The entire exercise is fuzzy accounting. Assumptions are built upon assumptions and all involve non-linear growth processes. The errors are huge in practice. But the basic concepts are simple. Estimate a future return of X from a given investment, and after comparing it with alternatives, pay Y today in the hopes of earning X in the future.
In manufacturing, this is relatively easy. A manufacturer's capacity and growth rates can be estimated with fair certainty. So can the average price per unit and expenses in terms of raw material costs.
In services, extreme scalability of capacity exists. To take an example, a five-person IT business can scale up to a turnover of $4 billion and a workforce of 72,000 in just 20-odd years (Infosys). However, tracking average revenues per man-hour (billing rates for IT/ ITES) and deriving future capacity estimates from
employee strengths, you can make some projections. Oddly enough the most tangible business of all - real estate - is the most difficult to track and value conventionally. A lot of guesstimates are required to value realty and the range of errors is truly extreme.
Estimates made on past earnings records don't offer much clue as to future earnings. And yields may vary year-on-year by 100 per cent or more. We've seen the positive side of this in windfall profits made over the last two years.
Let's say, you know last year's yield per square foot for a real estate company. That doesn't say much about next year's yield. You must also know the size and quality of the realtor's land-bank. You must know average acquisition costs for that land bank. You need to guess the yield per square foot in future for that land-bank.
This number may vary massively from past yields. There are macro policy and infrastructure issues totally beyond the realtor's control. So, you have also to guess the time frames in which a land-bank can be developed.
These uncertainties make realty a very high-risk business. And, for any high-risk business to survive, it must offer high returns. Realty has patently survived. It has actually generated fantastic growth rates in the last few years because land has been developed in a benign environment.
Firstly, commercial demand has been driven by the IT/ITES industry expansion. Secondly, retail demand has been driven by low interest rates and easy availability of home loans. Thirdly, yields have been unsustainably high because realtors released land that had been acquired at much lower cost. Fourthly, policy encouraged investment in commercial and retail development.
Factor one remains - the IT/ITES business is expanding quickly and so are other land-intensive industries such as retail. But factor two and three are no longer as favourable as they were over the past three years. Interest rates could harden further. Land banks, which will come to market after 2008, would cost more. So, yields are almost certain to be lower. Factor four has turned unfavourable after the SEZ mess in West Bengal.
The market has recognised these uncertainties by sharply downgrading real estate stock-valuations. Since February, the industry has lost a lot of market capitalisation. Many stocks have lost more than 30 per cent. Can there be further losses in market cap over the next year? Probably - both interest rates and real estate prices may peak out over the next six months to a year.
The sector has a negative correlation with rates - it's almost as sensitive as banking and housing finance. While interest rates continue to rise, real estate valuations are likely to fall.
Once interest rates do start falling, the sector will see a second big bull run. The only sane way to invest in real estate stocks would be to factor that in and implement a sort of systematic investment plan linked to rates. Use a casino "doubling up" model. If you invest X amount at the current rates, be prepared to invest 2X when rates fall by 1 per cent and another 2X or even 3X if rates fall by 2 per cent.