When I started doing equity research we were told to make ‘projected profit and loss account’, calculate the ratios, make the cash flow and then the balance sheet. Today it has a far sexier name – it is called ‘Financial Modelling’. Wow.

I was taking a class on equity research and I was asked ‘how does one do Financial Modelling for say Tata Motors. I said ‘you do not’ – you just accept what is available online – and tons of work is available especially if you can access Bloomberg. Sir, but how do WE do it. I said Financial Modelling does not happen in a closed room. You need to do the following – and do it well….

1. Know why are you doing the modelling: is it to assess the cash flow to find whether a loan will be repaid, whether the share is worth buying, or…normally the reason to do a financial model (I succumb) – one that forecasts future earnings, balance sheets and cash flows – is not just to find out if the company will beat or miss the next quarter. It is to understand a company’s complete business model. Where does it really make its money? What geographies, what products, what channels? How much does it take to keep the business stable? How much to grow it? And how does all this compare to other companies in the industry? And if they are leaders, what gives them the leadership position?
2. Do it yourself (I really mean it):  One analyst I know would take a very long time to start work on a company – even if it was an easy company like say Carborundum Universal. His logic was ‘I have to overcome bias..you spoke too much about it’. So true. It is easy enough to just pull together other analysts’ models – imagine our plight in the pre Internet era!! Today  you can get revenues and earnings estimates off of  Bloomberg, Yahoo!, Msn money, Moneycontrol… and other sources. Even if you do want to look at those numbers, do that AFTER you’ve done your own work.
3. Focus on sales: get all possible details on sales. Client-wise, salesman wise, region wise, product wise, channel wise – that is the only way to ensure that you are near the accurate figure. Just adding 15% to last year’s numbers is amazingly easy and pretty foolish. Most start ups come with numbers which is just proof that excel works. Not the result of some detailed data crunching. I guess if I were to do the report for Ta Mo today it would take me 2-3 days to collect sales projections for a year…and I would love numbers month wise and not quarter wise even though it is a quarter that I would be looking at. Longer term forecasts is amazing guess work but when you learn to break up in detail..and you talk to the capex team, dealer search team, sales satisfaction survey team,…you start respecting your own guess work. Ultimately it is sales. If you cannot see a dedicated, almost mad kinda sales team, run way – the sales ain’t gonna happen!
4. Focus on costs – slightly easier during our times, but these days more global data is available easily. However, let it not take away your focus. Raw material costs, marketing costs, and of course people costs – all are notorious to judge across countries – and Ta Mo is across continents. So guessing people costs (sure not a very imp component, but nevertheless) in China, US, Europe and India is just guess work based on inflation in these markets and using past trends. Finance costs have to be tempered with currency fluctuations – impact of Brexit for example. JLR has US $ borrowings, Euro costs, Yuan costs and multi currency revenues – try putting this in Excel !!
5. Focus on competition: Sales break up is very important – because the luxury car segment competes with Audi, Mercedes, BMW and the trucks compete with Volvo, Mercedes, in the local market with Ashok leyland…so only, only, only when you have the detailed – Hcv, Mcv, Lcv, luxury cars, ordinary cars…can you know which segment has to be compared to whom. Extreme detailing required – and clearly puts it beyond the one man research team which we all are today. I would not even attempt a company like Ta Mo given the fact that I will just not have access to such detailed data. Ignoring data from competitors will make the whole exercise futile.
6. Every model – be it for a quarter, half yearly or more…should have an impact on the PnL, balance sheet, and cash flow. So the robust model has to have the capability of seeing the impact of that quarter on the total balance sheet. You might notice a working capital crunch if the sales projections are too rosy – FORCING you to look at the funding requirement. Never, ever, make a quarterly projection and thinking you do not need the B/S. The ‘final impact’ has to be seen. Every time.
7. The ratios are obviously an out come of your results..but do take a look at EVERY ratio, EVERY time you make a significant change. In the longer run currency movement may not matter, but a company may decide to source items from a country with a weaker currency – and sell in a country with a stronger currency. Such ‘strategic decisions’ have to be built into the costs consideration!
8. Many of the things which are macro are at best guesses, do not fret and fume over it. Make sure you understand that all experts are guessing too.
9. Do not expect accuracy beyond 1 or 2 quarters..but make the detailed quarter wise projection for 2 years…and keep updating on a DAILY BASIS…and see how accurate you can get.

Personally I know I am now challenged, so for Ta Mo, I will trust consensus figures of analysts nicely copy pasted from Bloomberg.