Asset-Specific Forecasting: Integrating Qualitative & Quantitative Approaches

Asset-Specific Forecasting: Integrating Qualitative & Quantitative Approaches

The Importance of Asset-Specific Forecasting

Real estate investment decisions require accurate forecasting, not only at the macro or market level but also, critically, at the asset-specific level. This granularity allows for better informed investment strategies, risk management, and ultimately, superior returns. Asset-specific forecasting acknowledges that each property possesses unique characteristics that influence its performance, irrespective of broader market trends.

Linking Strategic and Asset-Level Forecasts

A robust approach involves a clear connection between the strategic (market-level) and asset-level forecasts. This can be achieved through a structured framework, such as the “four-column approach” presented in the file:

1. Benchmark Market: The overall market (e.g., national office market).
2. Specific Sub-Market: The specific sub-market where the asset is located (e.g., downtown office market).
3. Specific Asset: The individual property itself.
4. Comment: Detailed qualitative justification for any deviation between the asset’s forecast performance and the benchmark or sub-market.

This structure provides transparency and allows for reasoned adjustments if the outlook for the benchmark market changes. The “Comment” column is particularly important as it enables a clear explanation of why an asset might outperform or underperform the wider market.

Example:

Forecasting Approaches: A Combined Arsenal

There isn’t a single “right” approach to forecasting. The optimal method depends on the forecaster’s objectives, available resources, and data. The two main categories are:

• Informal Approaches: Rely on market experience, intuition, and sentiment, commonly used within real estate’s entrepreneurial traditions. These forecasts can often be heavily influenced by market sentiment and gut feeling.

• Formal Approaches: Can be broadly divided into:

  • Quantitative: Methods using numerical data and statistical techniques.

    • Time Series/Trend-Based: Analyze historical patterns without necessarily identifying causal relationships.
    • Causal/Structural: Build models with strong theoretical underpinnings to explain relationships between variables.
      • Qualitative: Incorporate expert opinions, surveys, and analogies (historical or geographical).

The most appropriate technique depends on what you want to forecast:

• Income Return: Emphasis on understanding trends and forecasting real estate occupational markets to forecast Net Operating Income (NOI). This is suited for quantitative approaches.
• Capital Return: Requires understanding the likely change in NOI and trends in capital markets. This requires a combination of quantitative and qualitative approaches.

Quantitative Modeling Approaches

Time Series/Trend-Based Models

• These models are empirical and aim to identify repeating patterns in historical data. They don’t necessarily explain why these patterns exist, but assume they will continue.
• Examples include identifying long-term trends, seasonal fluctuations, cyclical patterns, or short-term momentum.
• Advantages: Limited data needs, quick and easy to develop, useful as a baseline.
• Disadvantages: Lack of explanatory power, accuracy decreases further into the forecast horizon.
• Key Assumption: The past is a good guide to the future.

Mathematical Representation:

• Smoothing Models (e.g., Moving Averages):

  • TR_t = α + ρ_t + β₁ρ_t-1 + β₂ρ_t-2 + … + β_qρ_t-q

    • Where:
      • TR_t = Total return in period t
      • α = Constant
      • ρ = Disturbance term
      • β = Coefficient
      • q = Period over which the model is calculated
    • Regression Models (e.g., Autoregressive):

  • TR_t = α + β₁TR_t-1 + β₂TR_t-2 + … + β_qTR_t-q + ε_t

    • Where:
      • TR_t = Total return in period t
      • α = Constant
      • β = Coefficient
      • ε = Error term
      • q = Period over which the model is calculated

Practical Applications & Experiments:

1. Moving Average Example: Calculate a 3-year moving average of property values in a specific submarket to predict the next year’s value. This is a simple smoothing technique.
2. Autoregressive Model Experiment: Use historical quarterly rental growth rates for an office building. Fit an AR(2) model (using the two previous quarters’ growth rates to predict the current one). Evaluate the model’s performance using metrics like Root Mean Squared Error (RMSE).

Causal/Structural Models

• These models are based on theory. They link real estate returns to fundamental independent variables (e.g., demographics, economics).
• Aim to identify causal relationships that drive real estate performance.
• Advantages: Strong theoretical basis, can provide insights into why returns are changing.
• Disadvantages: Require more data, rely on forecasts of independent variables, can be complex.
• Key Assumption: The past relationships between variables will hold in the future.

Mathematical Representation:

• Single Equation (Multiple Regression):

  • TR_t = α + β₁X_1t + β₂X_2t + β₃X_3t + β₄X_4t + ε_t

    • Where:
      • TR_t = Total return in period t
      • α = Constant
      • β_i = Coefficient for variable X_i
      • X_i = Independent variable (e.g., GDP growth, population growth, interest rates)
      • ε = Error term

Practical Applications & Experiments:

1. Office Demand Model: Develop a regression model where office demand is a function of employment growth and business investment. Collect historical data and estimate the model’s coefficients.

2. Systems of Equations (More Complex): A nested set of individual equations.

   • The results from one equation feed into another as an input.

   • Requires that the nested system of equations should be stable
     (namely that a change in an input variable in one of the equations leads only to a change in the
     output of that equation and outputs of other equations in the system but does not lead to a change in the specification of
     the other equations).

   • Simplicity is an advantage.

Qualitative Approaches: The Human Element

• Qualitative methods are vital due to the limitations of quantitative modeling in real estate (e.g., data quality, forecasting independent variables).

• These methods can be used as:

  • A pure alternative to quantitative models.
  • A check on quantitative results.
  • In combination with quantitative techniques.

  • Examples:

  • Expert Opinion/Surveys: Gather insights from experienced market participants.

  • Delphi Methods: A structured process for gathering and refining expert opinions.
  • Historical/Geographical Analogy: Draw parallels to similar situations in the past or in other geographic locations.

Qualitative Modeling of Yields/Capitalization Rates:

• Yields and cap rates reflect market sentiment and risk appetite.

  • K = RFR + RP – G + D

    • Where:
      • K = Yield or capitalization rate
      • RFR = Risk-free rate (e.g., long-term government bond rate)
      • RP = Risk premium for real estate investment
      • G = Long-term average rental growth rate
      • D = Long-term average depreciation rate (or annual investment to maintain quality)

        Expert opinion or surveys of expert opinion and Delphic
        methods all provide a sense that a forecast is grounded in a
        ‘market view’ and hence a degree of reassurance that an
        investment decision is being rationally made.

Practical Applications & Experiments:

1. Risk Premium Survey: Conduct a survey among real estate investors to assess their required risk premium for different property types in different locations. Analyze the results to identify key factors influencing risk perception.
2. Delphi Method for Cap Rate Prediction: Assemble a panel of real estate experts. Conduct multiple rounds of anonymous questioning and feedback to converge on a consensus cap rate forecast for a specific asset.

Pitfalls of Qualitative Methods:

• Anchoring: Over-reliance on current values or trends.
• Confirmation Bias: Over-interpreting data that confirms existing beliefs.
• Groupthink: Pressure to conform to a dominant opinion.

Forecasting in Practice: A Pragmatic Approach

The “best” forecasting system is the one that works consistently. Organizations often adopt a pragmatic view, focusing on models that have demonstrated a strong correlation with actual results over time.

A Conceptual Framework for Forecasting Real Estate Returns:

1. Model Occupational Markets: Use causal/structural quantitative techniques to forecast determinants of NOI and derive income return.
2. Forecast Yield/Cap Rate: Use qualitative techniques (expert opinions, historical analysis) to forecast yield or cap rate.
3. Calculate Capital Return: Use the forecasted cap rate and NOI to project future property values and capital return.
4. Sum Income and Capital Return: Calculate the total return.

Integrating Demand and Supply Dynamics

• Demand Forecasting:

  • Best suited for causal/structural econometric techniques.

  • Demand for real estate is a product of the number of people in a defined trade area and their average wealth, and
    inversely related to the price of real estate in the previous period.

  • Key generic variables and their
    interrelationships: demographic, economic and
    employment.

• Supply Forecasting:

  • Short term (two to three years):
    • monitoring real estate
      projects under construction and building a granular database or
      simple causal/ structural techniques.
  • Long term (more than three years):
    • projecting completions using time-series/
      trend-based techniques.

The Future of Forecasting

The availability of high-quality, long-term data sets for real estate investment performance and independent variables is increasing. This will improve the accuracy and reliability of quantitative models. Integration with big data and machine learning techniques offer the most promising way for improving Real Estate Forecasting.