & I & & 1 \\
Average Home Price & & & I & & & & & & & & & 1 \\ \midrule
\multicolumn{1}{c}{{\ul Company Fundamentals}} & & & & & & & & & & & & \\
Total Revenue & & & & & & I & & & & & I & 2 \\
Operating Revenue & I & & I & & & & I & & & & & 3 \\
Net Income & & & & & & & & & & & & 0 \\
Basic EPS & & & & & I & & & & & & & 1 \\
EBITDA & & I & & I & & & & & & & I & 3 \\
EBIT & & & & & & & & & I & & & 1 \\
Current Ratio & & & & & & & & & & I & & 1 \\
Quick Ratio & & & & & & & & & & & & 0 \\
Gross Profit Margin & & I & I & I & I & I & & I & & & I & 7 \\
Net Profit Margin & & & & & & & & & I & & & 1 \\
Cash Flow Margin & & & I & & & & & & & & & 1 \\
Operating Margin & & & I & I & I & I & I & I & I & & I & 8 \\
ROA & I & & & & & I & & & & & & 2 \\
CROA & & & & & I & & & & & & & 1 \\
ROE & & & & I & & & & I & & & & 2 \\
Efficiency Ratio & I & & & & & & & & I & I & & 3 \\
Inventory Turnover & & & & & & & & & & & & 0 \\
Debt to Equity & & & & & & & & & & & & 0 \\
Debt Ratio & & I & & & & I & & & & & & 2 \\
Interest Coverage & & & & I & & & & & & & & 1 \\
Cash Flow to Debt & I & I & & & & & & & & & & 2 \\
Assets to Equity & & & & & & & & & & & & 0 \\
R\&D to Revenue & & & & & & & & I & & & & 1 \\
Investment CF to OCF & & & & & I & & & & & & & 1 \\
Financing CF to OCF & & & I & & & & & & & I & & 2 \\
Free Cash Flow & I & I & I & I & I & & I & I & I & I & & 9 \\
Capex & & & & & & & I & & & & I & 2 \\
BV per Share & & & & & & & & & & & & 0 \\
Tangible BV per Share & & & & & & & & & & & & 0 \\ \bottomrule
\end{tabular}%
}
\footnotetext{I: Identified as a top driver for the sector.}
\footnotetext{Financial Services: Industry Group}
\end{minipage}
\end{table}
\endgroup
Oil and cotton emerged as primary drivers for 9/11 and 6/11 sectors, respectively, highlighting their important impact on equity valuations and the economy at large. Oil's importance is unsurprising given its pervasive role in the economy, influencing production costs, transportation, and consumer spending. Cotton's prominence is likely more indirect, rising as an important indicator of overall consumer demand and economic activity. Gold was never identified as a top driver, however, it is highly correlated with silver (2/11) and platinum (3/12). These other metals tended to explain shifts in valuations more effectively than gold. Among the soybean (1/11), wheat (3/11), and corn (0/11) group of correlates, a preference was given to wheat.
The fear index, VIX, was a top driver in 9/11 sectors, underscoring its critical role in reflecting market sentiment and risk perceptions. The federal funds rate was a top driver in 8/11 sectors as an important economic variable with influence on borrowing costs, consumer spending, and investment decisions. Inflation and housing starts were also common drivers, both identified in 6/11 sectors, drawing attention to their broad impacts on the economy.
Free cash flow was the most commonly identified company fundamental, emerging as a top driver in 9/11 sectors. Profitability margins (e.g. gross profit margin, operating margin) were also prevalent as measures of financial health and operational efficiency.
A variable from the net income, ROA, EPS, ROE, and net profit margin group of correlates was identified in 6/11 sectors. However, which variable emerged varied among sectors; notably, net income never emerged as a primary driver, and the others surfaced in only 1 or 2 sectors each. This underscores the nuanced interplay between these metrics and equity valuations within the unique context of each sector.
\subsection{Energy Sector}
\label{sec:energy}
The Energy Sector model was the best-performing, exhibiting a median adjusted $R^2$ of 50.74\% on the LightGBM model (Table \ref{tab:Energy Sector-model-stats}). Top drivers of positive stock returns in the Energy Sector:
\selmet{Energy Sector}{
1 & Oil $\uparrow$ & 2 & VIX $\downarrow$ & 18 & ROA $\uparrow$ \\
3 & Cotton $\uparrow$ & 8 & 3 Month Treasury Yield $\uparrow$ & 19 & Operating Revenue $\uparrow$ \\
7 & Natural Gas $\uparrow$ & 11 & 10 Year - 2 Year Treasury Yield Spread $\downarrow$ & 22 & Cash Flow to Debt $\uparrow$ \\
9 & Palladium $\downarrow$ & 12 & CPI $\downarrow$ & 23 & Efficiency Ratio $\downarrow$ \\
& & & & 24 & Free Cash Flow $\uparrow$
}
Unsurprisingly, commodities take center stage, with oil as the most important driver and natural gas in the top 10. Given the energy sector's heavy reliance on oil/gas production and exploration, fluctuations in oil/gas prices directly impact the profitability and valuations of energy companies. Other commodities such as cotton and palladium also play a role, reflecting broader economic conditions and industrial demand. Likely due to its usage in vehicle manufacturing and petroleum refining, the sector is sensitive to palladium prices. Increases in palladium prices may impact costs for energy companies and their bottom line.
Key indicators such as market volatility (VIX), interest rates (Treasury yields), and broader macroeconomic metrics like the Consumer Price Index (CPI) are influential. Market volatility can signal economic uncertainty, impacting investor sentiment and consequently, valuations. The sector appears to do well in periods of higher short-term interest rates and low inflation.
Return on assets (ROA) measures how efficiently a company is utilizing its assets to generate profits. In the energy industry, which requires large investments in exploration, drilling, refineries, and infrastructure, efficient usage of assets is crucial for long-term sustainability and attracting investor capital.
Operating revenue directly reflects the top-line performance of a company, which is heavily influenced by commodity prices and demand for energy products. Consistently growing operating revenue is a positive signal for valuations.
The cash flow to debt ratio is vital as energy projects often require significant debt financing. A higher cash flow to debt ratio indicates a company's ability to service its debt obligations and maintain financial flexibility. This is especially important during periods of economic uncertainty and low energy price environments.
Efficiency ratio measures how well a company is managing its costs and operating expenses. Energy companies with lower efficiency ratios are better positioned to weather periods of low commodity prices and maintain profitability.
Free cash flow is the cash remaining after a company has paid for its operating expenses and capital expenditures. In the energy sector, where capital expenditures are high, robust free cash flow generation is essential for funding new projects, paying down debt, and driving long-term value.
Overall, the sector values more efficient asset utilization and debt management as well as increasing operating revenue with free cash flow growth.
\subsection{Materials Sector}
\label{sec:materials}
The Materials Sector exhibited a median adjusted $R^2$ of 35.97\% on the LightGBM model (Table \ref{tab:Materials Sector-model-stats}). Top drivers of positive stock returns in the Materials Sector:
\selmet{Materials Sector}{
2 & Oil $\uparrow$ & 1 & VIX $\downarrow$ & 7 & Gross Profit Margin $\uparrow$ \\
3 & Soybean $\uparrow$ & 10 & Federal Funds Rate $\downarrow$ & 11 & Debt Ratio $\downarrow$ \\
4 & Cotton $\uparrow$ & 15 & 10 Year - 2 Year Treasury Yield Spread $\downarrow$ & 17 & Cash Flow to Debt $\uparrow$ \\
5 & Palladium $\uparrow$ & 16 & CPI $\downarrow$ & 18 \& 22 & EBITDA $\uparrow$ \\
& & & & 28 & Free Cash Flow $\uparrow$
}
Commodities take center stage as key drivers for the sector, with oil as the most influential commodity much like the energy sector. Oil plays a pivotal role as a material input and energy source across various industries within the materials sector. Many commodities move in tandem with oil prices, reflecting broader economic conditions and industrial demand. High soybean and cotton prices are also reflective of high demand for raw materials.
Turning to economic factors, market volatility (VIX) stands out as an important driver. Heightened market volatility often signals economic uncertainty, which can impact demand for materials and disrupt supply chains, ultimately affecting the sector's performance.
The federal funds rate and the 10-year to 2-year Treasury yield spread also emerged as influential factors. Changes in interest rates can impact borrowing costs for materials companies, particularly those with substantial debt financing for capital-intensive projects or operations. Decreasing short-term interest rates has a positive impact on the sector, reducing borrowing costs.
The Consumer Price Index (CPI) can impact input costs for materials companies, potentially squeezing profit margins and affecting profitability.
Regarding company fundamentals, the gross profit margin stands out. It is ranked in the top 10 of importance despite its quarterly reporting frequency. Gross profit margin is a key metric that reflects a company's ability to generate profits after accounting for production costs. In the Materials Sector, where capital expenditures can be significant, maintaining healthy gross profit margins is crucial for profitability and long-term sustainability.
Lower debt (debt ratio) and the cash flow to debt ratio are vital metrics, as materials companies often rely on debt financing for capital-intensive projects and operations. A higher cash flow to debt ratio indicates a company's ability to service its debt obligations and maintain financial flexibility, which is essential during economic downturns or periods of low demand.
EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) is a widely used measure of a company's operating profitability, providing insights into its core business performance without the impact of financing and accounting decisions. In the materials sector, EBITDA is a valuable metric for assessing a company's earnings power.
In capital-intensive sectors like materials, robust free cash flow generation is essential for funding growth opportunities, paying down debt, and driving long-term value. Overall, the sector values higher profit margins and more efficient debt management with EBITDA and free cash flow growth.
\subsection{Industrials Sector}
\label{sec:industrials}
The Industrials Sector exhibited a median adjusted $R^2$ of 38.10\% on the LightGBM model (Table \ref{tab:Industrials Sector-model-stats}). Top drivers of positive stock returns in the Industrials Sector:
\selmet{Industrials Sector}{
2 & Copper $\uparrow$ & 1 & VIX $\downarrow$ & 11 & Operating Revenue $\uparrow$ \\
5 & Cotton $\uparrow$ & 4 & Federal Funds Rate $\downarrow$ & 17 & Gross Profit Margin $\uparrow$ \\
12 & Natural Gas $\uparrow$ & 7 & 10 Year - 2 Year Treasury Yield Spread $\downarrow$ & 19 & Operating Margin $\uparrow$ \\
15 & Coffee $\uparrow$ & 13 & Average Home Price $\downarrow$ & 20 & Free Cash Flow $\uparrow$ \\
& & 18 & Housing Starts $\uparrow$ & 21 & Financing CF to OCF $\downarrow$ \\
& & & & 22 & Cash Flow Margin $\uparrow$
}
Commodities greatly impact the sector, with copper, cotton, natural gas, and coffee playing key roles. Copper is essential in construction, manufacturing, and infrastructure projects, serving as an indicator of economic activity and industrial demand. Cotton is a key agricultural output, with higher prices signaling increased demand for textiles and in turn, industrial and farming equipment. Coffee is a consumer staple with global demand, reflecting broader economic conditions and industrial activity.
Market volatility (VIX) remains a crucial driver, with lower volatility typically associated with positive stock returns. Heightened market volatility can signal economic uncertainty, impacting investor sentiment and industrial demand.
Interest rates, as measured by the 10-year to 2-year Treasury yield spread and federal funds rate, are highlighted as influential factors. Higher interest rates can impact borrowing costs for industrial companies, particularly those with significant debt financing for capital projects and operations.
Real estate metrics such as average home prices and housing starts are primary drivers. Increases in home construction are a boon for the sector, signaling economic growth and demand for construction materials and equipment. Declining home prices are potentially the result of additional home supply, signaling the high demand for new projects. These metrics provide insights into housing market conditions, construction activity, and infrastructure development, which can impact industrial companies' performance.
Key company fundamentals include operating revenue, gross profit margin, and operating margin. These serve as vital indicators of a company's performance and profitability. Maintaining healthy revenue growth and profit margins is essential for industrial companies to drive shareholder value and sustain long-term growth.
Additionally, metrics like free cash flow, cash flow margin, and the financing cash flow to operating cash flow ratio deserve attention. Free cash flow offers insights into a company's financial robustness and capacity to fuel growth endeavors. The cash flow margin delineates the percentage of revenue converted into cash flow, while the financing cash flow to operating cash flow ratio assesses a company's reliance on external financing for its operations, with a diminishing ratio indicating greater self-sufficiency in funding operations. These metrics play a crucial role in evaluating a company's liquidity, financial flexibility, and ability to sustainably generate future cash flows.
Overall, the sector values operating revenue growth, higher profit margins, and more efficient financing operations with free cash flow growth.
\subsection{Consumer Discretionary Sector}
\label{sec:consumer-dis}
The Consumer Discretionary Sector exhibited a median adjusted $R^2$ of 38.42\% on the LightGBM model (Table \ref{tab:Consumer Discretionary Sector-model-stats}). Top drivers of positive stock returns in the Consumer Discretionary Sector:
\selmet{Consumer Discretionary Sector}{
2 & Oil $\uparrow$ & 1 & VIX $\downarrow$ & 16 & Operating Margin $\uparrow$ \\
3 & Platinum $\uparrow$ & 5 & Federal Funds Rate $\downarrow$ & 18 & Debt Ratio $\downarrow$ \\
9 & Cotton $\uparrow$ & 6 & Housing Starts $\uparrow$ & 19 & Gross Profit Margin $\uparrow$ \\
& & 10 & CPI $\downarrow$ & 20 & EBITDA $\uparrow$ \\
& & 12 & Federal Budget Surplus or Deficit $\downarrow$ & 23 & Free Cash Flow $\uparrow$ \\
& & & & 24 & ROE $\uparrow$
}
Commodities such as oil, platinum, and cotton are important drivers in the Consumer Discretionary Sector. Oil, platinum, and cotton are essential inputs for various consumer goods and services such as apparel, electronics, and automobiles. Higher prices of these commodities may reflect increasing consumer demand.
Market volatility (VIX) remains a crucial driver, with lower volatility typically associated with positive stock returns. Heightened market volatility can signal economic uncertainty, impacting consumer sentiment and discretionary spending.
The sector is sensitive to interest rates, as indicated by the federal funds rate. Lower short-term interest rates can stimulate consumer spending by reducing borrowing costs, benefiting firms in the sector.
Housing starts stand out as a key driver with a ranking in the top 5 despite its less frequent reporting frequency. New home builds can indicate consumer confidence and economic growth. Additionally, new homes can boost demand for home-related consumer goods and services such as furniture, appliances, and home improvement products which are included in the sector.
Lower inflation is a positive driver for the sector, as measured by the Consumer Price Index (CPI). Low inflation can boost consumer purchasing power and confidence, leading to increased spending on discretionary items. Higher government spending (a declining federal budget surplus), also appears to be a positive driver for the sector as this can stimulate economic growth and consumer spending.
Key company fundamentals such as operating margin, debt ratio, gross profit margin, EBITDA, free cash flow, and return on equity (ROE) are essential indicators of a company's financial health and profitability. Maintaining healthy profit margins, low debt relative to assets, strong returns on equity, EBITDA growth, and robust free cash flow generation are crucial for consumer discretionary companies to drive shareholder value and sustain future growth. Overall, the sector values higher profitability margins, lower debt, and higher returns on equity with EBITDA and free cash flow growth.
\subsection{Consumer Staples Sector}
\label{sec:consumer-stap}
The performance of both models was relatively weak, yet still statistically significant as measured by F-stat, in the Consumer Staples Sector, with a median adjusted $R^2$ value of 19.13\% for the LightGBM model (Table \ref{tab:Communication Services Sector-model-stats}). Poor performance in the sector may be attributed to the wide array of goods and services included in the sector, which includes the production, distribution, and retail of various consumer goods. Top drivers of positive stock returns in the Consumer Staples Sector:
\selmet{Consumer Staples Sector}{
2 & Platinum $\uparrow$ & 1 & VIX $\downarrow$ & 5 & Gross Profit Margin $\uparrow$ \\
4 & Oil $\uparrow$ & 7 & 10 Year - 2 Year Treasury Yield Spread $\downarrow$ & 10 & Free Cash Flow $\uparrow$ \\
6 & Coffee $\downarrow$ & 9 & Housing Starts $\uparrow$ & 17 & Investment CF to OCF $\downarrow$ \\
8 & Palladium $\downarrow$ & 14 & Federal Funds Rate $\downarrow$ & 24 & CROA $\uparrow$ \\
12 & Wheat $\uparrow$ & & & 25 & Basic EPS $\uparrow$ \\
& & & & 26 & Operating Margin $\uparrow$
}
Many commodities are essential inputs for various consumer goods, including food, beverages, and household products. Changes in commodity prices can be suggestive of demand and impact production costs as well as profit margins for consumer staples companies.
Market volatility (VIX) remains a crucial driver, with lower volatility typically associated with positive stock returns. Heightened market volatility can signal economic uncertainty and tempered expectations of economic growth. Declining short-term interest rates (federal funds rate) can stimulate consumer borrowing and spending, benefiting the sector.
Increased housing starts can boost demand for consumer staples products. Housing starts is also an indicator of economic growth and consumer confidence, which can drive demand for consumer staples.
Higher profit margins, as indicated by gross profit margin and operating margin, are drivers of higher valuations in the sector. Companies with higher profit margins can generate more revenue from each dollar of sales, leading to expectations of higher profitability in the future. Cash flow growth as measured by free cash flow, efficient usage of cash in financing operations (financing cash flow to operating cash flow ratio), and a more efficient cash return on assets (CROA) are primary drivers. EPS growth is another key metric for the sector.
\subsection{Health Care Sector}
\label{sec:health}
Much like the Consumer Staples Sector, the Health Care Sector had relatively weak performance, however, the model was statistically significant as measured by the F-statistic. The residual plots (Figure \ref{fig:Health Care Sector-resid}) and RMSE reveal even worse performance in the sector compared to consumer staples despite similar $R^2$ values.
The LightGBM model had a median adjusted $R^2$ of 19.38\% on the LightGBM model (Table \ref{tab:Health Care Sector-model-stats}). The sector includes a wide range of companies, from pharmaceutical/biotech firms to medical device manufacturers and healthcare providers, which may be contributing to the difficulty in predicting returns. Much of the sector is potentially driven by regulatory changes, drug approvals, and healthcare policies that are not captured in the dataset. Top drivers of positive stock returns in the Health Care Sector:
\selmet{Health Care Sector}{
2 & Copper $\uparrow$ & 1 & VIX $\downarrow$ & 11 & Operating Margin $\uparrow$ \\
4 & Oil $\uparrow$ & 6 & Federal Funds Rate $\downarrow$ & 12 & Total Revenue $\uparrow$ \\
7 & Sugar $\downarrow$ & 13 & BAA - AAA Corporate Bond Yield Spread $\downarrow$ & 15 & Gross Profit Margin $\uparrow$ \\
9 & Wheat $\uparrow$ & & & 19 & Debt Ratio $\downarrow$ \\
& & & & 23 & ROA $\uparrow$
}
Commodities such as copper, oil, sugar, and wheat are drivers in the sector with copper and oil used in medical devices and equipment.
Market volatility (VIX) remains a crucial driver, with lower volatility typically associated with positive stock returns. Heightened market volatility can signal economic uncertainty and impact investor sentiment. Lower short-term interest rates and a stronger corporate environment as indicated by the declining BAA to AAA corporate bond yield spread, appear to benefit the sector.
Higher profit margins, revenue growth, lower debt, and higher returns on assets are key drivers for the sector.
\subsection{Financial Services Industry Group}
\label{sec:fin-serv}
Banking and insurance industry groups were omitted from the analysis of the Financials Sector due to a lack of data. The LightGBM model had a median adjusted $R^2$ of 41.21\% (Table \ref{tab:Financial Services Industry Group-model-stats}). Top drivers of positive stock returns in the Financial Services Industry Group:
\selmet{Financial Services Industry Group}{
2 & Oil $\uparrow$ & 1 & VIX $\downarrow$ & 16 & Free Cash Flow $\uparrow$ \\
3 & Aluminum $\uparrow$ & 5 & CPI $\downarrow$ & 17 & Operating Margin $\uparrow$ \\
4 & Cotton $\uparrow$ & 14 & Housing Starts $\uparrow$ & 22 & Capex $\uparrow$ \\
& & & & 23 & Operating Revenue $\uparrow$
}
Commodities such as oil, aluminum, and cotton are important drivers in the Financial Services Industry Group. Increases in commodity prices can indicate economic growth, resulting in higher demand for financial services.
Market volatility (VIX) remains a crucial driver, with lower volatility typically associated with positive stock returns. Heightened market volatility can signal economic uncertainty, impacting investor sentiment and financial markets. Lower inflation, as measured by the Consumer Price Index (CPI), can boost purchasing power, leading to increased spending and investment in financial services. Housing starts can indicate economic growth and consumer confidence, which can drive demand for financial services.
Free cash flow growth, higher profit margins, investments in projects (capital expenditures), and operating revenue are key drivers in the sector. Free cash flow growth is essential for financial services companies to fund operations, invest in growth opportunities, and return capital to shareholders. Higher profit margins and operating revenue indicate strong financial performance and inform anticipation of future growth. Capital expenditures reflect investments within the company, signaling potential growth, expansion, and innovation.
\subsection{Information Technology Sector}
\label{sec:info-tech}
For the Information Technology sector, the median adjusted $R^2$ was 33.67\% on the LightGBM model (Table \ref{tab:Information Technology Sector-model-stats}). Top drivers of positive stock returns in the Information Technology Sector:
\selmet{Information Technology Sector}{
1 & Oil $\uparrow$ & 6 & BAA - AAA Corporate Bond Yield Spread $\uparrow$ & 13 & Free Cash Flow $\uparrow$ \\
2 & Platinum $\uparrow$ & 12 & Housing Starts $\uparrow$ & 18 & ROE $\uparrow$ \\
9 & Aluminum $\uparrow$ & 22 & Federal Funds Rate $\downarrow$ & 19 & R\&D to Revenue $\downarrow$ \\
& & & & 20 & Operating Margin $\uparrow$ \\
& & & & 21 & Gross Profit Margin $\uparrow$
}
Commodities such as oil, platinum, and aluminum are drivers of the Information Technology sector. Rising commodity prices can indicate economic growth resulting in increased demand for technology products and services.
Intriguingly, overall market volatility (VIX) was not identified as an important driver by the LightGBM model. The sector is sensitive to interest rate changes, as indicated by the BAA to AAA corporate bond yield spread and federal funds rate. Increased housing starts can also boost demand for technology products and services as an indicator of economic growth and consumer confidence.
Key metrics encompass growth in free cash flow, return on equity (ROE), and profit margins. Notably, a reduction in research and development expenditure relative to revenue (R\&D to revenue) serves as a detrimental factor for the sector. This decline could dampen prospects for innovation and product development, critical components for technology companies to sustain a competitive advantage and growth.
\subsection{Communication Services Sector}
\label{sec:comm}
The Communications Services Sector had somewhat weak performance with an adjusted $R^2$ of 25.15\%, however, with statistical significance as measured by the F-statistic (Table \ref{tab:Communication Services Sector-model-stats}). The relatively poor performance in the sector may be attributed to the diverse range of companies within the sector, including telecommunications, advertising, broadcast, cable, and entertainment. Top drivers of positive stock returns in the Communication Services Sector:
\selmet{Communication Services Sector}{
1 & Oil $\uparrow$ & 3 & 30 Year - 15 Year Mortgage Rate Spread $\uparrow$ & 12 & Free Cash Flow $\uparrow$ \\
2 & Silver $\uparrow$ & 4 & CPI $\downarrow$ & 17 & Efficiency Ratio $\downarrow$ \\
7 & Cotton $\uparrow$ & 9 & Housing Starts $\uparrow$ & 18 & Operating Margin $\uparrow$ \\
8 & Aluminum $\uparrow$ & & & 21 & EBIT $\uparrow$ \\
& & & & 22 & Net Profit Margin $\uparrow$
}
Commodities such as oil, silver, cotton, and aluminum are drivers in the sector. Rising commodity prices can indicate economic growth, resulting in increased demand for communication services.
Interest rate changes, as indicated by the 30-year to 15-year mortgage rate spread, are drivers in the Communication Services sector. Increased housing starts can boost demand for communication services as an indicator of economic growth and consumer confidence.
Free cash flow, efficiency ratio, profit margins, and EBIT were identified as primary drivers of valuations within the sector. Free cash flow growth, more efficient operations, and higher profitability ratios with EBIT growth are expected to drive up valuations.
\subsection{Utilities Sector}
\label{sec:utilities}
The Utilities Sector model was one of the best-performing, with a median adjusted $R^2$ of 44.84\% on the LightGBM model (Table \ref{tab:Utilities Sector-model-stats}). Top drivers of positive stock returns in the Utilities Sector:
\selmet{Utilities Sector}{
11 & Wheat $\uparrow$ & 1 & VIX $\downarrow$ & 21 & Free Cash Flow $\uparrow$ \\
18 & Palladium $\downarrow$ & 3 & 3 Month Treasury Yield $\downarrow$ & 22 \& 29 & Efficiency Ratio $\downarrow$ \\
& & 9 & Real GDP $\uparrow$ & 24 & Financing CF to OCF $\downarrow$ \\
& & 10 & Federal Funds Rate $\downarrow$ & 25 & Current Ratio $\uparrow$
}
Commodities such as wheat and palladium are drivers in the Utilities sector. Rising wheat prices can indicate economic growth and increased demand for utility services. Palladium is an important input for emissions control systems, higher prices could potentially lead to higher costs for utility companies.
Market volatility (VIX) is a crucial driver, with lower volatility typically associated with positive stock returns. Heightened market volatility can signal economic uncertainty and impact investor sentiment. The sector is sensitive to interest rate changes, as indicated by the 3-month Treasury yield and the federal funds rate. Lower interest rates can reduce borrowing costs, leading to higher valuations.
Key metrics include free cash flow, efficiency ratio, financing cash flow to operating cash flow, and current ratio. A declining financing cash flow to operating cash flow ratio can indicate further self-sufficiency in financing operations and growth. Overall, the sector is driven by more efficient operations, self-sufficiency in financing, and strong liquidity positions with free cash flow growth.
\subsection{Real Estate Sector}
\label{sec:real-estate}
The Real Estate Sector had one of the best-performing models, with a median adjusted $R^2$ of 49.26\% on the LightGBM model (Table \ref{tab:Real Estate Sector-model-stats}). Top drivers of positive stock returns in the Real Estate Sector:
\selmet{Real Estate Sector}{
2 & Silver $\uparrow$ & 1 & VIX $\downarrow$ & 16 & Operating Margin $\uparrow$ \\
10 & Oil $\uparrow$ & 3 & Unemployment Rate $\downarrow$ & 17 & Capex $\downarrow$ \\
& & 4 & Federal Funds Rate $\downarrow$ & 18 & Total Revenue $\uparrow$ \\
& & 6 & CPI $\downarrow$ & 20 \& 29 & Gross Profit Margin $\uparrow$ \\
& & & & 22 & EBITDA $\uparrow$
}
Commodities such as silver and oil are drivers in the sector. Rising commodity prices can indicate economic growth resulting in increased confidence and demand for real estate properties, driving up prices.
Market volatility (VIX) is a crucial driver, with lower volatility typically associated with positive stock returns. Heightened market volatility can signal economic uncertainty and impact investor sentiment in real estate. The sector is sensitive to interest rate changes, as indicated by the federal funds rate. Lower short-term interest rates can reduce borrowing costs for real estate companies, leading to higher valuations. Low inflation and a declining unemployment rate are positive drivers for the sector, as they indicate a healthy economic environment with increased consumer purchasing power and confidence.
Key financial metrics include profitability ratios, capital expenditures, revenue, and EBITDA. Capital expenditures are indicative of investments in growth opportunities and expansion. Overall, the sector values higher profitability margins, revenue growth, investments in projects, and EBITDA growth.
\section{Robustness}
\label{sec:robust}
\newcommand{\tabmstats}[5]{
\begin{table}[H]
\centering
\begin{minipage}{\textwidth}
\centering
\caption{#1: Model Statistics}
\label{tab:#1-model-stats}
\begin{subtable}{0.695\linewidth}
\centering
\caption{Linear: Cross-Validated}
#2
\end{subtable}
\hfill
\begin{subtable}{0.255\linewidth}
\centering
\caption{Linear: Full Dataset}
#3
\end{subtable}
\\[2ex]
\centering
\begin{subtable}{0.695\linewidth}
\centering
\caption{LGBM: Cross-Validated}
#4
\end{subtable}
\hfill
\begin{subtable}{0.255\linewidth}
\centering
\caption{LGBM: Full Dataset}
#5
\end{subtable}
\footnotetext{Refer to Appendix \ref{appendix:#1} for additional model figures and tables. \vspace{-\baselineskip}}
\end{minipage}
\end{table}
}
The LightGBM model outperformed a linear model in all sectors on an adjusted $R^2$ basis, averaging a difference of 12.27\%. The LightGBM model's ability to capture non-linear relationships through gradient-boosted decision trees was more effective in elucidating the variance in stock returns overall compared to a linear approach. Additionally, the LightGBM model consistently outperformed the linear model across all sectors in MSE and MAE, revealing better predictive accuracy and lower error rates. The F-statistic demonstrated higher values for the LightGBM model across 9 out of 11 sectors. Furthermore, all models exhibited statistical significance, as evidenced by F-stat p-values nearing 0 in all cases. However, it should be noted that both models regularly suffered from heteroscedasticity and non-normality of residuals (see the \nameref{sec:limitations} section).
\tabmstats{Energy Sector}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lccccc@{}}
\toprule
& Median & Mean & STD & Min & Max \\ \midrule
$R^2$ & 40.05\% & 37.95\% & 4.09\% & 29.93\% & 40.88\% \\
Adjusted $R^2$ & 39.81\% & 37.70\% & 4.11\% & 29.65\% & 40.65\% \\
MSE & 1.56\% & 1.55\% & 0.13\% & 1.37\% & 1.74\% \\
RMSE & 12.50\% & 12.45\% & 0.51\% & 11.72\% & 13.18\% \\
MAE & 8.92\% & 8.90\% & 0.21\% & 8.60\% & 9.24\% \\
F-stat & 151.04 & 157.21 & 19.20 & 136.26 & 190.99 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lc@{}}
\toprule
& Score \\ \midrule
$R^2$ & 38.88\% \\
Adjusted $R^2$ & 38.68\% \\
MSE & 1.53\% \\
RMSE & 12.39\% \\
MAE & 8.86\% \\
F-stat & 196.92 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lccccc@{}}
\toprule
& Median & Mean & STD & Min & Max \\ \midrule
$R^2$ & 51.01\% & 50.20\% & 2.68\% & 45.01\% & 52.41\% \\
Adjusted $R^2$ & 50.74\% & 49.93\% & 2.69\% & 44.71\% & 52.14\% \\
MSE & 1.23\% & 1.24\% & 0.08\% & 1.13\% & 1.34\% \\
RMSE & 11.08\% & 11.15\% & 0.35\% & 10.62\% & 11.59\% \\
MAE & 7.84\% & 7.82\% & 0.16\% & 7.55\% & 8.04\% \\
F-stat & 161.79 & 164.64 & 12.54 & 149.71 & 187.10 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lc@{}}
\toprule
& Score \\ \midrule
$R^2$ & 55.70\% \\
Adjusted $R^2$ & 55.50\% \\
MSE & 1.11\% \\
RMSE & 10.55\% \\
MAE & 7.37\% \\
F-stat & 234.70 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}
The LightGBM model outperformed the linear model by 10.93\% in the Energy Sector, with median adjusted $R^2$ values of 50.74\% and 39.81\%, respectively. Additionally, the LightGBM model surpassed the linear regression model across the other metrics including MSE, RMSE, MAE, and F-stat.
\tabmstats{Materials Sector}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 27.62\% & 27.42\% & 0.53\% & 26.51\% & 28.09\% \\
Adjusted $R^2$ & 27.39\% & 27.18\% & 0.53\% & 26.27\% & 27.86\% \\
MSE & 0.83\% & 0.84\% & 0.06\% & 0.74\% & 0.94\% \\
RMSE & 9.09\% & 9.14\% & 0.35\% & 8.62\% & 9.72\% \\
MAE & 6.58\% & 6.60\% & 0.20\% & 6.27\% & 6.87\% \\
F-stat & 117.95 & 118.28 & 9.03 & 108.10 & 131.85 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 27.81\% \\
Adjusted $R^2$ & 27.62\% \\
MSE & 0.83\% \\
RMSE & 9.12\% \\
MAE & 6.58\% \\
F-stat & 147.62 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 36.35\% & 36.77\% & 1.12\% & 35.60\% & 38.86\% \\
Adjusted $R^2$ & 35.97\% & 36.39\% & 1.13\% & 35.22\% & 38.50\% \\
MSE & 0.73\% & 0.73\% & 0.06\% & 0.63\% & 0.83\% \\
RMSE & 8.52\% & 8.53\% & 0.37\% & 7.95\% & 9.09\% \\
MAE & 6.12\% & 6.10\% & 0.21\% & 5.74\% & 6.36\% \\
F-stat & 97.11 & 96.36 & 7.73 & 86.60 & 109.59 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 53.80\% \\
Adjusted $R^2$ & 53.58\% \\
MSE & 0.53\% \\
RMSE & 7.30\% \\
MAE & 5.15\% \\
F-stat & 180.76 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}
The LightGBM model outperformed the linear model by 8.58\% in the Materials Sector, with median adjusted $R^2$ values of 35.97\% and 27.39\%, respectively. It also outperformed across all other metrics except F-stat.
\tabmstats{Industrials Sector}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 26.96\% & 25.89\% & 1.63\% & 23.30\% & 27.49\% \\
Adjusted $R^2$ & 26.86\% & 25.78\% & 1.63\% & 23.19\% & 27.39\% \\
MSE & 0.72\% & 0.73\% & 0.02\% & 0.71\% & 0.76\% \\
RMSE & 8.50\% & 8.55\% & 0.11\% & 8.43\% & 8.72\% \\
MAE & 6.11\% & 6.14\% & 0.07\% & 6.07\% & 6.23\% \\
F-stat & 247.12 & 250.87 & 6.54 & 244.06 & 261.42 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 26.25\% \\
Adjusted $R^2$ & 26.17\% \\
MSE & 0.73\% \\
RMSE & 8.53\% \\
MAE & 6.13\% \\
F-stat & 314.24 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 38.22\% & 37.86\% & 1.43\% & 35.54\% & 39.69\% \\
Adjusted $R^2$ & 38.10\% & 37.74\% & 1.43\% & 35.41\% & 39.57\% \\
MSE & 0.61\% & 0.61\% & 0.02\% & 0.60\% & 0.63\% \\
RMSE & 7.79\% & 7.83\% & 0.10\% & 7.73\% & 7.95\% \\
MAE & 5.43\% & 5.48\% & 0.08\% & 5.41\% & 5.62\% \\
F-stat & 293.52 & 292.22 & 4.79 & 285.07 & 298.84 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 48.91\% \\
Adjusted $R^2$ & 48.83\% \\
MSE & 0.50\% \\
RMSE & 7.10\% \\
MAE & 5.04\% \\
F-stat & 462.92 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}
The LightGBM model outperformed the linear model by 11.24\% in the Industrials Sector, with median adjusted $R^2$ values of 38.10\% and 26.86\%, respectively, outperforming across all other metrics as well.
\tabmstats{Consumer Discretionary Sector}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 25.46\% & 24.53\% & 1.17\% & 22.91\% & 25.50\% \\
Adjusted $R^2$ & 25.34\% & 24.41\% & 1.17\% & 22.79\% & 25.39\% \\
MSE & 1.01\% & 1.02\% & 0.02\% & 1.00\% & 1.06\% \\
RMSE & 10.06\% & 10.12\% & 0.10\% & 10.02\% & 10.27\% \\
MAE & 7.44\% & 7.47\% & 0.06\% & 7.39\% & 7.58\% \\
F-stat & 209.13 & 210.96 & 6.21 & 204.74 & 222.92 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 24.80\% \\
Adjusted $R^2$ & 24.71\% \\
MSE & 1.02\% \\
RMSE & 10.10\% \\
MAE & 7.45\% \\
F-stat & 264.32 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 38.55\% & 38.16\% & 1.12\% & 36.20\% & 39.20\% \\
Adjusted $R^2$ & 38.42\% & 38.04\% & 1.12\% & 36.07\% & 39.08\% \\
MSE & 0.83\% & 0.84\% & 0.02\% & 0.83\% & 0.88\% \\
RMSE & 9.10\% & 9.16\% & 0.11\% & 9.08\% & 9.37\% \\
MAE & 6.65\% & 6.67\% & 0.10\% & 6.57\% & 6.85\% \\
F-stat & 291.83 & 293.51 & 3.29 & 291.25 & 300.00 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 49.63\% \\
Adjusted $R^2$ & 49.55\% \\
MSE & 0.68\% \\
RMSE & 8.27\% \\
MAE & 6.08\% \\
F-stat & 473.23 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}
The LightGBM model for the Consumer Discretionary Sector outperformed the linear model across all metrics. Outperforming on an adjusted $R^2$ basis by 13.08\%, with median adjusted $R^2$ values of 38.42\% and 25.34\%, respectively.
\tabmstats{Consumer Staples Sector}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 11.40\% & 11.19\% & 2.27\% & 8.87\% & 15.15\% \\
Adjusted $R^2$ & 11.11\% & 10.89\% & 2.28\% & 8.56\% & 14.87\% \\
MSE & 0.67\% & 0.66\% & 0.03\% & 0.61\% & 0.71\% \\
RMSE & 8.19\% & 8.15\% & 0.20\% & 7.82\% & 8.41\% \\
MAE & 5.90\% & 5.87\% & 0.10\% & 5.71\% & 5.98\% \\
F-stat & 39.91 & 39.96 & 2.10 & 37.42 & 43.73 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 11.93\% \\
Adjusted $R^2$ & 11.70\% \\
MSE & 0.66\% \\
RMSE & 8.12\% \\
MAE & 5.85\% \\
F-stat & 49.52 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 19.55\% & 18.67\% & 2.14\% & 15.81\% & 21.29\% \\
Adjusted $R^2$ & 19.13\% & 18.24\% & 2.15\% & 15.37\% & 20.87\% \\
MSE & 0.61\% & 0.61\% & 0.03\% & 0.57\% & 0.65\% \\
RMSE & 7.83\% & 7.80\% & 0.18\% & 7.53\% & 8.09\% \\
MAE & 5.61\% & 5.59\% & 0.07\% & 5.50\% & 5.69\% \\
F-stat & 45.72 & 45.00 & 3.71 & 39.32 & 49.96 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 28.10\% \\
Adjusted $R^2$ & 27.80\% \\
MSE & 0.54\% \\
RMSE & 7.34\% \\
MAE & 5.28\% \\
F-stat & 64.92 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}
The LightGBM model outperformed the linear model by 8.24\% in the Consumer Staples Sector, with median adjusted $R^2$ values of 19.13\% and 10.89\%, respectively.
\tabmstats{Health Care Sector}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 11.24\% & 10.51\% & 1.21\% & 9.00\% & 11.91\% \\
Adjusted $R^2$ & 11.09\% & 10.35\% & 1.22\% & 8.85\% & 11.76\% \\
MSE & 1.13\% & 1.13\% & 0.03\% & 1.09\% & 1.17\% \\
RMSE & 10.62\% & 10.62\% & 0.14\% & 10.44\% & 10.83\% \\
MAE & 7.49\% & 7.51\% & 0.09\% & 7.38\% & 7.65\% \\
F-stat & 72.00 & 70.33 & 3.89 & 64.09 & 74.19 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 10.85\% \\
Adjusted $R^2$ & 10.73\% \\
MSE & 1.12\% \\
RMSE & 10.61\% \\
MAE & 7.49\% \\
F-stat & 87.08 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 19.59\% & 19.95\% & 1.34\% & 17.92\% & 21.95\% \\
Adjusted $R^2$ & 19.38\% & 19.74\% & 1.34\% & 17.71\% & 21.75\% \\
MSE & 1.00\% & 1.01\% & 0.04\% & 0.96\% & 1.07\% \\
RMSE & 9.98\% & 10.05\% & 0.19\% & 9.79\% & 10.35\% \\
MAE & 6.98\% & 7.03\% & 0.11\% & 6.93\% & 7.23\% \\
F-stat & 82.73 & 81.19 & 3.70 & 75.31 & 85.42 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 23.37\% \\
Adjusted $R^2$ & 23.22\% \\
MSE & 0.97\% \\
RMSE & 9.83\% \\
MAE & 6.88\% \\
F-stat & 107.22 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}
}
The LightGBM model outperformed the linear model by 8.02\% in the Health Care Sector, with median adjusted $R^2$ values of 19.13\% and 11.11\%, respectively.
\tabmstats{Financial Services Industry Group}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lrrrrr@{}}
\toprule
& \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Mean} & \multicolumn{1}{c}{STD} & \multicolumn{1}{c}{Min} & \multicolumn{1}{c}{Max} \\ \midrule
$R^2$ & 27.92\% & 28.13\% & 1.49\% & 25.72\% & 29.89\% \\
Adjusted $R^2$ & 27.61\% & 27.83\% & 1.49\% & 25.41\% & 29.60\% \\
MSE & 0.63\% & 0.63\% & 0.05\% & 0.56\% & 0.70\% \\
RMSE & 7.93\% & 7.95\% & 0.31\% & 7.51\% & 8.36\% \\
MAE & 5.86\% & 5.87\% & 0.17\% & 5.63\% & 6.12\% \\
F-stat & 97.43 & 96.08 & 6.86 & 87.80 & 106.23 \\
P\textgreater{}F & 0.0000 & 0.0000 & 0.0000 & 0.0000 & 0.0000 \\ \bottomrule
\end{tabular}%
}
}{
\resizebox{\textwidth}{!}{%
\begin{tabular}{@{}lr@{}}
\toprule
& \multicolumn{1}{c}{Score} \\ \midrule
$R^2$ & 28.79\% \\
Adjusted $R^2$ & 28.55\% \\
MSE & 0.63\% \\
RMSE & 7.93\% \\
MAE & 5.85\% \\
F-stat & 120.01 \\
P\textgreater{}F & 0.0000 \\ \bottomrule
\end{tabular}%
}