Can anyone tell me why origination spike at the beginning of the month for LC? Seems that the 5th is a hot spot.

http://www.nickelsteamroller.com/chart/combined/issued#daily

dontknow

And I can't buy a hit on my filter. You lucky dog!

dontvote

If you ran the prosper version of this chart it would bounce from zero to one at those times.

dont
joke

from: rajuabju on February 13, 2013, 06:49:55 PM

It is telling that I asked the prosper salespeople if they had done any work on this exact question and they weren't able to point me to their own blog. And this was the 25K+ account size salespeople.

In reality this graph you linked to looks at portfolios with 400+ notes which may indicate more sophisticated or better investors or some other 'larger portfolio' bias. The right calculation with existing data would be to take the return distribution of every portfolio you could make with 5-5000 notes.  You will see that there is a point where adding an additional note doesn't appreciably affect your diversification. The classical example is the one of asset classes in your portfolio. You get limited 'diversification benefit' after 16 or so assets classes.

Without spending time on a calculation I will probably get wrong I think that number is around 150-200 notes in your given strategy (criteria set or filter). By definition, notes in a given filter (esp a really restrictive one) will have a higher relative correlation to other notes in that strategy. You can designate another strategy that can form a portfolio of notes that do not overlap your current filter and but have a similar risk return profile to improve your diversification. With 2 return streams like that you would need fewer notes to diversify each one.  So I would think a better investment strategy would be to pick 2-10 strategies (based on your portfolio size and ability to find awesome return groupings) that have the least overlap and invest 100 notes in each one. Remember what we're trying to diversify out - in the 'same strategy' think you're looking for unique risk - that of two guys with the same 'profile' one is a deadbeat. You also should diversify out the chance that you are an idiot and pick stupid loan profiles by curvefitting (with no curve) the data using too many filters.

Tl;dr: If you have one filter, you can be around 150 notes and be diversified, so knock yourself out with the $1000 per note thing. If you have multiple strategies that do not have significant overlap of loans you should chop your investments down across them.

Note: I know I'm countering a graph with an opinion.

from: graceful on January 06, 2013, 12:45:54 PM