Hi @saulius. I am tagging you because I updated this comment. @Michael St Jules đ¸, you may also be interested in the update.
Stop the farms
When local residents successfully protest against a planned chicken farm, production will likely increase elsewhere to meet demandâbut how quickly? I found no clear methodology to estimate this and received no definitive answers when I asked on an economics forum. As far as I know, it could take anywhere from a week to 20 years, and the choice massively impacts cost-effectiveness.
Have you looked into the price elasticity of the supply of and demand for poultry meat in Poland? I did not find âelasticitâ in the post or sheet. For E_D = âprice elasticity of demandâ (negative because demand decreases with price), and E_S = âprice elasticity of supplyâ (positive because supply increases with price), the reduction in supply/âdemand as a fraction of a leftwards shift in the demand curve is âcumulative elasticity factor (CEF)â = E_S/â(E_SâE_D). Swapping _D and _S, the reduction in supply/âdemand as a fraction of a leftwards shift in the supply curve is E_D/â(E_DâE_S) = -E_D/â(E_SâE_D) = 1 - CEF[1].
CEF is 76 % for chicken meat according to Figure 8.2 of Norwood and Lusk (2011). I do not know the time horizon to which this refers to, and would not be surprised if it generalised badly to stopping farms in Poland. However, if it did, shifting the supply curve of chicken leftwards by 1 kg would decrease its demand/âsupply by 0.24 kg (= 1*(1 â 0.76)). Did you account for this when coming up with your guesses for the years of impact? I would explicitly include the cumulative elasticity factor in the estimation of the reduction in demand/âsupply, and then guess the leftwards shift in the supply curve caused by blocking and delaying the construction of farms.
It makes sense that âreduction in supply/âdemand for a leftwards shift in the supply curve as a fraction of the shiftâ = 1 - CEF = 1 - âreduction in supply/âdemand for a leftwards shift in the demand curve as a fraction of the shiftâ. For small shifts in the demand and supply curves, the change in supply/âdemand is CEF*âshift in demandâ + k*âshift in supplyâ. If the shift in demand matches the shift in supply, the change in supply/âdemand is equal to the shift. In this case, âshiftâ = CEF*âshiftâ + k*âshiftâ, which means k = 1 - CEF.
CEF is 76 % for chicken meat according to Figure 8.2 of Norwood and Lusk (2011). I do not know the time horizon to which this refers to, and would not be surprised if it generalised badly to stopping farms in Poland. However, if it did, shifting the supply curve of chicken leftwards by 1 kg would decrease its demand/âsupply by 0.24 kg (= 1*(1 â 0.76)). Did you account for this when coming up with your guesses for the years of impact?
To clarify my question above, are these estimates supposed to account for adjustments broiler producers can make to expand output along the existing supply curve (with the same farms and technology):
Higher stocking density â place more chicks per m² of barn space (within legal or contractual limits).
Increase batch frequency â shorten downtime between flocks (reduce cleaning/âresting days).
Longer grow-out period â delay slaughter so birds reach heavier weights.
Optimized feed formulation â shift to higher-energy or higher-protein rations if the higher output value outweighs feed costs.
Reduced mortality through management tweaks â e.g. stricter biosecurity, better litter quality, ventilation adjustments.
Use of feed additives or growth promoters (where legal) â probiotics, enzymes, coccidiostats to improve feed conversion.
Labor allocation â more intensive monitoring of flocks to reduce disease losses and increase uniformity.
Energy use adjustments â e.g. heating/âcooling more aggressively to maintain optimal bird growth conditions.
Extending usable facilities â keep older barns in service longer or run temporary housing (e.g. tents, converted sheds) to squeeze a bit more capacity.
Tightening contract terms with integrators â e.g. accepting higher chick placements per house if integrators provide them.
My understanding is that your estimates are just supposed to account for the decrease in the number of broiler farms, not the above. If so, I think the impact of Stop the Farms is 24 % (= 1 â 0.76) as large as you estimated based on the cumulative elasticity factor for chicken meat of 76 % presented in Figure 8.2 of Norwood and Lusk (2011).
No, I did not think about the effects you listed when choosing these numbers, at least not explicitly. I donât remember what exactly went through my head when I imputed these numbers. I think I was just trying to imagine what a chicken production graph would look like with or without campaign. Naively thinking, blocking a farm would postpone the production by at least 1-2 years, because thatâs how long it probably takes to get planning permits and build a farm. But 1-2 years felt like too optimistic though, so I was conservative, but probably not conservative enough.
Either way, those figures of years of impact are guesses in the spirit of If Itâs Worth Doing, Itâs Worth Doing With Made-Up Statistics, not estimates. I was supposed to finish the project and had no idea how to estimate these things, so I entered somewhat random numbers. Please donât take them seriously. I think you would be much better off ignoring them, and coming up with a new estimate from scratch. Clearly you are thinking about this much more deeply than I was.
Btw, another effect to consider is anticipation. Investors in Poland already know that new farms face a high risk of being blocked or delayed by protests. Given this, they may (a) decide not to build farms at all (but someone else might build them instead), (b) shift their plans to other countries where protests are less likely, or (c) submit more applications than they really need, expecting some to be blocked. Since the campaign has been active for years, itâs possible the market has already adapted to the reality that building new farms in Poland is unusually difficult, and has found alternative ways to meet the demand.
It might not be a genuine supply shift, if the same business/âcompany would just want to set up elsewhere. You could model the probability that they wouldnât, so there would be a supply shift, and then use elasticities for that.
Thanks, Michael. I meant â[expected] leftwards shift in the supply curveâ = âprobability of a farm being built by the target company elsewhere (p)â*âleftwards shift if a farm is built by the target company elsewhere (R1)â + (1 - p)*âleftwards shift if no farm is built by the target company elsewhere (R2)â. You are saying that R1 is 0, but I do not think this has to be the case. The farm would tend to start operations later if it had to be built elsewhere, and therefore there would be a temporary shift in the supply curve.
Thanks Vasco. I still donât see how such a multiplier would solve the core issue for me. Say a chicken costs $1 and 1,000 of them are produced and sold. Market is at equilibrium. We close down a farm, and now 900 are produced and sold, and the price goes up. The real question for me is how quickly the market adjusts backâhow soon someone else builds another farm to fill that gap. I have no idea. Iâve never seen an economic metric that directly measures that speed of recovery. If an economist were to estimate it, I imagine elasticities would be part of the picture, but I donât know how that estimate would actually look. So at this point, Iâd rather not introduce a multiplier that might confuse me and readers without solving the problem Iâm trying to get at.
I saw cumulative elasticity factor being used for impact estimations of veganism. I get it: if some people stop buying chicken, and now only 900 are sold at $1, the price might drop to $0.9. Then the question of how much production contracts is where elasticities come in. But with production shifts it feels different, messier. If we assumed that vegans would start eating chicken again once the price falls, then even the veganism case would be just as confusing.
Thanks, @saulius. I am tagging you because I updated the comment.
Below is a sketch of how I am imagining the change in demand/âsupply would evolve over time. It would initially be equal to the shift in supply Delta_Q_S (negative of the annual production of the affected farm), then increase to Delta_Q at equilibrium after Delta_t_eq (where Delta_Q = (1 - CEF)*Delta_Q_S), then remain at Delta_Q until Delta_t_S after the initial shift (where Delta_t_S is the duration of the shift in the supply curve), then increase by -Delta_Q_S to Delta_QâDelta_Q_S (as a result of a new farm starting operations), and then decrease to 0 over Delta_t_eq. The total change in demand/âsupply is the integral of the line below over time. The shaded areas are equal, so the total change in demand/âsupply is Delta_Q*Delta_t_S = (1 - CEF)*Delta_Q_S*Delta_t_S. This directly depends on CEF = E_S/â(E_SâE_D), so I would say the elasticities very much matter.
I would estimate Delta_t_S from âprobability of i) farm being built in the original placeâ*âdelay of the start of the farmâs operations given i)â + âprobability of ii) farm being built elsewhereâ*âdelay of the start of the farmâs operations given given ii)â + âprobability of iii) farm not being builtâ*âlifetime of the farmâ. It looks like your estimates for the years of impact refer to Delta_t_S. If so, for CEF equal to 76 % (as in Figure 8.2 of Norwood and Lusk (2011)), I believe the impact is 24 % (= 1 - CEF) as large as you estimated.
For CEF equal to 76 %, and Delta_t_S equal to 0.695 years (= (20.9*1.0 + 67.5*0.6)/â(20.9 + 67.5)), which is your estimate for the years of impact weighted by the number of broilers affected, the reduction in the demand/âsupply of chicken meat is 16.7 % (= (1 â 0.76)*0.695) of the annual production of the targeted farms.
I am taking a break from research and I wonât try to understand what you wrote here because itâs currently over my head and thatâs not the type of thing I want to focus on right now in life. You can talk with Jakub Stencel if you are interested on improving my estimates, he would tell you whether itâs worth it (he is on the forum you can just tag him). But you might also need to talk to someone who say works on the stop the farms campaign for some context. They say that there will be protests against new farms no matter where they are built within Poland because of the network Anima created, but in most other countries itâs not happening. To me the bigger question is if the campaign is even net-positive, because it might just shift production to countries where itâs more difficult to improve conditions for farmed animals.
Some random thoughts about your message: * I did look into elasticities during my project and other elasticities of chicken can be found in this very old ACE spreadsheet and in this paper which analyzes elasticity some years ago in Turkey and says âAccording to the supply and demand functions for chicken meat, supply elasticity is 0.377 and demand elasticity is 0.030â. I remember comparing the two and getting wildly different numbers. * I will say that tI still donât understand what you mean that âthe reduction in the demand/âsupply of chicken meat is 16.7 % (= (1 â 0.76)*0.695) of the annual production of the targeted farms.â I mean, if we say closed a farm of a million broilers, in the very short term at least, surely the reduction of the number of broilers farmed in the world is one million. Itâs not like those other chickens instantly appear somewhere else. So to me the question is still how quickly market goes back to equilibrium. Your variable Delta_t_S and my guess of 0.695 years of impact seem like two different things by the way. Maybe the reduction in production for a closed farm is 16.7% Ă âthe lifetime of a farmâ? Plus a bit more because it would take the market some time to adjust?
Thanks, Saulius. I was looking into this because I started working with Animal International 2 weeks ago, but I like public conversations (as long as confidential information is not shared).
this paper which analyzes elasticity some years ago in Turkey and says âAccording to the supply and demand functions for chicken meat, supply elasticity is 0.377 and demand elasticity is 0.030â. I remember comparing the two and getting wildly different numbers.
These elasticities imply a cumulative elasticity factor of CEF = 92.6 % (= 0.377/â(0.377 + 0.030)). In this case, 1 - CEF = 7.4 %, which is 30.8 % (= 0.074/â0.24) of the value implied by Figure 8.2 of Norwood and Lusk (2011).
I will say that tI still donât understand what you mean that âthe reduction in the demand/âsupply of chicken meat is 16.7 % (= (1 â 0.76)*0.695) of the annual production of the targeted farms.â
If the operations of a farm are delayed by Delta_t_S years, I would estimate the integral of the reduction in demand/âsupply over all time to be Delta_Q*Delta_t_S = (1 - CEF)*Delta_Q_S*Delta_t_S = (1 â 0.76)*âannual production of the delayed farmâ*Delta_t_S.
I mean, if we say closed a farm of a million broilers, in the very short term at least, surely the reduction of the number of broilers farmed in the world is one million. Itâs not like those other chickens instantly appear somewhere else.
I agree. This is represented in my drawing by the change in demand/âsupply starting at â1.
So to me the question is still how quickly market goes back to equilibrium. Your variable Delta_t_S and my guess of 0.695 years of impact seem like two different things by the way. Maybe the reduction in production for a closed farm is 16.7% Ă âthe lifetime of a farmâ? Plus a bit more because it would take the market some time to adjust?
Here is a description of my graph which may help:
At t = 0, there is a leftwards shift in the supply curve (less supply for the same price) because the operations of a new farm were supposed to start then, but they are delayed due to Stop the Farms.
From t = 0 to t = Delta_t_eq, the market adjusts, with other farms (not necessarily in Poland) producing more, and stocks (not necessarily in Poland) being spent to offset the reduced supply.
At t = Delta_t_eq, the market reaches the new equilibrium.
From t = Delta_t_eq to t = Delta_t_S, the market stays at the new equilibrium.
At t = Delta_t_S, there is a rightwards shift in the supply curve (greater supply for the same price) because the operations of a new farm start then (not necessarily in Poland).
From t = Delta_t_S to t = Delta_t_S + Delta_t_eq, the market adjusts, with other farms (not necessarily in Poland) producing less, and stocks (not necessarily in Poland) being saved to offset the increased supply.
My understanding isthat your estimates for the years of impact refer to Delta_t_S, which is the expected duration of the delay of the start of the operations of a new farm. So it seems to me that the impact should be 24 % (= 1 - CEF) of what you estimated for CEF = 76 %.
Hi @saulius. I am tagging you because I updated this comment. @Michael St Jules đ¸, you may also be interested in the update.
Have you looked into the price elasticity of the supply of and demand for poultry meat in Poland? I did not find âelasticitâ in the post or sheet. For E_D = âprice elasticity of demandâ (negative because demand decreases with price), and E_S = âprice elasticity of supplyâ (positive because supply increases with price), the reduction in supply/âdemand as a fraction of a leftwards shift in the demand curve is âcumulative elasticity factor (CEF)â = E_S/â(E_SâE_D). Swapping _D and _S, the reduction in supply/âdemand as a fraction of a leftwards shift in the supply curve is E_D/â(E_DâE_S) = -E_D/â(E_SâE_D) = 1 - CEF[1].
CEF is 76 % for chicken meat according to Figure 8.2 of Norwood and Lusk (2011). I do not know the time horizon to which this refers to, and would not be surprised if it generalised badly to stopping farms in Poland. However, if it did, shifting the supply curve of chicken leftwards by 1 kg would decrease its demand/âsupply by 0.24 kg (= 1*(1 â 0.76)). Did you account for this when coming up with your guesses for the years of impact? I would explicitly include the cumulative elasticity factor in the estimation of the reduction in demand/âsupply, and then guess the leftwards shift in the supply curve caused by blocking and delaying the construction of farms.
It makes sense that âreduction in supply/âdemand for a leftwards shift in the supply curve as a fraction of the shiftâ = 1 - CEF = 1 - âreduction in supply/âdemand for a leftwards shift in the demand curve as a fraction of the shiftâ. For small shifts in the demand and supply curves, the change in supply/âdemand is CEF*âshift in demandâ + k*âshift in supplyâ. If the shift in demand matches the shift in supply, the change in supply/âdemand is equal to the shift. In this case, âshiftâ = CEF*âshiftâ + k*âshiftâ, which means k = 1 - CEF.
To clarify my question above, are these estimates supposed to account for adjustments broiler producers can make to expand output along the existing supply curve (with the same farms and technology):
Higher stocking density â place more chicks per m² of barn space (within legal or contractual limits).
Increase batch frequency â shorten downtime between flocks (reduce cleaning/âresting days).
Longer grow-out period â delay slaughter so birds reach heavier weights.
Optimized feed formulation â shift to higher-energy or higher-protein rations if the higher output value outweighs feed costs.
Reduced mortality through management tweaks â e.g. stricter biosecurity, better litter quality, ventilation adjustments.
Use of feed additives or growth promoters (where legal) â probiotics, enzymes, coccidiostats to improve feed conversion.
Labor allocation â more intensive monitoring of flocks to reduce disease losses and increase uniformity.
Energy use adjustments â e.g. heating/âcooling more aggressively to maintain optimal bird growth conditions.
Extending usable facilities â keep older barns in service longer or run temporary housing (e.g. tents, converted sheds) to squeeze a bit more capacity.
Tightening contract terms with integrators â e.g. accepting higher chick placements per house if integrators provide them.
My understanding is that your estimates are just supposed to account for the decrease in the number of broiler farms, not the above. If so, I think the impact of Stop the Farms is 24 % (= 1 â 0.76) as large as you estimated based on the cumulative elasticity factor for chicken meat of 76 % presented in Figure 8.2 of Norwood and Lusk (2011).
No, I did not think about the effects you listed when choosing these numbers, at least not explicitly. I donât remember what exactly went through my head when I imputed these numbers. I think I was just trying to imagine what a chicken production graph would look like with or without campaign. Naively thinking, blocking a farm would postpone the production by at least 1-2 years, because thatâs how long it probably takes to get planning permits and build a farm. But 1-2 years felt like too optimistic though, so I was conservative, but probably not conservative enough.
Either way, those figures of years of impact are guesses in the spirit of If Itâs Worth Doing, Itâs Worth Doing With Made-Up Statistics, not estimates. I was supposed to finish the project and had no idea how to estimate these things, so I entered somewhat random numbers. Please donât take them seriously. I think you would be much better off ignoring them, and coming up with a new estimate from scratch. Clearly you are thinking about this much more deeply than I was.
Thanks for the context, Saulius!
Btw, another effect to consider is anticipation. Investors in Poland already know that new farms face a high risk of being blocked or delayed by protests. Given this, they may (a) decide not to build farms at all (but someone else might build them instead), (b) shift their plans to other countries where protests are less likely, or (c) submit more applications than they really need, expecting some to be blocked. Since the campaign has been active for years, itâs possible the market has already adapted to the reality that building new farms in Poland is unusually difficult, and has found alternative ways to meet the demand.
Thanks, Saulius! That makes sense.
It might not be a genuine supply shift, if the same business/âcompany would just want to set up elsewhere. You could model the probability that they wouldnât, so there would be a supply shift, and then use elasticities for that.
Thanks, Michael. I meant â[expected] leftwards shift in the supply curveâ = âprobability of a farm being built by the target company elsewhere (p)â*âleftwards shift if a farm is built by the target company elsewhere (R1)â + (1 - p)*âleftwards shift if no farm is built by the target company elsewhere (R2)â. You are saying that R1 is 0, but I do not think this has to be the case. The farm would tend to start operations later if it had to be built elsewhere, and therefore there would be a temporary shift in the supply curve.
Yes, good point. Itâs worth checking if the delay could have a significant impact.
Thanks Vasco. I still donât see how such a multiplier would solve the core issue for me. Say a chicken costs $1 and 1,000 of them are produced and sold. Market is at equilibrium. We close down a farm, and now 900 are produced and sold, and the price goes up. The real question for me is how quickly the market adjusts backâhow soon someone else builds another farm to fill that gap. I have no idea. Iâve never seen an economic metric that directly measures that speed of recovery. If an economist were to estimate it, I imagine elasticities would be part of the picture, but I donât know how that estimate would actually look. So at this point, Iâd rather not introduce a multiplier that might confuse me and readers without solving the problem Iâm trying to get at.
I saw cumulative elasticity factor being used for impact estimations of veganism. I get it: if some people stop buying chicken, and now only 900 are sold at $1, the price might drop to $0.9. Then the question of how much production contracts is where elasticities come in. But with production shifts it feels different, messier. If we assumed that vegans would start eating chicken again once the price falls, then even the veganism case would be just as confusing.
Thanks, @saulius. I am tagging you because I updated the comment.
Below is a sketch of how I am imagining the change in demand/âsupply would evolve over time. It would initially be equal to the shift in supply Delta_Q_S (negative of the annual production of the affected farm), then increase to Delta_Q at equilibrium after Delta_t_eq (where Delta_Q = (1 - CEF)*Delta_Q_S), then remain at Delta_Q until Delta_t_S after the initial shift (where Delta_t_S is the duration of the shift in the supply curve), then increase by -Delta_Q_S to Delta_QâDelta_Q_S (as a result of a new farm starting operations), and then decrease to 0 over Delta_t_eq. The total change in demand/âsupply is the integral of the line below over time. The shaded areas are equal, so the total change in demand/âsupply is Delta_Q*Delta_t_S = (1 - CEF)*Delta_Q_S*Delta_t_S. This directly depends on CEF = E_S/â(E_SâE_D), so I would say the elasticities very much matter.
I would estimate Delta_t_S from âprobability of i) farm being built in the original placeâ*âdelay of the start of the farmâs operations given i)â + âprobability of ii) farm being built elsewhereâ*âdelay of the start of the farmâs operations given given ii)â + âprobability of iii) farm not being builtâ*âlifetime of the farmâ. It looks like your estimates for the years of impact refer to Delta_t_S. If so, for CEF equal to 76 % (as in Figure 8.2 of Norwood and Lusk (2011)), I believe the impact is 24 % (= 1 - CEF) as large as you estimated.
For CEF equal to 76 %, and Delta_t_S equal to 0.695 years (= (20.9*1.0 + 67.5*0.6)/â(20.9 + 67.5)), which is your estimate for the years of impact weighted by the number of broilers affected, the reduction in the demand/âsupply of chicken meat is 16.7 % (= (1 â 0.76)*0.695) of the annual production of the targeted farms.
I am taking a break from research and I wonât try to understand what you wrote here because itâs currently over my head and thatâs not the type of thing I want to focus on right now in life. You can talk with Jakub Stencel if you are interested on improving my estimates, he would tell you whether itâs worth it (he is on the forum you can just tag him). But you might also need to talk to someone who say works on the stop the farms campaign for some context. They say that there will be protests against new farms no matter where they are built within Poland because of the network Anima created, but in most other countries itâs not happening. To me the bigger question is if the campaign is even net-positive, because it might just shift production to countries where itâs more difficult to improve conditions for farmed animals.
Some random thoughts about your message:
* I did look into elasticities during my project and other elasticities of chicken can be found in this very old ACE spreadsheet and in this paper which analyzes elasticity some years ago in Turkey and says âAccording to the supply and demand functions for chicken meat, supply elasticity is 0.377 and demand elasticity is 0.030â. I remember comparing the two and getting wildly different numbers.
* I will say that tI still donât understand what you mean that âthe reduction in the demand/âsupply of chicken meat is 16.7 % (= (1 â 0.76)*0.695) of the annual production of the targeted farms.â I mean, if we say closed a farm of a million broilers, in the very short term at least, surely the reduction of the number of broilers farmed in the world is one million. Itâs not like those other chickens instantly appear somewhere else. So to me the question is still how quickly market goes back to equilibrium. Your variable Delta_t_S and my guess of 0.695 years of impact seem like two different things by the way. Maybe the reduction in production for a closed farm is 16.7% Ă âthe lifetime of a farmâ? Plus a bit more because it would take the market some time to adjust?
Thanks, Saulius. I was looking into this because I started working with Animal International 2 weeks ago, but I like public conversations (as long as confidential information is not shared).
These elasticities imply a cumulative elasticity factor of CEF = 92.6 % (= 0.377/â(0.377 + 0.030)). In this case, 1 - CEF = 7.4 %, which is 30.8 % (= 0.074/â0.24) of the value implied by Figure 8.2 of Norwood and Lusk (2011).
If the operations of a farm are delayed by Delta_t_S years, I would estimate the integral of the reduction in demand/âsupply over all time to be Delta_Q*Delta_t_S = (1 - CEF)*Delta_Q_S*Delta_t_S = (1 â 0.76)*âannual production of the delayed farmâ*Delta_t_S.
I agree. This is represented in my drawing by the change in demand/âsupply starting at â1.
Here is a description of my graph which may help:
At t = 0, there is a leftwards shift in the supply curve (less supply for the same price) because the operations of a new farm were supposed to start then, but they are delayed due to Stop the Farms.
From t = 0 to t = Delta_t_eq, the market adjusts, with other farms (not necessarily in Poland) producing more, and stocks (not necessarily in Poland) being spent to offset the reduced supply.
At t = Delta_t_eq, the market reaches the new equilibrium.
From t = Delta_t_eq to t = Delta_t_S, the market stays at the new equilibrium.
At t = Delta_t_S, there is a rightwards shift in the supply curve (greater supply for the same price) because the operations of a new farm start then (not necessarily in Poland).
From t = Delta_t_S to t = Delta_t_S + Delta_t_eq, the market adjusts, with other farms (not necessarily in Poland) producing less, and stocks (not necessarily in Poland) being saved to offset the increased supply.
My understanding is that your estimates for the years of impact refer to Delta_t_S, which is the expected duration of the delay of the start of the operations of a new farm. So it seems to me that the impact should be 24 % (= 1 - CEF) of what you estimated for CEF = 76 %.
Best wishes for your new projects!