The nature of Research

Real estate research has its roots back in the 1970s when the big agents set up teams to

collect data on rents, yields and other measures of property market activity

Morphing from data to analysis It was a big step, but perhaps an obvious one, to start to use the raw data being collected to analyse that data, comparing locations, and trends over time, as well as turning points. The three biggest research teams were in Hillier Parker, Richard Ellis (now both assimilated into CBRE) and Jones Lang Wootton (now JLL). The raw data had a time series starting in the early 1960s, and was originally used just for market intelligence and as a means for the agents to market their knowledge using presentational material. By the 1970s, and largely in advance of the equities markets, there was an increasing demand to use the data with economics data to help in understanding market dynamics and, in particular, to forecast market movements. At this time, it was not just the agents who had assembled research teams, but some of the larger institutions were building small teams, not necessarily to replicate the research by the agents, but very often to use the research as a resource for internal purposes. An exception to that was the Prudential Insurance Company, which assembled an high-powered team. The focus at that time was ‘econometric modelling’. The concept of econometric modelling was to try to identify the independent variables (typically, but not necessarily, economic) that were the drivers for dependent variables like rents and yields. By using the respective histories of these variables and software (PCs were just coming into the working environment, although they were rare), it was possible to establish formula which were good approximations of historic movements. Using these formal and forecasting the independent variables, it was then possible to produce a forecast iof the dependent variables like rental values. These were heady days for real estate research, and there was some quite strong competition between the research teams. For the firms with research teams and the accumulated data, this became one of their unique selling points, and the teams were effectively used as both a marketing tool and for giving advice to clients, particularly to those who were new entrants into the market - such as foreign investors (London has always been the first point of entry to the European markets and also for Europeans outside their own countries). Different parts f the research teams or different research teams within the same organisation would concentrate on differenty parts of the markets, typically dividing between occupier (landlord/tenant) and investment (buyers/sellers), and between the three main sectors of offices, retail and industrail. In addition to the agents’ and the institutions’ research teams, there were also a small number of independent research teams that either collected specialist data or were based on building econometric models. Today, the most notable survivor is Propert Mqarket Analysis. But there was also a data analyst, Investmnet Propery Database, taht grew into something much larger. It positioned itself so that it was not so much in competition with the inhouse reserch teams but was providing performance data of real estate portfolios that was avalable to both researchers and the fund managers. This has now been taken over by MSCI. It is, I think, fair to say, that demand for research, while stillstrong, is not as strong as it was in those early days. Competition between research teams and services has reduced, pasrtly because of consolidation in almost every area: agents, institutions, independent houses. Therre was always a problem within the agents, which would occasionally flare up: the agent might be marketing an asset for a client but the research from the sale agents might be forecasting poor returns for that type of property. To be fair, most agents dealt with it well, but some clients were not happy giving instructions to an agent whose research advice was in conflict. Econometric modelling appeared to give much better forecasts in those days, which suggests that cycles wrre somewhat more stable and predictable. The other factor was that holding periods for property became increasingly longer, because of increased transaction taxes, and the flexibility to re- arrange portfolios in response to changing expectations became weaker. A final pint, which might prove controversial, was that the market became more international and more fragmented, with a separation of principal and agent. Fund managers became an emerging group, running co-mingled funds, who primary objective was to invest in the markets rather than to optimise performance. In effect, there developed an excess of naive and inexperienced investors, who did not necessarily tap into the research, actively participating and driving market pricing. Typically, such players develop a requirement for research after they have been active in the market for some time, have gained some experience, and need to be more selective in their purchases. This is more or less where we are today. Research relies somewhat less on econometric modelling and more about understanding the variables in both quantitative and qualitative terms. This provides a service which helps investors understand what is currently happening. They are truly researchers, assembling data, and building out a picture that forms the backdrop of investment decisions. They forecast less using econometric modelling and more in broad terms, understanding investor and market psychology. The tools available - PCs, internet news services, social media, software, and modelling enable the researchers to respond rapidly to changes in the drivers and this is all to the benefit of their clients. Teams are smaller, but they are more expert, specialising in particular aspects of the market, and undertaking more original research. Ten problems experienced by property (and other) researchers

Having produced and read much research over the years, we set out some of the pitfalls

and problems that researchers have in producing meaningful output

1. Confusing will with may It is amazing sometimes how arrogant researchers can be. Instead of writing that ‘I believe that rental value growth could be as high as 10% next year’, the researcher writes ‘Rental value growth will be 10% next year’. A slight variation on the theme is when the researcher assumes that the market knows and accepts what the searcher says, so that ‘Because of the forecast of 10% rental value growth next year, investors will be driving yields down this year’. Assuming projections are true carries the danger of taking you into ‘Lala land’ and, particularly when no explanation is proffered for the 10% forecast (I was once told that it is proprietary black box), readers will very quickly dismiss the numbers as nebulous musings. 2. Confusing supply with demand It might be thought that this is simply not possible. Well, we can excuse lay people, who may not have seen the basic economics chart showing a supply line and a demand line crossing, with that crossing point representing the ‘strike price’ at which supply and demand are in balance. In my experience, however, it is not lay people who get confused, but those property researchers who have never properly experienced the operation of markets. Many of them seem to become confused by the apparent simplicity of the relationship, suggesting that a low strike price actually implies ‘attractiveness’ for tenants or investors. In theory, the crossover point is actually the ‘neutral point’, although in relative terms – between markets – a difference in rent or price more typically reflects a difference in demand. Yes, I accept that there are many times and places – The Netherlands in recent years is a good example – when it is really a supply issue that is keeping rental values low, but these are typically transitional or cyclical factors. Long-term, supply lags demand. A recent property research article illustrates the confusion well. In talking about relative house prices in US cities, it is suggested that Austin is more attractive as a place to buy than New York or San Francisco - because its average house price is lower. Strangely, even a lay person would see the fallacy of that argument: higher prices suggest that occupier demand is higher. That usually occurs when the location is more attractive. Central London’s high house prices indicate high demand and a geographically-constrained supply, not low attractiveness. 3. Using an ‘average’ inappropriately The vast majority of workers in the UK earn less than the average wage – because it is skewed by a small minority who earn a lot more. In this context, talking about the median (the point at which 50% of workers are above and 50% below) is much more helpful. Similarly, markets are not driven by averages, but by activity at the margin. Most owner-occupiers in the UK could not afford the house that they occupy if they were to buy it today: at some time in the past there was a convergence between the price and their ability to purchase, but that does not necessarily persist for very long. But it would be erroneous to suggest – as some researchers might do in a different context – that house prices have to fall because they are ‘unaffordable’. The real question to be addressed is whether there are enough buyers at the margin who can afford properties for sale (which is, itself, also at the margin of the total stock). Neither of these two groups, the buyers nor the properties, are representative of the market as a whole. Economists love talking of averages, but that is often because that is what their data typically provides and they do not really understand how markets function. In the example above about the Austin market, average house prices were quoted. But that only reflects the weighting of the stock or transactions. It tells you nothing about range. It is very conceivable that a market with highest average price may encapsulate some of the cheapest property available (as well, of course, as some of the most expensive). As expensive properties imply high earnings and/or wealth, these can represent ‘attractive’ areas in which to be economically active. Medians and ranges are more useful in really understanding what is going on. Statistics should not be used blindly or as a marketing exercise, but as a means of examining a scenario and understanding it better. 4. Overweighting survey results Are you underpaid? I thought so. Having checked with other people, this means that virtually everybody in the country is underpaid. We should expect mass resignations, strikes, outward migration, uncooperative workers, etc., very soon. That sort of conclusion is what comes from listening too much to what people say and not placing enough weight on what they think or do. Confidence surveys are a classic example. They purport to be a leading indicator, but those analysts who have tried to use them for such purposes quickly realise that they are more coincident indicators (and even then, they are often misleading). Yet naïve researchers will typically turn to them when they try to identify what is going to happen next in the retail sector. Similarly, when you ask investors where they intend to invest next, the answers are often telling you where they were investing last month (please can everybody else buy these now and push up the prices) or where they are definitely not buying (why should I ramp up the prices in advance of my purchases?). Many of the other, typically not active in the market, will just follow the received wisdom in their responses. Instead, researchers should be asking themselves about investor motivations to discover why what is happening is happening. Factor in some reasonable expectations for the future and you may even have a realistic forecast. 5. Thinking outliers are frequent Another article that I saw recently, discussing how to hedge property investments, suggested that outliers are a fairly frequent event. No, actually, they are fairly infrequent: that is why they are outliers. If you continuously hedge against the possibility of them, you will probably slowly go bankrupt. That is not to say that you should take precautions against them: but looking before you cross the road is far more sensible than paying somebody to cross it for you. A few years ago, I heard that one new weather record is established on average every month, in the UK. When you consider the number of meteorological factors (wind speed, temperature, rainfall, etc.), methods of measuring (highs, lows, averages, peaks, rates, sequences, etc.), and the number of geographic locations, that may not be as surprising as it sounds. But the vast majority of such records do not matter. The same with other outliers: as long as you come through relatively unscathed or can recover in a reasonably quick period, the effect is minimal over the long-term – and is usually just a small blip on the chart. The big ones, such as the 2008/9 recession, are a once in a sixty-year event. The frequency can change, but that is a different argument, and we have seen no case being made for that. Of course, if you really believe that outliers are more frequent than the statistics suggest, then I suggest trying the national lottery. 6. Using misleading charts I regularly see property researchers being confused by charts. That is not surprising as the different forms of representing the same raw data, such as growth rates, real growth rates, indices, relative lines, log/linear scales, frequency, etc., all require some mental interpretation. No chart can provide a fully comprehensive picture, so it is understandable that the perspective provided will be partial. But what is inexcusable is a deliberate attempt by researchers to mislead. The most common misleading practices are to (a) start the chart at a particular point in the cycle (high, low, mid) to distort what then purports to be a trend to make a point (b) use a linear y-axis to demonstrate that the trend rate of growth is accelerating or is stable when it is actually slowing (c) refer to one or two new data points as a ‘trend’ when it is really just noise (d) omitting data which contradicts the argument being made to boost the credibility of the case (e) not explain or even understand the definition of the data series (in my opinion definitions are as important as the data itself) or (f) use poor quality data sources (e.g. Argentinian official inflation rates) knowing them to be defective and not say anything. In falling markets, which are perceived to be ‘unhelpful’ to fee earners, it is very common to see forecasts being tacked on to the end of the data so that the current data point becomes the inflection point. Look back at the interest rate or other economic forecast produced over the last five years and be amazed at the number of turning points – just about to happen – which are shown. I have seen all of these techniques being used. Users of charts need to be wary because the onus is on the viewer to interpret them. 7. Discovering something the market knows There is a real danger for younger or less experienced researchers in suddenly having an ‘Eureka’ moment when it all becomes clear to them. This may come after an extended period of confusion and is, therefore particularly welcome. The problem is that such a revelation may well be fully known and understood by most market practitioners. It is just that the researcher did not know. That does not, of course, make it worthy of propagation. The corollary is assuming that almost nobody else knows and then suggesting a strategy/tactic of buying or selling (or developing), as appropriate, on the basis of the analysis. Of course, if the market really does know, then it will be more or less priced in by it, and there will be little or no advantage to be gained from the researcher’s ‘intelligence’. Researchers, be warned that your target audience also reads the Financial Times, The Economist, etc. Repeating what the journals say, but with less authority, is going to annoy your readers. Similarly, merely rehashing agents’ data is will not be news to the market. This sort of thing happens frequently, as a recent article by a fund manager on the office development pipeline in Central London illustrates. Almost everybody in the market can see the cranes. 8. Not believing that clients can differentiate between propaganda and research

Ltd