This dissertation consists of three essays examining household energy decisions and behavior. The first essay examines the adoption of energy efficient Energy Star home appliances by U.S. households. Program effectiveness requires that consumers be aware of the labeling scheme and also change their purchase decisions based on label information. The first essay examines the factors associated with consumer awareness of the Energy Star label of recently purchased major appliances and the factors associated with the choice of Energy Star labeled appliances. The findings suggest that eliminating identified gaps in Energy Star appliance adoption would result in house electricity cost savings of \$164 million per year and associated carbon emission reductions of about 1.1 million metric tons per year. The second essay evaluates household energy security and the effectiveness of the Low-Income Home Energy Assistance Program (LIHEAP), the single largest energy assistance program available to poor households within the United States. Energy security is conceptually akin to the well-known concept of food security. Rasch models and household responses to energy security questions in the 2005 Residential Energy Consumption Survey are used to generate an energy insecurity index that is consistent with those found in the food insecurity literature. Participating in LIHEAP is found to significantly reduce household energy insecurity score in the index. Further, simulations show that the elimination of the energy assistance safety net currently available to households increases the number of energy insecure households by over 16 percent. The third essay develops a five equation demand system to estimate household own-price, cross-price and income elasticities between electricity, natural gas, food at home, food away from home, and non-durable commodity groups. Household cross-price elasticities between energy and food commodities are of particular importance. Energy price shocks reduce food expenditures for low-income households, as indicated by negative cross-price elasticity estimates for food and energy commodities. Additionally, low-income households reduce energy expenditures more than other households, further indicating ``heat or eat'' behavior. Results from all three essays provide policy makers with helpful information to shape future federal energy programs. / Ph. D.

Res-IRF is an energy-economy model of heating consumption in French dwellings developed at CIRED and calibrated against 2012. It is meant to project the evolution of the building stock and the heating demand by 2050 in response to socio-economic parameters, such as energy price and population, and public policies, such as thermal regulations or incentives for renovation. Res-IRF captures the relevant determinants of household decisions related to energy efficiency improvements and energy demand (sufficiency). The aim of the work presented in this report is to calibrate the model against a past year, to run it from this start date to 2012, and to compare the simulation results with observed data on this period. After an overview of the French residential sector in the last 40 years, this report aims at presenting the model and how it was calibrated against year 1984 and adjusted to the past situation of the building stock. Then, the results of a sensitivity analysis on key parameters of the model are compared to reality and it is discussed how the model can be improved to fit the data better. The main results show that the model accurately replicates the evolution of the building stock until 2012. However, the results do not fit well the data of repartition of heating fuels, especially for fuel oil and natural gas. This may be due to the structure of the model which allows fuel switch only for renovating dwellings; then it could miss possible fuel switches from fuel oil to natural gas without renovation due to the expansion of the natural gas network in France between 1980 and 2000. Furthermore, the actual unit consumption, which is a key output of the model, is well replicated by the model, with an error of 5 to 10%. / Res-IRF är en energi-ekonomi modell av värmebehovet i franska byggnader utvecklad av CIRED och kalibrerad mot data för 2012. Det är avsett att förutsäga utvecklingen för byggnadsbeståndet och värmebehovet fram till 2050 med utgångspunkt från socio-ekonomiska parametrar såsom energipriser och befolkningsmängd, politiska beslut som regleringar rörande uppvärmningssektorn och incitament för renoveringar. Res-IRF fångar upp de relevanta faktorer som påverkar hushållens beslut relaterade till förbättringar av energieffektiviteten och energibehoven. Målet med arbetet som presenteras i denna rapport är att kalibrera modellen mot ett redan passerat år, att köra modellen från startåret till 2012, och att jämföra simuleringsresultaten med verkliga observationer för denna period. Efter en översikt över den franska bostadssektorn de senaste 40 åren, följer i rapporten en presentation av modellen och hur den kalibrerades mot året 1984 och sedan anpassats till det dåvarande läget i byggnadsbeståndet. Därefter jämförs resultaten av en känslighetsanalys av nyckelparametrar i modellen med verkligt utfall och en diskussion följer om hur modellen kan förbättras för att bättre passa verkliga data. Huvudresultaten visar att modellen på ett korrekt sätt avbildar utvecklingen av byggnadsbeståndet fram till 2012. Däremot ger resultaten inte god överensstämmelse vad gäller fördelning av bränslen, speciellt inte fördelningen mellan olja och naturgas. Detta kan bero på modellens struktur, som tillåter bränslebyte bara vid renovering; därmed missar den bränslebyten som görs utan samtidig renovering, som tillkommit på grund av utbyggnaden av naturgasnäten i Frankrike mellan 1980 och 2000. Vidare visar modellen god överensstämmelse vad gäller energitillförsel per enhet, vilket är en nyckelparameter bland resultaten från modellen. Denna parameter predikteras med ett fel av 5 till 10%.