Investing in Deflation Inflation and Stagflation Regimes

Financial Analysts Journal
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/ufaj20
Investing in Deflation, Inflation, and Stagflation

Regimes
Guido Baltussen, Laurens Swinkels, Bart van Vliet & Pim van Vliet
To cite this article: Guido Baltussen, Laurens Swinkels, Bart van Vliet & Pim van Vliet (2023)
Investing in Deflation, Inflation, and Stagflation Regimes, Financial Analysts Journal, 79:3, 5-32,
DOI: 10.1080/0015198X.2023.2185066
To link to this article: https://doi.org/10.1080/0015198X.2023.2185066
© 2023 The Author(s). Published with

license by Taylor & Francis Group, LLC.
Published online: 11 Apr 2023.
Submit your article to this journal
Article views: 10279
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ufaj20
Financial Analysts Journal | A Publication of CFA Institute Research
https://doi.org/10.1080/0015198X.2023.2185066
Investing in Deflation,
Inflation, and Stagflation
Regimes
Guido Baltussen , Laurens Swinkels , Bart van Vliet, CFA , and
Pim van Vliet
Guido Baltussen is a professor at Erasmus University Rotterdam, and head of factor investing and co-head of quant fixed income at Robeco
Institutional Asset Management, Rotterdam, the Netherlands. Laurens Swinkels is an associate professor at Erasmus University Rotterdam,
and head of quant strategy for sustainable multi asset solutions at Robeco Institutional Asset Management, Rotterdam, the Netherlands.
Bart van Vliet, CFA, is a PhD candidate at Erasmus University Rotterdam, and manager of client reporting at Robeco Institutional Asset
Management, Rotterdam, the Netherlands. Pim van Vliet is head of conservative equities and chief quant strategist at Robeco Institutional
Asset Management, Rotterdam, the Netherlands. Send correspondence to Guido Baltussen at [email protected].
We examine asset class and factor pre- Introduction

miums across inflationary regimes. As igh inflation has always been a worry for investors, but inflation
periods of deflation, high inflation, and
especially stagflation are relatively
uncommon in recent history, we use a
deep sample starting in 1875. Moderate
H has been very stable at around 2% and has never exceeded 4%
during the past 30 years. Investors have become accustomed to
low and stable inflation, in which real returns are close to nominal
inflation scenarios provide the highest returns. However, the strong global spike in consumer price inflation
returns across asset class and factor pre- after the coronavirus crisis of 2020 and the following fears of stagfla-
miums. During deflationary periods, nom- tion has substantially increased questions about the impact of inflation
inal returns are low, but real returns are on risk premiums and investment strategies. Relatedly, deflation com-
attractive. By contrast, real equity and
monly follows periods of high inflation. An important question, hence,
bond returns are negative during a high
inflation regime and especially so during is how equity, bond, and factor style strategies behave across inflation-
times of stagflation. During these “bad ary regimes like periods of high inflation, deflation, or stagflation. In this
times,” factor premiums are positive, paper, we examine such relationships robustly using the deepest sample
which helps to offset part of the real available for a relatively broad cross-section of asset classes and factors
capital losses. starting in 1875.
Keywords: bond returns; deflation; equity There are two key challenges when it comes to examining the inflation
returns; factor premiums; inflation;
sensitivity of investment returns. First, there have been few periods of
stagflation
elevated inflation rates in the past 50 years. Most evidence is derived
Disclosure: The authors are employed from the high inflation period following the oil price shocks in the 1970s,
by Robeco, a firm that offers various We would like to thank Campbell Harvey, David Blitz, Martin Martens, and Olaf
investment products. The construction of Penninga for valuable contributions and discussions. The views expressed in this
these products may, at times, draw on paper are not necessarily shared by Robeco Institutional Asset Management.
insights related to this research. That said, This is an Open Access article distributed under the terms of the Creative Commons
the research article is independent of the Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/
views of Robeco and solely attributable to licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and repro-
the authors. duction in any medium, provided the original work is properly cited, and is not altered,
transformed, or built upon in any way. The terms on which this article has been pub-
lished allow the posting of the Accepted Manuscript in a repository by the author(s) or
PL Credits: 2.0 with their consent.
Volume 79, Number 3 © 2023 The Author(s). Published with license by Taylor & Francis Group, LLC. 5
Financial Analysts Journal | A Publication of CFA Institute
a period that spans relatively few independent observa- Next, we consider annual investment performance
tions. This holds especially for regimes that combine over various inflationary regimes. In our core analy-
inflation and macroeconomic growth, such as stagfla- ses, we divide our sample in various ex-post infla-
tion, with less than a handful of annual observations tion regimes and examine the average returns
since the 1970s. In a similar spirit, deflation could be as during each regime. For our base case, we divide
much of a risk for investors as inflation, but deflation our sample in four economically motivated global
has been virtually absent the past 50 years. Second, inflation regimes: (1) below 0%, or deflation, (2)
limiting the menu of assets to long-only investments in between 0% and the current central bank target of
conventional asset classes and ignoring factor premiums 2%, (3) a mild inflation overshoot, between 2% and
does not show the full potential of asset allocation to 4%, and (4) high inflation, above 4%. Note that this
deal with large economic shocks; see, for example, is more granular than the high/low inflation regime
Ilmanen and Kizer (2012).1 with an entry threshold of 5% as used in Neville
et al. (2021).3 Each of our four regimes constitutes
To overcome these challenges, we start by building about 20% to 30% of the observations.
an extensive historical database that includes a wide
range of inflationary regimes while still having high- Our findings show that asset class premiums vary sig-
quality data on asset class and factor returns. For nificantly across these inflationary regimes in both
asset class premiums, we combine several global nominal and real terms. Equities and bonds yield, on
datasets on equities, bonds, and cash returns. For average, lower nominal returns during periods of high
factor premiums, we use data from three key studies inflation, causing negative real returns. This is in line
that construct equity, bond, and global factor premi- with Neville et al. (2021), who use a 5% inflation
threshold and a 95-year sample period. Further,
ums (GFPs) back to the 19th century, which we
equity returns tend to be relatively low in nominal
update to the end of 2021.2 We consider four
terms in periods of deflation, but average in real
investment factor premiums: Value, Momentum, Low
terms. By contrast, equity, bonds, and GFPs are gen-
Risk, and Quality or Carry applied in three markets:
erally positive across inflationary regimes, displaying
equities, bonds, and across global markets. Some of
generally no significant variation across the inflation-
these 12 factor-return series start as early as 1800,
ary regimes, while they enhance nominal and real
but the availability of monthly global inflation series
asset class returns in (approximated) long-only asset
limits the start of our sample to 1875. Using this
class implementations. We show that these results
147-year sample provides the most testing power
are robust across different definitions of inflationary
and robustness to examine the performance of asset
regimes, including a 3% or 5% (instead of 4%) high
classes and factors premiums during various inflation-
inflation cutoff, the use of annual changes in inflation
ary regimes. (or unexpected inflation), the use of only U.S. (instead
We start our analysis by examining asset class (equi- of global) inflation, or the use of 3-year (instead of 1-
ties, bonds, and cash) returns and factor returns year) horizons.
within and across asset classes over our deep sam- We continue by dividing inflationary regimes in sev-
ple. We show that the global equity return has eral sub-regimes based on macroeconomic or market
been on average 8.4% (in arithmetic terms) between performance and inflation dynamics, most notably
1875 and 2021, while the global bond market for high inflation. This includes especially stagflation-
return (currency risk hedged) has been 4.5%. For ary episodes with both high inflation and economic
comparison, global inflation has been on average downturns (i.e., recessions). We find that the periods
3.2% per annum over the same period. Value, of stagflation are truly bad times for investors, as for
Momentum, Low Risk, or Quality/Carry factors also example nominal equity returns average –7.1% per
returned attractive and significant returns above 4%, annum, yielding double digit negative returns in real
providing substantial alpha over traditional asset terms. During these bad times, equity, bond, and
classes. The average multi-factor combination deliv- GFPs remain consistently positive. As such, factors
ers significant alphas on top of asset returns with t help to offset some, but not all, of the negative
values above 7.9 over this full sample period. In impact of high inflation in recessionary times. We
other words, asset class and factor returns are consider a variety of “stagflationary” sub-regimes
strong and consistent “empirical facts,” and factor definitions, including recessions, falling earnings
premiums offer material added value to asset class growth, or falling equity markets, and find mostly
premiums. consistent results across these times. Further, we
6
Investing in Deflation, Inflation, and Stagflation Regimes
examine sub-regimes based on increasing and dominated by currencies that were tied to gold or sil-
decreasing long-term interest rates or increasing and ver. High-inflation episodes in times of the gold
decreasing inflation rates. Overall, stagflationary epi- standard generally correspond to times during which
sodes, high-inflation bear markets, or rising inflation- the convertibility to gold was suspended to meet
ary times and, to a lesser extent, deflationary bear demand for additional government revenue, after
markets are bad times for investors, and factor pre- which the convertibility was reinstated and prices
miums alleviate some of the pain during these deflated to initial levels. For the United States, the
regimes. period until 1900 shows one short-lived inflation
peak up to 10% but was on average deflationary.
The following period until World War I showed mild
A Long History of Inflationary and positive inflation around 3% on average.5 The period
Deflationary Times in between both World Wars was deflationary. After
World War II, the 1970s and early 1980s are further
In this section, we examine inflation dynamics and
periods with high and persistent inflation.
regimes over time, starting in 1875 and ending in
2021. To measure inflation, we primarily use year- It is unlikely that investors care about transitory
on-year changes in Consumer Price Indices (CPIs).4 spikes in inflation that are expected to reverse in the
The primary source of inflation data is Datastream, next year when they price assets with long-term cash
which we backfill with inflation data from Global flows such as equities and bonds. Therefore, Figure
Financial Data (GFD) and MacroHistory before avail- 1B shows the rolling 3-year global inflation rate in
ability in Datastream. The monthly inflation data ser- addition to the annual global inflation rate. This fig-
ies starts in January 1875 for each of the following ure again shows that before World War I, inflation
countries that we consider: the United States, the was noisier and averaging over longer periods of
United Kingdom, Germany, France, and Japan. time reduces the volatility of inflation substantially.
Following Cagan (1956) and Baltussen, Swinkels, and However, after World War I, averaging has limited
Van Vliet (2021), we exclude hyperinflation periods effect, as inflation and deflation spikes tend to be
from our sample by excluding periods for which the more persistent.
annual inflation number is above 50% and start
including again 12 months after the hyperinflation Table 1 summarizes the distribution of inflation over
period has ended. We choose to exclude these our sample. Panel A buckets global inflation into four
hyperinflation periods as they are rare and special inflation regimes: (1) below 0%, or deflation, (2)
episodes that come with large measurement noise between 0% and the current central bank target of
and risks for investors that they typically choose to 2%, (3) a mild inflation overshoot, between 2% and
exclude. This affects especially Germany during the 4%, and (4) high inflation, above 4%, and reports the
post–World War I period from 1920 to 1926 and number of years of inflation observations in each
Japan during the post–World War II period from bucket. The midpoint between (2) and (3) was chosen
1946 to 1950. From the resulting series, we con- because 2% is the current target of many central
struct a global inflation measure by equally weighting banks in major developed markets. As such, we focus
across markets (but later verify robustness to using on periods of negative inflation, inflation that realizes
only U.S. inflation numbers). low relative to “target” (0–2%), inflation that over-
shoots the target (2–4%)—in 2021 an inflation level
Figure 1A shows the U.S. and global inflation series of 3% was seen by several investors as a material
over the sample period from January 1875 to worry6—and inflation that is substantial (>4%).
December 2021 together with National Bureau of
Economic Research (NBER) recession periods. Several Over the entire sample period (1875–2021), there
observations are noteworthy. First, U.S. and global have been 23.1 years of deflation and 46.1 years with
inflation behave very much alike. Second, inflation inflation above 4%. The subperiod analysis also
varies substantially over time, with several periods of shows that examining the most recent 30 years does
high inflation—such as during the 1880s, after World not yield much information about deflationary or
War I, around World War II, during the 1970s, and high-inflation periods, as inflation has almost always
more recently—but also periods of deflation—typically been in the range of 0% to 4%. Including the 1970s
after periods of high inflation (especially before gives a period of high inflation, but one needs to
World War II). Third, inflation is more volatile in the include the 1930s to also have a substantial number
period before the 1970s, which for a large part was of deflationary periods. Extending the sample to
Volume 79, Number 3 7

Figure 1. (A) Annual Consumer Price Inflation and (B) Triannual Global Consumer Price Inflation
A 30%
20%
Inflation rate (yoY)
10%
0%
-10%
-20%
1875
1880
1885
1890
1895
1900
1905
1910
1915
1920
1925
1930
1935
1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
2010
2015
2020
Recession US Global
B
30%
20%
Inflation rate (yoY)
10%
0%
-10%
-20%
1875
1880
1885
1890
1895
1900
1905
1910
1915
1920
1925
1930
1935
1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
2010
2015
2020
Recession Global 3Y
Notes: The figure shows the historical time-series behavior of global and U.S. year-over-year (YoY) inflation (panel A) or Global YoY versus
3-year (annualized) inflation (panel B). Periods highlighted in grey represent NBER recessions. The sample period is 1875 to 2021.
1875 increases both the number of years with high of observations fall within each inflation regime
inflation during and after World War II and with when looking at U.S. inflation numbers.
deflation at the end of the 19th century. Hence,
extending the period to 1875 gives a more reliable Central banks existed for most of our sample period.
assessment of what investors can expect during peri- The U.S. Federal Reserve was founded in 1913, the
ods of deflation or high inflation. Panel B confirms Bank of England in 1694, the German Bundesbank in
these insights for U.S. inflation. Panel C contains 1957 (its predecessor in 1948 and before World War
more information about the distribution of inflation. II the Reichsbank in 1876), and the Bank of Japan in
The median value of annual inflation equals 2.3% for 1882. Their mandate to curb inflation to a predeter-
global and 2.2% for U.S. inflation. At the 10th per- mined level is a relatively recent phenomenon, mostly
centile, there is a –1.4% and –2.4% deflation, introduced after the inflation shocks in the 1970s.
respectively. At the other end, the 90th percentile is Explicit inflation targeting only started in the 1990s
an inflation of 8.9% and 7.2%, respectively. Further, (see Svensson, 2011). Hence, we have to leave the
between 19% (deflation) and 29% (0%–2% inflation) impact of inflation targeting policies on the
8
Table 1. Inflation Frequencies

A. Global inflation episodes
Inflation bucket 1992–2021 1950–2021 1926–2021 1875–2021
<0% 0.9 1.6 7.7 23.1

0%–2% 20.3 25.5 27.6 42.9
2%–4% 8.6 22.7 25.9 34.9
>4% 0.3 22.3 34.8 46.1
All 30.0 72.0 96.0 147.0
B. U.S. inflation episodes

Inflation bucket 1992–2021 1950–2021 1926–2021 1875–2021
<0% 1.2 2.7 11.2 27.6

0%–2% 10.3 21.5 26.3 42.0
2%–4% 16.8 27.5 32.3 40.3
>4% 1.8 20.3 26.3 37.2
All 30.0 72.0 96.0 147.0
C. Inflation distribution
Inflation (YoY) Percentile
0.05 0.10 0.25 0.33 0.50 0.67 0.75 0.90 0.95

Global –3.0% –1.4% 1.0% 1.5% 2.3% 3.8% 4.6% 8.9% 12.1%
U.S. –4.1% –2.4% 0.2% 1.2% 2.2% 3.2% 4.1% 7.2% 11.2%
Notes: The table represents the historical distribution of inflation. Panel A (B) shows the average number of
annual observations over different sample periods falling in each year-over-year (YoY) inflation level bucket
based on global (U.S.) inflation. Panel C contains the distribution of global and U.S. inflation over the full
sample, 1875–2021.
relationship between inflation and asset returns indeed aware of increases and decreases of consumer
open. The lack of inflation targeting by central banks prices and linked it to asset prices.
does not mean that investors did not care about
deflation or inflation when judging stock and bond
markets. This follows from a simple count of words Long-Run Evidence on Asset Class
measure of inflation- and deflation-related words as and Factor Premiums: 1875–2021
mentioned in the Abreast of the Market columns of In this section, we examine asset class returns and
The Wall Street Journal and related market commen- factor returns within and across asset classes over
tary columns in The New York Times as used by Garcia our deep sample. Our dataset is at the monthly fre-
(2013), which we acquire between 1899 and 2021. quency and includes global equities, bonds, and cash
Figure 2 shows the relative importance of inflationary returns; equity factor returns (Value, Momentum,
minus deflationary words per rolling annual window, Low Risk, and Quality); government bond factor
computed by the standardized relative word frequen- returns (Value, Momentum, Low Risk, and Carry); and
cies.7 A value of 1 (–1) indicates inflation (deflation) global factor returns (Value, [Time-Series]
words were commonly mentioned during that year in Momentum, Low Risk, and Carry) all expressed in
the market commentaries (i.e., they were “topical”, USD. Our sample of returns and inflation numbers
while a value of 0 indicates less investor consideration starts in January 1875, the first year we have global
of inflation. This figure shows that before the inflation inflation series, and ends in December 2021. We
spikes in the 1970s, inflation was an important finance include data on Value, Momentum, Low Risk, and
topic during periods with elevated inflation, and defla- Quality equity factors in the cross-section of U.S.
tion was topical in deflationary periods. When stocks. These factors are the common factors
regressed on U.S. inflation, the R2 equals 51%. This employed in the industry, being the key motivation
evidence indicates that market participants were of our choice.8 Further, we consider U.S. stocks, as

Figure 2. Inflation Topicality
1.5 20%
15%
1
10%
Inflation rate (YoY)

Inflation topicality
0.5
5%
0 0%
-5%
-0.5
-10%
-1
-15%
-1.5 -20%
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
Inflation topicality US Inflation
Notes: The figure shows the historical time-series behavior of inflation topicality (the relative mentioning of inflation versus deflation
words in the market columns in The Wall Street Journal and/or The New York Times) versus U.S. year-over-year inflation. The sample
period is 1899 to 2021.
data on equity factors across the globe only starts these contain less powerful information about infla-
toward the end of the 1980s. For global government tionary regimes.
bond market factors, we use data from Baltussen,
Panel A contains the returns on the conventional
Martens, and Penninga (2021) on Value, Momentum,
long-only asset classes equities, government bonds,
Low Risk, and Carry factors, which we update until
and cash, and a multi-asset portfolio that consists of
the end of December 2021. Finally, global (“cross-
60% equities and 40% government bonds.10 Long-
asset”) factor returns on Value, Momentum, Low
term nominal average returns for global equity
Risk, and Carry are taken from Baltussen, Swinkels,
investors have been on average 8.4% per annum, a
and Van Vliet (2021), which we also update until the
number that is high and significant (t value ¼ 7.4).11
end of 2021.9 They construct GFPs using bond,
Hence, equities offered an attractive return over
equity, currency, and commodity market data at the
147 years. For sample periods that start later, aver-
country level, so no individual stocks. Appendix A.1 age returns are also consistently positive ranging
describes the data and factors that are used through- from 8.9% (1992–2021) to 12.1% (1950–2021) per
out the paper in detail. annum.
We start our analysis by examining the long-run evi- Global (currency risk hedged) government bond
dence on global asset class and factor returns. Table returns have been substantially below global equity
2 contains the average returns on each of these asset market returns with 4.5% per annum, but still signifi-
classes and factor premiums over the long-run sam- cantly above zero (t value 14.0). This compares to an
ple from 1875 to 2021, as well as several subsamples average return on cash of 3.4% per annum, while
that start later. The reason to include these shorter global inflation has been on average 3.2% per annum
subsamples is that they are often employed in earlier over the same period (see Table 1).12 Hence, the glo-
studies, thereby proving a form of “in-sample” evi- bal term premium is 1.1% per annum. Although this
dence, while the longer-run evidence reveals the may appear to be modest, this may be partially due
“out-of-sample” robustness of the asset class and to the lack of a good proxy for the short-term risk-
factor premiums (although we leave a full out-of- free rate over the long historical period, which may
sample study to the original papers with deep his- overstate the cash returns. For example, U.S.
tory). Moreover, those who find that the recent his- Treasury bills were not regularly issued until
tory is more representative of the future may be December 1929, when the auctioning of 13-week
interested in more recent subsamples, even while bills started. Most studies use commercial paper with
10
Table 2. The Long-Run Evidence on Asset Class and Factor Premiums

1992 1950 1926 1875
2021 2021 2021 2021 t value Alpha t value
A. Asset classes
Equities 8.9 12.1 10.6 8.4 7.4 – –
Bonds 5.4 5.6 5.1 4.5 14.0 – –
Cash 2.3 4.0 3.2 3.4 40.4 – –
60/40 7.5 9.5 8.4 6.8 9.5 – –
B. Equity factors
Value 1.6 3.3 4.0 3.5 3.0 3.9 3.3
Momentum 5.4 8.2 7.7 6.9 4.8 8.5 6.0
Low Risk 6.4 6.9 6.0 6.5 6.7 5.0 4.9
Quality 3.2 2.9 2.5 2.5 4.9 3.2 6.0
MFE 4.1 5.3 5.2 5.1 7.4 5.5 7.9
C. Bond factors
Value 0.8 4.0 3.4 2.4 2.5 2.1 1.8
Momentum 2.4 2.6 1.7 3.9 3.9 4.2 3.6
Low Risk 0.9 4.3 4.5 4.4 4.0 6.1 5.0
Carry 5.5 7.9 6.6 7.4 7.9 7.4 6.8
MFB 2.4 4.7 3.9 4.6 9.4 4.9 9.2
D. Global factors
Value 2.6 2.7 2.4 2.3 4.8 2.6 4.8
Momentum/Trend 4.8 7.3 7.4 7.4 13.8 7.6 12.6
Low Risk 0.7 2.2 2.1 1.6 4.3 1.4 3.5
Carry 5.5 6.5 4.9 5.5 11.4 5.3 9.7
GFP 3.4 4.7 4.2 4.2 17.7 4.2 15.7
Notes: The table shows historical returns on asset class and factor premium portfolios over various sample periods. Equity and
bond returns are based on global portfolios, cash on short-dated U.S. cash investments. 60/40 represents a 60% equity, 40%
bond portfolio. Equity factors: Quality starts in 1940, all other series in 1875. MFE takes an equally weighted combination of the
equity factor premiums. Bond factors: Low Risk starts in 1922, all other series in 1875. MFB takes an equally weighted combin-
ation of the bond factor premiums. Global factors: All series start in 1875, composed of equal volatility weighted combinations of
the factor premiums in equity indices, government bonds, currencies, and commodity futures. GFP takes an equally weighted com-
bination of the global factor premiums. Alphas are relative to equity and bond asset class returns. Significant average returns at
the 5% significance level are denoted with a and significance at the 1% significance level with . Numbers within parentheses
represent t values. Standard errors are computed using the Newey–West procedure.
2 to 3 months’ maturity to proxy the short-term economically and statistically significant in all sub-
interest rate before the issuance of Treasury bills or samples and is 4.1% over the most recent 30 years.
certificates, but these contain a small credit premium During this more recent period, the Quality factor is
relative to government issued securities.13 somewhat stronger, with 3.2% per annum versus
Consequently, we choose to focus our analysis on 2.5% in the longest sample that for Quality goes
total asset class returns (instead of excess returns back to 1940 because no reliable accounting data are
over cash). available before that time. Furthermore, the combin-
ation of Value and Momentum gives more consistent
In panel B, the four equity factors show strong and results than the individual strategies profiting from
statistically significant performance over the long-run negative correlation between these two factors.14
sample, with average returns ranging between 2.5% When we correct for asset class risks, all equity fac-
(Quality) to 6.9% (Momentum). Low Risk has the tors remain significant in economic and statistical
highest t statistic (6.7). The overall multi-factor equity terms. Momentum and Low Risk have the highest
(“MFE”) strategy, constructed as an equally weighted alphas of 8.5% and 5.0% per annum, respectively,
combination of the individual factors available each while Value and Quality have alphas exceeding 3%.
period, gives a robust and significant outperformance The combined MFE alpha equals 5.5%, with a high t
over each sample period. The average return since value of 7.9. In other words, equity factors are eco-
1875 equals a statistically significant 5.1% with a nomically sizable and statistically robust phenomena
high t value of 7.4. The average return is over the last 147 years.

Panel C shows that results for bond factor premiums regimes. Our key focus is on the four inflationary
are similar. Value yields a statistically significant 2.4% regimes (<0%, 0%–2%, 2%–4%, and >4%) defined
per annum return (t value, 2.5), while Carry is the previously, but we like to stress that we consider
strongest factor, with an average return of 7.4% per robustness to other regime classifications as well.
annum (t value, 7.9). Further, the individual bond fac- The first question to answer is the impact of inflation
tors are generally sizable and statistically significant on returns of the three major asset classes (equities,
over most subsamples. The major exception is the bonds, and cash). Panels A1 and A2 of Table 3 con-
most recent subperiod, in which Carry factor is the tain the nominal and real returns of these asset
strongest factor with 5.5% return per annum, while classes over the full period as well as during the four
the three other factors are positive but statistically inflation regimes defined above. We also include a
indistinguishable from zero. However, due to diversi- multi-asset portfolio that consists of 60% equities
fication benefits across factors, the equally weighted and 40% bonds. This global portfolio returned 6.8%
multi-factor bond (“MFB”) strategy is statistically sig- per annum in nominal terms (see also the previous
nificant, with a return of 2.4% over the most recent section) and in 3.5% in real terms over the full sam-
30 years. Over the full sample period, the MFB strat- ple period 1875 to 2021.
egy returns a sizable 4.6% return, with again a high t
value of 9.4, highlighting its robustness and statistical Deflationary periods, where inflation is below 0%,
significance. As for equity factors, market risk adjust- coincide with relatively low nominal returns for equi-
ments in the final columns indicate that market risk is ties of 2.4% per annum, well below the 8.4% uncon-
unable to explain any of the factors to a meaningful ditional average return. Bonds and cash show a 5.2%
extent, and alpha is both sizable and significant for and 2.8% per annum nominal return during deflation-
the MFB combination (4.9% per annum with a t value ary periods, respectively, which is slightly above the
of 9.2). unconditional averages of 4.5% and 3.4%.
Consequently, a multi-asset 60/40 investor achieved
Finally, GFPs are also sizable and highly significant, as a 3.5% nominal return during deflationary periods.
shown in panel D. The Low Risk factor is the weak- However, an investor that does not suffer from
est among the factors, but it still shows an economic- money illusion realizes that even though the nominal
ally and statistically significant premium of 1.6% per return is low, the real return is decent because of the
annum (t value, 4.3).15 Momentum is the strongest deflation. Adjusted for purchasing power, the multi-
over the sample since 1875 with an average return asset investor earns 6.7% per annum during defla-
of 7.4% per annum (t value, 13.8). The equally tionary periods.
weighted multi-factor GFP strategy returns 4.2% per
annum, with an extremely high t value of 17.7. The High inflation periods, where inflation is above 4%,
four GFPs combined are economically and statistic- show a positive nominal return for equities (6.9%)
ally significant over the various subsamples highlight- and bonds (3.9%). At face value, it may seem that
ing their robustness. Asset class risks cannot explain the multi-asset investor does quite well with a 5.7%
the global factor returns, witnessing alphas between nominal return. However, the real return is substan-
1.4% (Low Risk) and 7.6% (Momentum) and t values tially negative at –2.9% per annum, leading to a
between 3.5 and 12.6. As for equity and bond factor severe reduction in purchasing power.
premiums, market risk is not able to explain any of
these GFPs, with an average alpha of 4.2% per The two periods in between, with a positive inflation
annum (t value, 15.7) between 1875 and 2021. In just below current central bank targets of 2% and a
sum, asset class (i.e., equities and bonds) and equity, mild overshoot of those targets to 4%, are both good
bond, and GFPs are economically sizable and statis- for equities and bonds, with nominal returns on equi-
tically significant and robust phenomena over the last ties of 11.0%, well above their unconditional average,
147 years. and nominal returns on bonds in line with their
unconditional average. In real terms, returns on these
asset classes are also good, with the undershoot
Investment Returns During scenario being even somewhat better with a 9.8%
real return for equities and 3.4% for bonds. The real
Deflationary and High-Inflation return on the 60/40 multi-asset portfolio is 7.2% and
Periods 5.6% per annum for the 0%–2% and 2%–4% inflation
In this section, we examine the performance of asset buckets. In other words, positive but low consumer
classes and factor premiums over inflationary price increases are good for nominal investment
12
Table 3. Investment Returns by Inflation Regime

Inflation Bucket
All <0% 0%–2% 2%–4% >4% Wald
Years 147 23.1 42.9 34.9 46.1 –

Likelihood (%) 100 15.7 29.2 23.7 31.4 –
Inflation (%) 3.2 –3.1 1.2 2.8 8.5 –
A1. Asset classes (nominal return)
Equities 8.4 2.4 11.1 11.0 6.9 3.9
Bonds 4.5 5.2 4.7 4.5 3.9 0.6
Cash 3.4 2.8 2.7 3.4 4.2 3.5
60/40 6.8 3.5 8.5 8.4 5.7 3.0
A2. Asset classes (real return)
Equities 5.1 5.5 9.8 8.2 –1.7 4.2
Bonds 1.2 8.4 3.4 1.6 –4.6 24.9
Cash 0.1 5.9 1.5 0.5 –4.3 37.0
60/40 3.5 6.7 7.2 5.6 –2.9 6.5
B. Equity factors
Value 3.5 4.2 3.0 2.4 4.5 0.5
Momentum 6.9 5.9 6.1 7.4 7.7 0.3
Low Risk 6.5 7.0 7.4 5.3 6.2 0.3
Quality 2.5 3.6 3.3 2.2 2.0 0.5
MFE 5.1 5.9 5.1 4.5 5.3 0.4
C. Bond factors
Value 2.4 0.8 2.0 0.9 4.8 1.7
Momentum 3.9 5.6 5.0 6.0 0.4 2.7
Low Risk 4.4 10.3 2.2 3.0 5.8 2.2
Carry 7.4 7.8 7.7 8.5 6.1 0.5
MFB 4.6 5.3 4.7 4.9 4.0 0.4
D. Global factors
Value 2.3 1.2 3.3 1.9 2.1 1.8
Momentum/Trend 7.4 6.8 7.4 5.7 8.9 2.9
Low Risk 1.6 1.0 1.1 2.1 2.0 1.2
Carry 5.5 4.7 6.0 7.1 4.3 2.4
GFP 4.2 3.4 4.4 4.2 4.3 1.3
E. Assets þ Factors (real return)
Equities þ MFE 7.7 8.5 12.4 10.4 1.0 4.1
Bonds þ MFB 3.5 11.0 5.7 4.1 –2.6 21.3
Cash þ GFP 4.3 9.3 5.9 4.7 0.0 24.5
60/40 þ All factors 6.9 10.2 10.6 8.7 0.4 6.3
Notes: The table shows historical returns on asset classes and factor premiums bucketed by four contemporan-
eous global inflation categories (from deflation, <0%, to high inflation, >4%). The sample period is 1875 to
2021. Returns are in percentages per annum. Significance at the 5% significance level is denoted with and sig-
nificance at the 1% significance level with . A Wald F test is used to test for significant average return differ-
ences across the different inflationary states. All standard errors are computed using the Newey–West
procedure. In panel E, we proxy a long-only investment with a 100% allocation to the market and a 50% alloca-
tion to the long-short multi-factor premium strategy in that asset class.
returns, as well as those adjusted for purchasing 12 months’ overlapping observations) that tests
power. whether the variation of returns across inflation
scenarios is statistically significant. Nominal equity,
The last column summarizes the results of a Wald cash, and multi-asset returns vary significantly across
test (adjusted for serial correlation due to the use of inflationary regimes, while the variation in nominal

Figure 3. Nominal and Real 60/40 Returns across Inflation Regimes, 1875–2021
10.0
8.5 8.4
8.0 7.2
6.7
5.6 5.7
6.0
Return (% p.a.)
4.0 3.5
2.0
0.0
-2.0
-4.0 -2.9
Deflation (< 0%) Low inflation (0-2%) Mild overshoot (2-4%) High inflation (> 4%)
Nominal Real
Notes: The figure shows the historical average nominal return (grey bars) and real return (black bars) on the 60/40 global equity/bond
portfolio across inflationary regimes. The sample period is 1875 to 2021.
bond returns is not significant. Because of the huge indicate that the observed variation across inflation
differential of average inflation during the inflation scenarios is not statistically significant, even for some
scenarios, the variation in real returns is significant factors sometimes believed to vary with inflation
for all asset classes. such as Momentum and Value. These findings for
Value, Momentum, and Quality generally align with
Figure 3 summarizes these insights by depicting the the findings of Neville et al. (2021) over a shorter
average nominal and real annual return on the multi- sample and with a different inflation definition.16
asset portfolio across the inflationary regimes.
Inflation just below or above the inflation target of Bond factors also deliver positive returns across all
2% is good for investors in both nominal and real four inflationary regimes, although they generally
terms. Deflation is relatively bad for nominal returns, perform in a more mixed way across the regimes.
but good in real terms. Nominal returns during peri- Value in bonds is performing well when inflation is
ods with high inflation seem acceptable but in real high (4.8%), but not in deflationary periods (0.8%).
terms are dramatically negative. That said, its return variations across regimes are
insignificant as reflected in the Wald test.
As outlined in the previous section, factor premiums Momentum in bonds, on the other hand, is perform-
offer significant and diversifying returns over asset ing well in each period (5.6%, 5.0%, and 6.0%) except
class premiums. Many investors distinguish factor when inflation is high (0.4%), and its variation across
premiums in their strategic asset allocation, and it inflationary regimes is significant. The Low Risk fac-
therefore is of interest to examine the variation in tor performs especially well in times of deflation
factor premiums across inflationary regimes. Panels (10.3%, compared to a 4.4% unconditional average)
B, C, and D in Table 3 show the factor returns for and inflation (5.8%), but again return variations
each of the four inflation buckets. Interestingly, the across regimes are not significant. Further, the Carry
factor performances do not seem to depend much factor is not much affected by inflation regimes. As
on the level of inflation, in contrast to the asset class for equities, the multi-factor bond portfolio performs
returns. Indeed, the multi-factor equity portfolio per- consistently high across inflation regimes, delivering
forms at 5.9% in deflation, 5.3% in inflationary peri- positive returns in each inflationary regime.
ods, and 5.1% and 4.5% when inflation is just below
or above the central bank targets, respectively. The GFPs also deliver positive and consistent returns
Variations across inflation buckets at the individual across all four inflationary regimes. Seemingly, Value
factor level are somewhat higher but remain quite has lower returns during deflationary periods with
close to their unconditional averages. The Wald tests 1.2% per annum, but the Wald test fails to reject the
14
Figure 4. Factor Premiums across Inflation Regimes, 1875–2021
7.0
6.4 6.4
6.1 6.2
5.9
6.0 5.7 5.7
5.4 5.3
5.0
Return (% p.a.)
4.2 4.1
3.9 3.8
4.0
3.0 2.8
2.1
2.0 1.7
1.0
0.0
Deflation (< 0%) Low inflation (0-2%) Mild overshoot (2-4%) High inflation (> 4%)
Value Momentum Low risk Quality/Carry
Notes: The figure shows the historical average return on Value, Momentum, Low Risk, and Quality/Carry factor premiums across
inflationary regimes. Factor premiums are computed as the equally weighted average of equity, government bond, and global factor
premiums. The sample period is 1875 to 2021.
null hypothesis of no significant return variations involved with different instruments, and the invest-
across the four inflationary regimes (Time-Series). or’s willingness and ability to deal with derivatives
Momentum varies significantly across the regimes, as what an optimal implementation in practice
it performs best during periods with high inflation would be.
(8.9% per annum), which is consistent with the find-
Panel E considers these portfolio implementations. For
ings of Neville et al. (2021) who examine the post-
equities, we replace its “passive” allocation with an
1926 period. However, it delivers positive returns on
approximate long-only allocation to the multi-factor
average across all four regimes. Low Risk performs
equity strategy. Similarly, for bonds we replace its
above average when inflation is above 2% but does
“passive” allocation with an approximate long-only
not display significant return variation across regimes.
allocation to the multi-factor bond strategy.17 We add
Carry also delivers positive and fairly similar average
the GFP strategy, which represents a long-short strat-
returns across the regimes. Similar to the equity and egy in derivatives across the major global markets, to
bond factors, the multi-factor strategy in the multi- the cash portfolio to generate a factor-based absolute
asset universe is virtually immune to inflation or return macro strategy. We see that the multi-factor
deflation shocks, at least on average, with excess long-only equity investor can increase their full sample
returns ranging from 3.4% to 4.4% per annum. real return from 5.1% to 7.7% per annum, the long-
The factor returns are the return differential between only bond investor from 1.2% to 3.5%, the cash
investor from 0.1% to 4.3%, and the 60/40 multi-
a long portfolio with the highest factor exposures
asset portfolio from 3.5% to 6.9% per annum, assum-
and a short portfolio with the lowest factor expo-
ing that the GFP is added via an overlay on 20% of
sures. The excess returns on these “zero investment”
the portfolio’s value. Further, the long-only factor
portfolios are therefore the same in nominal and real
strategies consistently add across inflationary regimes,
terms. In practice, an investor would need to hold
with for example the factor-enhanced equity (multi-
either a cash position to fund these long-short strat-
asset) allocation managing to make up for the losses
egies, for example, as a derivatives-based overlay on
of the conventional portfolio in the high-inflation
top of the conventional asset allocation, or replace
regime, as its average real return equals 1.0% (0.4%).
the conventional asset allocations to equities and
bonds with the long side of each factor. It depends The factor premiums averaged across equity, govern-
on the nature of the factor, the trading costs ment bonds, and the GFPs for each of the four

Table 4. Inflation Regimes with 3-Year Investment Horizon

Inflation Bucket
All <0% 0%–2% 2%–4% >4% Wald
Years 147 21.0 47.8 36.3 41.9 –

Likelihood (%) 100 14.3 32.5 24.7 28.5 –
Inflation (%) 3.2 –2.2 1.4 2.9 8.3 –
Equities 8.2 4.7 9.5 12.3 5.5 2.5
Bonds 4.5 4.6 4.1 5.1 4.3 0.6
Cash 3.4 3 2.7 3.8 4.1 1.8
60/40 6.7 4.6 7.3 9.4 5.0 2.1
Equities 5.0 6.9 8.1 9.4 –2.7 3.1
Bonds 1.3 6.8 2.6 2.2 –3.9 17.1
Cash 0.2 5.2 1.3 0.9 –4.2 25.4
60/40 3.5 6.9 5.9 6.5 –3.2 4.5
B. Equity factors
Value 3.6 6.5 2.2 2.8 4.5 1.2
Momentum 6.9 7.2 5.4 8.4 7.1 1.2
Low Risk 6.4 7.0 6.7 7.3 4.8 0.6
Quality 2.5 – 3.1 2.5 1.8 0.6
MFE 5.2 7.0 4.6 5.5 4.6 1.0
C. Bond factors
Value 2.4 1.1 0.9 2.4 4.7 2.2
Momentum 4.0 5.6 5.2 5.7 0.3 2.1
Low risk 4.3 8.9 1.3 4.7 5.8 3.0
Carry 7.5 7.8 7.4 9.0 6.1 0.8
MFB 4.6 5.3 4.2 5.7 3.7 1.5
D. Global factors
Value 2.3 0.8 2.3 2.5 2.8 1.6
Momentum/Trend 7.4 7.7 6.8 6.5 8.7 1.1
Low Risk 1.6 0.7 0.7 2.7 2.0 2.7
Carry 5.6 4.4 5.8 7.2 4.4 2.2
GFP 4.2 3.4 3.9 4.7 4.5 2.2
Equities þ MFE 7.6 10.4 10.4 12.2 –0.4 3.1
Bonds þ MFB 3.6 9.5 4.7 5.1 –2.1 15.2
Cash þ GFP 4.4 8.6 5.2 5.6 0.3 7.8
60/40 þ All factors 6.8 10.7 8.9 10.3 –0.2 4.4
Notes: The table shows historical returns on asset classes and equity, bond, and global macro-factor premiums
bucketed by four global inflation categories (from deflation, <0%, to high inflation, >4%) using a 3-year inflation
classification and investment holding period. The sample period is 1875 to 2021. Returns are in percentages per
annum. Note that for equity Quality, which starts in 1940, no 3-year deflationary observations are available.
Significance at the 5% significance level is denoted with and significance at the 1% significance level with .
A Wald F test is used to test for significant average return differences across the different inflationary states.
All standard errors are computed using the Newey–West procedure.
inflationary regimes are displayed in Figure 4. The performs relatively well during high-inflation periods
Value factor is the weakest stand-alone (while diver- when conventional asset classes do poorly. On the
sifying well to the other factors; see, for example, other hand, Quality/Carry perform slightly worse dur-
Baltussen, Swinkels, and Van Vliet 2021), but ing inflationary times, but better in each of the other
16
inflationary regimes. Low Risk performs well espe- In sum, our findings show that asset class premiums
cially in the extremes, that is, deflationary, or high- vary significantly across the inflationary regimes in
inflation regimes, and is weaker in the middle two nominal and especially real terms. Equities and bonds
that are goldilocks scenarios for equities and bonds. on average yield lower nominal returns during peri-
Finally, Momentum performs consistently and well ods of high inflation, causing negative real returns. By
across the inflationary regimes. contrast, equity, bond, and GFPs are generally posi-
tive across high-inflation regimes, displaying generally
Over longer-term horizons, the risk of equities, limited variation across, while they enhance nominal
bonds, and factor premiums can be different. For and real asset class returns in (approximated) long-
example, Ely and Robinson (1997) and Schotman and only asset class implementations. These results sug-
Schweitzer (2000) show that equities offer better gest that equity and bond market premiums could be
inflation protection over the long term. Table 4 explained by offering a compensation for inflation
shows the same results as Table 3, but now moving risk, while factor premiums cannot. This leaves room
from a 1-year to a 3-year evaluation horizon. This for further research into other explanations as to
hardly affects the observation frequency of each of why factor premiums exist. A portfolio implication is
the regimes. The deflation regime occurs 14% of the that a balanced exposure to multiple factor premiums
time (16% for 1 year), in which the 60/40 portfolio is difficult to beat with dynamic factor timing based
offers the highest annualized real return across each on inflationary cycles.19
of the regimes. During a high-inflation regime (>4%)
the annualized real equity return is still negative,
even somewhat lower than on a 1-year evaluation Dissecting Inflationary Regimes:
horizon. This is contrary to the hypothesis that equi- Stagflation Hurts
ties offer higher real returns on longer investment
In the previous section, we observed that real returns
horizons.18 The equity, bond, and GFPs remain stable
during inflationary periods are negative for investors
across the different inflation regimes.
in stocks and bonds but that equity, bond, and GFPs
In Appendix A.2, we show that the results are robust are mostly resilient during different inflation scen-
across (i) different definitions of inflationary regimes, arios. In this section, we further split up the “bad
including a 3% or 5% (instead of 4%) high inflation times” of high-inflation episodes as well as deflation-
cutoff, (ii) the use of annual changes in inflation rates ary periods based on other economic or financial
(as measure of unexpected inflation), (iii) the use of market circumstances. Not all inflationary regimes
only U.S. (instead of global) inflation, and (iv) across are alike, and hence the question is how investment
the first (i.e., 1875–1948) and second half (i.e., returns behave over various sub-regimes within high-
1949–2021) of our sample period. Noteworthy inflation and deflationary regimes.
observations include poorer nominal and real global
To this end, we divide the periods based on five dif-
equity returns the higher the high inflation cutoff
ferent characteristics: (1) recession or expansion, (2)
and stronger returns on equity factors (especially
falling or rising earnings growth, (3) bear or bull
Momentum and Low Risk) and stronger returns on
equity markets, and (4) increasing or decreasing inter-
two value factors (Bond Value and Global Value) dur-
est rates, and (5) increasing or decreasing inflation.
ing times of falling inflation. During times of strongly This includes stagflationary episodes with both high
rising inflation, Equity Momentum and Global inflation and recessions. We date recessions based
Momentum (Trend) have higher returns. Finally, the on business cycle data from the National Bureau of
results are robust across two subsamples with some Economic Research20 by requiring that at least six of
noteworthy observations. During the early period the months in a rolling 12-month window are classi-
from 1875 to 1948, bonds are doing even worse dur- fied as recession.21 In total, there have been 30
ing inflationary periods, both in nominal and real recessions over the period 1875 to 2021 with an
terms, whereas Bond Momentum shows low but average duration of 17 months. We define increasing
positive returns. Note that in our second sub-period, or decreasing interest rates based on the sign of the
1949 to 2021, there are less than 2 years with defla- 12-month change in global government bond rates
tion observations, which is not sufficient for report- and a bull (bear) equity market based on the sign of
ing meaningful average realized returns. In general, global equity returns.
factors premiums are robust across the two subsam-
ples offering positive returns across the different The average returns per high inflation sub-regime are
inflationary regimes. displayed in Table 5. Panel A1 (A2) of Table 5 shows

that nominal (real) returns on equities are particularly returns on the multi-asset portfolio remain poor,
bad during recessions with high inflation, or stagfla- although better than during NBER stagflations, with
tionary episodes. The nominal (real) return on the –5.8% versus –11.7%. The MFB premiums are posi-
multi-asset portfolio is –2.2% (–11.7%) per annum, tive, but relatively low with 2.5%, because Momentum
compared to 8.2% (–0.1%) during expansionary peri- and Carry have relatively low returns. MFE and GFP
ods with high inflation. Recessions likely lead to lower show average returns, leading to a –3.2 real return on
expected corporate cash flows, which dominate a a multi-asset portfolio that includes factors.
decrease in the discount rate, leading even to negative
nominal stock returns of –7.1% per annum. Decreasing equity markets in periods with high inflation
Recessionary periods are somewhat better for nominal are especially disastrous with an annualized real return
bonds than expansionary periods with nominal returns of –28.8%. At the same time, bonds suffer in real terms
of 5.1% versus 3.6% per annum. Decreasing interest during high inflation and falling equity market episodes
rates lead to positive marked-to-market gains. The with –9.0% per annum negative real returns, yielding a
negative real return for equities of –16.6% suggests –20.9% real return on the multi-asset portfolio. For
that equities are a particularly bad inflation hedge dur- both bear and bull equity markets in times of inflation,
ing stagflationary periods. While expansionary periods a diversified portfolio of factor premiums yields robust
in general tend to be good for investors, this does not performance enhancements, thereby alleviating the pain
hold when inflation is high. Real returns on stocks are of high inflation. Again, all factor premiums yield posi-
marginally positive (2.9% per annum) and real returns tive returns, except for Momentum in bonds and Low
on bonds are deeply negative (–4.7%). Even though Risk across assets during high inflation bear markets.
the economy is doing well, times of high inflation are When we sub-condition on changes in interest rates,
generally not good for investors. it becomes clear that increasing interest rates cause
Panels B, C, and D contain the result of the factor pre- more pain (real –6.8% per annum) to a conventional
miums and give a completely different picture. The multi-asset portfolio than decreasing interest rates
multi-factor equity portfolio has a return of 5.4%, the (real 2.2% per annum), as both equities and bonds suf-
multi-factor bond portfolio 4.7%, and the GFP portfolio fer in real terms (–6.0% and –8.0% per annum,
a return of 4.1% during stagflation periods. If anything, respectively). By contrast, during decreasing rate peri-
the stagflation performances for factor premiums in ods equities and bonds experience materially better
equities and bonds are even better than during expan- real returns (3.9% and –0.3% per annum). Average
sions in high-inflation periods, albeit not statistically sig- returns on factor premiums are again good across
nificantly different (unreported). All factor premiums sub-regimes, but generally a bit better during when
perform well during stagflations except for Momentum rates increase. Especially Momentum in equities and
in bonds, which returns –2.4% on average. Similarly, Trend following stand out during these episodes, while
GFPs perform well, with only Low Risk returning zero Low Risk in equities and Value in bonds benefit more
and all other factors returning positively. from declining rates in times of inflation. Again, factor
premiums materially help to soften to burden of high
Panel E of Table 5 also show that investors who are, inflation for traditional portfolios (Table 5).
for example, worried about achieving negative real
returns during stagflation periods may improve their The final sub-regime is that of increasing inflation dur-
asset allocation by including factors across asset ing periods in which inflation is above 4%. This hap-
classes. This would help their portfolio to a certain pens in 28.5 out of 46.1 years. Nominal and real equity
extent from these adversary business cycle conditions. and bond returns are worse than during the average
Even though –8.3% per annum in real terms can hardly year in a high-inflation regime. Consequently, real
be called a success, it is substantially better than the returns on the multi-asset portfolio are –5.0%, worse
alternative of –11.7% per annum. Note that the factor than with decreasing inflation but not as bad as during
series are all long-short factors that do not include stagflation periods. Again, factor premiums perform
trading costs, and investors need to construct efficient consistent and well, with government bond factor pre-
ways to exploit these factor strategies in practice. miums being particularly high, mostly due to Low Risk
and Carry outperforming. As before, a diversified port-
The NBER definition of a recession is based on nega- folio of factor premiums yields robust performance
tive GDP growth, but we also examine declining enhancements, thereby alleviating the pain of high
(annual) earnings of the stock market (based on data inflation. The robust results on factor investing do not
from Shiller’s website), as that part of the economy mean that investing in factor premiums is without risk.
may be more relevant for equity investors. Real For example, Blitz (2021) suggests that quant equity
18
Table 5. A Deeper Look at Returns During High-Inflation Periods

High Inflation (>4%)
NBER Earnings Equity Rates Inflation
# " # " # " " # " #
Years 10.9 35.3 16.8 29.3 12.8 33.4 25.9 20.1 28.5 17.6
Likelihood 23.4 76.6 36.5 63.5 27.8 72.2 56.2 43.8 61.8 38.2
Inflation (%) 9.5 8.3 9.8 7.8 10.8 7.7 9.0 8.0 9.1 7.6
Equities –7.1 11.2 3.2 9.0 –18.0 16.4 3.0 11.9 5.6 9.0
Bonds 5.1 3.6 5.0 3.3 1.8 4.7 1.0 7.7 2.4 6.4
Cash 5.4 3.8 4.6 4.0 4.3 4.2 4.7 3.6 3.8 4.8
60/40 –2.2 8.2 4.0 6.2 –10.1 11.7 2.2 10.2 4.1 7.5
Equities –16.6 2.9 –6.7 1.2 –28.8 8.7 –6.0 3.9 –3.5 1.4
Bonds –4.4 –4.7 –4.8 –4.5 –9.0 –3.0 –8.0 –0.3 –6.8 –1.1
Cash –4.1 –4.4 –5.2 –3.8 –6.5 –3.5 –4.3 –4.4 –5.3 –2.8
60/40 –11.7 –0.1 –5.8 –1.6 –20.9 4.0 –6.8 2.2 –5.0 0.0
B. Equity factors
Value 3.0 4.9 3.0 5.3 6.2 3.8 3.4 5.8 4.3 4.8
Momentum 7.2 7.9 7.3 8.0 8.0 7.6 10.0 4.8 8.7 6.1
Low Risk 7.6 5.8 6.1 6.4 5.6 6.5 4.9 8.0 4.7 8.8
Quality 4.6 1.4 2.7 1.6 3.1 1.6 2.5 1.4 1.9 2.2
MFE 5.4 5.2 4.8 5.5 5.7 5.1 5.3 5.1 5.1 5.5
C. Bond factors
Value 5.8 4.5 6.1 4.1 5.9 4.4 2.8 7.4 1.9 9.6
Momentum –2.4 1.3 –3.8 2.9 –5.9 2.8 1.0 –0.2 1.7 –1.7
Low Risk 14.3 3.3 10.2 3.5 1.4 7.8 7.0 4.9 6.1 5.3
Carry 5.5 6.2 1.4 8.8 9.5 4.4 7.0 4.3 7.4 3.9
MFB 4.7 3.7 2.5 4.8 1.6 4.9 4.2 3.7 4.1 3.7
D. Global factors
Value 3.9 1.6 3.7 1.2 4.0 1.4 1.4 3.1 1.4 3.3
Momentum/Trend 7.3 9.4 8.0 9.4 9.6 8.6 10.4 6.9 9.1 8.5
Low Risk 0.0 3.3 1.4 2.3 –0.3 2.8 1.9 2.1 1.7 2.5
Carry 5.2 4.1 3.1 5.0 2.8 4.9 5.0 3.5 5.2 3.0
GFP 4.1 4.6 4.0 4.5 4.0 4.4 4.7 3.9 4.3 4.3
E. Assets and Factors (real return)
Equities þ MFE –13.9 5.5 –4.3 4.0 –26.0 11.3 –3.4 6.5 –1.0 4.2
Bonds þ MFB –2.1 –2.9 –3.6 –2.1 –8.2 –0.6 –5.9 1.6 –4.8 0.8
Cash þ GFP 0.0 0.2 –1.2 0.7 –2.5 0.9 0.4 –0.5 –1.0 1.5
60/40 þ All factors –8.3 3.1 –3.2 2.4 –18.1 7.4 –3.4 5.3 –1.6 3.7
Notes: The table shows historical returns on asset classes and factor premiums during high-inflation regimes, sub-bucketed based
on recessions or expansion, increasing or decreasing corporate earnings, equity bull or bear markets, increases or decreases in
long-term rates, or increases or decreases in 1-year inflation rates. The sample period is 1875–2021. Returns are in percentages
per annum. NBER: National Bureau of Economic Research.
factors follow their own “quant cycle,” which is unre- on a 0 to 100 scale in percentage. We have seen that
lated to business cycle variables that we examine here. average nominal equity returns are negative in a stag-
flation scenario, with also most observations returning
To test for consistency in returns, Table 6 shows the
negative and only 40.8% of all annual equity returns
probability of a positive annual nominal return for
being positive. Panel B shows that the annual positive
each high inflation sub-regime. The probabilities are
returns of the combined MFE equity factors vary

Table 6. A Deeper Look at Returns During High Inflation Periods: Frequencies

High Inflation (>4%)
# " # " # " " # " #
Years 10.9 35.3 16.8 29.3 12.8 33.4 25.9 20.1 28.5 17.6
Likelihood 23.4 76.6 36.5 63.5 27.8 72.2 56.2 43.8 61.8 38.2
Inflation (%) 9.5 8.3 9.8 7.8 10.8 7.7 9.0 8.0 9.1 7.6
A. Asset classes (nominal return)
Equities 40.8 82.0 65.3 76.4 0.0 100.0 67.2 78.9 74.3 69.2
Bonds 82.3 78.0 85.1 75.5 65.4 84.3 63.3 99.2 73.4 88.2
Cash 98.5 95.5 98.0 95.2 95.4 96.5 98.4 93.4 96.8 95.3
60/40 54.6 83.7 72.8 79.2 18.3 99.3 69.8 86.0 76.3 77.7
B. Equity factors
Value 57.7 67.1 62.9 66.1 69.3 63.3 59.8 71.5 64.0 66.4
Momentum 70.8 79.4 73.3 79.8 77.1 77.5 82.0 71.5 80.4 72.5
Low Risk 70.8 73.8 73.3 72.9 75.8 72.0 69.5 77.7 69.3 79.1
Quality 47.7 47.8 51.5 45.6 45.8 48.5 47.9 47.5 40.9 58.8
MFE 76.2 85.1 77.7 86.0 82.4 83.3 85.2 80.2 83.0 82.9
C. Bond factors
Value 56.9 64.3 60.4 63.8 58.8 64.0 60.1 65.7 54.7 75.4
Momentum 50.0 54.4 44.6 58.4 42.5 57.5 54.7 51.7 54.1 52.1
Low Risk 56.2 51.1 51.5 52.7 60.1 49.3 55.9 47.5 46.5 61.6
Carry 66.2 71.9 62.9 74.9 61.4 74.0 73.3 66.9 75.4 62.6
MFB 70.0 76.1 63.4 81.2 62.7 79.3 78.8 69.4 74.9 74.4
D. Global factors
Value 71.5 64.5 74.8 61.3 77.1 62.0 60.1 64.0 60.8 74.9
Momentum/Trend 89.2 91.7 89.1 90.2 87.6 92.5 95.5 85.5 92.7 88.6
Low Risk 29.2 72.6 63.4 61.8 41.8 70.3 58.5 67.4 61.4 74.0
Carry 77.7 72.3 69.8 75.8 66.0 76.5 79.1 66.5 81.0 61.6
GFP 87.7 95.5 92.6 94.3 89.5 95.3 93.6 93.8 93.9 93.4
Notes: The table shows historical frequencies of positive annual returns on asset classes and factor premiums during high-inflation
regimes, sub-bucketed based on recessions or expansion, increasing or decreasing corporate earnings, equity bull or bear markets,
increases or decreases in long-term rates, or increases or decreases in 1-year inflation rates. The sample period is 1875 to 2021.
Frequencies are expressed in percentages per annum. NBER: National Bureau of Economic Research.
between 76% and 85%. This means that the average strategies offer good and consistent protection during
premiums are positive and consistent across all high- high-inflation periods. In this more granular analysis
inflation regimes. On an individual factor level, Quality within a wide range of high-inflation periods, Trend is
is somewhat less consistent, with positive return rates again the most persistent factor, which extends and
below 50%. Momentum is most consistent, with posi- supports these earlier findings.
tive return rates around 75%. Still, due to diversifica-
tion benefits, the combination of factors offers more Next, we examine performances over deflationary
consistent positive returns than the best individual sub-regimes, as shown in Table 7. The total defla-
equity factor (Momentum). Panel C shows a similar tionary period is 23.1 years, about half of the
consistent picture for bond factor returns: The aver- 46.1 years that have inflation above 4%. This
age annual combined MFB returns are positive, with means that the sub-regimes are sometimes small,
probabilities varying between 63% and 81%. The most with several of them having less than 10 years of
consistent bond factor is Carry. Finally, the GFPs also observations, reducing the accuracy of the displayed
show consistently positive annual returns during high- averages. Deflationary expansions are relatively
inflation regimes, with probabilities ranging between good for investors, with a 10.4% real return per
87% and 95%. Neville et al. (2021) find that trend annum for the 60/40 portfolio, while deflationary
20
Table 7. A Deeper Look at Returns During Deflationary Periods

Deflation (<0%)
# " # " # " " # " #
Years 15.8 7.3 14.6 8.5 8.2 14.9 6.4 16.7 6.1 17.0
Likelihood (%) 68.2 31.8 63.2 36.8 35.4 64.6 27.8 72.2 26.4 73.6
Inflation (%) –3.3 –2.7 –3.7 –2.2 –3.9 –2.6 –2.7 –3.3 –2.7 –3.2
Equities –1.3 10.2 –0.6 7.5 –12.1 10.3 –4.3 4.9 11.1 –0.7
Bonds 4.9 5.9 5.5 4.8 4.1 5.9 1.7 6.6 5.9 5.0
Cash 3.0 2.6 2.9 2.8 2.9 2.8 2.8 2.8 2.6 2.9
60/40 1.6 7.7 2.2 6.0 –4.1 7.8 –1.2 5.4 8.2 1.9
Equities 2.0 13.0 3.0 9.7 –8.2 13.0 –1.6 8.2 13.8 2.5
Bonds 8.2 8.6 9.1 7.0 8.1 8.5 4.4 9.9 8.6 8.2
Cash 6.3 5.3 6.5 4.9 6.9 5.4 5.5 6.1 5.3 6.2
60/40 4.9 10.4 5.8 8.1 –0.2 10.4 1.5 8.7 10.9 5.2
B. Equity factors
Value 8.3 –4.5 7.8 –2.0 8.0 2.1 14.1 0.4 7.1 3.2
Momentum 7.9 1.5 7.2 3.7 19.1 –1.4 13.1 3.1 –14.2 13.1
Low Risk 2.6 16.1 3.7 12.6 0.8 10.6 6.3 7.2 7.8 6.6
Quality 5.3 –0.1 3.0 4.2 6.6 2.4 4.4 3.0 – 3.6
MFE 6.7 4.3 6.7 4.6 9.7 3.9 11.2 3.9 0.2 8.0
C. Bond factors
Value 1.3 –0.3 1.9 –1.2 4.8 –1.4 2.2 0.3 –2.7 2.0
Momentum 5.6 5.5 5.4 5.9 6.4 5.1 10.1 3.9 1.6 7.0
Low Risk 12.2 7.2 12.5 6.1 15.0 6.8 2.7 11.9 7.6 11.3
Carry 8.6 6.2 7.7 7.9 9.0 7.2 9.3 7.2 6.1 8.4
MFB 5.7 4.4 5.7 4.6 7.6 4.0 7.0 4.6 2.1 6.4
D. Global factors
Value 2.0 –0.4 1.5 0.8 3.3 0.1 1.8 1.0 –3.0 2.7
Momentum/Trend 7.7 4.9 7.6 5.3 9.2 5.5 5.8 7.2 7.5 6.5
Low Risk 0.5 1.9 1.1 0.6 –0.6 1.7 0.8 1.0 3.6 0.0
Carry 5.5 2.9 4.3 5.3 4.3 4.9 4.2 4.9 3.3 5.2
GFP 3.9 2.3 3.6 3.0 4.1 3.0 3.1 3.5 2.9 3.6
Equities þ MFE 5.4 15.2 6.4 12.0 –3.4 15.0 4.0 10.2 13.9 6.5
Bonds þ MFB 11.1 10.8 12.0 9.3 11.9 10.5 7.9 12.2 9.7 11.4
Cash þ GFP 10.2 7.6 10.1 7.9 11.0 8.4 8.6 9.6 8.2 9.8
60/40 þ All factors 8.4 13.9 9.3 11.5 3.6 13.8 6.2 11.7 12.8 9.2
Notes: The table shows historical returns on asset classes and factor premiums during high-inflation regimes, sub-bucketed based
on recessions or expansion, increasing or decreasing corporate earnings, equity bull or bear markets, increases or decreases in
long-term rates, or increases or decreases in 1-year inflation rates. The sample period is 1875 to 2021. Returns are in percentages
per annum. NBER: National Bureau of Economic Research.
recessions are slightly positive (1.6%) in nominal that coincide with equity bear markets. The equity,
terms, but better in real terms (4.9%). For the multi- bond, and global factors multi-factor combinations
asset investor, interest rate increases are worse do well for each of the sub-regimes during defla-
than decreases (with equities returning negatively) tionary periods, especially equity and bond factors
and equity market bear markets worse than bull during the states with poor equity or bond returns—
markets. Most troubling are deflationary episodes deflationary recession or bear markets. Further,

Table 8. A Deeper Look at Returns During Deflationary Periods: Frequencies

Deflation (<0%)
# " # " # " " # " #
Years 15.8 7.3 14.6 8.5 8.2 14.9 6.4 16.7 6.1 17.0
Likelihood (%) 68.2 31.8 63.2 36.8 35.4 64.6 27.8 72.2 26.4 73.6
Inflation (%) –3.3 –2.7 –3.7 –2.2 –3.9 –2.6 –2.7 –3.3 –2.7 –3.2
A. Asset classes (nominal return)
Equities 50.8 94.3 50.9 88.2 0.0 100.0 50.6 70.0 86.3 56.9
Bonds 94.2 100.0 94.3 99.0 89.8 99.4 87.0 99.5 98.6 95.1
Cash 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
60/40 66.7 97.7 66.3 94.1 33.7 100.0 67.5 80.0 94.5 70.1
B. Equity factors
Value 61.9 40.9 60.6 46.1 66.3 49.2 73.5 47.5 57.5 54.5
Momentum 77.2 58.0 77.7 59.8 88.8 61.5 76.6 69.0 41.1 81.9
Low Risk 54.5 88.6 57.7 78.4 49.0 74.3 64.9 65.5 64.4 65.7
Quality 7.9 4.5 5.1 9.8 6.1 7.3 11.7 5.0 0.0 9.3
MFE 74.6 73.9 74.9 73.5 78.6 72.1 83.1 71.0 52.1 82.4
C. Bond factors
Value 49.7 52.3 53.1 46.1 61.2 44.7 55.8 48.5 34.2 56.4
Momentum 72.0 67.0 72.6 66.7 66.3 72.6 88.3 63.5 50.7 77.5
Low Risk 28.0 37.5 33.1 27.5 37.8 27.4 15.6 37.0 30.1 31.4
Carry 83.6 80.7 80.0 87.3 83.7 82.1 85.7 81.5 74.0 85.8
MFB 75.7 80.7 76.0 79.4 80.6 75.4 85.7 74.0 56.2 84.8
D. Global factors
Value 61.4 37.5 56.0 50.0 68.4 45.8 61.0 51.0 30.1 62.3
Momentum/Trend 84.7 87.5 82.9 90.2 92.9 81.6 83.1 86.5 82.2 86.8
Low Risk 54.5 54.5 60.0 45.1 45.9 59.2 51.9 55.5 75.3 47.1
Carry 76.7 75.0 72.6 82.4 68.4 80.4 66.2 80.0 65.8 79.9
GFP 89.4 75.0 86.9 81.4 91.8 81.0 85.7 84.5 74.0 88.7
Notes: The table shows historical frequencies of positive annual returns on asset classes and factor premiums during deflation
regimes, sub-bucketed based on recessions or expansion, increasing or decreasing corporate earnings, equity bull or bear markets,
increases or decreases in long-term rates, or increases or decreases in 1-year inflation rates. The sample period is 1875 to 2021.
Frequencies are expressed in percentages per annum. NBER: National Bureau of Economic Research.
individual factor performances vary, although we recession, equities have positive annual returns
have to be cautious as some samples are fairly small. 50.8% of the time, whereas equities almost always
For example, equity Value does rather poorly during go up with deflationary expansion periods (94.3%).
deflationary expansion periods (–4.5%) when equity The highest return on 60/40 is made during the
Low Risk does particularly well (16.1%). The same deflation scenario when prices start to rise again.
holds for bond Value and, to a lesser extent, for During these “good times,” the 60/40 portfolio has
Value across assets. The GFP portfolio does not positive annual returns 94.5% of the time. The com-
vary a lot across sub-regimes, ranging between 2.3% bined equity factors (MFE) tend to do well across
and 4.1%. Overall, we observe that especially equity most deflation regimes, as shown in panel B. The
and bond factors perform tend to do well during the same applies for bond factors and global factors, as
bad times in deflationary cycles but that these bad shown in panels C and D. This again confirms the
times are much milder than during stagflations. diversification benefits of factor investing across
inflationary regimes.
To test for consistency in returns, Table 8 shows the
probability of a positive annual nominal return for Figure 5, panel A summarizes the results of the high-
each deflation sub-regime. During a deflationary inflation bad times presented in Table 4. Clearly, for the
22
Figure 5. Stagflationary Times and Investment Returns, 1875–2021
A. High inflation “bad times”
-5
Real return (% p.a.)
-10
-15
-20
-25
NBER ↓ Earnings ↓ Equity ↓ Rates ↑ Inflation ↑
60/40 60/40 + All factors
B. Deflationary “bad times”
15
10
Real return (% p.a.)
-5
-10
NBER ↓ Earnings ↓ Equity ↓ Rates ↑ Inflation ↑
60/40 60/40 + All factors
Notes: The figure shows the historical average return on (i) the 60/40 global equity/bond portfolio and (ii) the 60/40 portfolio aug-
mented with equity, bond, and global factor premiums across high-inflation “bad time” regimes (panel A) or deflationary bad times
(panel B). High-inflation (deflationary) times and defined as high- (>4%; low, <0%) inflation periods in combination with an eco-
nomic recession (National Bureau of Economic Research [NBER]), falling corporate earnings, falling equities, rising rates, or inflation
(changes over a 12-month period). The sample period is 1875 to 2021.
multi-asset investor stagflationary and other bad times summarizes the results during deflationary bad times.
episodes are challenging, while the same Although returns on traditional portfolios are substan-
asset allocation including factor premiums perform bet- tially better than during high-inflation bad times, defla-
ter in each of these sub-regimes. Figure 5, Panel B tionary bear markets also present a challenge for

investors. Again, factor premiums consistently improve Appendix: Dataset Construction

on traditional portfolios. Overall, we can conclude that
the most severe bad times for investors in traditional
and Robustness Results
asset classes are times of high inflation with either eco-
nomic or earnings downturns, rising rates, falling equity
A.1. Dataset Construction. In the section
below, we describe the data sources and factor defi-
markets, rising inflation, or deflationary bear markets,
nitions used in this study, mostly building upon
and factor premiums on average help to alleviate the
Baltussen, Van Vliet, and Van Vliet (2022), Baltussen,
pain during these periods.
Martens, and Penninga (2021), and Baltussen,
Swinkels, and Van Vliet (2021). For further informa-
Concluding Remarks tion, we refer to the studies.
We examine investment returns across inflationary
regimes. As more extreme inflation regimes—like high
A.1.1. Recession and Earnings Growth
inflation, stagflation, and deflation—are relatively
Data. We construct our global recession data from
uncommon, we utilize a deep sample between 1875
the National Bureau of Economic Research (NBER)
and 2021. Inflation has varied considerably over time,
U.S. recession indicator from the NBER website
fluctuating between deflationary, moderate-inflation, or
(1875–2021).22 The corporate earnings data are from
high-inflation (including stagflationary) regimes.
the website of Robert Shiller.23
Moreover, inflation has generally been an important
topic for investors over the past 147 years. Asset class
A.1.2. Asset Class Returns. Equity and bond mar-
and factor premiums are strong and consistent ket returns are sourced from Datastream or
“empirical facts,” with attractive significant average Bloomberg, spliced with GFD, as in Baltussen,
returns over time. However, asset class premiums vary Swinkels and Van Vliet (2021). We use global equity
substantially across inflationary regimes. Deflationary market returns expressed in USD from MSCI World,
and moderate inflation scenarios generally provide posi- and before existence use global value-weighted equity
tive nominal and real equity and bond returns, while market returns from GFD (ticker TRWLDM) and, in
real returns especially suffer during times of high infla- case not available, weighted market returns across key
tion. Splitting up inflationary regimes into sub-regimes developed markets. For the global government bond
reveals that stagflationary episodes, inflationary bear returns we use returns on the Bloomberg Barclays
markets, or rising inflationary times and, to a lesser Global Treasury index, which we splice inception
extent, deflationary bear markets are bad times for before with GDP-weighted bond returns across U.S.,
investors. During these “bad times,” multi-factor equity, U.K., German, France, and Japanese bond markets.
multi-factor bond, and GFPs are consistent and attract- Bond return data are from Bloomberg and GFD, while
ive as they are across inflationary regimes. Such factors GDP data are from the Maddison Project Database
help to alleviate the pain during bad times, offsetting (Maddison 2006). All bond returns are hedged to
some of the negative impact of high inflation. USD, as common in practice as currency risk is a very
substantial driver of relative risk in bond portfolios.
These results have several important implications. First, Cash returns are returns on short-dated U.S. Treasury
for investors, times of high inflation, and especially bills obtained from Kenneth French website and
stagflation and inflationary bear markets, are challeng- before its inception from Jeremy Siegel.24
ing, which suggests that long-term asset class premi-
ums may be a compensation for bearing risks during A.1.3. Equity Factors. For the pre-1927 data, we
these bad times. Second, as equity, bond, and GFPs are use data from Baltussen, Van Vliet, and Van Vliet
generally consistent across inflationary regimes, they (2022); for the post-1926 data, we use the Kenneth
provide diversification to traditional asset class invest- French online data library. Baltussen, Van Vliet, and
ments. Still, these factor premiums do not take away Van Vliet (2022) have compiled their data from sev-
inflation risk, as their returns do not substantially eral sources in order to obtain a reliable and historic-
increase during bad times. Third, our results suggest ally extensive dataset. The sample covers 61 years of
that factor premiums in equities and bonds and across data on monthly stock prices, dividend yields, and
asset classes are not a compensation for bearing infla- market capitalizations for all major stocks traded on
tionary risks. Finally, our results indicate that portfolio the NYSE, NY Curb, and regional exchanges. The
managers may be better off with a balanced exposure sample spans the period from January 1866 through
to multiple factor premiums instead of trying to dynam- December 1926 and is at the monthly frequency.
ically rotate across factors based on inflation. They build their dataset from the Commercial and
24
Financial Chronicle (CFC, which was also used to A.1.5. Global Factor Premiums. For global (“cross-
build the CRSP sample as of 1926) and Global asset”) factor premium, we use the dataset and defi-
Financial Data (GFD). Note that we use their series nitions from Baltussen, Swinkels, and Van Vliet
from 1875 until 1926. (2021). They source price and return data of equity
The equity factors are constructed as follows: futures and indices from Bloomberg, Datastream, and
Global Financial Data. Their primary source is the
Value: futures from Bloomberg, with gaps filled in by
1875–1926: Dividend yield. Source: Baltussen, Van Datastream data and spliced before futures inception
Vliet, and Van Vliet (2022).
1927–2021: Book-to-market (HML). Source. Kenneth
with index-level data, as in Baltussen et al. (2019).
French on-line data library. They backfill these data with equity index–level data
Momentum: downloaded from Global Financial Data and obtain
1875–1926: Past 12-1 month total return. Source: dividend yields from the same sources. For Carry,
Baltussen, Van Vliet, and Van Vliet (2022). they use the spot, front futures, and second futures
1927–2021: Past 12-1 month total return (MOM).
Source. Kenneth French on-line data library. prices. Before they have data on futures, they recon-
Low risk: struct the monthly implied carry as if these markets
1875–1926: Past 36-month beta. Beta neutral long- had listed futures using the regression methodology
short portfolio. Source: Source: Baltussen, Van Vliet, on the difference between total return and price indi-
and Van Vliet (2022).
ces. The markets considered are spread around the
1927–2021: Past 36-month volatility. Volatility neu-
tral long-short portfolio.25 Source. www.paradoxinves- globe and cover the major developed markets with
tor.com. substantial data history.
Quality (50% Profitability & 50% Investment):
1875–1939: Not available, because companies did Bond futures price and return data are sourced from
not have (standardized) accounting data, see Wahal Bloomberg and splice these with bond index–level
(2019) and Baltussen, Van Vliet, and Van Vliet (2022). data from Datastream, backfilled before inception
1940–1962: Operating profitability, defined as reve- with Global Financial Data. From the same sources,
nues minus cost of goods sold, minus selling, general,
they obtain yields and inflation data, the latter
and administrative expenses, minus interest expense,
scaled by book equity. Investments, defined as the extended when possible with data from
change in total assets from the fiscal year ending in Macrohistory.net. The markets considered are the
year t-2 to the fiscal year ending in t-1, divided by t-2 major developed bond markets around the globe.
total assets. Source: Wahal (2019).
1963–2021: Operating profitability (RMW) and Currency forward and spot prices are primarily from
Investments (CMA), defined the same as above. Datastream, spliced with Bloomberg data and Global
Source. Kenneth French on-line data library. Financial Data. Purchasing power parity data are
obtained from the OECD website and, before 1971,
A.1.4. Bond Factors. For bond market factors, we with data from Macrohistory.net. The main measure
use data from Baltussen, Martens, and Penninga for the financing rates is short-term London Inter-
(2021) on global government bond factor premiums. bank Offered Rate (LIBOR) rates (sourced from
They have compiled data from 31 December 1799 Bloomberg and Datastream), spliced with (in order of
through 31 December 2020, which in this paper we usage) Eurodollar rates from Datastream, short-term
extend by 1 year. The paper sources bond futures pri- Treasury bill rates and commercial paper yields from
ces and return data from Bloomberg and splice these Global Financial Data, and short rates from
with bond index-level data from Datastream, backfilled Macrohistory.net and, for the United States, with
before inception with Global Financial Data (GFD). data from Jeremy Siegel. When all are unavailable,
From the same sources, they obtain yields and infla- they splice with lagged Treasury bill returns.
tion data, the latter extended where possible with data
from Macrohistory.net. They apply a 2-month lag to Commodity futures price and return data are sourced
inflation numbers to mimic their real-time availability. from Bloomberg, spliced with monthly commodity
futures data from CBOT annual reports (1877–1962)
Value: Real yield (bond yield minus inflation). Available obtained from TwoCenturies.com. For the commodity
1875–2021. Value measure, they use commodity spot prices from
Momentum: Past 12-1 month total return. Available Bloomberg and Datastream, spliced with spot data
1875–2021. from Global Financial Data and TwoCenturies.com.
Low Risk: 36-month beta. Beta neutral long-short port-
folio on the U.S. bond curve. Available 1922–2021. For Carry, they use the front futures and second
Carry: Term spread (bond yield minus short-term interest futures prices. The main contracts based on their
rate). Available 1875–2021. general usage and liquidity are included. Due to

restrictions on tradability, they exclude gold as a combination of absolute and relative purchasing power par-
speculative asset during the currency gold standard ity, and for commodities, the 5-year reversal in spot prices.
Momentum/Trend: Time-series Momentum that takes a
up to the end of the Bretton Woods system (which
long (short) position if the 12-1 month return is positive
was effectively a gold standard) in 1971. (negative).
The factors in their dataset are constructed as Low Risk: Past 36-month beta with respect to asset class.
follows: Beta neutral long-short portfolio.
Carry: For equities the excess implied dividend yield priced
Value: For equities, the dividend-to-price ratio (D/P) into the futures versus spot contract, for government bonds
defined as past 12-month dividend payment by the current the slope of the yield curve defined as the 10-year yield
price, for government bonds the Real yield (bond yield minus the short rate, for currencies the short-term yield dif-
minus inflation), for currencies an equally weighted ferential, and for commodities the slope of the futures curve.
A.2. Robustness Results.
Table A1. Robustness Results: Inflation Classifications

A. Global equity, bond, and cash returns
Nominal Returns Real Returns
Inflation Bucket N (ann.) Inflation (avg.) Equities Bonds Cash 60/40 Equities Bonds Cash 60/40
<3% 87.7 0.4 9.0 4.7 2.9 7.3 8.6 4.3 2.5 6.9
>3% 59.3 7.4 7.4 4.2 4.1 6.1 0.0 –3.2 –3.3 –1.3
Wald – – 0.5 0.4 8.4 0.6 11.0 46.4 58.5 20.1
<5% 113.3 1.2 9.3 4.9 3.2 7.5 8.1 3.7 2.0 6.3
>5% 33.8 10.1 5.4 3.1 4.0 4.5 –4.6 –6.9 –6.1 –5.5
Wald – – 1.3 4.4 1.7 2.0 11.6 62.3 70.3 21.0
All 147 3.2 8.4 4.5 3.4 6.8 5.1 1.2 0.1 3.6
B. Factors returns
Inflation Bucket <3% >3% Wald <5% >5% Wald
Equity factors Value 3.0 4.2 0.5 3.0 5.2 1.1

Momentum 6.3 7.8 0.7 6.6 7.9 0.5
Low Risk 6.8 6.0 0.2 7.4 3.4 5.1
Quality 3.0 1.9 1.1 2.9 1.4 2.9
MFE 5.1 5.1 0.0 5.3 4.6 0.6
Bond factors Value 1.2 4.3 4.9 2.0 4.0 1.3
Momentum 5.3 1.8 5.1 5.1 0.0 8.0
Low Risk 3.2 5.8 1.8 3.8 6.2 1.6
Carry 7.6 7.1 0.1 7.5 7.0 0.1
MFB 4.6 4.5 0.0 4.8 3.8 0.9
Global factors Value 2.4 2.1 0.2 2.3 2.2 0.0
Mom (Trend) 6.9 8.0 1.1 6.9 8.9 2.2
Low Risk 1.3 2.0 1.1 1.6 1.4 0.1
Carry 5.8 5.2 0.4 6.0 3.9 3.7
GFP 4.1 4.3 0.0 4.2 4.1 0.5
Notes: The table shows historical returns on asset classes (panel A), and equity, bond, and global macro factor premiums (panel B)
bucketed by two different global inflation category definitions (<3% vs. >3% and <5% vs. >5%). The sample period is 1875 to
2021. Returns are in percentages per annum. Significance at the 5% significance level is denoted with and significance at the 1%
significance level with . A Wald F test is used to test for significant average return differences across the different inflationary
states. All standard errors are computed using the Newey–West procedure.
26
Table A2. Robustness Results: Annual Changes in Inflation

Change in N Inflation
Infl. Bucket (ann.) (avg.) Equities Bonds Cash 60/40 Equities Bonds Cash 60/40
<0% 74.9 –2.5 7.7 5.5 5.2 6.8 10.2 8.0 7.8 9.3
0–2% 44.9 0.8 10.0 4.2 3.3 7.7 9.2 3.4 2.5 6.9
2–4% 10.7 2.9 7.5 2.7 2.7 5.6 4.7 –0.2 –0.2 2.7
>4% 16.5 7.4 7.7 1.6 5.9 5.3 0.3 –5.8 –1.5 –2.2
All 147.0 3.2 8.4 4.5 3.4 6.8 5.1 1.2 0.1 3.6
Wald – – 1.1 8.1 0.9 1.2 2.9 55.7 51.3 8.3
B. Factors returns
Change in Infl. Bucket <0% 0%–2% 2%–4% >4% All Wald

Momentum 9.0 4.4 0.0 8.5 6.9 3.3
Low risk 8.3 4.3 5.5 4.8 6.5 2.7
Quality 2.7 2.4 2.3 1.2 2.5 0.5
MFE 6.4 3.4 2.8 5.7 5.1 3.8
Bond factors Value 3.9 1.9 –1.8 –0.3 2.4 3.6
Momentum 4.2 4.4 3.7 1.0 3.9 0.8
Low risk 7.4 6.6 8.6 8.7 7.4 1.4
Carry 4.6 3.6 8.9 2.8 4.4 0.6
MFB 5.1 4.2 4.5 3.3 4.6 1.3
Global factors Value 3.3 2.3 –1.4 0.2 2.3 6.7
Mom (Trend) 7.2 6.7 8.3 9.5 7.4 1.5
Low risk 1.3 2.0 1.7 1.5 1.6 0.6
Carry 5.4 6.6 4.4 4.2 5.5 2.2
GFP 4.3 4.4 3.2 3.8 4.2 2.0
Notes: The table shows historical returns on asset classes (panel A) and equity, bond, and global macro factor premiums (panel B)
bucketed by four global inflation categories (from deflation, <0%, to high inflation, >4%) using annual changes in inflation. The
sample period is 1875 to 2021. Returns are in percentages per annum. Significance at the 5% significance level is denoted with
and significance at the 1% significance level with . A Wald F test is used to test for significant average return differences across
the different inflationary states. All standard errors are computed using the Newey–West procedure.
Table A3. Robustness Results: U.S. Inflation Analysis

N Inflation
Inflation Bucket (ann.) (avg.) Equities Bonds Cash 60/40 Equities Bonds Cash 60/40
<0% 27.6 –3.6 5.2 5.2 4.1 5.2 8.8 8.8 7.7 8.8
0%–2% 42.0 1.0 8.9 3.9 3.9 6.9 7.9 2.9 2.9 5.9
2%–4% 40.3 2.9 10.1 5.4 5.2 8.2 7.2 2.5 2.3 5.3
>4% 37.2 7.9 8.2 3.6 4.8 6.4 0.4 –4.3 –3.1 –1.5
All 147.0 3.2 8.4 4.5 3.4 6.8 5.1 1.2 0.1 3.6
Wald – – 0.9 3.0 7.9 1.0 2.0 30.7 37.7 5.5
(continued)

Table A3. Robustness Results: U.S. Inflation Analysis (continued)

B. Factors returns
Inflation Bucket <0% 0–2% 2–4% >4% All Wald

Momentum 4.3 5.9 8.5 8.3 6.9 1.2
Low Risk 6.9 8.2 6.9 3.9 6.5 1.9
Quality 2.4 2.0 2.8 2.5 2.5 0.2
MFE 5.3 4.6 5.6 5.2 5.1 0.3
Bond factors Value 0.9 2.5 1.1 5.0 2.4 2.1
Momentum 5.8 5.7 2.9 1.6 3.9 2.0
Low Risk 4.2 3.6 3.4 6.5 4.4 0.8
Carry 7.5 8.2 6.8 7.0 7.4 0.3
MFB 4.7 5.3 3.6 4.8 4.6 1.5
Mom (Trend) 7.9 6.1 6.8 9.0 7.4 2.1
Low Risk 0.5 1.0 1.9 2.7 1.6 2.0
Carry 4.2 5.8 6.1 5.5 5.5 1.0
GFP 3.5 4.0 4.3 4.8 4.2 2.4
Notes: The table shows historical returns on asset classes (panel A), and equity, bond, and global macro factor premiums (panel B)
bucketed by four inflation categories (from deflation, <0%, to high inflation, >4%) using U.S. inflation data only. The sample period
is 1875 to 2021. Returns are in percentages per annum. Significance at the 5% significance level is denoted with and significance
at the 1% significance level with . A Wald F test is used to test for significant average return differences across the different
inflationary states. All standard errors are computed using the Newey–West procedure.
Table A4. Robustness Results: Subsamples 1875 to 1948 and 1949 to 2021
A. 1875–1948
N Inflation
Inflation Bucket (ann.) (avg.) Equities Bonds Cash 60/40 Equities Bonds Cash 60/40
<0% 21.3 –3.3 2.9 5.4 3.0 3.9 6.2 8.7 6.3 7.2
0%–2% 17.3 1.0 8.2 3.3 3.3 6.3 7.2 2.4 2.4 5.3
2%–4% 12.2 2.8 8.6 3.0 3.2 6.4 5.8 0.2 0.4 3.6
>4% 23.3 10.0 2.5 1.7 2.1 2.2 –7.5 –8.2 –7.9 –7.8
All 74.0 3.2 4.9 3.4 2.8 4.3 1.7 0.1 –0.4 1.1
Wald – – 2.0 4.8 4.8 1.4 3.2 32.9 60.4 6.0

Momentum 7.6 5.0 4.6 4.9 5.6 0.1
Low Risk 8.0 5.7 4.0 5.5 6.0 0.5
Quality – – 2.9 –0.8 –0.7 –
MFE 5.3 4.1 1.2 3.5 3.8 0.8
Bond factors Value 0.1 2.4 –2.2 2.2 0.9 0.8
Momentum 6.4 7.9 8.6 0.4 5.2 2.7
Low Risk 11.2 3.6 0.9 1.4 4.6 3.2
Carry 8.1 9.6 10.7 1.8 6.9 4.3
MFB 5.5 6.7 5.7 1.4 1.4 5.9
Mom (Trend) 7.5 7.6 7.5 7.6 7.5 0.0
Low Risk 1.1 0.7 0.5 1.3 1.0 0.4
Carry 4.8 6.7 7.5 1.5 4.6 6.5
GFP 3.6 4.6 4.0 3.2 3.8 2.1
(continued)
28
Table A4. Robustness Results: Subsamples 1875 to 1948 and 1949 to 2021 (continued)
B. 1949–2021
Inflation Bucket N (ann.) Inflation (avg.) Equities Bonds Cash 60/40 Equities Bonds Cash 60/40
<0% 1.8 –0.9 – – – – – – – –

0%–2% 25.7 1.4 13.0 5.5 2.4 10 11.5 4.1 0.9 8.6
2%–4% 22.8 2.8 12.3 5.3 3.5 9.5 9.4 2.4 0.6 6.6
>4% 22.8 7.1 11.4 6.2 6.4 9.3 4.3 –0.9 –0.7 2.2
All 73.0 3.2 11.9 5.6 3.9 9.4 8.6 2.4 0.7 6.1
Wald – – 1.7 0.6 19.4 2.3 2.1 4.2 0.5 2.9
Equity factors Value – 2.2 3.1 5.4 3.2 3.1

Momentum – 6.8 9.0 10.7 8.2 1.7
Low Risk – 8.6 6.1 7.0 7.0 1.3
Quality – 3.3 2.2 3.0 2.9 0.8
MFE – 5.2 5.1 6.5 5.3 2.7
Bond factors Value – 1.7 2.6 7.4 3.9 2.4
Momentum – 3.0 4.6 0.5 2.6 1.8
Low Risk – 2.1 3.3 7.7 4.3 1.6
Carry – 6.3 7.3 10.5 7.8 1.3
MFB – 3.3 4.4 6.5 4.6 2.4
Global factors Value – 3.2 2.6 1.9 2.6 0.4
Mom (Trend) – 7.4 4.8 10.2 7.2 9.9
Low Risk – 1.3 3.0 2.7 2.2 6.2
Carry – 5.5 6.9 7.3 6.5 0.8
GFP – 4.4 4.3 5.5 4.6 2.8
Notes: The table shows historical returns on asset classes, equity, bond, and global macro factor premiums bucketed by four global
inflation categories (from deflation, <0%, to high inflation, >4%). Panel A includes the first sample period 1875 to 1948, and panel
B includes the second sample period 1949 to 2021. Returns are in percentages per annum. Significance at the 5% significance
level is denoted with and significance at the 1% significance level with . A Wald F test is used to test for significant average
return differences across the different inflationary states. All standard errors are computed using the Newey–West procedure.
Editor's Note
Submitted 9 November 2022
Accepted 23 February 2023 by William N. Goetzmann
Notes
1. Several authors have examined the inflation-hedging (2021) on global government bond factor premiums. For
characteristics of a variety of asset classes over different global (“cross-asset”) factor premium, we use the dataset
investment horizons; see, for example, Froot (1995), compiled by Baltussen, Swinkels, and Van Vliet (2021).
Schotman and Schweitzer (2000), Martin (2010),
Crawford, Liew, and Marks (2013), and Podkaminer, 3. More specifically, Neville et al. (2021) roughly define
Tollette, and Siegel (2022). These studies typically focus inflationary times as those periods when year-over-year
on the United States instead of global markets and realized inflation rises above 5% or has not fallen back
below 50% of its peak over a rolling 24-month window.
examine considerably shorter sample periods than we do.
4. We choose to focus on CPI as core measure of inflation
2. For equity factors, we use the dataset on U.S. equity factors
instead of GDP deflators, as the former measures
compiled by Baltussen, Van Vliet, and Van Vliet (2022),
inflation average prices of the typical expenditure basket
combined with data from Wahal (2019). For bond market
of an urban consumer and is typically considered as the
factors, we use data from Baltussen, Martens, and Penninga

headline number by investors, while GDP deflators focus than their market value after issuance. Snowden (1990)
on average prices of domestically produced final goods argues that it seems more important to include a short-
and services within an economy. term instrument that carries some default risk rather than
not using one at all. Moreover, it is not the case that data
5. Barsky and De Long (1991) investigate U.S. inflation on default-free securities are not available, but rather that
expectations in the period before World War I and conclude commercial paper was the closest to default-free present
that they may have been related to increased gold production. in financial markets. Siegel (1992) attempts to remove this
default premium from U.S. commercial paper rates by
6. See, for example, The Guardian (24 June 2021) “Recovery
using the term structure observed in the United Kingdom,
likely to push inflation above 3% by end of year, says Bank”.
which was more likely the global risk-free rate of the
7. More specifically, we employ a bag of words approach by time; see Friedman and Schwartz (1982). Similar issues
counting the relative frequency of words each day related may also be at play for the other government bond
to inflation (including inflation, inflationary, and stagflation) markets outside the United States in our sample, leading
and deflation (i.e., including deflation, deflationary, to term premia that are lower than we have observed in
disinflation, anti-inflationary, and noninflationary), see Garcia more recent markets where an entire term-structure of
(2013) for more specifics about the bag of words approach (near) default-free government securities can be traded.
and information about The New York Times columns. We
utilize the approach of Garcia (2013) but apply it to inflation 14. Also see, for example, Asness, Moskowitz, and Pedersen
or deflation by taking all inflation-related words in the TIAA- (2013) for evidence that a combination of Value and
CREF financial glossary and the General Inquirer dictionary. Momentum is significant.
Further, in line with an annual inflation horizon, we compute
15. As discussed in Baltussen, Swinkels, and Van Vliet (2021),
an annual average and compute its rolling 10-year
the Low Risk factor works better within asset classes than
standardized Z score. We assign a resulting value of 1 (–1)
across markets, which is confirmed by the fact that Low
when the Z score exceeds 1 (falls below –1).
Risk is the strongest factor in equities (panel B) and the
8. We do not include the size factor. Baltussen, Van Vliet, second strongest in bonds (panel C), but the weakest
and Van Vliet (2022) show that there is no long-run factor in global markets (panel D).
evidence for an alpha relative to the capital asset pricing
16. Note that Cohen, Polk, and Vuolteenaho (2005) find that
model in the United States, confirming the observation of
the risk-return relation becomes inverse during inflationary
Blitz and Hanauer (2020) that size is a weak stand-alone
factor in international equity markets. In (unreported) times, indicating a larger Low Risk premium in equities. By
robustness results, we have verified that conclusions on contrast, Neville et al. (2021) find that Betting-Against-
size across inflationary regimes are not materially Beta (BAB) is significantly weaker during times of inflation.
different from other equity factors. With this extended series based 2 3 sorted portfolios
with value weighting (which prevent an extreme size bias
9. The time-series we use are updated versions of the data that or large exposure to illiquid stocks), we do not find a clear
have been made available online at https://doi.org/10.25397/ relation between Low Risk equities and inflation.
eur.14237024.v1. We do not include cross-sectional
Momentum, as its returns are highly correlated with those of 17. We proxy a long-only investment with a 100% allocation
time-series Momentum (also known as Trend). We also do to the market and a 50% allocation to the long-short
not include the monthly seasonal factor, which is difficult to multi-factor premium strategy in that asset class.
line up with our annual inflation measure.
18. Extending the horizon to 5 years still gives negative real
10. Equities and bonds have historically been the most equity returns.
important asset classes making up the invested market
19. Factor timing involves additional risk and additional
portfolio; see Doeswijk, Lam, and Swinkels (2020). For
transaction costs. Besides inflation, one could consider
example, commodities make up less than 2.5% of the
factor value spreads (Cohen, Polk, Vuolteenaho, 2003),
market portfolio over the period 1960 to 2017. In this
factor momentum (Ehsani and Linnainmaa 2022), and
study, we do not include corporate bonds as a good
other macro indicators if the aim is to improve the multi-
historical database on the returns of investing in
factor mix.
corporate bonds going back to 1875 is to the best of our
knowledge not available. 20. See https://www.nber.org/research/data/us-business-
11. This is fairly similar to the arithmetic (geometric) average cycle-expansions-and-contractions.
of 10.65% (8.45%) reported in Jord a et al. (2019) over the 21. We have verified that results are similar when using the
period 1870 to 2015.
Global Recession Indicator from the OECD when available
12. Jorda et al. (2019) report an arithmetic (geometric) 6.06% (and before availability backfilled with NBER data), which
(5.71%) return for government bonds and a 4.58% dates back to the 1960s, motivating our choice of NBER
(4.53%) return for cash over the period 1870 to 2015. dates. Historical business cycle dating exercises for other
countries stretching back to the 19th century are not
13. Relatedly, Garbade (2008) explains that from 1920 readily available. As alternative we have also evaluated
onward, Treasury certificates had been brought to the the sign of the annual changes in nominal GDP across the
market for a fixed price. This resulted in massive United States, United Kingdom, Germany, France, and
oversubscriptions because the fixed issue price was lower Japan, finding again similar results (for example the
30
nominal equity return (60/40) during stagflationary a large part of our sample, making it in our view less
episodes equals –14.9% (–8.9%), and factors materially suited as an aggregate commodity market proxy. Further,
improve this number). we do not include Treasury Inflation-Protected Securities
(TIPS), assets that are designed to protect the investor
22. https://www.nber.org/research/data/us-business-cycle- against inflation shocks. Investing in them sometimes
expansions-and-contractions. requires locking in negative real yields, for example since
the beginning of 2020 (see Podkaminer, Tollette, and
23. http://www.econ.yale.edu/shiller/data.htm.
Siegel 2022).
24. In this study, we choose to not include other asset classes
25. Note that we choose to focus on 2 3 sorted portfolios
like aggregate returns on global real estate, commodities, to circumvent issues of high tilts to smaller caps and
or credit markets due to a lack of index data in the
illiquid stocks, as is the case in the Betting-Against-Beta
beginning of our sample. Commodities futures data are factor in stocks.
available as of 1877 (see Baltussen, Swinkels, and Van
Vliet 2021), but only includes agricultural commodities for
References
Asness, C. S., T. J. Moskowitz, and L. H. Pedersen. 2013. Doeswijk, R., T. Lam, and L. Swinkels. 2020. “Historical Returns
“Value and Momentum Everywhere.” The Journal of Finance 68 of the Market Portfolio.” The Review of Asset Pricing Studies 10
(3): 929–85. doi:10.1111/jofi.12021. (3): 521–67. doi:10.1093/rapstu/raz010.
Baltussen, Guido, Martin Martens, and Olaf Penninga. Ehsani, S., and J. Linnainmaa. 2022. “Factor Momentum and
2021. “Factor Investing in Sovereign Bond Markets: the Momentum Factor.” The Journal of Finance 77 (3): 1877–
Deep Sample Evidence.” The Journal of Portfolio 919. doi:10.1111/jofi.13131.
Management 48 (2): 209–25. doi:10.3905/jpm.2021.1.311.
Ely, D., and K. Robinson. 1997. “Are Stocks a Hedge against
Baltussen, G., L. Swinkels, and P. Van Vliet. 2021. “Global Inflation? International Evidence Using a Long-Run Approach.”
Factor Premiums.” Journal of Financial Economics 142 (3): Journal of International Money and Finance 16 (1): 141–67. doi:
1128–54. doi:10.1016/j.jfineco.2021.06.030. 10.1016/S0261-5606(96)00039-3.
Baltussen, G., S. Van Bekkum, and Z. Da. 2019. “Indexing and Friedman, M., and A. Schwartz. 1982. Monetary Trends in the
Stock Market Serial Dependence around the World.” Journal of United States and the United Kingdom. Chicago, IL: University of
Financial Economics 132 (1): 26–48. doi:10.1016/j.jfineco.2018. Chicago Press.
07.016.
Froot, K. 1995. “Hedging Portfolios with Real Assets.” The
Baltussen, G., B. Van Vliet, and P. Van Vliet. 2022. “The Cross- Journal of Portfolio Management 21 (4): 60–77. doi:10.3905/
Section of Stock Returns before 1926 (and beyond).” SSRN
jpm.1995.409527.
Working Paper 3969743.
Garbade, K. 2008. “Why the U.S. Treasury Began Auctioning
Barsky, R., and B. De Long. 1991. “Forecasting pre-World War
Treasury Bills in 1920.” Federal Reserve Bank of New York
I Inflation: The Fisher Effect and the Gold Standard.” The
Economic Policy Review, 14.
Quarterly Journal of Economics 106 (3): 815–36. doi:10.2307/
2937928. Garcia, D. 2013. “Sentiment during Recessions.” The Journal of
Finance 68 (3): 1267–300. doi:10.1111/jofi.12027.
Blitz, D. 2021. “The Quant Cycle.” The Journal of Portfolio
Management 48 (2): 26–43. doi:10.3905/jpm.2021.1.304. Ilmanen, A., and J. Kizer. 2012. “The Death of Diversification
Has Been Greatly Exaggerated.” The Journal of Portfolio
Blitz, D., and M. Hanauer. 2020. “Settling the Size Matter.” The
Journal of Portfolio Management 47 (2): 99–112. doi:10.3905/ Management 38 (3): 15–27. doi:10.3905/jpm.2012.38.3.015.
jpm.2020.1.187. Jord
a, O., K. Knoll, D. Kuvshinov, M. Schularick, and A. Taylor.
Cagan, P. 1956. “The Monetary Dynamics of Hyperinflation.” 2019. “The Rate of Return on Everything, 1870–2015.” The
In: Studies in the Quantity Theory of Money, edited by M. Quarterly Journal of Economics 134 (3): 1225–98. doi:10.1093/
Friedman, 25–117. Chicago, IL: University of Chicago Press. qje/qjz012.
Cohen, R., C. Polk, and T. Vuolteenaho. 2003. “The Value Maddison, A. 2006. The World Economy, A Millennial
Spread.” The Journal of Finance 58 (2): 609–41. doi:10.1111/ Perspective. Paris: OECD Publishing.
1540-6261.00539. Martin, G. 2010. “The Long-Horizon Benefits of Traditional and
Cohen, R., C. Polk, and T. Vuolteenaho. 2005. “Money Illusion New Real Assets in the Institutional Portfolio.” The Journal of
in the Stock Market: The Modigliani-Cohn Hypothesis.” Alternative Investments 13 (1): 6–29. doi:10.3905/jai.2010.13.1.
Quarterly Journal of Economics 120 (2): 639–68. doi:10.1162/ 006.
0033553053970133.
Neville, H., T. Draaisma, B. Funnell, C. Harvey, and O. Van
Crawford, G., J. Liew, and A. Marks. 2013. “Investing under Hemert. 2021. “The Best Strategies for Inflationary Times.” The
Inflation Risk.” The Journal of Portfolio Management 39 (3): 123– Journal of Portfolio Management 47 (8): 8–37. doi:10.3905/jpm.
35. doi:10.3905/jpm.2013.39.3.123. 2021.1.274.

Podkaminer, E., W. Tollette, and L. Siegel. 2022. “Protecting Snowden, K. 1990. “Historical Returns and Security Market
Portfolios against Inflation.” Journal of Investing 31 (3): 23–44. Development, 1872–1925.” Explorations in Economic History 27
doi:10.3905/joi.2021.1.207. (4): 381–420. doi:10.1016/0014-4983(90)90022-Q.
Schotman, P., and M. Schweitzer. 2000. “Horizon Sensitivity of Svensson, L. 2011. “Inflation Targeting.” In Handbook of
the Inflation Hedge of Stocks.” Journal of Empirical Finance 7 Monetary Economics, edited by M. Friedman and M. Woodford,
(3-4): 301–15. doi:10.1016/S0927-5398(00)00013-X. Volume 3b, 1237–302. Amsterdam, Netherlands: Elsevier.
Siegel, J. 1992. “The Real Rate of Interest from 1800 to 1900: Wahal, S. 2019. “The Profitability and Investment Premium:
A Study of the U.S. and U.” Journal of Monetary Economics, 29 Pre-1963 Evidence.” Journal of Financial Economics 131 (2):
(2): 227–52. doi:10.1016/0304-3932(92)90014-S. 362–77. doi:10.1016/j.jfineco.2018.09.007.
32

Investing in Deflation Inflation and Stagflation Regimes

Uploaded by

Copyright:

Available Formats

Investing in Deflation Inflation and Stagflation Regimes

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Investing in Deflation Inflation and Stagflation Regimes

Uploaded by

Copyright:

Available Formats

Financial Analysts Journal

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/ufaj20

Investing in Deflation, Inflation, and Stagflation

To link to this article: https://doi.org/10.1080/0015198X.2023.2185066

© 2023 The Author(s). Published with

Published online: 11 Apr 2023.

Submit your article to this journal

Article views: 10279

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

We examine asset class and factor pre- Introduction

Volume 79, Number 3 7

Table 1. Inﬂation Frequencies

<0% 0.9 1.6 7.7 23.1

B. U.S. inflation episodes

<0% 1.2 2.7 11.2 27.6

0.05 0.10 0.25 0.33 0.50 0.67 0.75 0.90 0.95

Volume 79, Number 3 9

Figure 2. Inﬂation Topicality

Inflation rate (YoY)

Inflation topicality US Inflation

Table 2. The Long-Run Evidence on Asset Class and Factor Premiums

Volume 79, Number 3 11

Table 3. Investment Returns by Inﬂation Regime

Years 147 23.1 42.9 34.9 46.1 –

Volume 79, Number 3 13

Figure 4. Factor Premiums across Inﬂation Regimes, 1875–2021

Value Momentum Low risk Quality/Carry

Volume 79, Number 3 15

Table 4. Inﬂation Regimes with 3-Year Investment Horizon

Years 147 21.0 47.8 36.3 41.9 –

Volume 79, Number 3 17

Table 5. A Deeper Look at Returns During High-Inﬂation Periods

Volume 79, Number 3 19

Table 6. A Deeper Look at Returns During High Inﬂation Periods: Frequencies

Table 7. A Deeper Look at Returns During Deﬂationary Periods

Volume 79, Number 3 21

Table 8. A Deeper Look at Returns During Deﬂationary Periods: Frequencies

Figure 5. Stagﬂationary Times and Investment Returns, 1875–2021

A. High inflation “bad times”

60/40 60/40 + All factors

B. Deflationary “bad times”

60/40 60/40 + All factors

Volume 79, Number 3 23

investors. Again, factor premiums consistently improve Appendix: Dataset Construction

Volume 79, Number 3 25

A.2. Robustness Results.

Table A1. Robustness Results: Inﬂation Classiﬁcations

Equity factors Value 3.0 4.2 0.5 3.0 5.2 1.1

Table A2. Robustness Results: Annual Changes in Inﬂation

Equity factors Value 3.4 2.9 4.0 5.1 3.5 0.2

Table A3. Robustness Results: U.S. Inﬂation Analysis

Volume 79, Number 3 27

Table A3. Robustness Results: U.S. Inﬂation Analysis (continued)

Equity factors Value 4.1 1.5 3.3 5.5 3.5 1.2

Inflation Bucket <0% 0–2% 2–4% >4% All Wald

Equity factors Value 5.3 4.1 1.2 3.5 3.8 0.5

<0% 1.8 –0.9 – – – – – – – –