Post

When the inflation data came out, I was crouched in front of the supermarket’s refrigerated meat case, looking at the beef price tags. That moment made things more real than watching the markets: the same cut of beef had gone from $6.99 to $8.49 per pound in a year, and the price tags were changing more often than candlestick charts. Lately, people on Twitter have been debating whether economic data is still reliable. My view hasn’t changed: the data itself isn’t wrong; the mistake is treating it as the only thing that matters. Data is an input, not a conclusion; a signpost, not the destination.



I have four rules. Just follow them.

First, when you look at the data, check the source and the revisions. When the initial and revised figures conflict, trust the revised figure, not the day’s emotions. For example, if the initial nonfarm payrolls report shows a gain of 250,000, the market might celebrate first, only for the figure to be revised down to 180,000 two months later, turning all that excitement into “it wasn’t as strong as we thought.” The same goes for GDP and retail sales. Initial figures often include estimation and seasonal-adjustment noise; revised figures are closer to the complete sample. People who trade on the initial reading are often paying for someone else’s revisions.

Second, the same data flows through to different markets in different ways. Stocks care about earnings; currencies care about interest-rate differentials. Don’t apply one set of logic to every asset. A higher-than-expected CPI could mean rising discount rates and lower valuations for stocks, wider rate differentials and a stronger dollar for FX, and changes to real rates and duration pricing for bonds. The same “high inflation” reading can affect tech stocks, bank stocks, the dollar, and long bonds in completely different ways. Apply one template to every asset, and you’ll only trap yourself.

Third, decide what you’ll do “if the data comes in above expectations” before it’s released, instead of looking for an explanation afterward. Hindsight is expensive. I usually write down three scenarios in advance: what to do if the data comes in above expectations, in line with expectations, or below expectations; how to adjust my position, where to place my stop-loss, and whether to buy protection. For example, before CPI, I’ll decide whether to reduce long-duration exposure or buy SPY protection if core month-over-month inflation comes in above 0.3%; and whether to chase or wait for a pullback if it comes in below 0.2%. After the release, I just execute. I don’t make up a story on the fly.

Fourth, it takes three consecutive data points in the same direction to count as a trend; treat a one-off anomaly as noise. The answer is in the details, but you need enough details to add up. For example, a PMI that rises for three consecutive months is a sign of recovery; a one-month spike could just be seasonal or sample variation. The same applies to employment, inflation, and retail sales. One data point is a dot; three make a line. Only when the line is long enough does it deserve to be called a trend.

There’s a counterexample, too: for a while last year, the data was broadly positive, but the market traded on the surprise, not the absolute numbers. The data was good, but not “good enough to change the rate-cut path,” so the good news was already priced in. The data was bad, but not “bad enough to trigger recession trades,” so the bad news didn’t push the market down. So now, when I look at the data, I ask one question first: Who is this number for? The Fed, earnings models, or market sentiment?

When CPI comes out next week, my bet is that core inflation will have a bigger impact on the stock-bond relationship than headline inflation. Headline inflation is easily affected by oil and food prices; core inflation is more closely tied to persistence and the policy path. Let’s see next week whether I’m right about $SPY.
View Original
post-image
This page contains third-party content and does not constitute any advice, nor does it represent Gate's endorsement of such views. For details, please see disclaimer.

  • 1

Add a comment
Add a comment

Comment 2
BitcoinBallast
$SPY See you next week.
PerpPessimist
Those trading on revised figures are also paying for someone else’s next revision; this game has no end.