Does more efficient lighting actually increase total energy consumption?

Historical Impact of New Lighting Technologies

A couple of weeks back I came across a story in the Economist‘s science section asking this exact question.   The Economist’s story originated from a recent article (Tsao et al 2010) which contained some rather interesting data on how the total amount of light generated/used has increased over the last 300 years with more efficient lighting technologies (e.g., candles to electric bulbs) as shown below:

From Tsao et al 2010 (who in turn adapted this from Fouquet and Pearson 2006) - showing three centuries of light consumption in the UK. Y-axis is teralumen-hours per year, and the separate lines represent different lighting technologies used. The black line at the top represents total light produced by all technologies active at any given time.

In addition,  the empirical relationship between the amount of light used and GDP per person-year  over the same 300 year time period is also presented in the paper – see the figure below:

From Tsao et al. 2010. Y-axis shows overall light consumption (petalumen hours per year) while the x-axis shows per-capita GDP ($/(person-year) divided by ownership cost of light in dollars per megalumen-hour. The filled circles represent actual data points for different countries or groups of countries at different times (e.g., UK 1750 is the United Kingdom in 1750). The white diamond at the top of the figure represents projected overall world light consumption in 2030 - assuming solid state lighting penetrates the market but a business-as-usual cost of energy.

A couple of conclusions are evident from the two figures above:

  • The introduction of new, more efficient lighting technologies over the last 300 years has NOT decreased the cumulative amount of light produced (top figure).
  • The number of lumens generated relative to the GDP per person year has remained largely constant even as GDP has increased – as illustrated by the straight line in the second figure.
  • Further, the associated amount of energy consumed for lighting shows the same linear increase as the second figure above – meaning that as GDP per person-year has increased over the last three centuries, so to has the cumulative amount of energy consumed for lighting.
  • The above data shows that the introduction of new lighting technologies DID NOT reduce the  cumulative energy consumption associated with lighting – but rather INCREASED IT!

Thus, one of the main points made by Tsao et al. is that the development and market penetration of highly energy-efficient solid state lighting today will likely lead to FURTHER INCREASES in the total energy consumption associated with lighting – not decreases in overall energy consumption.   Indeed the white diamond at the top of the second figure gives their “business as usual” prediction of total energy consumption for lighting in 2030 assuming broad penetration of solid state lighting into the market place.

Jevons Paradox: Increasing Efficiency Leads to Higher Consumption

The reason that more efficient lighting technologies introduced over the last 300 years have not reduced the cumulative  energy consumed is that the lower costs associated with new lighting technologies mean people could (and did) use more of it, thereby completely offsetting any efficiency gains offered by the new technology.   The more challenging question is why we humans keep using more light?   Explanations involve factors such as:

  • Humans do not like the dark and much prefer light.
  • Light is what enables us to be productive 24/7, thereby increasing our productivity and economic condition.  With light, we can be productive inside buildings and are not constrained to simply work outdoors during seasonal daylight hours.  Thus, light has been an enabler of GDP growth, and increased GDP  has enabled greater deployment of lighting technologies.

The case above of increased lighting efficiency leading to larger energy consumption is hardly unique.   It is well documented that major household appliances (fridge, stove, freezer) were much less efficient in the 1950s than they are today.  Yet overall household electricity consumption is much greater today than in the 1950s because homes today not only have fridges, stoves and freezers – but also microwaves, more lights, coffee-makers, toasters, dishwashers, washing machines, dryers, large screen TVs, computers, video games, cell-phone chargers, etc…

This phenomena of efficiency gains leading to increased consumption is known today as Jevons’ Paradox.  In the 1860s, William Stanley Jevons wrote about how the increasing efficiency of coal fired engines incrased, rather than decreased, the overall consumption of coal.   Similarly, the increasing productivity of agriculture (i.e., more tonnes of grain per hectare of land) has not lead to reductions in the amount of agricultural land.   Instead, efficiency gains contribute to the lowering of prices which increases demand and overwhelms the energy efficiency gains achieved with the new technology.  This relation between efficiency improvements leading to increased energy consumption is known as the Khazzoom-Brookes postulate and has been shown to be consistent with neoclassical economic growth theory.   While the impacts of efficiency gains in a technology can have disruptive effects on particular industries (e.g., coal displaced charcoal and in turn coal was displaced by petroleum all of which lead major changes in industries)  the resulting productivity increases for the overall economy are what really drive such efficiency innovations forward.

Avoiding Jevons’ Paradox

A key implication of Jevons’  paradox is that simply simply improving efficiency will have little long-term impact on overall energy consumption unless demand is not allowed to rise.   A number of mechanisms can be used to limit demand, but these will likely have to be government interventions:

  • Caps on outputs – e.g., a cap on overall levels of CO2e emissions.
  • Caps on inputs – e.g., a cap on the amount of land that can be disturbed, or the amount of water that can be consumed from a watershed.
  • Taxes or licenses fees which keep the cost of use of the new technology the same or even higher.

As I see it, economy-wide energy efficiency gains are only possible if such mechanisms are in place at the larger scale.  Thus, there must be an interplay between the scale over which technological innovation is occurring (e.g., improvement in jet engine technology) and the larger scale over which the impacts of this technology interacts with other technologies, markets and human choice (e.g., a cap on a country’s CO2e emissions).    This is a “top down” approach to achieve a much stronger version of sustainability.  It is an open question (for me anyways) on when “bottom-up” approaches  will achieve such larger scale environmental objectives.


Fouquet, R., and P. J. G. Pearson.  2006. “Seven Centuries of Energy Services: The Price and Use of Light in the United Kingdom (1300-2000).  The Energy Journal, vol 27, #1, pp 139-177.

Tsao, J.Y., H.D. Saunders, J.R. Creighton, M.E. Coltrin, and J. A. Simmons. 2010.  “Solid-state lighting: and energy-economics perspective”.  Journal of Physics D:  Applied Physics.  vol 43, p.1-17

One thought on “Does more efficient lighting actually increase total energy consumption?

  1. Pingback: renewable energy doesn’t offset fossil fuel use | Energy and the Future

Leave a Reply

Your email address will not be published. Required fields are marked *