Monday, April 23, 2007

A View From The Style Box: What's Performing Best And What That Tells Us


A while back we looked at the style box and found that value was handily outperforming growth. If we take a fresh look at investment styles and performance, however, a different picture emerges. In this look, we're using the Vanguard ETFs: VUG (large cap growth; dark blue line); VTV (large cap value; pink line); VBK (small cap growth; yellow line); and VBR (small cap value; light blue line).

What we find is that, since the start of 2007, growth has been outperforming value. Indeed, small cap growth is beating the pants off the other investment styles.

Investor preference for small caps is not exactly what I'd expect if the market were expecting recession. Nor would I expect a preference for growth stocks.

But then, again, there are numerous indications that the market is not pointing toward recession:

* Breadth Strength - Remember how beat up the advance-decline stats were during the late February and early March drop? They have recovered that ground and then some, rising steadily to bull market highs.

* Dollar Volume Flows - Twelve of the last fifteen trading sessions have shown above average money flows into Dow stocks. Institutions are putting money to work in the market, not taking funds away.

* All Time Price Highs - It's not just the Dow making all-time highs. The broad NYSE Composite is doing the same. Over 86% of S&P 500 stocks are trading above their 50-day moving averages--and that proportion has been rising. That's not exactly the weakness we'd expect to see in a topping market.

Our look at the investment styles suggests that we have transitioned from a more defensive market (in which value leads growth and large caps lead small caps) to a more speculative one (in which we see the reverse). At this juncture, institutions are committing funds to equities. And that is benefiting the most entrepreneurial segments of the market.

4 comments:

Damian said...

Formula Research did an interesting piece in 2003 regarding what happens when growth leads value. They compared equity returns during periods of out-performance of the Russell 2k Value and Growth indices.

The results were very interesting:

When R2k growth was the leader:
- R1k Growth: 20.6% ARR/25% DD
- R1k Value: 15.1% ARR/22% DD
- R2k Growth: 40.4% ARR/20% DD
- R2k Value: 24.3% ARR/17% DD

When R2k value was the leader:
- R1k Growth: -1.3% ARR/65% DD
- R1k Value: 7.2% ARR/40% DD
- R2k Growth: -20.4% ARR/78% DD
- R2k Value: 1.7% ARR/44% DD

So if the Rk2 Growth is leading, then the entire market is raised with it.

Test period was 1995-2003.

Out-performance by R2k value/growth was calculated as the average of the 5/15/25/35 day rate-of-change for the Rk2 value and growth.

I don't know if this relationship has continued since 2003 but it would be interesting to look at the idea further and see if it still works.

Thoughts appreciated!

Damian

Brett Steenbarger, Ph.D. said...

Hi Damian,

*Very* interesting. I really like the Formula Research work. My best guess is that allocation to growth vs value is a kind of sentiment indicator, capturing one's optimism about the economy. That's why I think the current shift is worthy of notice. It's a real shift from the past couple of years.

Brett

Damian said...

This is the kind of testing I want to learn to do but I really don't know how to approach it. A while back I suggested you do a post on how to put together your measurements/stats - any help would be great! I'm assuming I'll start by downloading data from 2003 to 2007 on the Rk2 growth and Rk2 value, then calculate the 5, 15, 25, 35 day returns. Average those. Then, when the Rk2 growth is greater than value, look at returns. Any suggestions/tips on how to do this? Anything I'm missing?

And a personal note: I find your blog very inspiring.

Brett Steenbarger, Ph.D. said...

Hi Damian,

The important thing is to get clean historical data on the right ETFs for those styles. Your basic approach seems sound; just depends on the time frame you're trying to predict--

Brett