Safeguarding Downside Risk in Portfolio Insurance: Navigating Swiss

Stock Market Regimes with Options, Trading Signals, and Financial

Products

Sylvestre Blanc

, Emmanuel Fragnière

, Francesc Naya

and Nils Tuchschmid

Grammont Finance, Derivatives and Portfolio Management, Switzerland

HES-SO Valais-Wallis, Business School, ITO, Sierre, Switzerland

HES-SO School of Management Fribourg, Fribourg, Switzerland

Keywords: Backward-Looking Signals, Financial Safety, Forward-Looking Signals, Portfolio Insurance, Options,

Portfolio Optimization.

Abstract: Our research uses options to safeguard equity portfolios from downside risk. Despite the cost challenges of

passive put protection, we explore leveraging diverse market signals, backward and forward-looking ones, to

enhance portfolio risk-return balance while maintaining acceptable safeguards. These signals aid in selecting

underlying assets for option positions, aiming to achieve protection while minimizing put premium

expenditure. Certain signals, like "trend" or "low volatility”, either empirical or implied, demonstrate added

value, although their effectiveness depends on market conditions (or regimes). We also evaluate whether a

set of trading rules can enhance the efficiency of such strategies. Our study highlights the importance of

financial product safety, akin to safety measures for industrial products. By doing so, we underline the

importance of portfolio insurance in finance. Further developments will aim at implementing a trading system

that offers greater adaptability to different market regimes, for example high volatility phases, and under real

market conditions.

1 INTRODUCTION

Options strategies, when implemented and

incorporated correctly into traditional equity

portfolios, can be a powerful tool to modify the risk

and return profiles of such portfolios, allowing

investors to express more accurately their investment

views, risk tolerance, and return objectives or

mandates. Options expand the universe of

opportunities available, that would be rather limited

in equity-only portfolios. For instance, a simple

covered call strategy can serve as a yield

enhancement instrument that allows the investor to

achieve a targeted return whereas capital remains at

risk. Likewise, a plain vanilla protective put limits the

downside risk of an investment and can serve as a tool

to efficiently protect the portfolio. However, these

strategies are not easily and efficiently implemented.

https://orcid.org/0000-0001-9628-4543

https://orcid.org/0000-0001-8773-3386

https://orcid.org/0000-0002-2046-0376

It is widely known that a “passive” plain vanilla

protective put strategy is too costly in terms of option

premia, and the resulting drag in long-term

performance likely does not compensate for the

protection during drawdown periods, worsening the

portfolio’s long-term risk-adjusted performance

(Ilmanen et al., 2021). In this article, we test whether

the introduction of an actively managed portfolio of

long put options into an equity portfolio can improve

the combined portfolio’s risk-adjusted performance.

This active management is performed by exploiting a

set of backward-looking trading signals coming from

the equities space (momentum, trend, or empirical

volatility) and forward-looking ones from the options

themselves (implied volatility or skew) that help in

selecting the right underlying for which to take option

positions.

The backward-looking signals from the equities

space have been widely tested and implemented as

Blanc, S., Fragnière, E., Naya, F. and Tuchschmid, N.

Safeguarding Downside Risk in Portfolio Insurance: Navigating Swiss Stock Market Regimes with Options, Trading Signals, and Financial Products.

DOI: 10.5220/0012524200003717

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 6th International Conference on Finance, Economics, Management and IT Business (FEMIB 2024), pages 33-41

ISBN: 978-989-758-695-8; ISSN: 2184-5891

trading strategies on equity portfolios across different

periods and geographies, both in long-only format

(e.g., smart-beta funds) or long-short (alternative beta

or alternative risk premia funds). Jegadeesh & Titman

(1993) were the first to formally test the “momentum”

factor, by showing that portfolios that buy past

winners and sell past losers generate positive returns

over the next 3 to 12 months, which was not explained

by their systematic risk exposures. Assness,

Moskowitz and Pedersen (2013) found this

momentum factor to be relevant not only in equities

but across asset classes and across geographies. A

similar strategy is the popular “Trend Following”

strategy, which is also exploited by practitioners in

various asset classes. While the momentum signal is

a relative one (ranking-based), trend (sometimes

named time-series momentum) is an absolute signal.

Incorporating trend strategies in equities portfolios is

desirable not only because it is expected to generate

abnormal returns over the long term, but also due to

its convex return profile, which helps at mitigating the

impact of equity market drawdowns. This

phenomenon was studied by Moskowitz, Ooi and

Pedersen (2012) and more recently by Dao et al.

(2016), Hurst et al. (2017), Babu et al. (2020), AQR

(2022) and Co (2023). The third backward-looking

signal used in this study is low empirical volatility.

Frazzini and Pedersen (2014) popularized the concept

of the Low Beta strategy, which puzzles traditional

finance theory by showing that lower beta stocks

consistently outperform higher beta stocks. Blitz and

Vidojevic (2017) confirmed the low-risk anomaly,

but found the mispricing of low volatility stocks to be

stronger than the one of low beta stocks.

Forward-looking signals that exploit information

from the options markets, such as implied volatility

and skew (or smile), have been less explored. The

motivation to also examine these signals is that they

represent investors’ expectations on future risk or

market movements. Baltussen (2012) found that

information extracted from options market in US

large-cap stocks could be used to build trading

strategies that outperform the benchmark, after

controlling for other known risk factors. Our use of

forward-looking signals was inspired by this article.

The relevance of these backward-looking and

forward-looking signals in the Swiss equity market,

which is the one used in this article, has been studied

by the authors. We aim to assess whether these

signals, which appear to be valuable when

constructing long-short equity portfolios, can also add

value when selecting the underlying for a long-put

position. In addition, we experiment with a set of

trading rules that could intuitively help to improve the

efficiency of the strategy, explained in more detail in

the following sections.

These trading strategies are implemented here into

two base equity portfolios: an equally weighted (EW)

and a global minimum variance (GMV) portfolio.

Results are evaluated using two different

approaches. First, we apply a series of out-of-sample

historical backtests, that help not only at testing

whether a strategy would have worked in the past or

not but are especially useful in understanding the

behavior of the strategies in different market regimes.

Second, we test the strategies over 500 bootstrapped

periods. This second method provides an estimate of

the return distribution of such strategies.

This paper is organized as follows. Section 2

provides a brief literature review. Section 3 presents

our dataset and the trading signals used select the

underlyings in which put option positions are taken.

Section 4 shows the bench test we have created to

apply our different trading strategies, specifying the

constraints and/or additional rules applied in each

case. Sections 5 and 6 show the results of the

backtests and bootstraps respectively, Section 7

concludes and provides direction for further research.

2 LITERATURE REVIEW

The landscape of portfolio risk management includes

various strategies and approaches to protect downside

risk and improve risk-adjusted performance. In this

literature review, we briefly review the key studies

that contribute to the understanding of portfolio

insurance such as the traditional passive put

protection and other alternative risk mitigating

strategies. While passive put options have their

merits, they also come with limitations. The cost

associated with maintaining put options can erode

portfolio returns over time. Moreover, the

effectiveness of passive puts in various market

conditions may be limited. Researchers have begun to

explore alternative strategies to mitigate these

limitations. Ilmanen et al. (2021) conducted a study

comparing a passive long-put strategy with a

long/short trend strategy. Their research highlights

the trade-off between cost and efficiency in protecting

an equity portfolio, shedding light on the efficacy of

passive put options in risk management but

concluding that trend is preferable, as the long-term

cost of a passive long put is simply too high.

Moreover, Israelov and Nielsen (2016), in their

work on portfolio protection in calm markets, shed

light on practical applications of portfolio insurance

strategies. Other studies, such as those by Boulier and

FEMIB 2024 - 6th International Conference on Finance, Economics, Management and IT Business

Kanniganti (2005), Annaert et al. (2009), and

Figlewski et al. (1993), have evaluated the

performance of protective put strategies, providing

valuable insights into their effectiveness. Lhabitant

(1998) highlighted the potential of enhancing

portfolio performance using option strategies,

emphasizing the feasibility of outperforming the

market with the right approach.

Incorporating options market signals into

portfolio management has also been explored by

Kostakis, Panigirtzoglou, and Skiadopoulos (2011),

who focused on market timing with option-implied

distributions, and Harper and Sarkar (2019), who

examined option market signals and the disposition

effect around equity earnings announcements. Dew-

Becker, Giglio, and Kelly (2017) and Mohanty (2018)

explored investors' perceptions of risks and forward-

looking indicators in portfolio allocation, offering

further avenues for research and practical application.

Finally, other authors focused on the impact of

different asset allocation techniques to portfolio

performance and risk. DeMiguel et al. (2009)

contributed to the discussion on portfolio allocation

methods, highlighting the inefficiencies of optimized

portfolios compared to the 1/N portfolio strategy.

Maillard, Roncalli and Teiletche (2009) proposed the

risk parity or Equal Risk Contribution allocation, with

portfolio volatility as the standard risk measure. They

showed analytically that resulting weights are

between the ones from the EW and GMV solutions.

Jurczenko and Teiletche (2015) extended the ERC

version to a portfolio tail-risk measure: the expected

shortfall (conditional VaR or CVaR). Rockafellar and

Uryasev (2000) first presented the approach to

optimize portfolios by minimizing their expected

shortfall. Brodie et al. (2009) and more recently

Kremer et al. (2020) emphasized the imposition of

constraints or penalties on the classic mean-variance

optimization problem in order to reduce the impact of

estimation errors and achieve robust out-of-sample

optimal portfolios.

In conclusion, the literature on portfolio and risk

management encompasses a wide range of strategies

and approaches, including allocation methods, option

strategies, equity market signals, and forward-

looking indicators from options markets. These

studies provide valuable insights into the trade-offs,

limitations, and potential enhancements associated

with different portfolio protection and management

techniques, informing our research's combined

approach in employing these strategies.

3 DATA EXPERIMENTS AND

TRADING SIGNALS REVIEW

The universe is composed of 24 large-cap stocks from

the Swiss market that belong, or belonged at some

point during the sample period, to the Swiss Market

Index (SMI Index). The sample period spans from

March 24, 2006, to January 7, 2022. For each of the

24 components, we have the time series of daily stock

values and a complete daily dataset of the options

available at each date, with different strikes and

maturities (corresponding to millions of data points).

Using an internal model from the commercial partner

(Grammont), we can also extract, for every option, an

implied volatility value and a skew (or smile) value.

This dataset allows us to estimate accurately the price

of any option at any point in time during the sample

period, for any combination of moneyness and

maturity. Using the options dataset, we apply

interpolation methods to create a daily time series of

implied volatility and skew values for 12-month

maturity ATM options, which will be used to

calculate the forward-looking signals. Smaller

capitalization stocks are not considered due to the

lack of availability in the options dataset. Options for

most smaller cap underlyings are either inexistent or

highly illiquid and thus it is not possible to construct

reliable time series for these securities, neither to

construct systematic trading strategies. A potential

concern with our dataset is that the 24 components

have data available until the end of the sample period.

As a rule, if there was a large-cap stock that stopped

trading during the sample period, it does not appear

in the sample. However, and to the best of our

knowledge, only Transocean (RIGN), which was

delisted from the SIX in March 2016, is not included.

Moreover, since the goal is to compare the relative

performance among strategies and portfolios, we

believe that the survivorship bias is mitigated, as it

would affect, even though not to the same extent, all

the strategies tested. The dataset includes the same

options data for the SMI Index. For a proxy of the

risk-free interest rate, we use the overnight SARON.

3.1 Backward-Looking Signals

The first group of signals takes information from

historical stocks’ data and is thus “backward-

looking”. It includes trend, momentum and low

empirical volatility.

Trend (TREND): A given underlying shows a

positive signal if the return over the most recent 12-

months, 6-months, and 3-months is positive.

Likewise, it shows a negative signal if the return over

Safeguarding Downside Risk in Portfolio Insurance: Navigating Swiss Stock Market Regimes with Options, Trading Signals, and Financial

Products

the same 3 periods is negative. Note that the sign of

the return must be the same for the three periods. If

the sign is different for at least one of the periods, then

the signal is neutral or, in other words, there is no

trend signal. At any rebalancing date, all underlying

can have positive, negative, or neutral signals.

Momentum (MOM): At each rebalancing date,

the signal is positive for the best-performing 25%

stocks and negative for the worst-performing 25%

stocks. The performance measure to rank the stocks

is the most recent 12-month return. As opposed to

trend, momentum is a cross-sectional signal. Thus,

even if a stock shows a negative return, it has a

positive signal if it falls in the best quartile. In the

same vein, stocks with positive returns can show a

negative signal. This implies that at every rebalancing

date in which the number of stocks is the same (e.g.,

24 underlyings), there are always the same number of

underlyings with positive and negative signals (e.g.,

6 underlyings).

Historical Volatility (HVOL): Like momentum,

at each rebalancing date this signal is positive for the

25% underlyings with the lowest empirical volatility

and negative for the 25% underlyings with the highest

empirical volatility. Volatility is calculated using the

most recent 12-month period.

3.2 Forward-Looking Signals

The second group is made of signals that use data

coming from the options prices, and thus are

understood to be “forward-looking”. These are

implied volatility and skew, both cross-sectional and

in time-series format. The signals are calculated using

our daily time-series dataset of estimated implied

volatility and skew for a 12-month ATM option. They

are seen as forward-looking because they express

investors’ expectations on future risk and market

variation (e.g., implied volatility can be interpreted as

the investors’ expected future market volatility).

Implied Volatility (IVOL): At each rebalancing

date, the underlyings are ranked by their implied

volatility value (using the implied volatility of the

hypothetical ATM 12-month maturity option) and the

signal is positive for the first quartile and negative for

the fourth one. This signal is expected to yield similar

results as the empirical volatility signal since

underlyings with high (low) empirical volatility tend

to exhibit high (low) implied volatility in their option

prices.

Option’s Skew (SKEW): At each rebalancing

date, the underlyings are ranked by their skew value

(using the skew of the hypothetical ATM 12-month

maturity option) and the signal is positive for the first

quartile and negative for the fourth one. A high skew

happens when the implied volatility of OTM options

is larger than the implied volatility of ATM options

for the same expiries and underlyings. This means

that investors demand more OTM options concerning

ATM ones and are willing to pay more for the former

ones, or that they expect large variations to be more

frequent than expected by an options pricing model,

such as Black-Scholes.

Implied Volatility Spike (IVOL-Spike): At each

rebalancing date, the signal is negative for an

underlying whose implied volatility on the previous

day, extracted from our time-series data, is at least

one standard deviation larger than its mean, taking the

previous one-year daily implied volatility values. The

signal is neutral (or no signal) otherwise.

Skew Spike (SKEW-Spike): At each rebalancing

date, the signal is negative for an underlying whose

skew on the previous day, extracted from our time-

series data, is at least one standard deviation larger

than its mean, taking the previous one-year daily

skew values. The signal is neutral (or no signal)

otherwise.

ALL: For comparison purposes, we will include

a portfolio that takes option positions on all stocks

available, regardless of signals.

4 PORTFOLIOS’

CONSTRUCTION PROCESS

We test the relevance of these signals on active long-

put strategies implemented into two base equity

portfolios. an equally weighted (EW) and a global

minimum variance (GMV) portfolio, rebalanced

every 6 months. We also tested the Equal Risk

Contribution (ERC) and CVaR minimization (CVaR-

min) allocation methods. While ERC showed results

in between EW and GMV, results from CVaR-min

portfolios were almost the same as GMV. To comply

with the article’s format requirements and size limits,

only EW and GMV results are presented in the paper.

Also, we test for a set of trading rules that intuitively

could improve the portfolios’ performance, detailed

below.

To determine the performance of the strategies

and the relevance of the signals and trading rules, we

perform historical backtests as well as tests on 500

one-year bootstrapped periods.

The portfolios are constructed as follows. We

assume an investor with CHF 100 million to allocate

at time t=0, which is April 4, 2007, for the historical

FEMIB 2024 - 6th International Conference on Finance, Economics, Management and IT Business

backtests, and the first date (randomly selected) of

each bootstrapped period. This capital can be used to

purchase stocks or put options.

In the portfolios where no option positions are

taken, the portfolio will be fully invested in equities.

To find the weights in GMV case, the parameters,

namely the covariance matrix, are estimated using the

underlyings’ prior 12-month returns. Also, we set the

constraint of maximum weight to an individual name

to 20% and, for all stocks whose weight in the

optimization is lower than 0.2%, we set their weights

to be zero and redistribute them proportionally across

the remaining components in the portfolio. The

weights are re-calculated on every rebalancing date

(i.e. every 6 months, for both the backtests and the

bootstrap periods).

For the strategies that are long put options (with

the exception of the “SMI” strategy, that is long puts

on the SMI index and thus acting as a portfolio

“macro hedge”, explained in more detail below), a

target budget of 2.5% of portfolio value to be spent

on option premia on every rebalancing is set.

However, for the “absolute value signals” (i.e. not

relative ranking), such as trend, it is desirable to allow

a varying budget that is a function of the number of

underlyings with a signal, while keeping a limit.

Therefore, the methodology applied is as follows: at

t=0 and at each rebalancing date, a global budget is

set equal to 10% of the portfolio value (e.g., CHF 10

mio. at t=0). Then, this global budget is divided

equally among all the underlyings available (even

those with no weight in the GMV portfolio), which

represents the individual budget that will be spent to

purchase put options of that underlying if it shows a

negative signal. Note that, for relative (ranked)

signals, in which a 25% of underlyings show a signal

at each rebalancing date, a fixed 2.5% of portfolio

value is spent on each rebalancing (in reality, this

value ranges between 2.07% and 3.37%, due to

rounding the 25% to the closest number of

underlyings available, yet in most cases is between

2.3% and 2.8%). For “absolute signals”, the range is

between 0% if no underlying shows a signal and 10%

if all underlyings show a signal. On the historical

backtests, on average is spent 2.44% of portfolio

value at each rebalancing period for the trend signal,

2.10% for the IVOL-TS signal and 2.90% for the

SKEW-TS signal, close to the 2.50% target.

The strike is set at 90% of the spot price (OTM)

and maturity is always 6 months, coinciding with the

rebalancing period. The individual budget divided by

each estimated option premium sets the number of put

options that are purchased at each rebalancing date.

Note that, by using this method, the combination of

equities and put positions can be partially protective,

fully protective or even speculative (negative delta),

as the number of puts depends on the budget and the

premium, and not on the number of stocks in the

portfolio from each underlying.

The remaining capital available (portfolio value

minus the amount spent in option premia) is used to

purchase stocks.

The following trading strategies are simulated and

compared:

BASE: Base portfolio. It is an equity-only

portfolio, with no options. It serves as the benchmark.

OPT: At t=0 and each rebalancing date, it buys

put options on individual names that show a negative

signal.

LEV: It is constructed as the OPT strategy but

with 20% leverage. Thus, the initial invested capital

is CHF 120 million. It assumes a leverage cost of

SARON + 40 bps, paid at each rebalancing period.

The idea is that the risk reduction obtained from the

options’ protection can be used to leverage the equity

portfolio and its long-term return.

TR1: It is built as the OPT strategy. Yet, during

the investment period (i.e., between one rebalancing

date and the next one), an additional trading rule is

added (TR1): If an option’s delta reaches a pre-

defined trigger (-0.9 or less), the put that was initially

bought is sold to close the position at a profit. The aim

of this rule is that, in the case of a short-term upside

reversal, the profit that is made in the options’

positions is not lost with the reversal.

TR2: It is a variation of TR1: once an option’s

delta reaches the predefined trigger, if the negative

signal is still present in the underlying, it won’t close

the positions. If the signal is positive or neutral, it will

close the positions as in TR1. This trading rule is

implemented in the backtests only.

TR3: It is built as the OPT strategy, in which put

options are bought on stocks that show a negative

signal, but with an additional rule: if the implied

volatility of any option is below 10% (i.e., low

implied volatility), it will purchase the put options,

regardless of whether the underlying shows negative

signal or not. Likewise, if the implied volatility is

above 30% (i.e., high implied volatility), it won’t

purchase the put options even if they show a negative

signal. This additional rule buys options when they

are cheap and avoids them when they are expensive,

regardless of the signal.

WEI.: This strategy does not take option

positions. Instead, it sets the weights of the

underlyings with a negative signal at zero and

redistributes these weights proportionally across the

remaining components in the portfolio.

Safeguarding Downside Risk in Portfolio Insurance: Navigating Swiss Stock Market Regimes with Options, Trading Signals, and Financial

Products

SMI: This portfolio takes option positions on the

SMI index, rather than on individual underlyings. To

calculate the number of options, at t=0 and each

rebalancing date it calculates the market beta of each

equity portfolio (EW, GMV) and purchases the

number of options proportional to each portfolio’s

beta. It also serves as a benchmark to compare

whether taking positions on individuals is more

effective than simply taking option positions on the

index.

5 HISTORICAL BACKTESTED

RESULTS

As mentioned previously, the historical backtests

begin on April 4, 2007, and are rebalanced every 6

months, coinciding with the options’ expiry dates.

Summary results are presented in Table 1 for the EW

allocation and Table 2 for GMV. These results

represent the average annualized return, the worst 6-

month period, and the ratio of average return over the

worst 6-month return. First, it is noticeable that all

GMV portfolios outperform EW ones, with or

without options. They show both higher average

return and a smaller loss in the worst 6-month period.

Focusing on the signals, trend appears to be the one

that provides the best results, largely outperforming

the Base portfolio regardless of the allocation method

or trading rule, except for the TR3 case. This is due

to the large profit of the put options during the 2008

GFC period and 2011 (EU sovereign debt crisis).

Another signal that outperforms the Base in both EW

and GMV, both at the simple portfolio with options

(OPT) and in the case of TR1, TR2, and LEV is the

IVOL-spike signal. IVOL, SKEW, and SKEW-Spike

signals outperform in the GMV case, but their results

are more mixed on EW portfolios.

Finally, momentum (MOM) and historical

volatility (HVOL) do not add value, except for the

WEI. strategy which does not take any option

positions. Overall, the added value of the trading rules

TR1, TR2, and TR3 is negative in most cases, and,

when positive, its effect is very minor. Interestingly,

the passive put strategy that takes options’ positions

on all underlyings (OPT portfolio with Signal 1 or

ALL) drops the average return from 3.39% (Base) to

-3.51% in the EW case and from 5.30% to -2.63% in

GMV, comparable values found by Ilmanen et al.

(2021).

Table 1: Historical backtested results: EW allocation.

Port.

SIGNAL

1 2 3 4 5 6 7 8

BASE

3.39

-29.5

0.12

SMI

0.00

-16.9

0.00

OPT

-3.51

-10.6

-0.33

3.77

-13.3

0.28

2.55

-25.6

0.10

3.01

-25.3

0.12

3.14

-25.3

0.12

2.26

-17.6

0.13

3.85

-20.2

0.19

2.95

-21.3

0.14

LEV

-4.53

-13.2

-0.34

4.27

-16.0

0.27

2.89

-31.3

0.09

3.41

-31.4

0.11

3.55

-31.5

0.11

2.56

-22.0

0.12

4.36

-24.2

0.18

3.35

-25.2

0.13

TR1

-0.97

-18.3

-0.05

3.58

-20.8

0.17

3.44

-27.9

0.12

3.53

-28.1

0.13

3.54

-28.1

0.13

2.39

-25.5

0.09

3.29

-20.2

0.16

2.35

-22.1

0.11

TR2

3.95

-13.0

0.30

2.92

-26.5

0.11

3.16

-25.3

0.12

3.39

-25.3

0.13

2.56

-20.2

0.13

3.29

-20.2

0.16

3.20

-21.3

0.15

TR3

-0.67

-29.5

-0.02

2.80

-29.5

0.09

-0.67

-29.5

-0.02

-0.67

-29.5

-0.02

3.48

-29.5

0.12

2.29

-29.5

0.08

2.71

-29.5

0.09

2.51

-29.5

0.09

WEI.

4.01

-29.2

0.14

4.14

-30.5

0.14

4.38

-29.9

0.15

4.06

-29.9

0.14

5.08

-24.0

0.21

3.94

-39.0

0.10

4.34

-22.5

0.19

Signals: 1=ALL, 2=TREND, 3=MOM, 4=HVOL, 5=IVOL, 6=SKEW, 7=IVOL-

SPIKE, 8=SKEW-SPIKE In each cell, results show the average annualized return

(%) above, worst 6-month period return (%) in the middle, and the ratio of avg.

return to worst 6M return below.

Table 2: Historical backtested results: GMV allocation.

Port.

SIGNAL

1 2 3 4 5 6 7 8

BASE

5.30

-18.8

0.28

SMI

2.84

-10.9

0.26

OPT

-2.63

-13.3

-0.20

5.15

-8.2

0.63

4.09

-15.2

0.27

4.55

-14.9

0.31

4.68

-14.9

0.31

3.81

-08.6

0.44

5.33

-11.2

0.48

4.34

-12.7

0.34

LEV

-3.34

-16.1

-0.21

5.80

-9.8

0.59

4.63

-18.3

0.25

5.14

-18.2

0.28

5.28

-18.3

0.29

4.32

-10.3

0.42

5.99

-13.1

0.46

4.92

-14.4

0.34

TR1

0.02

-13.0

0.00

5.03

-10.8

0.46

5.01

-17.5

0.29

5.12

-17.7

0.29

5.13

-17.7

0.29

4.04

-15.1

0.27

4.86

-11.2

0.43

3.87

-12.4

0.31

TR2

5.32

-8.2

0.65

4.46

-16.1

0.28

4.71

-14.9

0.32

4.93

-14.9

0.33

4.13

-09.8

0.42

4.86

-11.2

0.43

4.61

-12.4

0.37

TR3

1.07

-18.8

0.06

4.66

-18.8

0.25

1.07

-18.8

0.06

1.07

-18.8

0.06

5.36

-18.8

0.28

4.14

-18.8

0.22

4.60

-18.8

0.24

4.35

-18.8

0.23

WEI.

4.78

-19.2

0.25

5.19

-18.8

0.28

5.23

-18.8

0.28

5.31

-18.8

0.28

5.93

-18.8

0.32

5.60

-14.0

0.40

7.30

-16.0

0.46

Signals: 1=ALL, 2=TREND, 3=MOM, 4=HVOL, 5=IVOL, 6=SKEW, 7=IVOL-

SPIKE, 8=SKEW-SPIKE In each cell, results show the average annualized return

(%) above, worst 6-month period return (%) in the middle, and the ratio of avg.

return to worst 6M return below.

6 BOOTSTRAPPED RESULTS

Bootstrap results, displayed in Table 3 and Table 4

for EW and GMV allocation methods respectively,

somehow differ from those of the backtests. In this

FEMIB 2024 - 6th International Conference on Finance, Economics, Management and IT Business

case, only empirical and implied volatility (HVOL

and IVOL) signals outperform the Base when no

additional trading rules are implemented. Both

signals are indeed closely related, as stocks whose

returns have been volatile the past year tend to show

a high implied volatility value. TR1 seems to add

value to all signals, making trend and momentum

signals outperform both EW and GMV allocations.

The disparity of results between backtest and

bootstrap suggests that the results are indeed sample-

dependent: either the backtested results are too reliant

on the one-time 2008 GFC gain, or the bootstraps are

overrepresented in a sample period that has been calm

in general for equities, for which they have

experienced a long rally with exceptionally low

volatility. We performed the same tests but using a

restricted sample period ending on February 14, 2014.

Using this restricted sample results closely align with

the backtests: historical and implied volatility signals

still show outperformance, as in the original

bootstrapped results, but it is the trend signal that

improves the portfolio’s performance the most.

These results suggest that classic signals such as trend

or low volatility (empirical or implied) can help

improve the risk-adjusted performance of equity

portfolios using put options: the signals help reduce

the long-term cost of the option strategies but still

offer partial protection on large equity market

drawdown periods. Yet, no signal works at every

period and market regime. A more dynamic strategy

that identifies, for instance, when trend or low

volatility signals are useful and when they are not,

could certainly improve the performance of the

strategies.

Table 3: Bootstrapped results: EW allocation.

Port.

SIGNAL

1 2 3 4 5 6 7 8

BASE

6.44

-24.2

0.27

SMI

2.25

-23.6

0.10

OPT

-8.62

-28.7

-0.30

4.59

-18.5

0.25

3.15

-17.8

0.18

4.69

-13.6

0.34

4.83

-14.9

0.32

2.12

-20.6

0.10

2.83

-25.0

0.11

1.53

-20.6

0.07

LEV

-10.39

-34.4

-0.30

5.47

-22.6

0.24

3.74

-21.2

0.18

5.59

-16.6

0.34

5.76

-17.8

0.32

2.50

-24.6

0.10

3.36

-30.1

0.11

1.80

-24.6

0.07

TR1

-4.10

-26.0

-0.16

5.53

-16.5

0.34

4.96

-15.4

0.32

5.55

-15.1

0.37

5.74

-15.2

0.38

3.33

-20.1

0.17

3.94

-23.9

0.16

3.77

-20.1

0.19

TR3

-2.70

-31.2

-0.09

5.49

-24.9

0.22

4.78

-24.9

0.19

6.72

-22.2

0.30

7.22

-23.7

0.30

3.69

-22.4

0.16

4.51

-23.7

0.19

2.97

-22.4

0.13

WEI.

5.45

-18.5

0.29

3.50

-16.6

0.21

4.98

-13.5

0.37

5.21

-14.9

0.35

2.46

-20.2

0.12

3.52

-24.7

0.14

2.17

-20.2

0.11

Signals: 1=ALL, 2=TREND, 3=MOM, 4=HVOL, 5=IVOL, 6=SKEW, 7=IVOL-

SPIKE, 8=SKEW-SPIKE In each cell, results show the average annualized return

(%) among the 500 bootstraps above, the annualized return (%) of the worst 5%

case in the middle, and the ratio of avg. return to worst 5% return below.

Table 4: Bootstrapped results: GMV allocation.

Port.

SIGNAL

1 2 3 4 5 6 7 8

BASE

6.67

-14.6

0.46

SMI

3.72

-15.6

0.24

OPT

-8.96

-29.1

-0.31

4.57

-11.7

0.39

3.22

-10.2

0.31

4.75

-09.1

0.52

4.91

-09.2

0.54

2.15

-12.4

0.17

2.84

-15.0

0.19

1.50

-12.4

0.12

LEV

-10.80

-34.8

-0.31

5.45

-14.0

0.39

3.82

-12.3

0.31

5.67

-10.9

0.52

5.85

-11.1

0.53

2.54

-15.1

0.17

3.37

-18.5

0.18

1.76

-15.1

0.12

TR1

-4.42

-27.7

-0.16

5.57

-11.4

0.49

5.02

-08.7

0.58

5.66

-08.4

0.67

5.86

-08.8

0.66

3.36

-11.6

0.29

3.97

-14.6

0.27

3.74

-11.6

0.32

TR3

-3.25

-26.4

-0.12

5.60

-14.2

0.39

4.86

-14.0

0.35

6.78

-12.4

0.55

7.27

-12.6

0.58

3.65

-13.6

0.27

4.66

-15.0

0.31

3.04

-13.6

0.22

WEI.

5.41

-11.7

0.46

3.55

-10.2

0.35

5.02

-08.9

0.56

5.26

-09.2

0.57

2.46

-12.6

0.19

3.53

-15.0

0.24

2.14

-12.6

0.17

Signals: 1=ALL, 2=TREND, 3=MOM, 4=HVOL, 5=IVOL, 6=SKEW, 7=IVOL-

SPIKE, 8=SKEW-SPIKE In each cell, results show the average annualized return

(%) among the 500 bootstraps above, the annualized return (%) of the worst 5%

case in the middle, and the ratio of avg. return to worst 5% return below.

7 CONCLUSIONS

Protecting the downside risk of equity portfolios is a

paramount aspiration for any risk-averse investor or

portfolio manager, and options can be very effective

instruments to achieve this objective, thanks to their

flexibility in their payoff profiles. However, the high

cost of passive put protection and its large negative

impact on a portfolio’s long-term performance makes

this type of strategy implausible in practice. In this

article, we test for a set of market signals that could

help take option positions more efficiently, reducing

the cost spent in options’ premia but still benefiting

from an acceptable degree of portfolio protection,

even though this one is partially offset. We test for a

set of backward-looking market signals widely used

in the equities space, such as trend, momentum, and

historical volatility, and forward-looking signals

coming from the options markets, which are less

explored by the industry. The latter are implied

volatility and skew, both cross-sectionally (relative

rank) and in time-series format (implied volatility or

skew rise). The benefit of these signals is that they

show investors’ expectations about future market

movements, as opposed to past information. We

perform backtest and bootstrap tests to determine

whether the signals add value to a portfolio of Swiss

Safeguarding Downside Risk in Portfolio Insurance: Navigating Swiss Stock Market Regimes with Options, Trading Signals, and Financial

Products

large-cap equities. Also, we test for different

strategies that implement a set of trading rules that

intuitively could improve the portfolios’

performance. Two equity allocations are simulated:

an equally weighted and a minimum variance

portfolio. The results suggest that it is possible to use

market signals to select the underlyings to which

purchase put options, similarly to the use of signals to

construct long-short portfolios, and improve the risk-

return relationship of a portfolio made of equities and

put options. However, the signals’ efficacy depend

largely on the market regime. For instance, trend

appears to add substantial value during large and

long-dated equity market drawdowns, but less clearly

during calm markets, when historical volatility or

implied volatility seems to work better. More

sophisticated trading signals do not improve the

results consistently. Finally, it is worthwhile noticing

that GMV portfolios systematically outperform their

EW counterparts in all strategies (with or without

options, with or without trading rules, with or without

leverage), both in backtests and bootstrapped results,

even though GMV portfolios tend to be heavily

concentrated in a few individual stocks.

The empirical results of this study should be

considered in the light of some limitations. First, the

limited dataset composed only by Swiss large-cap

underlyings results into rather concentrated

portfolios. Practically, equity portfolios would likely

be more diversified, by holding a larger number of

different names and including smaller-capitalization

stocks. Also, according to the authors’ own backtests,

the signals appear to be more relevant in a larger

universe that includes smaller-capitalization stocks.

Moreover, the strategies’ relevance in markets other

than Switzerland, or even international portfolios,

remains to be tested. Another shortcoming is the

uniqueness and simplicity of the strategy

implemented to purchase put options that is presented

in this article. Certainly, alternative methods exist and

their efficiency is worthwhile being tested. The

authors have tested the incorporation of other option

strategies with different risk-return objectives into the

same Swiss equity portfolios. For instance, the classic

protective put (fully protecting the underlying equity

position below some strike) on negative-signal

underlyings, selling OTM covered call options (e.g.

at 110% of strike) on those same underlyings for yield

enhancement, or by going long OTM calls on

underlyings with a positive signal, thus exploiting the

potential of the positive side of the signals. However,

due to space limitations, only this simple long-put

strategy that exploits the negative side of the signals

is presented. In the same way, the strategies’

sensitivities to parameter modifications remain to be

tested. For instance, strikes can be set at different

levels (e.g. 80% or 95% instead of the 90% used),

signals could be defined using other lookback

windows (e.g. 6 months instead of 12 months).

Finally, the results of this current study show that the

effectiveness of the different signals on the put option

strategies are dependent on different market regimes.

Thus, further research is necessary to test the

effectiveness of the signals conditional to market

regimes and whether these regimes can be

consistently and timely identified.

In conclusion, this paper, primarily focused on the

field of financial engineering, aims to demonstrate

that the complex concepts of portfolio insurance,

options, and trading signals, which have been widely

discussed in the financial risk management literature,

should be subject to a comprehensive, global

benchmark test. The intention behind this undertaking

is to emphasize that these complex notions should not

simply remain in the realm of financial theory but

should also serve as valuable tools for effective

portfolio management. Our project, supported by

Innosuisse and conducted in collaboration with an

options trading house, highlighted a key point. We

confirmed that the optimal portfolio protection

strategy, despite its potential cost, should be

adaptable to different market regimes.

ACKNOWLEDGEMENTS

The authors acknowledge Innosuisse, the Swiss

Innovation Agency, for financially supporting the

project 34627.1 IP-SBM; A forward-looking risk

management tool for Swiss pension funds for the

period 2020-2022. They are also grateful to the two

anonymous referees that with their comments helped

improve this article.

REFERENCES

AQR (2022). Trend-Following: Why Now? A Macro

Perspective. White paper.

Annaert, J., Van Osselaer, S., Verstraete, B. (2009).

Performance evaluation of portfolio insurance

strategies using stochastic dominance criteria. Journal

of Banking & Finance, 33(2), 272-280.

Asness, C. S., Moskowitz, T. J., Pedersen, L. H. (2013).

Value and momentum everywhere. The journal of

Finance, 68(3), 929-985.

Babu, A., Hoffman, B., Levine, A., Ooi, Y. H., Schroeder,

S., Stamelos, E. (2020). You can’t always trend when

FEMIB 2024 - 6th International Conference on Finance, Economics, Management and IT Business

you want. The Journal of Portfolio Management, 46(4),

52-68.

Baltussen, G., van der Grient, B., de Groot, W., Hennink,

E., Zhou, W. (2012). Exploiting option information in

the equity market. Financial Analysts Journal, 68(4),

56-72.

Blitz, D., Vidojevic, M. (2017). The profitability of low-

volatility. Journal of Empirical Finance, 43, 33-42.

Boulier, J. F., Kanniganti, A. (2005). Expected performance

and risk of various portfolio insurance strategies.

Proceedings of the 5th AFIR International Colloquium

(pp. 1093-1124).

Brodie, J., Daubechies, I., De Mol, C., Giannone, D., Loris,

I. (2009). Sparse and stable Markowitz portfolios.

Proceedings of the National Academy of Sciences,

106(30), 12267-12272.

Co, D. (2023). A new approach to trend-following.

Unpublished manuscript.

Dao, T. L., Nguyen, T. T., Deremble, C., Lemperiere, Y.,

Bouchaud, J. P., Potters, M. (2016). Tail protection for

long investors: Trend convexity at work. arXiv preprint

arXiv:1607.02410.

DeMiguel, V., Garlappi, L., Uppal, R. (2009). Optimal

versus naive diversification: How inefficient is the 1/N

portfolio strategy? The Review of Financial Studies,

22(5), 1915-1953.

Dew-Becker, I., Giglio, S., Kelly, B. (2017). How do

investors perceive the risks from macroeconomic and

financial uncertainty? Evidence from 19 option

markets. Working paper.

Figlewski, S., Chidambaran, N. K., Kaplan, S. (1993).

Evaluating the performance of the protective put

strategy. Financial Analysts Journal, 49(4), 46-56.

Frazzini A., Pedersen, L. H. (2014). Betting against beta.

Journal of Financial Economics, 111(1), 1-25.

Harper, A. C., Sarkar, S. K. (2019). Option Market Signals

and the Disposition Effect Around Equity Earnings

Announcements. Journal of Behavioral Finance, 20(4),

471-489.

Hurst, B., Ooi, Y. H., Pedersen, L. H. (2017). A century of

evidence on trend-following investing. The Journal of

Portfolio Management, 44(1), 15-29.

Ilmanen, A., Thapar, A., Tummala, H., Villalon, D. (2021).

Tail risk hedging: Contrasting put and trend strategies.

Journal of Systematic Investing, 1(1), 111-124.

Israelov, R., Nielsen, L. N. (2015). Still not cheap: Portfolio

protection in calm markets. The Journal of Portfolio

Management, 41(4), 108-120.

Jegadeesh, N., Titman, S. (1993). Returns to buying

winners and selling losers: Implications for stock

market efficiency. The Journal of Finance, 48(1), 65-

91.

Jurczenko, E., Teiletche, J. (2015). Risk-Based Investing

but What Risk (s)? Risk-Based and Factor Investing,

pp. 147-171. Elsevier.

Kostakis, A., Panigirtzoglou, N., Skiadopoulos, G. (2011).

Market timing with option-implied distributions: A

forward-looking approach. Management Science

57(7), 1231-1249.

Kremer, P. J., Lee, S., Bogdan, M., Paterlini, S. (2020).

Sparse portfolio selection via the sorted ℓ1-Norm.

Journal of Banking & Finance, 110, 105687.

Lhabitant, F. S. (1998). Enhancing portfolio performance

using option strategies: why beating the market is easy

(No. rp1). International Center for Financial Asset

Management and Engineering.

Maillard, S., Roncalli, T., Teïletche, J. (2010). The

properties of equally weighted risk contribution

portfolios. The Journal of Portfolio Management,

36(4), 60-70.

Mohanty, S. S. (2018). Enhancing Portfolio Performance in

Global Equity Allocation with a Forward-Looking

Indicator. The Journal of Investing, 27(4), 81-97.

Moskowitz, T. J., Ooi, Y. H., & Pedersen, L. H. (2012).

Time series momentum. Journal of Financial

economics, 104(2), 228-250.

Rockafellar, R. T., Uryasev, S. (2000). Optimization of

conditional value-at-risk. Journal of risk, 2, 21-42.

Safeguarding Downside Risk in Portfolio Insurance: Navigating Swiss Stock Market Regimes with Options, Trading Signals, and Financial

Products