IRR is more a cashflow analysis which coincides with the phrase "cash is king". DCF is fine too but too many difficulties estimating WACC. Anyway I've always used IRRs when I'm looking at exits (as that's how I look at MBOs and LBOs) and where it's a going concern I use a DCF.
Actually I bought a place last year and I'm yielding 7.5% (don't ask me how). I punched out a DCF (assumed 2% growth in rent, 3% growth in most of the operating costs and used a terminal growth rate of 3% assumming it's a going concern for an infinite amount of time) and got an NPV above my purchase price with a WACC of 8.5% (high gearing, hence much has been captured by 7% cost of debt). Anyway you're absolutely right. Twig the differentials (ie assumptions) and you can change the outcome dramatically.
I absolutely agree the spread has to be lowered otherwise numbers don't make sense. That said I'm also a believer that low yielding places (eg Toorak and Double Bay) don't necesarilly go by these figures... people buy them because it's more than just an investment asset. That's the conumdrum I have with these property investments . For investment class property, I believe spread will catch up through increased rental demand. I'm a believer in market within markets, a big believer in high yield + high growth, hence why I debate with the less financially-attuned members here about inner vs outer suburbs and so forth. I personally believe yields will catch up over time and will be towards the higher end of inflation (relative to other commodities, such as food etc).
But yea getting paid to do this stuff is pretty good. Maybe you should do this on a contracting basis. The other day I heard some oil and gas analyst who was contracted to build some oil and gas models (which is pretty easy compared to other resource commodities) was charging $3k per day. That's pretty high in my opinion.
There's not much discussion at this level on SS DB. I appreciate your input.
Several years ago, there was some good analysis being done. Guys would disagree about some point, then go away and come back with a xls model showing timing was superior to b&h, or some such thing. Timing consisted of switching most or all capital between property and shares or cash.
See_Change was good at this stuff, as was Pete from Perth (an xls guru contractor who the last I heard was driving a new ferrari), Mark_B (ex economist with RBA), GreatPig, and Sim (the forum administrator). It's worth a squiz at some of the threads from 2003-4, when many of us thought the market had peaked for the time being.
BTW, when talking about spreads, I presume the higher the spread the higher the risk, which is why I put it as rate:yield, rather than vv.
We better agree on terminology.
What I am moving towards here is a form of cap rate forecasting for resi IP, but using gross yield rather than net. I'll use the figures to guide offers on IPs. I think we are moving into a structurally different leverage climate, and it is going to become more important to focus on yield.
I'll post a link to the xls I am messing with to model this stuff. APpreciate your opinion on how to better develop it. You obviously have a handle on the 'greeks'.
Re me working as a modeler, I wouldn't have the confidence. I've met some of the best in Australia and the US (maths PhD's), and I know how quickly and confidently they can complete a project. At this point, I am more interested in minimizing gigo, by improving my undestanding of bank credit and floating exchange rate mechanism on the economy, current account, balance of trade, BOP, etc. Hard to model well without realistic interpretations of this stuff. I talk about this stuff with two mates (an ex trade commissioner to the UK and a CFO for a large insurance company) and am surprised they don't understand it at a deeper level. Presume BHP has a large room full of guys to analyse/hedge risk 24/7. I wonder how much Rudd has changed their beta gradient?
edit: btw, when I first read up on the greeks, I got the impression everyone used parametric stats, including capm. I couldn't believe it. I presume things have moved on and non parametrics is more the go. Presuming normal distributions in the presence of fat tails and high kurtosis can kill.