Hi Guys, can you help me to understand why sometimes I observe in the OIS market (both Euro and US), the 3 week quote would be higher than 1m quote? Does this mean we do expect the overnight rate goes lower during the 4th week? Or it's just a supply/demand issue regarding to the huge liquidity difference between the two contracts? This behaviour happens about 1/3 of time on each month from my database. Several standard paper about OIS discounting also shows this pattern in their sample data.Since I am trying to use them to build my OIS curve and it always gives out a kink point and I am trying to decide if that point is ideal and better to smooth. Thanks!

Funding rates exhibit some very specific and somewhat regular calendar effects, which mkts price. I believe that's what you're observing, although it's hard to be sure w/o specifics.For instance, month-ends/quarter-ends generally produce higher fixings. In the past, in EUR there were some regular effects arnd the last week of the reserve maintenance period (higher or lower fixings, depending on regime). If I had to guess, you're observing something like that.

Thank you so much for the explanation Martingalehoul!Yeah, I am observing sometimes the 1,2,3 week OIS rate is higher than 1 month OIS rate. For example, on 5/3, the mid-quote of 1,2,3 week OIS rate is 0.365% but the 1 month OIS was quoted as 0.362%. I guess this is due to special reasons like you said about the reserve maintenance period but not so sure. Since weekly OIS contract are way more illiquid than 1 month contract so I also feel the mid quote may deviate. The problem is should I include weekly contract into bootstrapping procedure. If it is due to fed behaviour, I should catch it in my curve. If it is due to liquidity noise, I should exclude them.

Since the "inversion" you're observing seems to be occurring with some regularity in the beginning of the month, I am inclined to think that it has to do with some sort of month end effect. In general, the US FedFunds mkt, in its current format, sorta has one foot in the grave (there's a LOT of gory mechanics, which I'll skip). This probably creates more volatility and strangeness than normal. This may go away once the Fed finally transitions to OBFR for everything.