Christian Kaps

In this paper, we aim to understand when private households invest in rooftop solar installations and battery storage, and how these investment decisions affect their electricity usage patterns as well as the market structure overall. We answer three main research questions: 1) What drives customers to combine solar power with storage installations; 2) How does privately-owned storage change consumer autonomy and the grid provider business model, and how heterogeneous is this effect across the population; 3) What effects do subsidies have on investments, demands, and carbon emissions. We develop a structural estimation model of residential electricity usage, that allows us to estimate a household's hourly consumption preferences, and a non-financial utility the household has for using self-generated solar power over grid-procured electricity; we call this utility greenness valuation. Applying this model to a novel, proprietary, big-data-set of German households, we find the median household to have a greenness valuation of 0.29€ per kilowatt-hour(kWh). We furthermore find this sustainability-related valuation in the population follows an exponential distribution and helps explain the early adoption of behind-the-meter batteries. We then show that, in the future, at electricity prices of 38 cents/kWh, a rate seen in Europe in 2022, investing in solar and some amount of storage is optimal for 72% of households, even without any greenness valuation - this would reduce the energy purchased from the grid provider by over a quarter. Lastly, we quantify the amount of carbon saved per dollar spent of subsidies for the households observed in the data-set to be 674€ per metric ton. We show how storage subsidies' additionality depends on technology prices and a household's greenness valuation, while solar subsidies are not needed to incentivize broad adoption at current electricity prices.

What is the impact on electricity prices if the sites for renewable energy projects are chosen based on a novel metric grounded in historic, geospatial data on wind renewable resource quality compared to the currently employed local cost minimization approach? In this paper, we propose such a new metric - which we term "Quality-Adjusted-Power". It takes into account not only the expected energy output of a renewable site, but the value of said energy to the grid at large, based on timing and variability of output. We show that employing this new metric to auction off 5GW of renewable capacity reduces counter-factual whole-sale electricity prices in the studied ERCOT market by 3% - mainly through leveraging the correlation between historic wind speed patterns and market prices and reducing transmission cost.

Solar power output follows a diurnal pattern with peak generation during mid-day while wind-power output has a much less predictable daily pattern, but is less variable than solar overall. When aiming for high renewable penetration in a grid, balancing the relative investments in solar and wind can help match the demand profile and thus reduce cost through reducing the necessity for energy storage. In this paper we develop a model to capture the trade-off between levelized cost of energy as well as renewable and demand correlation. This allows us to study how solar and wind capacities should be balanced to reach certain levels of renewable penetration goals. In particular we are interested in how this balance changes with varying availability of hydro-power as a natural form of multi-hour storage. In collaboration with a Latin-American corporate partner, we then numerically compare our model results to advanced, long-term capacity planning tools for Chile and other countries. Preliminarily, we find that for moderate levels of renewable penetration, the renewable technology with the cheapest expected cost will be used predominantly (solar in most scenarios). However, when one targets renewable shares above 50%-60%, solar capacities almost stagnate and the share of wind increases, as the slightly higher costs of wind generation are off-set by matching the base-load demand well and having less pronounced generation spikes reduces curtailments or additional storage investments. The presence of hydro-power mitigates said effect, but only in regions where it is available reliably and expected to be so in the future. We find considerable heterogeneity between countries with similar hydropower capacity, depending on idiosyncratic weather patterns.

Globally, 1.5 billion people live off the grid, their only access to electricity often limited to operationally-expensive fossil fuel generators. Solar power has risen as a sustainable and less expensive option, but its generation is variable during the day and non-existent at night. Thanks to recent technological advances, which have made large-scale electricity storage economically viable, a combination of solar generation and storage holds the promise of cheaper, greener, and more reliable off-grid power in the future. Still, it is not yet well-understood how to jointly determine optimal capacity levels for renewable generation and storage. Our work aims to shed light on this question by developing a model of strategic capacity investment in both renewable generation and storage to match demand with supply in off-grid use-cases, while relying on fossil fuel as backup. Since the exact model is intractable, we develop two newsvendor-like approximations that are analytically tractable, yield precise values for the optimal investment decisions and profit in some cases, and provide bounds to the optimal investment decisions and profits in all other cases. We use these approximations to obtain additional insights into the problem. First, we find that solar generation and storage capacity levels are strategic complements, except in cases with very high penetration of either technology, when they surprisingly turn into strategic substitutes, with implications for long-term investment decisions. Second, we develop a simple heuristic to determine which technology, within a given portfolio, can turn a profit in the broadest set of market conditions, and thus is likely to be adopted first. We find that currently, low-efficiency, cheap technologies such as thermal can more easily turn a profit in off-grid applications than high-efficiency, expensive ones such as lithium-ion batteries. To conclude, we calibrate our models to measure the accuracy of our solutions utilizing real-life data from three geographically-diverse islands, and then use our approximations to provide high-level insights on the role that large-scale storage will play in the years ahead as technology improves, carbon taxes are levied, and solar becomes cheaper.

In this paper method and practice of cross-efficiency calculation is discussed. The main methods proposed in the literature are tested not on a set of artificial data but on a realistic sample of input-output data of European warehouses. The empirical results show the limited role which increasing automation investment and larger warehouse size have in increasing productive performance. The reason is the existence of decreasing returns to scale in the industry, resulting in sub-optimal scales and inefficiencies, regardless of the operational performance of the facilities. From the methodological perspective, and based on a multidimensional metric which considers the capability of the various methods to rank warehouses, their ease of implementation, and their robustness to sensitivity analyses, we conclude to the superiority of the classic Sexton et al. (1986) method over recently proposed, more sophisticated methods.