Why not just let the credit spreads you've sold expire worthless? Why spend $0.15 or $0.20 to buy in a credit spread you sold for $0.70, when it's still twenty days before expiration? It might be cheaper closer to expiration or worthless at expiration.
Moderating risk, that's why.
Other reasons include freeing up margin for other trades, trades that might make more than the $0.15 or $0.20 still left to eke out of the sold credit spreads you're holding. Still, in my eyes, eliminating risk ranks as the most important reason to consider buying in sold credit spreads once they've narrowed to $0.15-0.20.
When I first started trading iron condors in size in 2005 or so, I did not always buy in my sold credit spreads. On November 18, 2005, an expiration Friday, I came to regret that decision to hold out-of-the-money credit spreads.
Here's what happened, explained via an excerpt from my April 8, 2006 "Trader's Corner" article. The SPX had closed Thursday, November 17 at 1242.80. Those who had written or sold the November 1250 call probably felt fairly confident, if perhaps slightly nervous, that their sold calls would expire worthless the next morning. I had sold 1250/1260 call credit spreads as part of my iron condor. OptionNet Explorer's backtesting abilities tell me that on the close that Thursday, November 17, 2005, the mid price for the NOV 1250/1260 call credit spread I'd sold was 0.225. I likely would have had to pay more than that to buy it back near the close that day, but the mid price had been as low as $0.10 at times during that day. However, I reasoned that I wouldn't lose money unless the SPX settled at a value equal to 1250 + the credit I'd taken in for the entire iron condor position (put credit spreads plus call credit spreads) minus commission costs. I had some leeway above 1250.
These days, no one would have felt "fairly confident" about that trade that Thursday, with the SPX closing only a little more than 7 points below the sold strike. However, at that time, a standard deviation for the SPX was approximately 6.98 points, and indices didn't tend to experience morning gaps as often as they do now. We at Options Investor used to counsel that one of the benefits in trading the indices over individual equities was that indices didn't tend to gap in the morning.
In fact, the SPX did not gap that next morning, settlement day for the (monthly, since no weeklies existed) SPX options. Nor did the SPX ever trade above 1250 that day. That didn't matter, however, because settlement values for the monthly SPX options are not determined by the opening price nor any price the index reaches the Friday of option-expiration day. They're based on the opening values of all the separate SPX component stocks. Settlement values can differ widely from any price reached that day. If stocks are rolling open, and the first few open higher, then the next few open higher, and the first batch starts down while the third batch opens higher, etc., you can see settlement values far higher than any price reached that day. You can also see settlement values lower than the open or any value reached that day.
The SPX's settlement value on November 18, 2005 was 1254.85. Also from my Trader's Corner article came this information: That weekend, those who had written the 1250 call got messages from their brokers that their sold calls had expired in the money, and that their accounts were being docked the appropriate $485 per contract, plus any commission for closing the position. I was one of the unlucky people getting that notice.
As far as I remember, that was the last time I went into expiration with open credit spreads. I changed my trading plan, so that for the big indices such as the SPX, OEX, and RUT, I closed any credit spread as soon as it narrowed to $0.20. Back in those days, that often happened rather soon in the trade if there had been a big move. I was often able to close one side and remove risk in that direction, only to see the index switch direction. I can't tell you how many times a credit spread that I'd bought in for $0.20 would have been in threatened when prices changed direction. If they'd been open, profitable trades often would have been transformed into losing ones. However, in my last days of trading iron condors in about 2010 before I began switching to butterflies, it seemed that premium didn't leak out of those credit spreads quite as soon.
Is $0.20 the magic number? No, of course not. At the time, I was trying to take in about $0.65-$0.85 per side on the iron condors for spreads 10 points wide. If you're trading a different kind of iron condor with wider spreads, say 20- or 30-point spreads, you're likely taking in more credit. You might want to lock in profit and remove risk at $0.30 or $0.40, depending on how much you took in when the trade was initiated. If you are trading an equity and the credit you took in was much less, you might be waiting until the spreads narrow to $0.05 or $0.10.
Of course, buying in credit spreads costs money and lowers your potential profit. However, if those credit spreads have been sold as a directional trade or if they're the last remaining side of an iron condor, buying them in not only locks in profit and lowers risk in that trade, but it also frees up funds for a new trade.