Nearly half of private capital fund managers rely exclusively on Excel – survey

Alternative reading: A company trying to sell portfolio management software at a premium calls the segment of their market not paying for their software luddites for using a cheap, powerful, ubiquitous, general purpose tool that's easy to find developers for -- that is currently meeting all their needs when they obviously should be paying for the shiny new thing.

Many many years ago I was at an investment bank. Our branch was in NYC, but the bank was based in Europe (it's not important which bank). Our department was principle finance, investing in synthetic CTOs. Our team was responsible for building the real risk reporting system....

.... porting from Excel.

Those spreadsheets had been so tortured. There were custom macros that loaded data from a multitude of other systems that had everything including quotes, ratings, LIBOR curves. The whole thing would be loaded and an XML file was generated. Then the quant tool ran against the XML generating a second XML file. This would be loaded back into the spreadsheet, where everything would be summed up and generate a series of CSV files that were then FTPed to some central reporting system. All of it was run by Excel macros, and the whole thing took about 2 hours, not including the quant work which could range up to 20 hours.

Historical records were just a series of copies of the file on a shared drive that everyone had write access to, so be careful, and make sure you copy the file locally if you need to look at it. Of course there were errors everywhere. The spreadsheets had formula copy/paste errors often. and sometimes you had to figure out when it was fixed or introduced by finding the correct copy of the file.

This was the system of record for about $50bn in assets.

We built a system to actually process this stuff in a real system. I was pretty proud of it. We got the imports running in under 15 minutes, and then we'd shard the work out to a farm of quant machines, so that we could generally get runs done and loaded in another 15 minutes. I learned a lot from the process, but in particular I learned that Excel is extremely powerful as a prototyping tool. And when I have a one-off job or a prototype to build, I'm much more likely to reach for a spreadsheet than an IDE.

What the hell? "Luddites"? Excel is fantastic, and I would bet on the competence/brightness of anybody using it for anything complex.

One thing overlooked by the article is that portfolio management and other similar applications often involve "trade secret" proprietary analysis and evaluation methods. It's way easier to build, manage, and protect that sort of IP in a spreadsheet than in other application environments. In particular, it's reasonable to be leery of software that requires paying the vendor to customize it, since at least some of your methodology is almost certain to leak to competitors. Excel has many faults, but it does allow you to keep everything behind your own walls, with no dependencies on vendors, consultants, the cloud, etc.

Meanwhile Institutional Investor continues to trot out thinly veiled marketing fluff pieces, produced by self interested parties, as news articles.

Excel is also used extensively by investment bankers and hedge fund bankers.

Because, it often is the right tool to use...