For evaluating the effectiveness ofefficiency improvements in domesticspaceheating, this study provides some important insights. First and foremost, we find that on average U.S. households ‘takeback’ about 30% ofthe energy savings from improvements in energy efficiency. That is, US households only manage to realize 70% of expected savings. Thus, an intervention that would have been planned based on assuming perfect elasticity would have the realized a 43% lower cost-effectiveness than initially estimated. Rebound in spaceheating is thus non- negligible in the US domestic sector. Consequently, including these behavioral responses into models of projected energy savings and use may provide more a realistic picture on the effectiveness of policy interventions and thus help to improve decisions over competing strategies. On the other hand, it needs to be pointed out that our empirical estimate ofthereboundeffect is well below the 60% mark quoted in earlier literature (Sorrell and Dimitropoulos, 2008). That is, despite a non-negligible degree of ‘takeback’, energy efficiency improvements in domesticheating yield considerable reductions in energy demand, suggesting that measures to improve domesticheatingefficiency are valuable policy instruments for curbing domestic energy use. This clearly underlines the argument by Gillingham et al. (2013) that the existence ofrebound effects alone is not sufficient to justify the rejection of energy efficiency policies.
October 16, 2018
Improvements in cars’ fuel efficiency may induce people to travel more, taking back some ofthe potential fuel savings. This behavior, known as the (direct) reboundeffect, has received much attention in the lit- erature. However, no consensus has been reached regarding its size or the methodology to measure it. In this paper, we estimate thereboundeffectfor private vehicle transportation in Switzerland using individ- ual household data. We estimate a system of equations that explain travel distance, fuel efficiency, and vehicle weight using seemingly un- related regressions. Our results point to substantial rebound effects ranging from 57% to 82%, which lie at the higher end ofthe estimates obtained in the literature, but concur with findings in other European countries. Importantly, our results also indicate that OLS estimates oftherebound tend to be under-estimated rather than over-estimated as usually assumed.
a limited number of respondents.
Concerning theeffectofthe scenario dummies, i.e., ofthe magnitude oftheefficiency improvement, we find that larger efficiency improvements lead to a lower directrebound. This finding can again be explained in the framework of hierarchical choices and non-infinite substitution. People do not rebound proportionally to efficiency improvements: Once they reach an optimal comfort level, they stop. In the perspective of policy implications, it implies that one-shot large efficiency improvements should be favoured, because less energy savings would be lost through thedirectrebound. Hypotheses on who will reboundthe most have often been made in previ- ous research, but very seldom tested due to the lack of data. Our findings highlight important facts: First, the magnitude ofthedirectrebound is very heterogeneous among households. Second, large improvements do not mechanically imply a higher rebound than smaller improvements. Third, accounting forthe tenant/owner status is not sufficient, because the struc- ture oftheheating bill matters, and often differs between the two groups. Fourth, measures targeting low-income households (for instance subsidies conditional on income) will be less effective in terms of energy saved, be- cause they are associated with larger directrebound effects.
A first group of assumptions concern solar radiation and the geometry ofthe buildings. The procedure described in GBG (2012) is only implemented for two different alternative inclinations, 90° and 45°. This implies that simulated buildings possess only perfectly vertical walls (90° inclination) and that there is just one type of roof (45° inclination) for buildings with non flat roofs. Moreover, it is assumed that the external walls of buildings are only to be aligned on the basic cardinal points: North, South, East and West and that they are equally distributed. These simplifications are made in order to deal with the fact that GIS information, including detailed characteristics of roofs and walls (LoD2), is still scarce. The characteristic of having a flat roof, or not, can be derived from the model buildings in the German domestic buildings´ typology, available in Loga et al. (2011) . Similarly, in the case that there is no further available information, the distribution of windows to the different orientations is to be derived from factors based on the model buildings ofthe TABULA typologies. The second relevant group of assumptions to be made corresponds to the internal temperature set- points. The proposed model does not include integration to a statistics based bottom-up model forthe determination ofthe internal temperature set-points according, for example, to the socio- economic characteristics ofthe occupants ofthe building. Nevertheless, it is expected that newer buildings present higher internal temperatures due to factors such as for example thereboundeffect generated due to the lower cost ofheating (Born et al., 2003; Haas et al., 1998; Schuler et al., 2000). Analogously, old and badly isolated houses usually have lower internal room temperatures due to the significantly higher costs ofheating (Born et al., 2003; Loga et al., 2003). In the German case, it has been found that actual average internal room temperatures can range between 15°C in old and large buildings, and 24°C in modern and compact buildings (Loga et al., 2003; Loga, 2005). Unfortunately, a well defined function to describe this situation could not be found and therefore fixed values forthe different construction year classes are used. 16° C for construction year classes A to E, 18° C for F and G, 20° C for H and I, and 23°C for J.
Without a doubt, environmental awareness has been increasing in the last decades. This trend is reflected in the growth of a number of environmentally-friendly products (see e.g., Chen, 2011 and Hunt and Dorfman, 2008), from fluorescent bulbs to organic food (see e.g., Torjusen et al. 2001). As the example on fluorescent bulbs shows, environmental consciousness should translate into socially better decisions and energy saving behaviours. Surveys shows that wealth and status are often associated with green knowledge and general concern towards environmental quality (Diamantopoulos et al., 2003). However, there is also an idea deeply rooted in public discourses that environmental awareness translates itself more into slogans than actions. A green discourse is not reflected by actual green actions, or at best, these actions are very marginal. A Google search ofthe terms such as “Green hypocrisy” or “Environmental hypocrisy” returns 31,000 pages, more than 9,000 of which from last year (Google accessed on 15 May 2012), which deal with conflicts between the lifestyles of (sometimes very wealthy and famous) members of green parties or groups advocating energy savings and carbon neutral policies while leaving large footprints behind.
of 1 K-1.5 Kfor the RSM case.
Applying the efficient user behavior, all rooms are again too warm while the user is absent. The rooms do not cool down as far as the set-point. If the user is present, the temperatures are below the set-points, with the bedroom once again as exception. As the valves where usually fully opened, theheating sys- tem is not powerful enough to reheat the rooms on time. The deviations from the set-point are much larger than in the standard case with up to 4 K and will affect the thermal comfort ofthe user. Simulations using the setup with low or high energy demand show the same results for standard and efficient behavior. The deviations are less pronounced in both cases because the rooms are better insulated resulting in warmer rooms even if it is unheated (low energy scenario) or because theheating system is more powerful (high energy scenario). In figure 4.3 a boxplot represents the different distributions ofthe simulated temperatures forthe six reference cases. In a boxplot, the box itself covers the second and third quartile ofthe data, i. e. 50 % ofthe data are present within this box. The line in the box represents the median, not the mean, ofthe data and separates the second and third quartile. The whiskers at the end ofthe box show the extend ofthe first and fourth quartile. The maximum length of these whiskers is 1.5 times the inner quartile distance. Values outside this range are considered as outliers and marked as single dots above the whiskers 1 . The up- per part of figure 4.3 shows the distribution ofthe simulated temperature for all rooms, the lower figure the distributions for all rooms except the bedroom. As mentioned in section 3.2.5 including the bedroom strongly influences the re- sults ofthe measured temperature distribution. The median of all simulated temperatures is lower if the bedroom is considered, due to its lower tempera- ture set-points. Theeffect on the median value of simulated temperatures is approximately 1 K throughout all reference cases.
Our identification in time has shortcomings since the conflict data is only available on a yearly basis. Therefore forthe early interviews we might count confrontations that had not yet happened (our indicator is forthe whole year of 2004 and some interviews started already in October) and for late interviews there might be confrontations we did not count (the interviews continued until the middle of 2005). There are also weaknesses in the spatial identification. Since we only count what happens in the district, the fighting in large districts could have taken place very far from the interviewed household, which would matter if theeffectof violence decreases with distance. On the other hand we underestimate the conflict intensity people are exposed to in small districts, where confrontations happening in neighbouring municipalities are still very close but not counted (often they would be only a few kilometres away). We use different approaches to try to account for this. Our findings are however robust to all those different specifications (not all are reported in the paper). There are also arguments for possible endogeneity issues like reverse causality and unobserved variable bias. Since we do not think that this is a major problem and the discussion is somewhat lengthy, it is not presented in the main results but separately in section 4.4.
The conditions in the reactor pressure vessel (RPV) in the late phase of a core melt accident can differ significantly depending on the failure history ofthe core. If a failure ofthe RPV lower head occurs, the condition ofthe molten core, i.e. temperature, composition, distribu- tion, and mass ofthe melt, together with the design ofthe lower head, determines the loca- tion, shape and size ofthe breach. The decisive parameter forthe mode of melt release is the system pressure at the time of RPV-failure. If the pressure is at the same level as the pressure in the containment, or only slightly above, the molten part ofthe core inventory will flow into the reactor pit by gravity. The release of radioactive aerosols will be small and the further development ofthe accident is determined by the specific plant provisions. The as- sumption that the system pressure at core melt accidents will be low (<20 bar) in German plants is justified, due to the pressure reduction system provided and the fact that a failure ofthe surge line is expected in such cases [Roth94]. However, if the pressure is between 5 and 20 bar, e.g. because of late reflooding and rapid steam generation, the melt will be ejected forcefully into the reactor pit and possibly beyond, even at these low pressures, accompanied by the blowdown ofthe reactor cooling system. In this case it depends on the cavity geome- try whether the melt will be trapped without severe consequences in places with long term cooling capabilities or whether it will be dispersed into the containment atmosphere with un- favorable effects upon the accident progress. The finely fragmented melt particles lead to efficient debris-to-gas heat transfer, hydrogen generation by metal/steam reactions in the reactor pit and hydrogen combustion in the containment. These processes, referred to as Direct Containment Heating (DCH), may cause a rapid increase in temperature and pressure in the containment and may have an impact on vital safety components (see sketch below).
4.1.1 Results forthe full sample
In the following, we present and discuss the findings forthe full sample in some more detail.
Socio-demographic characteristics: The results presented in table 8 show that socio-demographic YDULDEOHV DUH LPSRUWDQW GHWHUPLQDQWV RI WKH KRPHRZQHUV¶ GHFLVLRQV ,Q SDUWLFXODU Income, Age and University seem to be relevant variables. For all types of RHS, Income is significant. Income has a positive impact on the probability to choose either GAS-ST (0.017 *** ) or HEAT-P (0.013 * ), for an otherwise equivalent household. For OIL-ST and PELLET, we observe a negative relationship. The probability to choose either OIL-ST (-0.016 *** ) or PELLET (-0.014 ** ) decreases with a higher income level. However, all income-related effects are minor. These findings correspond to the findings from other studies. For instance, Braun (2010) finds forthe case of Germany that richer households avoid oil- and solid fuel-based RHS, while they favor to use gas-fired RHS. 35 Vaage (2000) finds for Norway that wealthier households tend to choose electricity as the only RHS, while combinations of electricity with solid fuels (biomass) or oil are rather unpopular. The variable Age is barely significant for all RHS except GAS-ST. Age has a positive impact on OIL-ST (0.004 *** ), while it has a negative influence on HEAT-P (-0.003 *** ) and PELLET (-0.001 ** ). This implies that older homeowners favor an oil-fired solution, while younger homeowners are more open towards innovative RHS, such as a heat pump or wood pellet-fired boiler. Age may also reflect differences in the risk aversion. The level of education (University) has a positive impact on GAS-ST (0.030 * ) and HEAT-P (0.046 ** ), while it has a negative influence on PELLET (-0.063 *** ). These findings also correspond to Braun (2010), who shows that households with higher education tend to apply gas-fired solutions, while households with a lower education tend to have solid fuel-fired systems installed. The findings for Income and University can DOVREHPRWLYDWHGZLWK%HFNHU¶VKRXVHKROGSURGXFWLRQWKHRU\FIVHFWLRQ7KHYDULDEOHFemale is
4.4 Fully separated interactions
For fully separated cases the hot spot is generally found ahead ofthe protuberance and its magnitude is independent ofthe state ofthe incoming boundary layer. While this could possibly be related to stagnation conditions as also hypothesized by Nestler (1985), no supporting evidence is found with the present dataset neither since the physical mechanisms that induce high heating ahead of protuberances with separated interactions do not directly correlate with the relevant relations considered in the classic stagnation heat transfer theory (Fay and Riddell 1958). Given that the highest heat flux ahead ofthe α=135° cases is strongly dependent on the location where the separated flow reattaches to the surface and also since there is a clear increase in heat flux as the deflection angle is increased, it seems obvious that the maximum heating is linked to the reattachment ofthe flow ahead ofthe protuberance also (Section 4.5). In this way, as the protuberance deflection angle is increased, the deflection that the incoming flow experiences before reattaching to the wall is lower and so its impact energy to the surface is higher. This is analyzed in more detail in Appendix A, where a further assessment ofthe results is performed. Based on this analysis, Fig. 19 shows a correlation ofthe present experimental dataset in terms ofthe dominant parameters: Reynolds number, Mach number and protuberance deflection angle and height. The actual thermal capacity ofthe flow is already taken into account in the definition of Stanton number. The correlation provided is subject to a total conservative uncertainty of ±25% while also considering the ±10% uncertainty inherent to the measurements. The maximum heat transfer to the side ofthe protuberance is as shown in Eq. 6. The hot spot will nevertheless be located ahead ofthe protuberance in most cases following the relation in Eq. 7. In these cases the hot spot will take place as shown in Eq. 8. Only in low-Reynolds low-deflection cases this will be located to the side ofthe protuberance as in unseparated interactions (Eq. 5).
When looking at the economic factors influencing rebound behavior, it quickly becomes clear that a restriction to the financial analysis is not sufficient to adequately explain rebound effects. For example, the needs of consumers must be taken into account. They reflect consumers' preferences and are therefore essential for rational decision-making. From a purely economic point of view, an increase in demand takes place until the satisfaction of needs reaches a maximum in compliance with the given restrictions. A reboundeffect in the economic sense thus goes hand in hand with a rational decision-making process. The overconsumption compared to the initial level thus reflects the new consumption choice ofthe individual, brought about by a change in the restrictions.
in or crowds out domestic investment for 12 countries in each of three developing regions (Africa, Asia and Latin America) from 1971 to 2000 and found that foreign direct investment crowds out domestic investment in Latin America and has generally left domestic investment unchanged. Ndikumana and Verick (2008) further revealed positive impact of foreign direct investment on growth and that private investment enhances foreign direct investment positively too. However, fixed effects models treats variables as if they are non-random since it has no control for variables that vary over time while introduction of more dummies may lead to over dampening ofthe model. In examining how foreign direct investment affects economic growth, Gui-Diby (2014) used 50 African countries from 1980 to 2009. The study employed system generalized method of moment estimator developed by Blundell and Bond (1998) and found that FDI inflows had significant impact on economic growth in African region. In a similar vein, Sukar, Ahmed and Hassan (2007) examined theeffectof foreign direct investment on economic growth in Sub-Sahara African countries using panel data spanning 1975 to 1999 from 12 Sub-Sahara African countries. They found that foreign direct investment and domestic investment advance economic growth positively. However, since the number of years (25) were relatively more than countries (12 countries), the application ofthe traditional panel techniques such as Fixed Estimator (FE), Instrumental Variables (IV), Pooled Effect, Random Effects, GMM estimators may produce inconsistent and potentially very misleading estimates ofthe average values ofthe parameters in dynamic panel data model.
Provided in Cooperation with:
Swiss Society of Economics and Statistics, Zurich
Suggested Citation: Gonseth, Camille; Thalmann, Philippe; Vielle, Marc (2017) : Impacts
of Global Warming on Energy Use forHeating and Cooling with Full Rebound Effects in
2.1.1 Cognitive and normative approaches towards technology adoption
Cognitive models and theories have their roots in the research around attitude formation and social psychology. In particular, this approach is often applied in research on the adoption of environmental- friendly products or health-related behavior. Two well-known examples for cognitive models include Ajzen and Fishbein’s (1980) theory of reasoned action (TRA) and Ajzen’s (1991) theory of planned behavior (TPB). Both theories assume that behavior is rationally driven and that there is a linear relationship between beliefs and behaviors. The TRA belongs to the class of expectancy-value models. According to Pollard et al. (1999, p. 443) these models “formalized the view that consumers’ anticipated satisfaction with a product (and hence the purchase of that product) is determined by their beliefs that the product fulfills certain functions and that it satisfies some of their needs”. The TRA can be modeled as follows: Behavioral intentions determine actual behavior and can serve as a proximal measure of behavior. Behavioral intention depends on an individual’s attitude towards performing the behavior, and subjective norms (i.e. the influence of peers). However, behavior is not always under an individual’s full control. In other words, this means “the performance of many behaviors depends not only on motivations but also on non-motivational factors like a person’s ability to actually perform the behavior” (Sanhi, 1994, p. 442). An implication is that attitudes and social norms do not provide a sufficient explanation of behavior whenever control over behavior is limited by external factors or personal capabilities. The TPB tries to overcome this problem. For this purpose, the TRA is extended by the component perceived behavioral control (PBC) in order to capture non-motivational factors, such as the availability of resources, the ability to carry out a certain action or environmental constraints to predict behavior more accurately. Ajzen (1991) defines PBC as a person’s belief with respect to how difficult or easy performance ofthe behavior is likely to be. In general, the TPB states that the stronger each ofthe three factors (attitude, subjective norm, perceived behavioral control) is, the stronger is the individual’s intention to perform the behavior. However, these components are not always weighted equally when predicting an individual’s behavior. According to Miller (2005), this means that depending on the individual and the context, these three factors might have very different effects on behavioral intention.
at the beginning of each sparking cycle that lasts ∼ 10µs. Towards the end of a cycle the potential drop will decrease significantly, due to separation ofthe large electron/positron densities which have been produced in the discharge (Rud- erman & Sutherland 1975; Melikidze et al. 2000), and by the thermal release of iron ions which cause an additional screen- ing ofthe potential gap (Gil et al. 2003). The secondary par- ticle electron/positron plasma produced either by curvature radiation, or inverse Compton scattering within the IAR is more dense but less energetic. Its γ-factor is about 4 orders of magnitude smaller than that ofthe primary particles. Sec- ondary charges produced outside the IAR don’t contribute to theheatingofthe real polar cap surface.
and disability than cancer, malaria, traffic accident and wars combined (Morrison and
Orlando, 1999 ) . Although most societies look down upon gendered violence, in India the
reality is that they are often endorsed under the garb of cultural practices, collective norms or religious beliefs. India is ranked 133 among 135 countries in terms of sex ratio in the Global Gender Gap Report (2011). The poor ranking is attributed to female infanticide and the systematic neglect of daughters relative to sons. The prevalence rate ofdomestic violence in itself occupies a large variation among differing reports, from 17% (Martin, Tsul et al., 1999) to 41% (Peedicayil, Sadowski et al., 2004). One possible explanation for this is the non- standardisation of survey questions regarding violence in the various reports and differences in the subjective interpretation of violence. However, a more plausible explanation is the under-reporting of incidences due to the social stigma attached to violence and the underestimation of violence in itself. Actual prevalence of violence in India is therefore at a risk of underestimation and is thought by experts to be much higher than reported.
The chemical sector is commonly identified as one ofthe sectors with the great potential for SHIP due to its large energy consumption in the low and medium temperature segment . For Jordan the chemical sector, including mining and production of bulk chemical commodities as well as fertilizer production is also one ofthe major sectors with potash and phosphate derivatives, nitrogenous fertilizers, chlorides, bromines and Sulphur derivatives as the most important products. Unfortunately there is neither specific data on the energy consumption within the chemical sector of Jordan available nor are there specific targets forthe use of renewable energy for industrial process heating.
products, reference priced and differentiated products that confirmed the hypothesis. Institutions matter most for trade in differentiated products, as do search costs.
The results on organized exchange products indicate that the analysis has not controlled sufficiently forthe role of comparative advantages between countries. The inclusion of GDP per capita captures some ofthe effects, though. To consistently analyse comparative advantages, a different sectoral disaggregation may be necessary as well. Another problem ofthe estimates, is the large number of zero-entried trade flows, especially in the organized exchange group. Rauch (1999) notes that if these observations tend to occur between distant countries that do not exhibit cultural or historical links, the comparison between product types is biased, because the omission of these observations tends to reduce the size ofthe effects related to search costs and insecurity transaction costs. A transformed Tobit gravity model has been estimated for each product group. The results again confirm the hypothesis on the impact of institutional quality, but do not unambiguously support the search cost hypothesis proposed by Rauch (1999). Whether this may be related to problems related to the extended sample of countries, their trade data or the transformed gravity model, remains open to further research. A further extension ofthe paper will link theeffectof governance on bilateral trade patterns to an analysis ofthe impact of variation in institutional effectiveness on export specialization and the importance of intra-industry flows. If good governance is important for exporting and importing differentiated goods, this may provide a new link between dynamic specialization, economic development and intra-industry trade.
The Bluetooth wireless communication system has proved very robust over a range from 30 to 100m, depending upon the devices used (eg type 2 or type 1 respectively). Battery life is limiting at the moment, but some of that is down to the GP board. A more readily wearable system would use one ofthe “nano-watt” series of PIC devices . This would go into a periodic sleep mode consuming literally nano-watts. Upon demand (or at timed intervals) it would then wake-up, turn on the sensors, take measurements, transmit the results and then go into hibernation again. It is the base station that takes decisions regarding events and actions. It would be possible forthe server version to allow the persons healthcare professional to set or modify rules that would trigger events such as “inform local health centre” or “call an ambulance”. The base station is mains fed and permanently powered, so there need be no loss in monitoring activity.
ofthe former study is checked by employing four additional waves of data forthe years 2006 to 2009. Second, expanding on the single-car focus of F RONDEL , P ETERS ,
and V ANCE (2008), the data set analyzed here includes multiple-vehicle households, thereby allowing us to explore the sensitivity ofthe estimates to their inclusion. Third, we add a fourth definition ofthereboundeffect relying on the fuel price elasticity of travel demand and argue that for empirical reasons, therebound should be preferably estimated on this basis. Finally, in addition to providing for average effects across all types of households, which serve as a reference point, the estimates using quantile re- gression indicate that the magnitude ofthe estimated reboundeffect depends inversely on the household’s driving intensity: Households with low vehicle mileage exhibit re- bound effects that are significantly larger than those for households with high vehicle mileage.