Skip to content

Did CFL Subsidies Crowd Out LED Purchases?

New research finds that timing matters for standards and subsidies.

Light bulbs were back in the news last month, with the Department of Energy announcing new tighter U.S. rules for energy efficiency. By July, manufacturers must stop making bulbs that produce less than 45 lumens per watt, effectively phasing out not only incandescents but also halogen incandescents. 

These new rules accelerate a technological transition that has already been going on for more than a decade. New research by economist Sarah Armitage shines a bright light on the timing of this historic transition. In particular, Armitage asks whether early support for CFLs crowded out later adoption of LEDs. Her results are fascinating and imply that when it comes to environmental policies, it is important not only to choose the right tools, but also to get the timing right.

Impressive Innovation

When we think about energy transitions we typically think of the growth of renewables, electrification of buildings and vehicles, and phasing out fossil fuels. But one of the most important recent energy transitions is the humble light bulb.

After being dominated by traditional incandescents for 100+ years, in the last two decades the light bulb market has transitioned to halogen incandescents and CFLs, and then, finally, LEDs, with the price of LEDs falling 90% over the last decade. 

For me, it is helpful to think about these technologies in terms of lumens per watt. Like other measures of energy efficiency, this tells us how much is produced per unit of energy. The new U.S. energy efficiency rules require bulbs to produce 45 lumens per watt, and neither incandescents nor halogens can meet this standard.


Can we just take a minute and appreciate these energy-efficiency gains? Rarely does a new technology come along which does the same thing, but using a small fraction of the amount of energy. Going from an incandescent to an LED is like replacing a car that gets 25 miles-per-gallon with another one that gets 130 mpg! 


Ok but the road from incandescents to LEDs has not been smooth or easy. Armitage’s research is aimed at understanding the timing of this transition, and the role played by standards and subsidies.

One of the key features of the analysis is that it takes durability seriously. In addition to being more energy-efficient, these new bulbs also have much longer lifetimes. Traditional incandescents burn out quickly, in part because they produce so much heat, but the later technologies last much longer. 

This durability matters. Interestingly, Armitage documents a 40% decrease in U.S. light bulb shipments over the last decade. Are people sitting around in the dark? No! They are buying fewer light bulbs because their bulbs last longer.

Armitage estimates a dynamic model of residential lighting demand that takes this durability into account, and then uses the model to evaluate alternative timing for standards and subsidies.

Crowd Out

The figure below shows U.S. market shares for general purpose light bulbs. Back in 2010, the market was dominated by incandescents and, to a lesser extent, CFLs. In part, this early market share for CFLs reflects widespread subsidies which CFLs received throughout the first half of this period.


Between 2012 and 2014, U.S. federal standards were implemented which effectively banned incandescents. As you can see in the figure, the market quickly transitioned. By far the biggest winner initially was the halogen incandescent, a bulb created to be just efficient enough for the U.S. standards. CFLs also increased market share in 2014 and 2015.

Armitage uses her model to show that these CFLs and halogens crowded out later adoption of LEDs. Why? It goes back to durability. Once you install a light bulb, it tends to be there for a long time. Halogens (about 2,000 hours) and CFLs (about 8,000 hours, depending on usage), in particular, tend to last a lot longer than incandescents (about 1,200 hours). 

People tend not to replace light bulbs as long as they are still working. So these CFLs and halogens kept being used year after year. Even today, there are hundreds of millions of CFLs and halogens still working and still being used in U.S. sockets. Ask yourself, do you still have CFLs and halogens installed in your home? I know I do.


Armitage then uses her model to simulate policy counterfactuals. That is, she asks how the transition would have been different with alternative timing for subsidies and standards.

She finds that if the U.S. had delayed these federal standards until 2017 or 2018, we would have sold millions *more* LED bulbs. LEDs were cheap enough by that time that many households would have jumped all the way to LEDs rather than adopting one of the intermediate technologies. Ironically, these early energy efficiency standards actually slowed later adoption of an even more energy-efficient technology. 


On the other hand, the standards did accelerate the retirement of traditional incandescents. Armitage finds that, to minimize energy consumption (and thus, environmental externalities) over this time period, you want to implement standards as early as possible and phase out CFL subsidies after 2014. This combination of policies gets the worst offenders out of the market quickly, while minimizing crowd out from long-lasting CFLs.

The key takeaway from the research is that early policy action generates immediate environmental benefits from increased adoption of available efficient products, but may crowd out later adoption of even better products.

What’s Next?

In the long run, LEDs are probably going to be the overwhelmingly dominant light bulb technology worldwide. They are so efficient, so durable, so versatile, that it seems hard to think otherwise. But Armitage’s research is an important reminder that the journey is just as important as the destination. If it takes another decade for LEDs to completely take over, that means a whole lot of additional carbon emissions in the meantime.

In future work, it will be interesting to study this next wave of federal standards. How quickly will halogens be phased out? How long will it take for the existing stock of CFLs and halogens to burn out? I wonder if it also might be worth thinking about a “cash for clunkers” style policy where we pay people for their old bulbs. There is sunk cost fallacy and other behavioral constraints at play here that might be ameliorated with a policy nudge.

Finally, I wonder about the rebound effect. Nobel laureate Bill Nordhaus showed that, as lighting has become cheaper, humans have massively increased their consumption, using thousands of times more lumens today than they did in previous centuries. Have we finally reached the saturation point, or is there more scope for increased usage?


Keep up with Energy Institute blogs, research, and events on Twitter @energyathaas.

Suggested citation: Davis, Lucas. “Did CFL Subsidies Crowd Out LED Purchases?” Energy Institute Blog, UC Berkeley, May 31, 2022,

Lucas Davis View All

Lucas Davis is the Jeffrey A. Jacobs Distinguished Professor in Business and Technology at the Haas School of Business at the University of California, Berkeley. He is a Faculty Affiliate at the Energy Institute at Haas, a coeditor at the American Economic Journal: Economic Policy, and a Research Associate at the National Bureau of Economic Research. He received a BA from Amherst College and a PhD in Economics from the University of Wisconsin. His research focuses on energy and environmental markets, and in particular, on electricity and natural gas regulation, pricing in competitive and non-competitive markets, and the economic and business impacts of environmental policy.

12 thoughts on “Did CFL Subsidies Crowd Out LED Purchases? Leave a comment

  1. Current LED ‘bulbs’ need additional electronics to work in the 110v system. This electronics burns a lot of watts, and makes the LED chip VERY HOT. Thereby reducing life. LED bulbs were ‘claimed/ rumoured’ to have a life of 25 years. Likely will be closer to 5 years. WHEN residential wiring is changed to the ~45v level LEDs will be a lot happier. And last a lot longer.

  2. To believe that “The key takeaway from the research is that early policy action generates immediate environmental benefits from increased adoption of available efficient products, but may crowd out later adoption of even better products. ” is actionable guidance for policymakers or regulators requires that they have a level of political and technical omniscience that has barely ever existed – and certainly not in the realm of light bulbs. Who knew in the 2012-14 timeframe that LEDs would become as efficient/efficacious as they are now or that the price per lumen would be as low and going lower as it is now. Most LED’s in the early 2010’s consisted of the cylindrical/domed devices with most of the light directed out of the domed top – a far cry from today’s surface mount devices (SMD) that can produce lots of lumens either as a single device or as a strip of such devices. Sure, one always can lament that if ‘we only knew then what we know now…’ – but a not uncommon outcome of that approach is ‘paralysis by analysis…’ So in 2012-14, encouraging folks to move to CFLs seemed like a good idea – and was, in terms of energy saved.

    Of course, who knew that Trump (and as a consequence, Rick Perry as head of DoE) would succeed Obama (and Ernie Moniz as Energy secy). So enacting this policy in 2012-14 looks brilliant compared to waiting until 2017+ — don’t forget Trump tried to undo the demise of the incandescent bulb.

    Seems to me that important question to ask is not whether current policy processes can affect later policies in the face of many technical and political unknowns, but whether the energy and cost savings of switching to LEDs should induce the government or utilities to establish programs to encourage folks to move to LEDs. As someone else noted, the local utility (PG&E) certainly had programs to encourage folks to move to CFLs, but I’m not aware of any current programs to encourage the use of LED through rebates, etc.

    I have replaced all of the (mostly CFL) lights in my house and garage workshop with LED’s for three reasons – first, the lighting quality is much better (to my aging eyes) than with either fluorescent tubes/CFLs or incandescents; second, I will likely never have to replace any of them again (at least not because they failed); and third, I don’t have to worry about any potential breakage of a fluorescent. Contrary to a previous poster, it is not true that replacing fluorescent tubes with equivalent length LED tubes requires rewiring the lamp fixture. There are three types of LED replacement tubes: those that can be used as direct replacements (meaning they are compatible with most electronic ballasts); hybrid tubes, which can be used as replacements without rewiring or can be used in a direct wiring mode (meaning simply removing the electronic ballast and rewiring/replacing the ‘tombstones’ — these are apparently more efficient when used in the rewired mode); and third, tubes that are direct wired (remove ballast and change/rewire tombstones). I have used both the first and third types – the latter only when I found that the electronic ballast wasn’t really compatible with the direct replacement tube I wanted to use. Rewiring is well documented in the instructions and easy for anyone who has the skills to replace a light fixture or a wall receptacle.

    • “So in 2012-14, encouraging folks to move to CFLs seemed like a good idea – and was, in terms of energy saved.”

      Unfortunately, there can be TOO much encouragement, which was the case in California in that period. As one commentor mentioned, he stocked up on a lifetime supply of overly subsidized CFLs in that period. It’s the same foolish approach that led to the California utilities signing too many renewable contracts for too long of terms in that same period so that California ratepayers are basically unable to enjoy the technological revolution that their dollars fomented. Making a management mistake periodically is understandable, but repeated gross mismanagement is the key symptom of a failed system.

  3. I wonder if the bold first mover approach in the Netherlands to require heat pumps, but allow them to be installed in “hybrid systems” with gas fired backup is an example of that kind of a compromise that reduces and extends emissions. Netherlands to require heat pumps in all new construction and
    when replacing furnaces in existing buildings starting 2026:
    I think policy makers often look for compromises and small steady steps that may no longer make sense in a climate emergency, especially when installing long lived devices against a short lived climate.
    Perhaps the best achievable policy path is using one small stepping stone for a very short duration (if needed) before stepping to the ultimate wise requirement. Of course the perfection solution would’ve been mustering full courage to skip the stepping stone and leap to the ultimate solution with no intermediate stranded assets.

    • I tried to find a dual fueled hot water heater that had both a Natural gas burner with pilot light and a backup electrical element that could be programed to either take the lead and turn on first or for temperature top offs or set as a backup if natural gas recovery dropped the temperature to low and could boost the temperature. I could not find anyone who make a dual powered hot water heater. It could also be programed to use up the excess summer generation of the solar panels on the roof thus using less natural gas.

  4. In 1973, PG&E, in California, had a program to swap out all of the residential incandescent bulbs for CFL light bulbs. The paid their customers $15.00 for every CFL that was replaced in a home. CFLs at that time cost $15.00 to $24.00 each depending on the manufacturer and point of sale. 30 years later they were buying CFL lamps and selling them for 4 for a dollar at certain grocery stores and many people, like myself would “Stock Up” They have been lasting 10 to 12 years, at my home, and i still have over 120 CFL lamps left yet to be used or a lifetime supply. LED lamps have not had any of the utility rebates or “Special Price” programs and yet I have been able to find LED lamps of all sizes on clearance or on sale at major hardware stores and Dollar Tree stores for about a dollar each. CFLs give off the same wattage as comparable incandescent with a 74% savings in electricity. LED lamps do the same but at an 86% savings over incandescent but only a 45% savings over CFLs. Most people do not replace a working lamp until it burns out. Many people do not like the bright while or blueish white from CFL or early LED Lamps here in the USA. Asia, where most of the lamps today are made, likes the blueish coloration rather than the yellowish coloration given off by incandescent lamps. I always shopped for 2800K coloration that is like the yellow coloration of the Sun through our atmosphere and I could find 2800K in both CFL and LED lamps.

    CFL lamps contain Mercury that is poisonous and when put into landfills, can leach down into the aquifer over time. High intensity Discharge mercury vapor and metal halide lamps also contain mercury also and the newer high intensity LED lamps are replacing those also but the whole fixture, costing hundreds of dollars, must also be replaced. Many homes in the 1960s and 1970s came with 4-foot-long tube type fluorescent light fixtures and lamps that now have LED 4-foot lamps, but the ballast must be removed, and the fixture rewired. Most people will not replace those until there stock of 4-foot f40-T12 or f40-T8 fluorescent lamps is exhausted because the difference between $3.00 Fluorescent lamps and a $16.00 LED lamps and the re-wiring is quite beyond most homeowners’ electrical qualifications. The chart in this article shows the slow adoption of CFLs over incandescent over the past 50 years and it could take another 50 years to finally get the Mercury filled lamps replaced with LEDs lamps without some kind of government intervention and financial incentives given to homeowners and businesses like what was given to get people to switch to CFL lamps.

    Utilities now are less motivated to have programs that cut usage because they make money selling energy, nor conserving it. Before it was “not overloading the grid” and reducing need for more power plants. Today they are worried that they will not sell enough electricity to keep the power plants, they already have, making a profit. Now that homeowners are generating their own electricity with solar panels, they are even thinking about “Taxing the Sun” legislation to force homeowners to “Pay Up” for conserving power or generating their own power.

    • On top of these impediments, most localities don’t have an easy means for disposing of CFLs. Many require residents to drive to the landfill (an emission of GHG that may offset any of the lighting savings) to legally throw away CFLs and other fluorescent tubes/bulbs. The lack of foresight, despite warnings from many. More uncoordinated state and local agencies’ bumbling.

  5. The California utilities gamed the state’s energy efficiency incentives to gain hundreds of millions by selling CFL bulbs that were not installed (perhaps never due to the rise of LEDs). The utilities were almost penalized for this gaming but the then CPUC president Mike Peevey wrote an alternative decision that overruled the CPUC staff and administrative law judge. Yet another example of the capture of California regulators by the regulated.

  6. Thanks. This is a good piece on an often overlooked aspect of the global energy picture. LED technology has also fundamentally shifted the landscape for those looking at energy access in the developing world. By dramatically lowering the amount of electricity needed for lighting and television, it is now typically cheaper to address those high priority needs with a small (<50 Wp) solar home system than to connect to an adjacent power pole. This has broad implications for the viability of grid extensions or mini-grids for many areas of sub-Sharan Africa.

  7. Fifty lumens per watt is an acceptable target, but at what count of lumens? Light from an LED element is directional. A luminaire as a plate, aimed in the direction of a task, is about twice as effective as a point source. A 500 lumens LED disk light matches the task illumination of a 100 watt incandescent bulb rated 1600 lumens. LED light tricked to look like a point source bulb is stupid regarding efficacy, and is wastefully short-lived. The L-Prize was absurd.

    • I wish hindsight was 20/20. Unfortunately, California drew several incorrect lessons from the 2000/01 energy crisis and much of the current energy system mismanagement comes from those distorted lessons.

%d bloggers like this: