Your cashflow tool isn’t a crystal ball, but you shouldn’t leave home without it

By Lora Benson | May 04, 2017
Last week, I was driving back from a family holiday in Lincolnshire. When I left, my satnav told me that, taking the quickest route, I’d be home at 12:14. I wasn’t. What the satnav didn’t take into account was roadworks on an A road, and two “incidents” on the M1.

So why do I need cashflow modelling?

Last week, I was driving back from a family holiday in Lincolnshire. When I left, my satnav told me that, taking the quickest route, I’d be home at 12:14. I wasn’t.

What the satnav didn’t take into account was roadworks on an A road, and two “incidents” on the M1. 

When we got fed up of sitting in a stationary queue of traffic, I did a U-turn. Ignoring my car’s increasingly insistent attempts to persuade me I was going the wrong way, it eventually calculated a new route down a B road, and we bypassed the big queue.  Hooray. ETA now 12:24.

My trusty satnav then told me about an incident on the motorway, and suggested a detour. Suddenly our ETA was 12:18. The route we’d been on when we left was no longer quickest. It had reassessed the data, and moved the goalposts.

We made up even more time, thanks to friendly traffic lights, and ended up arriving home at 12:16. Two minutes off our original ETA, despite taking a totally different route to the one we’d originally planned!

Feed your computer the very best garbage possible

There’s a programming truism that simply says, “garbage in, garbage out”. Put simply, if you feed a programme bad data, it will give you bad results. Conversely, if you feed a programme GOOD data, it will give you GOOD results.

What I effectively did was to fed my satnav the very best garbage I could find (my chosen “fastest” route). True to form, it gave me garbage back – it told me with absolute certainty that I would be at home at 12:14. I wasn’t.

But it didn’t just give me a list of directions and a time, and shove me out of the door - it constantly reassessed the route and adjusted the ETA accordingly. In the end, it was pretty darn close.

And therein lies the rub. When you plug your destination into your satnav, and it tells you that you will arrive at 12:14, you don’t ACTUALLY expect to arrive at 12:14. But a better satnav will give you a more accurate guesstimate at the start (the better models take into account traffic information and existing roadworks). 

I like to think of deterministic modelling as the “illusion of certainty”, while stochastic modelling gives the “illusion of uncertainty” - both are illusions! You will neither retire on the 3rd of June 2023, nor buy 75% of a new car. But you will retire, and you might buy a new car.

If we can accept that a computer can’t accurately predict something 90 minutes away, how can we expect it to be right about something 6 months, 12 months, or even 40 years in the future? If we start out with the assumption that the computer is wrong (as we do with our satnav), then we see the value of feeding in the very best garbage possible. If you want to get a very good idea of what the future might hold, make the most accurate assumptions possible. But it will still be wrong. Deterministic modelling – you will retire on the 3rd of June 2032

Deterministic modelling works based on the assumption that every assumption you key in is correct. Unfortunately, they aren’t. None of them!

Garbage in, garbage out!

Deterministic modelling works on the assumption that every assumption you key in is correct, although we now know it isn’t! Your client will not spend exactly £2,000 on wine. Nor will they spend exactly £8,000 on holidays. Their income won’t escalate at exactly 5% for the next 20 years. 

The better the assumptions you feed in, the closer your model will be to reality. If your client completes a simple fact find in 5 minutes, you can show them a simple cashflow.  If they go to the trouble of in-depth factfinding, you can show them an in-depth cashflow – one that has a pretty good chance of being somewhere near reality. But it will still be wrong!

Stochastic modelling – there’s a 75% chance you will buy a new car

There’s a popular argument that stochastic (or probabilistic) modelling is inherently superior to a deterministic model. This is because it gives a probability of success, rather than claiming to predict the future. 

Sadly, this is wrong too. 

Let’s look at an example. There is a 75% chance you will be able to buy your new car in 18 months’ time.  So there’s a 25% chance that you won’t have enough money, right?  Wrong.  This suggests a totally passive attitude – i.e. you do nothing for 18 months and then suddenly wake up and check your bank balance. 

“Do I have enough money to buy the car today?  No?  I guess I’ll take the bus!”

If your goal is to save enough money to buy a new car, you’re probably keeping a close eye on your bank balance. And you’re probably tweaking your inflows and outflows on a regular basis to enable you to do it. The probability may be 75% today, but your efforts will all be geared towards increasing this. 

Buying a car is a binary decision – you either buy it or you don’t. You can’t 75% buy it.  The probability will either rise or fall until, on the day of purchase, it’s either 100% (you buy the car) or 0% (you don’t). 

Move the goalposts regularly!

The value of advice comes from reviewing the assumptions regularly, and moving the goalposts.

If it gets to 6 months from when you want to buy your new car and you still don’t have enough money, you make changes. If it gets to 5 years before you want to retire, and you don’t have enough to see you through, you reassess. You don’t blindly set out on the assumption that the initial model will continue to be correct, because it definitely won’t. There’s a chance (as with my journey home), that you’ll get there round about the right time, but the route may well have been entirely different. 

Nobody can predict the future, but just like a satnav will give you a more accurate idea of your arrival time than a map and a piece of string, so cashflow modelling software will give you a more accurate idea of your client’s future finances. 

In a month when we’ve seen plans for satnavs to be included in the practical driving test in the UK, can you really afford to be using a map and a piece of string to predict your client’s future?

Prestwood Software’s Adam Leci BA (Hons) MA (Oxon) Cert CII (FS) Software 01384 273736

www.truthsoftware.co.uk

 


 

Prestwood are the leaders in client facing financial planning software. In 2007 they launched 'Truth®' which has since helped hundreds of advisers move to more profitable financial planning models. Truth is easy to use with clients thereby creating powerful client relationships. It moves the focus off products and investments to what matters most – helping clients identify, achieve and maintain their desired lifestyle.