Monday, March 21, 2016

Buyback Performance Demystified

Earlier this month, in my post Stock Buybacks Demystified I attempted to remove some of the mystery surrounding buybacks, showing they are no different from an economic perspective (if you ignore the impact of taxes and the effects of signaling) than dividends. Given the recent outperformance of dividend paying stocks (as defined by those in the S&P 500 Dividend Aristocrat index) vs. stocks engaged in buybacks (as defined by those in the S&P 500 Buyback index) over the last year, I thought it might be helpful to demystify what has driven the recent outperformance of dividend stocks, share the historical performance of each, and outline some forward expectations for relative performance given where we currently sit.


Background: Buybacks have consistently outperformed since inception 

The first chart shows the growth of $1 invested in dividend stocks vs $1 invested in buyback stocks going back to the inception of the S&P 500 buyback index in early 1994. What we see is pretty consistent underperformance of dividend paying stocks over time that has compounded to 2% / year outperformance of the buyback stocks since inception (note both have outperformed the S&P 500). 


Background: Relative performance is highly mean-reverting
The second chart shows the relative performance of dividend stocks less buyback stocks over 12-month rolling periods going back to the inception of the S&P 500 buyback index. What we see is:
  • Pretty consistent underperformance of dividend paying stocks (most of the relative return series is negative) 
  • Mean-reversion characteristics (when one outperforms the other materially, it tends to bounce the other direction pretty quickly)
  • Dividend outperformance during periods of market stress 

Why have dividend stocks outperformed recently? Valuation differences
In addition to market stress that favors dividend stocks, the change in relative valuations have driven dividend stocks (i.e. it's been seemingly more technical than fundamental). The chart below shows valuations (i.e. P/E) of the dividend and buyback indices as of month-end February and as of a year ago. While dividend stocks were richer by this measure even a year ago, valuations among dividend stocks have held up pretty well. On the other hand, valuations among buyback stocks have gotten materially cheaper, driving the relative underperformance of buyback stocks and creating a huge valuation gap between the two. 


What now? 
The final chart shows the impact of shorter-term (12-month) relative performance between dividend and buyback indices on longer (3-year) forward relative performance. We can see:
  • Pretty consistent underperformance of dividend paying stocks over most three year periods 
  • Mean-reversion characteristics kicking in when dividend stocks have outperformed by the 10% level they have over the last 12-months (when dividend stocks have outperformed by 10% or more over a 12-month time frame, they have underperformed by an average of 6.4% / year for the next three years)

Takeaway
Neither dividend stocks or buyback stocks outperform in all periods or market environments, but under the view that the underlying economy appears to be holding up, current valuations, the historical outperformance of buyback stocks, the mean-reversion characteristics of buybacks (following recent dividend outperformance), and certainly the tax efficiency of buybacks all seem to support a tilt toward buybacks. 

Wednesday, March 9, 2016

Stock Buybacks Demystified

Based on my Twitter feed, stock buybacks seem broadly misunderstood in terms of what they are meant to accomplish (to redistribute excess capital back to shareholders) and the impact they have relative to dividends. As an aside, I also don't understand the following typical complaints:
  1. Buybacks are done when stocks are rich: if the stock of a company performing buybacks is "rich", then why are you owning it to begin with?
  2. Buybacks are often done during bull markets and stop during bear markets: that is due to the fact companies often have higher earnings and more excess cash available to distribute during bull markets

As a result, I wasn't too surprised when the below chart made the rounds yesterday, along with the following implications:
  • Stock performance is higher only because of financial engineering
  • Households are scared of equities and are poor market timers

In this post I'll provide the case for why (ignoring taxes and the effects of signaling):
  • Buybacks and dividends are economically identical
  • Buybacks are an incremental driver of the outflows we've seen from households
  • Why flows (both inflows and outflows) do not relate to demand / why households almost always have outflows

Buybacks and Dividends are Economically Identical

Excluding the potential signaling or tax effects of buybacks vs dividends, buybacks and dividends are identical in terms of overall economic impact. For example, assuming the following:
  • Corporation X has $2.5 billion in excess cash to distribute back to shareholders
  • Corporation X has a market cap of $5 billion ($2.5 billion enterprise value + $2.5 billion cash)
  • Corporation X has 100 million shares outstanding, priced at $50 / share (100 million shares x $50 / share = $5 billion market cap)
  • Corporation X will earn $1 billion ($10 / share) the following year
  • Shareholder X owns 100 shares and has spending needs of $1000
  • 100% of Shareholder X's wealth and income comes from their stock ownership

Situation 1: Corporation X distributes the excess cash to their shareholders via a $2.5 billion buyback
  • Corporation X buys back $2.5 billion of their shares at $50 (i.e. they retire 50 million shares)
  • Corporation X now has 50 million shares outstanding at $50 = $2.5 billion market value (all made up of their enterprise value)
  • With no dividend payment, shareholder X will need to sell 20 shares (x $50) to meet their $1000 spending need, meaning they will have 80 shares x $50 = $4000 in Company X stock
  • Corporation X will earn $1 billion next year or $20 / share given their 50 million shares
  • So... next year Shareholder X will be entitled to a $20 x 80 = $1600 of those earnings

Situation 2: Corporation X distributes the excess cash to their shareholders via a $2.5 billion dividend
  • Corporation X distributes $2.5 billion via dividend ($25 / share)
  • Corporation X still has 100 million shares outstanding, but at an enterprise value of $2.5 billion each share is now worth $25 / share
  • Shareholder X can meet all of their spending needs through the dividend distribution (1000 x $25 = $2500 dividends) and after spending $1000 still has $1500 cash remaining
  • Shareholder X can use the excess $1500 to buy 60 more shares at $25 share, meaning they now own 160 shares at $25 = $4000 in Company X stock
  • Corporation X will earn $1 billion next year, which is $10 / share given 100 million shares
  • So...next year Shareholder X will be entitled to $10 x 160 shares = $1600 of those earnings

A table summarizing the above example (click for larger image)


What has changed? 
  • # of shares outstanding
  • Price of shares outstanding
  • Earnings per share
  • Shareholder X becomes a net seller of shares in the buyback scenario

What has stayed the same?
  • Enterprise value of the firm ($2.5 billion)
  • The overall level of earnings ($0.5 billion)
  • Earnings Shareholder X is entitled to next year ($1600)
  • Overall market value / demand for Company X stock from Shareholder X ($4000)
  • Overall net purchases of the stocks ($1.5 billion)*
* In the buyback scenario, $2.5 billion is bought back by Company X, but if all shareholders acted like Shareholder X, they would sell $1 billion for their spending needs ($1.5 billion net purchases); in the case of dividends, of the $2.5 billion distributed, $1 billion is spent, and the same $1.5 billion is used to buy back shares with the excess cash.

Household Flows Don't Matter / Should be Negative

As highlighted above under 'what has changed', household outflows are in fact impacted on the margin by the form of capital distribution (i.e. whether it is received via buyback or dividend). In the case of a buyback, households are simply creating their own dividend through the sale of shares. Given the two situations are identical in terms of overall demand for Company X stock (demand at time 0 was $5000 worth of stock, post spending it was $4000), we can see why flows really don't matter.

In fact, while the initial chart circulating through Twitter highlights the negative flows from the household sector for stocks from 2008-2015, what may be a surprise is that the overall level of stocks held by the household sector (i.e. a better measure of demand) jumped from $5.4 trillion at the end of 2008 to more than $12.7 trillion over that same time frame (as of 9/30/15  - the latest z.1 report), a normalized increase of 37% of GDP to 70% of GDP.

But a key point is that household net flows for a mature / functioning economy should be negative... when markets have positive returns, investors put in less money today than what that investment should compound to when they make withdrawals in the future. Thus, it should be no surprise that net flows from the household sector have historically been negative over all longer periods of time going back 60 years, while the amount of stock held by the household sector has continued to move higher.



To summarize... the form of distribution really does not matter and buybacks are not evil... the next time you hear someone state buybacks are the cause of the run up in stocks, try replacing the word buyback with dividend.

"Stocks are up because of a huge increase in dividends" sounds a lot less controversial than "stocks are up because of a huge increase in buybacks", though they are both identical signs that the performance has been driven by an improvement in fundamentals and an increase in cash flows.