Hacker News new | ask | show | jobs
by compumike 1251 days ago
Figures 6 and 9 look a lot like https://totalrealreturns.com/ , especially with the trendlines on these figures, logarithmic y-axis, (disclosure: my side project, recomputed daily at market close)

It would be cool to merge in some longer-term historical data as the article author has done, instead of just using actually-tradable assets like I've done.

Per the article, the author considers these charts "misleading":

> Charts such as Figure 9, with their accompanying commentary, and combined with the distinctive behavior of the product function, may lead investors to mis-anchor their expectations about the future performance of bond and stock investments. Faced with a yawning visual gap, and apprised of the numerical dominance of stock returns (in Siegel 2014, estimated at 6.6% real versus 3.6% for bonds), an investor readily infers that bonds are never going to out-perform stocks over any lengthy period.

I'm not sure I agree with the conclusion. A log scale hides a lot of volatility, but it's still fairly obvious that the stocks line has a lot more volatility, prolonged periods of substantial drowdowns (painful!)...

As another commenter points out, this article's use of price-only data (even if adjusted for inflation) is intellectually dishonest, ignoring returns from dividends. And yes, your typical price-only, non-inflation-adjusted charts from Yahoo Finance / Google Finance / Apple Stocks should probably be considered intellectually dishonest, or at least confusing, in my opinion...

1 comments

Yahoo uses the CRSP method (see: factor to adjust price) to back-adjust old prices when dividends and splits occur, so it's not as misleading as you're probably thinking. I'd be surprised if the others didn't do something like this too.
Yahoo does have this “Adjusted Close” column available in their historical data downloads, but they do not use it for charts.

Their charts are price-only. Same with Google Finance and Apple Stocks.

Definitely not seeing Alphabet's 20:1 split in July 2022 when I glance at Yahoo Finance charts - you sure about that? Or are you saying they're adjusted for some corporate actions but not for others?

EDIT: OK, I see there's a note that close is split-adjusted but not dividend-adjusted.

That’s correct: their charts are split-adjusted but that’s all.
Dividend adjustment wouldn't really matter in this study. The price discount at dividend ex date should match the price increase post earnings announcements, so it's netted and disappears once you compute >=quarterly returns.
This is not correct.

The price increase at earnings announcements is (to a first order approximation) indicative of the extent to which the company's earnings exceeded the market's expectations. It's as likely to be positive as negative. Otherwise you could buy the company's stock the day before the earnings announcement and get yourself some free money.

I use TradingView for most of my charting and at the bottom right side there's a button labeled "adj". Clicking it will toggle whether the chart you're viewing is dividend adjusted or not. Very convenient for quick off the cuff comparisons of returns.

Using SPY I can easily see the non adjusted returns from the bottom of the GFC to the top in 2021/2022 were a little over 600% but then accounting for dividends see it's actually a bit over 800%.

For SPY, from 2009-03-09 to 2022-01-03, I get +588.5% total real return, +16.24%/year, after adjusting for dividends and inflation:

https://totalrealreturns.com/s/SPY?start=2009-03-09&end=2022...

I'm pretty sure Tradingview doesn't take inflation into account.