What a famous chart, trading the gold price around the daily "fix".
The fix is in!
Academic financial research calls charts like this "anomalies". A trading strategy that earns steady profits over decades? Arbitrage profit. Theoretically, there's not much room for that in an "efficient market".
And, yet, there it is.
As Ronan Manly explains at Bullion Star,
"Starting in 1970 when the price of gold was $35 per ounce, if every day you bought at the price of the afternoon Gold Fixing and sold 19.5 hours later at the price of the next day’s morning Gold Fixing price, your $35 would now be worth $15,843. That’s the cumulative value of the Overnight strategy. On the other hand if every trading day since 1970 you bought at the morning Gold Fixing price and sold 4.5 hours later at the afternoon Gold Fixing price, your $35 would now be worth, wait for it, just $3.33. That’s the cumulative value of the Intraday strategy. The key point is the the London Gold Fixings suppress the gold price intraday and that the current gold price of around $1500 is far lower than it should be because of the prices in these Gold Fixes. None of which is mentioned by the LBMA in its commentaries of the 100th anniversary of the Gold Fixing this month in London." Ronan Manly, 16 September 2019.
Rothschild emerges from the shadows for the Centenary of the London Gold Fixing, https://www.bullionstar.com/blogs/ronan-manly/rothschild-emerges-from-the-shadows-for-the-centenary-of-the-london-gold-fixing/
So what?
I wrote a theoretical paper in 2013 titled "New Testing Procedures to Assess Market Efficiency with Trading Rules". Original idea was from trading quant to show what really good trading stratgies looked like, in theory. Compare them with strategy that's not particularly good.
Key point was that theoretical model I made had an inefficiency, by design. It was a "mean reverting" process. There was a bunch of trading strategies that could earn an arbitrage profit from exploiting the pattern and I showed how one of them performed in the paper. The returns to that arbitrage strategy were "reliably positive". Think Citadel, circa 2010s.
The other trading strategy I described in the paper isn't an arbitrage strategy. It tries to exploit the same mean reverting property, but the underlying price simulation I created doesn't have the "mean reversion" property. It's a random walk.
See the theory I came up with in this graph showing the steady profits for one trading strategy and unreliable returns for the other one.
Did you know Eugene Fama wrote an article in the Journal of Finance in 1970 on "Efficient Capital Markets"? Apparently trading rules played a role in his thinking on the topic, too.
The theory is fun. A computational technique to test market efficiency based on profits for trading rules? Formal testing criteria based on each technique and uses simulation to establish existence results about the tests for efficiency? Market simulations with inefficiencies built-in?
Again, the theory is great, but I never thought I'd see one in the wild. And to think that it's in the gold price, of all things!
