Humans and not algorithms are real deal in property investments

We see this in international markets as well, where distress at in the mid-market - where job losses have been highest - has not affected the luxury side (which, in Dubai has moved higher after years of a declining trend). In a sense, this not only reflects international trends, but is also a reversion to the zeitgeist of the city that has set new standards for luxury over the years.

Whether the trend will continue becomes the $6 million question.

Human emotions matter

Where real estate differs from other asset classes is that it is not dominated by algorithmic trading, even though numerous startups have started implementing auto valuation models such as ‘Vector Auto-regression’ and ‘Error Correction’ to generate bids that isolate thousands of variables and search for “alpha” opportunities.

Clearly, the capital is there for such experimentation, but thus far, investment and trading in the real estate market seems to be grounded in human activity. And therefore, still ruled predominantly by human emotions.

Underneath the surface, the asset allocation towards affordable housing revealed the vulnerability this segment had to job losses, which is what has dominated the zeitgeist over the last year. Even as prices have trended lower in this segment, transactions have increased in suburban areas - where people have looked for more space - as well as more densely populated areas such as Jumeirah Village, with the quality of build as well as work-from-home initiatives gathering pace.

Deciphering buyers

All of this suggests that allocation of assets is being skewed by a few variables: 1) replacement value, 2) preferences for space 3) build quality and 4) zonal flexibility. These dominate the middle-income segment.

At the high-end, a different set of dynamics have taken root, which predominantly have been to do with capitalizing on pricing anomalies, preferences for old build structures that can be upgraded, and capitalizing on distress sales.

The rental structure has been easier to read, as job losses have translated into lower rents across the board, regardless of segment. To some extent, this has increased the popularity of short-term rental models, and has increased preference for furnished spaces as the employment market becomes more flexible.

This has resulted in lower yields but more pricing pressure on investment-oriented areas rather than end-user ones. Of course, the lines become blurred quickly as replacement value factors take over very quickly.

Not capturing all

Models for the most part are not capturing these sets of data in Dubai. In more advanced economies like the US, where there have been attempts to trade off computer algorithms, initial performance sets reveal the performance of such funds is far inferior.

This suggests that a) there is much room for improvement for data capture and hence the various approaches to cluster modeling in this space, and b) not all the variables are being captured in the data set, as could be done by the friendly neighbor.

Chatter driven

Similar experiences have been seen recently in the equity capital markets, where chaos has reigned because of chatter generated in Reddit rooms that have outfoxed most hedge fund models. Data capture is always critical and more work needs to be done to more accurately paint an investment landscape that can be deciphered by the median investor.

But to live under the illusion that data can build algorithms that strip away the human element of investing is sheer lunacy. Recent events in equity markets have shown that to be the case, even as the “everything rally” comes to an end. In real estate more than ever, the data suggests - oh the irony - that the human element intertwined with the investment decision can never be separated. To quote the famous bard again: “That way madness lies…”.

- Sameer Lakhani is Managing Director at Global Capital Partners.