Classic Bernstein 6 — The Value Premium and Inflation

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
SimpleGift
Posts: 3608
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Classic Bernstein 6 — The Value Premium and Inflation

Post by SimpleGift » Tue May 10, 2016 12:25 pm

PREFACE: This is the sixth in a series of posts highlighting the classic investing insights of William Bernstein from the 1990s and early 2000s. Many new Bogleheads have never been exposed to his early writings — and while the data sets used may seem antiquated, his portfolio concepts and novel analyses are still helpful to investors today, new and old alike.

Previous topics in the series: 1-Asset Allocation & Time Horizon, 2-Choosing Portfolio Bond Duration, 3-Diversifying Portfolio Equities, 4-Stocks Always Beat Bonds?, 5-What’s a Thing Worth?.
------------------------------------------------------------------------------------------

In the last post, Mr. Bernstein laid the theoretical foundation for why low interest rates are more beneficial to growth stocks than to value stocks. In this analysis, he examines the real world data for the relationship between the value premium and inflation over the 74-year period from 1926-1999.

Note: This analysis uses the HmL data series from Fama-French, which represents the return of the top third of stocks sorted on book value/price, minus the return of the bottom third — the value-minus-growth premium. A positive number signifies higher returns for value stocks, and vice-versa. Also note that this premium represents a long-short portfolio that’s unattainable in actual practice, as even the hardcore, upper-third value funds only obtain a portion of the full value premium.

The Study: In the chart below, Mr. Bernstein regresses annualized inflation versus HmL, by decade. Not only is the positive correlation between the value premium and inflation visually impressive, but statistically as well, with a t-stat of 3.41 and p-value of 0.014. The monthly regression was also significant, with a t-stat of 2.91 and p-value of 0.0037.
  • Image
    Note: The 20s data point is only a three-and-a-half year period, starting July 1926.
    Source: William Bernstein, Efficient Frontier (7/01)
What observations does Mr. Bernstein draw from this short study?
  • a) The original Fama-French paper on the value effect covered the high-inflation decades of the 60s, 70s and 80s (specifically, 1963-1990), and consequently showed a very robust value premium. Since then, interest rates and inflation have been in a long decline — just the environment to favor growth stocks.

    b) His formulation also explains why value investing was such a life-saver in the 1970s, with its raging inflation — and why growth investing dominated in the 1930s Great Depression, the worst period of deflation in U.S. history.

    c) Finally, he notes that the slope of HmL on inflation in the chart is 1.1, so each percent of inflation adds about one percent of HmL. Thus, as long as there is inflation over the long-term (almost a given with modern fiat currency), there should also be a corresponding value premium.
Thoughts?
Last edited by SimpleGift on Sat May 28, 2016 8:09 pm, edited 2 times in total.

User avatar
saltycaper
Posts: 2650
Joined: Thu Apr 24, 2014 8:47 pm
Location: The Tower

Re: Classic Bernstein 6 — The Value Premium and Inflation

Post by saltycaper » Tue May 10, 2016 2:23 pm

Some things it makes me wonder:

1) If value companies have higher debt levels, and if inflation makes it easier for them to service that debt, it makes sense they will do well during periods of high inflation. But what about real interest rates? Is there any reliable correlation there that might counteract the benefits of inflation?

2) It would be nice to see data points representing shorter time periods, perhaps one data point per year. I wonder how the correlation between value and inflation may differ in the short term versus the long term.

3) I've seen other studies on the small and value premiums being pervasive worldwide. Can we say the same about the relationship between value and inflation, or has it been mostly a U.S. phenomenon?
Quod vitae sectabor iter?

larryswedroe
Posts: 16022
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Classic Bernstein 6 — The Value Premium and Inflation

Post by larryswedroe » Tue May 10, 2016 2:46 pm

the problem IMO is that there is NOT a linear relationship with inflation and value premium. I studied it long ago and the logic follows the data.
First, value companies tend to have more debt and deflation really bad even of itself as it raises the real cost of debt. But deflation usually comes with bad economic times and value companies tend to have much more irreversible capital and cannot cut costs. So they have both high financial and high operating leverage.

Second, when have relatively low inflation that is typically good economic times when the risks of value stocks doesn't show up. Then for a while with rising inflation you have the benefit of falling real costs of debt. But then you get to period when inflation is higher than Fed wants and they slam on the breaks and the real rate rises and economy slows and value stocks get hit. For example check out 78080 when HmL was highly negative and inflation basically in double digits.

The following data is from a piece I did long ago, but when inflation was negative HmL averaged -5.6%. It was largest when it was in the 4-6 % range, so moderate, rising inflation

Hope that helps
Larry

User avatar
Topic Author
SimpleGift
Posts: 3608
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Classic Bernstein 6 — The Value Premium and Inflation

Post by SimpleGift » Tue May 10, 2016 9:03 pm

saltycaper wrote:Some things it makes me wonder...
Just a reminder: There is much more data available today, compared with 2001 when Mr. Bernstein did his study in the OP. Interested Bogleheads with an analytical bent are encouraged to replicate any of Mr. Bernstein’s analyses — using more current data sets or alternative time periods — and share their results in these threads. It would definitely add a more current perspective to many of these topics.

User avatar
Topic Author
SimpleGift
Posts: 3608
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: Classic Bernstein 6 — The Value Premium and Inflation

Post by SimpleGift » Wed May 11, 2016 8:26 am

Just to add more current charts to this thread, updated to 2015, showing the decline in the value premium (at left), corresponding with the decline in inflation since the 1970s (at right). If Mr. Bernstein’s formulation is correct, we’re not likely to see a very robust value premium again (or any value premium at all) until the current low-inflation environment changes.

Image
Note: The value premium is trailing 20-year average, from Fama-French data.
Sources: Bogleheads Forum and JPMorgan Chartbook.

dkturner
Posts: 1563
Joined: Sun Feb 25, 2007 7:58 pm

Re: Classic Bernstein 6 — The Value Premium and Inflation

Post by dkturner » Wed May 11, 2016 9:50 am

Simplegift wrote:Just to add more current charts to this thread, updated to 2015, showing the decline in the value premium (at left), corresponding with the decline in inflation since the 1970s (at right). If Mr. Bernstein’s formulation is correct, we’re not likely to see a very robust value premium again (or any value premium at all) until the current low-inflation environment changes.
This is one reason why I prefer the Vanguard Equity Income, Wellington and Wellesley funds. They are all managed by Wellington Management, which seems to have understood this relationship for a very long time and gradually changes the makeup of their portfolios to add more "growthy" stocks as the value premium wanes. I feel confident that if or when inflation becomes a problem again the Wellingtom Management people will adjust their portfolios to add more of the traditional value stocks.

Theoretical
Posts: 1505
Joined: Tue Aug 19, 2014 10:09 pm

Re: Classic Bernstein 6 — The Value Premium and Inflation

Post by Theoretical » Wed May 11, 2016 10:13 am

Japan is a major counterexample to this, as it's suffered deflation and a flat to negative equity market with a positive value premium over a long haul (I think it was something like 4-4.5%)

User avatar
LadyGeek
Site Admin
Posts: 61204
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Classic Bernstein 6 — The Value Premium and Inflation

Post by LadyGeek » Sun May 29, 2016 1:27 pm

A complete list of Simplegift's series is now in the wiki: Classic Bernstein
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

Post Reply