Essays in Applied Microeconomics

Abstract

Chapter 1: We study the effects of limited attention on property prices and energy-efficiency (EE) investments in the housing market. Using a novel dataset, we analyse over 5 million residential property sale transactions in England and Wales, each containing information about sale price, property and location characteristics, and a mandatory energy performance certificate (EPC). The EPC includes a continuous energy cost rating (SAP) which is mapped into seven colour-coded rating bands (ranging from green A to red G). Applying a Regression Discontinuity Design (RDD), we document significant price discontinuities at the rating band thresholds. We estimate that – holding the underlying SAP score equal – being in a higher rating band increases the final sale price of a property between 0.8% and 2.5% depending on the threshold crossed. These price discontinuities suggest that individuals are attentive to the simpler colour-coded rating band and partially inattentive to the more precise SAP rating. We present a simple model for estimating the degree of inattention and show that, for a given level of attention to EE, rating bands reduce attention to the SAP rating by 25% on average. Importantly, the price discontinuities appear to influence market behaviour: Sellers whose property receives an EPC rating just below a threshold are between 0.4% and 11% more likely to make EE investments before marketing their property. We discuss recommendations of how to best leverage these threshold effects to improve policy design, which can be extended to other settings where the provision of simplified information creates reference thresholds.

Chapter 2: We study the effects of weather conditions on the economic valuation of energy-efficiency (EE) in the UK housing market. Individuals make valuations based on the total utility that they expect to derive from a product and its features. In the UK housing market, utility is almost exclusively realised in the future as buyers make a priced offer over a month before the sale transaction is complete and they are able to either move in or rent the property. While the value of EE features (e.g. insulation) depends directly on the expected weather over the ownership time frame (e.g. for maintaining heat during cold periods), due to its notorious unpredictability, present weather conditions provide little to no additional information about future weather conditions (beyond common knowledge such as seasonal temperature differences). Using transaction-level data of over 5 million residential property sales in England and Wales, we find that weather conditions during the month the buying decision is made can disproportionately influence the EE valuation of properties. In line with intuitive predictions, during rough weather (i.e. cold and rainy) the EE rating of a property has a stronger influence on its sale price than during favourable weather (i.e. warm and dry). We show that these results are unlikely to be driven by energy-cost optimisation behaviour. We model and discuss potential mechanisms for these findings within an intertemporal decision framework (to investigate projection bias and over-inference) and within a limited attention framework (to investigate salience). We provide policy considerations for educational interventions.

Chapter 3: We study the market learning process triggered by the introduction of mandatory energy-efficiency (EE) disclosure policies in the UK housing market. The polices require sellers to provide a standardised energy performance certificate (EPC) to potential buyers. The EPC includes a salient EE rating graph with a numerical energy-cost rating (the SAP rating - ranging from 1 to 100) mapped into seven arbitrary colour-coded letter rating bands (from red G to green A). Analysing over 7 million sale transactions from England and Wales we find that, following the introduction of additional legislation requiring EE rating graphs to be included in all marketing materials (printed and online), the effect of EE ratings on sale prices appears to have reached a near-equilibrium level, providing evidence that both buyers and sellers have learnt to incorporate EE information in their decision processes. Nonetheless, we find that price discontinuities at the rating band thresholds have appeared and are not decreasing over time, suggesting that limited attention can create boundaries for the learning process. We also document that the proportion of properties with low rating bands (e.g. red G) that receive EE investments before a sale is increasing. Conversely, the proportion of properties with high rating bands (e.g. green C) that receive investments has remained mostly constant. Since EE investment costs are convex, these results suggest that, at least in part, sellers rely on EE rating information to optimise their investment decisions. Importantly, we document that the proportion of sellers that base their EE investment decisions on rating band thresholds is increasing over time, suggesting sellers are learning to exploit the price discontinuities described above. We argue that limited attention may prevent information asymmetries from fully self-correcting even when disclosure policies are in place. We discuss the welfare trade-off implications of our results for policy design.