Gas Prices are High, But Don’t Blame the Usual Suspects: Implications for Policy

A.  Introduction

Gasoline prices in the US (and indeed elsewhere) are certainly high.  Given that in the US much of the voting population views cheap gas as much of a right as life, liberty, and the pursuit of happiness, this has political implications.  It is thus not surprising that politicians, including those in the Biden administration, are considering a range of policy measures with the hope they will bring these gas prices down.  And while fuel prices have indeed come down some in the last few weeks from their recent peak, they remain high, and their path going forward remains uncertain.

One of the most common such measures, already implemented in six states (as of July 6) and under consideration in many more, has been to reduce or end completely for some period state taxes on fuels.  And President Biden on June 22 called on Congress to approve a three-month suspension of federal gas and diesel taxes.  The political attraction of such proposals is certainly understandable.  A Morning Consult / Politico public opinion poll in March found that 72% of those surveyed would favor “a temporary break from paying state taxes on gasoline”, and 73% would favor a similar “temporary break from paying federal taxes on gasoline”.  It is hard to find anything these days that close to three-quarters of the population agree on.

But would this in fact help to reduce what people are paying at the pump?  The answer is no.  One has to look at what led to the recent run-up in gas and other fuel prices, and only with a proper understanding of that can the appropriate policy response be worked out.  Cutting taxes on fuels should not be expected to lead to a reduction in what people pay at the pump for their gas.  Indeed, what could lower these prices would be to raise fuel taxes, and then use the funds generated to cover measures that would, in the near term, reduce the demand for these fuels.

This post will first examine the recent run-up in fuel prices, putting it in the context of how that market has functioned over the last decade and what is different now.  Based on this, it will then look at what the impact would be of measures such as cutting fuel taxes, releasing crude oil from the nation’s Strategic Petroleum Reserve, encouraging more drilling for oil, and similar.  None of these should be expected, under current conditions, to lead to lower prices at the pump.

Rather, one could raise fuel taxes and use these funds to support measures that would reduce the nation’s usage of gas.  For example, an immediate action that would be effective as well as easy to implement would be to encourage ridership on our public transit systems by simply ending the charging of fares on those systems.  One could stop charging those fares tomorrow – nothing special is needed.  Some share of those driving their cars for commuting or for other trips would then switch to transit, which would lead to a reduction in fuel demand and from this a reduction in fuel prices.  The lower price will benefit all those who buy gas, including those in rural areas who have no transit options.  And as will be discussed, the cost to cover what is being collected in fares would be really quite low.

A note on usage:  All references to “gas” in this post are to gasoline.  They are not to natural gas (methane) nor indeed any other gas.  Fuels will refer to gasoline and diesel together, where statements made with a specific reference to gas will normally apply similarly to diesel.

B.  The Rise in Fuel Prices and the Factors Behind It

Fuel prices have certainly gone up in the first half of 2022.  As shown in the chart at the top of this post, despite the fall in recent weeks fuel prices (the line in red) are still 75% above where they were in early-December (in June they were more than double), with those December 2021 prices double what they had been in October / November 2020.  Crude oil prices (the line in black) have also been going up, and have been since late 2020 (following the dip earlier in 2020 due to the Covid lockdowns).  This rise in the price of crude oil can explain the rise in the retail prices for fuels up through early this year.  But as we will discuss, the factors behind the more recent rise in fuel prices changed in late February 2022 – coinciding with Russia’s invasion of Ukraine.

First, some notes on the data.  The figures all come from the Energy Information Administration (EIA), part of the US Department of Energy, and weekly averages are used.  For reasons to be discussed below, the price of “fuel” is a 2:1 weighted average of the prices of regular unleaded gasoline (unleaded) and diesel (ultra low-sulfur no. 2), both wholesale FOB spot prices and for delivery at the US Gulf Coast.  While it is an average, this does not really matter much in practice as the wholesale prices of gas and diesel have not, at any point in time, differed by all that much from each other.  They move together.  Nor have their average prices over time differed by all that much.  For the period since the start of 2014, the average wholesale cost of gas was $1.81 per gallon while that for diesel was $1.90 – a difference of just 9 cents.  While there can be larger differences at various points in time, for the purposes here the distinction between the two fuels is not central.

The cost of crude oil (the line in black) is for West Texas Intermediate (FOB spot price, for delivery at Cushing, Oklahoma), the benchmark crude most commonly used in the US and also the basis for the main financial contracts used to hedge the price of oil in the US.  It is presented here on a per-gallon basis to make it comparable to the other prices, where one barrel of oil is equivalent to 42 gallons.

A refinery will purchase crude oil and then through various processes refine that oil into gasoline, diesel, and other petroleum products that can then be used as fuels by our cars and trucks as well for other purposes.  The difference in price between what the refinery can sell these finished products for and the cost of the crude it buys as the primary input is called the “crack spread”.  While the crack spread will be unique for each refinery, as it will depend on the technology it has (how modern and efficient it is), what types of crude it has been designed to process most efficiently (as different crudes have different characteristics, such as viscosity and sulfur content), the mix of specific products it produces (the share ending as gas or diesel, but also jet fuel, heating oil, etc.), and the location of the refinery (as the crude oil must be delivered to it, and it then must arrange for the delivery of its products to the ultimate purchasers), a simplified standard spread is often calculated to provide an indication of how market prices are moving.  The most common such standard spread is called the “3-2-1 crack spread”.

The 3-2-1 crack spread is calculated for a refinery that would process 3 barrels of crude oil into 2 barrels of gasoline and 1 barrel of diesel.  For the calculations here, all were expressed on a per-gallon basis, and the specific fuels and delivery locations are as specified above.  The 3-2-1 crack spread is then simply calculated as the value of two gallons of gasoline plus one gallon of diesel, minus the cost of three gallons of crude oil, with that total then divided by three as three gallons of fuel are being produced.  It is a gross spread, as a refinery will of course have other operational costs (including the cost of labor), plus the refinery will need to generate a return on the capital invested for it to be viable in the long term.  But this simple gross spread is often used as an indicator of what is happening in the market.

That calculated 3-2-1 crack spread is presented as the blue line in the chart at the top of this post.  From 2014 through 2021, it rarely moved above $0.50 per gallon, and it averaged just $0.36 per gallon over the period.  In 2021 it was not much higher, averaging $0.42 per gallon over the year.  But from late February 2022, coinciding with the Russian invasion of Ukraine, it has shot upward.  As of the week ending June 24 it had reached $1.46 per gallon, but as of the week ending July 8 it had come down to $1.02.  That is still high – it is still close to three times what it had averaged before.

To understand the factors that led to this jump in the crack spread this year, one should first consider how prices are determined in these markets.  The key is that the crack spread is not itself an independently determined price, but rather a spread between the price of the final product (gasoline and diesel fuels) and the price of crude oil, both of which are determined independently.

Start with the final products – gasoline and diesel:  These are sold in highly competitive markets of numerous gas stations pricing their product to sell at the best prices they can get, but where for the nation as a whole, stocks of the fuels are kept within a narrow range.  One can calculate (again from EIA data), that in recent years (2017 through 2022H1), the nation’s stocks of motor gasoline have averaged 236 million barrels, with no clear upward or downward trend.  While the stocks will vary over the course of the year due to seasonality, at comparable weeks in the year they have been kept in a relatively narrow range, with a standard deviation of just 2.1% of the weekly averages over this period.  This means (assuming a normal distribution, which is reasonable) that in about two-thirds of the weekly cases, the stocks will be within +/- 2.1% of the average for those weeks (one standard deviation), and in 95% of the cases will be within +/-4.2% of the averages (two standard deviations).  That is, the stocks are managed to stay within a relatively narrow range, although at a target level that depends on the season of the year.

In such a market, if producers (either directly or through the gas stations they contract with) price their gasoline at too low a price for the conditions of the time, they will find that their stocks will be running down – soon to unsustainable levels.  They would need to ration what they sell, either by long lines at the pumps or by some direct rationing system.  And if they price their gasoline at too high a price, they will find their stocks accumulating to levels that exceed what they can store.  They sell their gas for the highest price they can get, but that price will be constrained to be such that they will be able to manage their inventories of refined gasoline (and similarly for diesel fuels) to within a certain range.  And as noted above, that range is a narrow one of normally just +/- 2% or so.

Crude oil prices are determined differently.  Here there is a world market, where OPEC producers (as well as a few producers who cooperate with OPEC, where the most prominent is Russia) set production ceilings by OPEC member (and cooperative partner) with the aim of achieving some price target.  They do not always succeed in achieving that target, as global conditions can change suddenly.  Recent examples include conditions triggered by the Covid crisis in 2020, or by the global financial crisis that began in the US in 2008.  OPEC also responds sluggishly to changes in the markets, particularly when crude oil prices are rising – which many OPEC members are rather pleased with – as the production quotas must be negotiated among the members.  But it is correct to say that the market for crude oil is a managed one, although often not a terribly well managed one due to the inherent difficulty in forecasting global demands and then responding on a timely basis to unexpected changes.

With the retail price of the fuels determined on the one side by conditions in the competitive markets for fuels, and the price of crude oil determined on the other side by the actions of OPEC and those who cooperate with it, the crack spread will be a margin that has now been determined.  That is, it is not a price that the refiners themselves will normally be able to set.  There is a lower limit, as a gross crack spread that is too low to cover their other operating costs (and is expected to stay that low for some time), will lead refiners to shut down their operations.  But based on what we observe for the period from 2014 in the chart at the top of this post, it appears that a crack spread of $0.36 per gallon (the average from 2014 through 2021) is sufficient to cover such costs as well as provide a return on the capital invested, as refineries stayed open and continued to produce over this period with such a spread.

This spread then jumped in late February of this year – coinciding with the Russian invasion of Ukraine – to a level that has been between three and four times what it was before.  What happened?  While the Russian invasion was clearly significant, one should look at this in the context of where the market was just prior to the invasion.  It was tight, and the Russian invasion should be seen as a tipping point where refinery supplies of these fuels could no longer meet the demand.

First of all, demand has been growing, both in the US and in the rest of the world, as economies have recovered from the lockdowns that were necessary at the start of the Covid crisis.  The US enjoyed a particularly strong recovery in 2021, with real GDP growing by 5.7% – the fastest such growth in any calendar year in the US in close to 40 years.  And the personal consumption component of GDP rose by 7.9% in 2021 – the fastest such growth in any year since 1946!  But it should be recognized that this was coming after the sharp falls in 2020 due to Covid (of 3.4% for GDP and 3.8% for personal consumption).  The rest of the world recovered similarly in 2021, although at various different rates.

This raised the demand for gas, diesel, and other fuels.  Petroleum refineries could keep up in 2021, as this followed the lower demands they had for their products in 2020.  But the lower demands (and hence lower refinery throughputs) in 2020 due to Covid did have an effect.  It led to decisions to close some of that refinery capacity, leading to a reduction in capacity in 2021 for the first time in decades.  Albeit small, worldwide, refinery capacity fell from 102.3 million barrels per day in 2020 to 101.9 million barrels in 2021 (a fall of 0.4%).  Refinery capacity in the US fell similarly, from 18.1 million barrels per day in 2020 to 17.9 million barrels in 2021 (a fall of 1.1%).  With the recovery in demand for fuel products in 2021, this placed producers at closer to their limits.

But the limit to how much petroleum refineries can produce is pretty rigid.  They normally operate on a continuous, 24-hours a day, basis – at a rate as close as possible to their design capacity.  Thus they cannot increase production by adding an extra work shift or by running processes at a faster rate.  They do need to shut down periodically for preventive maintenance, as their systems are complex and they must deal with flammable liquids that are being processed at often high temperatures and pressures, where a failure of some part can lead to a catastrophic explosion.  They must also shut down on occasion for safety reasons, such as when a hurricane or other major storm threatens (an increasingly frequent occurrence in recent years in the US Gulf Coast, where much of the US refinery capacity is located, due to climate change – such weather-related shutdowns are discussed further below).  In general, then, refinery throughput is highly constrained in the short run by existing available capacity, which is being run continuously at as high a rate as they can.

Over the longer term, refinery capacity will depend on what investments are made to expand that capacity.  But new refineries cost billions of dollars, are rare, and when undertaken take many years to plan and then build.  Significant expansions in existing refineries are also very costly, and also require significant time to plan and then build.  Thus such investments are very carefully considered and are only made when they expect there will be a demand for the products of those refineries for many years to come – at least a decade or more.  It is not something they rush into.  Even if capacity is tight right now, such investments will not be made unless the owners expect those conditions to last for an extended time.  And even if the decision is made to make such an investment to expand capacity, it will normally take years before the added capacity will become available.

Thus in the near term, when one is already operating at close to the design limits of the refineries it will not be possible to supply much more than what the existing available capacity will allow.  Economists call this “inelastic supply”, as the percentage increase in supply of some product for some given percentage increase in the price that would be paid for that product (an “elasticity”) is low.  For refineries that are already operating at close to their technical limits, it will be very low.

The other factor in price determination is demand.  And for fuels such as gas or diesel, many will say the price elasticity of demand for such fuels is also low.  Indeed, a common view in the general population is that the price elasticity of demand for gas is zero – that they will have to buy the same number of gallons each week whatever the price is.  This is not really true (and contradicted by the assertion that they also cannot “afford” to pay more – if true, then at a higher price they will have to buy less).  But studies have found that while not zero, it is low.

For example, the Energy Information Agency in 2014 estimated the price elasticity of demand for gasoline in the US was just -0.02 to -0.04.  That is tiny.  It implies that if the price of gas were to rise by 10% (say from $4.00 to $4.40 per gallon), the demand for gas would decline only by 0.2 to 0.4%.  Other estimates that have been made have often been somewhat higher, although still low.  A widely cited review in 1998 by Molly Esprey, for example, examined 300 published studies, and found that the median estimate of this elasticity across those studies was -0.23.  This is still low.  It implies that a 10% increase in the price will be met by only a 2.3% fall in demand.

With a demand for fuel that does not go down by much when prices rise, and a supply for fuel that does not go up by much when prices rise (i.e. when refineries are already operating at close to their capacity), one should expect prices for fuels to be volatile.  And they are.  Even small shifts in the available supply or in the demand can lead to big changes in prices.

In these already tight markets of early 2022, Russia then invaded Ukraine on February 24.  The crack spread rose from $0.49 per gallon for the week ending February 25, to $0.64 the following week and to $0.74 the week after that.  It reached $0.88 by the end of March and $1.35 by the end of April.  As of the week ending June 24 it had reached $1.46, but then came down to $1.02 two weeks later.

The Russian invasion not only affected production at refineries in Ukraine, but international sanctions on Russia meant a significant share of Russian refineries would also no longer supply global markets.  While refineries in Ukraine are not a significant share of global capacity (just 0.2% in 2021), refineries in Russia are significant, with a 6.7% share of global capacity in 2021.  As a comparison, US refineries account for 17.6% of global capacity.

One should note that this does not mean that global capacity was effectively reduced by 6.7% of what it was.  Russian refineries continued to produce for their own markets, while also supplying others.  But the sanctions have reduced the volume effectively available by a significant amount.

In a market that was already tight, with refineries operating at close to capacity following the strong recovery demand in 2021 in the US and much of the world, such a reduction in effective supply acted as a tipping point.  The 3-2-1 crack spread shot up immediately.

C.  Policy Implications

What, then, can be done to reduce fuel prices?  I will take it as a given that that is the objective.  A case could well be made that to address climate change and the consequent need to reduce the burning of fossil fuels, high prices are good.  But while important, that is a separate issue I am not trying to address in this post.

First, where are gas prices now?:

The figures here are based on data gathered by the Bureau of Labor Statistics (BLS) for its calculations of the monthly CPI.  The figures are a consistent series going back to 1976 (further back than any other consistent series I have been able to find), are available in current price terms per gallon, and are not (here) seasonally adjusted so they reflect the actual prices paid that month.  And like the overall CPI that is commonly cited, it is an estimate of prices in urban areas.

As of June 2022, the average retail price of regular unleaded gasoline in the US was $5.058 per gallon.  For the chart, I have then shown what the historical prices would have been when adjusted for general inflation to the prices of June 2022 (based on the overall CPI).  The June prices are not the highest gas prices have been – they hit $5.51 a gallon in July 2008 – but they are close.  Although declining in recent weeks as I am writing this, it remains to be seen whether gas prices might resume their upward trend sometime soon.  The markets continue to be volatile, and prices could soon set a new record.

Whether that will happen will depend in part on what the policy response now is.  There are measures that can be taken that will reduce prices, but also measures that are being discussed that would likely have little effect, or might even raise prices. In this section, I will first discuss why, given the underlying causes of the price increases this year discussed above, some of the measures being discussed will likely do little and might indeed be counterproductive.  I will then discuss measures that could help lead to a reduction in prices.

1)  What Not To Do

First, some policies that will not lead to lower prices, or might even lead to higher prices:

a)  Perhaps the most widespread assumption is that if OPEC produced more crude oil, gasoline prices would then fall.  But that should not be expected given the current situation.  As seen in the chart at the top of this post, the crack spread widened sharply starting in late February, as a certain share of global refining capacity became not usable.  In the already tight markets refinery capacity became the effective binding constraint, not the price of crude oil.

More crude oil production by OPEC (or indeed by anyone) could well lead to lower crude oil prices – and indeed likely would.  But unless more of that crude oil can be refined into final fuel products such as gasoline, the available supply of gas in the market would not be affected.  Retail prices would remain the same.  What would change is that if crude oil prices decline by some amount with the increased supply of crude, the crack spread would widen.  That is, refiners would gain by this.  Consumers would not.

b)  For the same reason, sale of crude oil out of the Strategic Petroleum Reserve should not be expected to lead to lower retail prices for gas either.  President Biden announced on March 31 that the US would start to sell one million barrels of crude oil per day (an unprecedented amount) out of the US Strategic Petroleum Reserve for at least six months.  This announcement may well have had some effect on crude oil prices:  Crude oil prices had been rising through late March and then fell a bit (before returning to March levels in late May, and then continuing to rise until mid-June).  But this did not affect retail prices for fuels, which continued to rise until the last few weeks.  Rather, the crack spread rose (as seen in the chart at the top of this post) as refiners were able to obtain a larger margin between what they could sell their products for and what they had to pay for their crude oil.

c)  Also popular has been the proposal to reduce or eliminate taxes on the sale of gas and other fuels.  The federal tax is 18.4 cents per gallon on gasoline and 24.4 cents on diesel, while state taxes are of varying amounts.

President Biden on June 24 called on Congress to approve a temporary suspension of federal taxes on gas and diesel for three months.  As of my writing this, Congress had not approved such a suspension (it would complicate infrastructure funding, as such funding is linked to fuel tax revenues), and it does not look likely that it will.  But one never knows.  And as of July 6, six states had suspended their state fuel taxes for varying periods, with many more considering it.

What effect would such a tax cut have?  First, consider the federal tax, as it applies across the entire country.  As discussed above, the supply of fuels such as gas and diesel is constrained by available refinery capacity.  Economists refer to this as operating where the supply curve is “vertical”, in that a higher price for the fuel cannot elicit a significant increase in the supply of the fuel in the near-term, due to the capacity constraint.  A lower tax will not then lead to a lower price, as a lower price (if one saw it) would lead to greater demands for the fuels and refiners cannot supply more.  In such a situation, refiners are earning a rent, and a lower tax to be paid on the fuels will just mean that the refiners will be able to earn an even larger profit than they are already.  The crack spread will go up by the amount the tax on fuels is reduced.

The situation would be different if refiners could supply a higher amount.  Retail prices would fall by some amount due to the reduction in the tax, supplies would rise by some amount, and in the end consumers and refiners would share in the near-term gains from the lower tax.  What those relative shares will be will depend on how responsive the supply of fuels would be from the refiners (the elasticity of supply).  In the extremes, if refiners are able and willing to supply the increase in demand at an unchanged price (the supply curve is flat), then retail prices will fall by the entire amount of the tax cut and consumers will enjoy all of the benefit.  But if refiners are unable to supply more due to capacity constraints, then retail prices will be unchanged by the tax cut and refiners will pocket the full amount of the tax cut.  Currently, we are far closer to the latter set of circumstances than to the former.

The situation is a bit different at the state level.  If one state cuts its taxes while the taxes remain the same elsewhere, refiners will be able to move product to meet the higher sales of fuels in the state where taxes were cut.  This would, however, be at the expense of lower supply in the states that did not cut their taxes.  Fuel prices in the state cutting its taxes (and not matched by others) will fall by some amount due to the now higher availability of fuels in that state.  But with the overall supply constrained by what the refineries can produce, the lower amounts supplied to the rest of the country will lead to higher prices in the rest of the country.

Overall there will be no benefit, and indeed on average prices (net of taxes) will rise.  But there will be some redistribution across the states.  The amount will depend on what share of the states decide to cut their taxes.  At one extreme, if only one state does it and that state does not account for a large share of the overall US market, then the retail price (inclusive of taxes) will fall in that state.  If that state is small, prices elsewhere in the country would only rise by a small amount, but they still would rise.  But if more and more states decide to cut their fuel taxes, then one will approach the situation discussed above with the cut in federal taxes on fuels.  The full benefits of the lower taxes will accrue to the oil refiners, not to any consumers.

Finally, one needs to recognize that there is no free lunch.  The states cutting their fuel taxes will need to make up for the revenues they consequently lose.  To fund the expenditures paid for by the fuel taxes (often investments in road and other infrastructure), those states would need to raise their taxes on something else.

2)  What To Do

So what would lead to lower fuel prices given the current conditions?  The simple fact is that for prices to go down, one will need either to increase the supply of the refined products, or reduce the demand for them.  Taking up each:

a)  As was discussed above, refineries normally operate at close to their maximum capacity, and there is not much margin to respond to unforeseen demands.  Refineries are expensive, hence are not designed with much excess capacity to spare, and when operating are operated on a continuous, 24-hour a day, basis.  They also need to be shut down periodically for scheduled maintenance, as well as when unscheduled maintenance is required or when a strong storm threatens.

Still, there might be some measures that can be taken to push refinery throughput at least a bit higher.  Refiners certainly have an incentive to do so, given how high the crack spread is now (three to four times higher in recent months than what it was on average between 2014 and 2021).  But the crack spread does not need to be anywhere close to that high to provide a strong incentive.  A spread that is double what it would be in more normal times should more than suffice to elicit refiners to do whatever they can to maximize refinery throughputs.

There will also be an element of luck, given the increasingly volatile weather conditions that climate change has brought.  One can see this in a simple snapshot of a chart available on the EIA website, showing idle US refinery capacity (which is more properly measured by and referred to as distillation capacity) by month going back to 1985:

Volatility rose significantly starting in 2005 (the year of Hurricanes Katrina and Rita) and has been high since.  The sharp peaks seen in the chart are all in September or October – the peak months of hurricane season for the US.  Especially prominent peaks in the capacity that had to be idled were in September 2008 (Hurricanes Gustav and Ike), September 2017 (Hurricane Harvey), and September 2021 (Hurricane Ida).  With hurricanes threatening, refineries must be shut down for safety.  How fast they can then reopen depends on how much damage was done, but will require some time even if there was only limited damage.

It is impossible to say what will happen in the upcoming hurricane season.  But with the market so tight, any closures could have a large impact on prices.

b)  The other side to focus on is demand.  This could also be more productive in the near term given that little more may be possible on the supply side (as well as subject to chance, given the uncertainty in what will happen in the upcoming hurricane season).  But progress on demand-side measures will depend on political will, and Americans have been historically averse to measures that would reduce the near-term demand for fuels.

But it is important to recognize that not much would be needed in terms of reduced demand in order to reduce fuel prices by a substantial amount.  This is precisely because the demand for fuels is so price inelastic, as discussed before.  That is, a substantially higher price for gas does not lead to all that much of a reduction in the quantity of it purchased.  What economists call the “demand curve” (the amount purchased at any given price) is close to vertical.  When this is coupled with an also close to vertical supply curve for refined products (as refineries are operating close to their capacity, and cannot produce more no matter what price they can get), small shifts in the amount demanded at any given price will have a major effect.

[An annex at the end of this post uses simple supply and demand curves to examine this graphically.]

Given this lack of sensitivity to price under current conditions for both supply and demand, it would not take all that much to get prices to fall by a substantial amount.  Supply of refined products is constrained by refineries operating at close to their maximum, while on the demand side, purchases of fuels do not adjust by much when prices change.  As was noted above, the EIA in 2014 published an estimate of the price elasticity of demand for gasoline of just -0.02 to -0.04.  That implies that a 10% rise in the price of gas would reduce demand by only 0.2 to 0.4%.  Others have estimated higher elasticities, but all still relatively low.

Suppose, for the sake of illustration, that the price elasticity of demand was -0.10, so that a 10% rise in the price would lead to a reduction in demand of 1%.  This relationship also tells us a good deal about the shape of the demand curve – specifically its slope (locally).  If facing a completely vertical supply curve, then it implies that a 1% reduction in the demand for gasoline at any given price (meaning a shift in that demand curve to the left by 1%) would lead to a new price that is 10% lower than before.  And a 2% shift would lead to a price that is 20% lower.  While extrapolating in this way from what might be true for small changes to something substantially larger is dangerous, a 20% fall in the price of gas that is at $5.00 per gallon would lead to a new price of $4.00 per gallon – all resulting from just a 2% shift in the demand.  This is substantial but depends, as noted above, on how responsive demand is to the price.  If truly not very responsive, as is commonly held by many, then it will not take much of a reduction in demand (at any given price) to lead to a very substantial reduction in the price.

How, then, might one reduce the demand for fuels?  One possibility would be to encourage more work from home.  One saw the effect of this on fuel demands (and hence prices) in 2020, when working from home was required for health reasons at the start of the Covid crisis.  Workers are now returning to the office, but perhaps our political leaders should encourage a delay in this, or at least a slower pace on the return.  But it probably could not be mandated, and indeed probably should not be simply for the sake of cutting the price of gasoline.  And while opinions differ on this, some would say that extending work-from-home even further will reduce worker productivity.

A better way to reduce fuel demands would be to provide a greater incentive to take public transit rather than drive a car for a higher share of the trips one undertakes.  One could do the following:  First, raise tax revenues that could be used for these measures by raising federal taxes on fuels by, say, $0.25 per gallon.  As was noted above, when one is operating with a vertical supply curve, as we are now, increasing taxes on fuels will not lead to higher prices for the consumer.  The crack spread would fall, but with that spread that has varied between $1.00 and $1.50 per gallon in recent months, a higher fuel tax of $0.25 per gallon would still leave that crack spread at two to three times the $0.36 it averaged before.

According to EIA data, the total supply of motor gasoline in the US averaged 9.3 million barrels per day between 2016 and 2019 (taking a four-year average, and excluding 2020 due to Covid), while diesel supply averaged 4.0 million barrels per day.  Mutliplying this sum of 13.3 million barrels per day by 42 gallons per barrel and 365 days per year, the annual supply of these fuels averaged 204 billion gallons.  Rounding this to 200 billion gallons, a tax of $0.25 per gallon would raise $50 billion on an annualized basis.

This could be used to support public transit.  Something that could be done instantly (starting literally the next day) would be simply to stop charging fares on public transit systems – including buses, rail (subways), commuter trains, and whatever.  According to the National Transit Database, in 2019 all these public transit systems generated a total of $16.1 billion in revenues, mostly from fares but including also other locally-generated revenues such as from the sale of advertising.  (Again, 2020 was an unrepresentative year due to Covid so it is better to use 2019 figures.)  The database does not separate out fares from other revenues, but even if one treated it all as fares, the $16 billion needed would be far below the $50 billion that would be generated (on an annualized basis) by increasing the federal tax on gasoline and diesel by $0.25.

Filling empty seats on buses and subways also does not cost anything.  Indeed, operating costs would in fact go down by not having to collect fares.  There are significant direct costs in collecting fares (and to ensure too much is not stolen), but one would also gain operational efficiencies.  Buses now take a relatively long time to cover some route in part because at each stop people have to line up and go one-by-one through the front door to pay their fares in some way.  Not having to take so long at each stop would allow the buses to cover their routes at a faster pace.  This would increase effective capacity or, if capacity were to be kept the same as before, one could provide that capacity with fewer buses and their drivers.

The aim is to shift people from driving their cars to taking public transit for a higher share of the trips they take.  To the extent this simply fills up some of the empty seats, there is then no additional cost.  But if ridership increases by a substantial amount (something to hope for), capacity would need to grow.  This could most easily be accommodated by additional buses.  This would cost something, but according to the National Transit Database figures, the total spent in 2019 from all sources (federal, state, and local), for all modes of public transit, for both operating and capital costs, was $79 billion.  With the $34 billion left after using $16 billion to cover fares (out of the $50 billion that the $0.25 per gallon would collect), one could cover an increase in spending on public transit of more than 40%.  This would be far more than what would be needed even with a huge increase in ridership.  But we are now going beyond the very short-term measures that could be taken to reduce fuel demand.  However, with the long-term need to reduce the burning of fossil fuels, it is good to see that even a relatively modest fee of just $0.25 per gallon of fuel could support such an expansion in public transit.

Such an approach would lead to a reduction in the demand for those fuels.  How much I cannot say with the information I have, but it should be substantial.  And as discussed before, even a small reduction in the demand for these fuels should lead to a substantial fall in their price.  That fall in price would also be of benefit to all those who purchase these fuels, including those in rural areas who are far from any public transit option.  It would be a mistake to presume that stopping the collection of fares on public transit systems would only be of benefit to the users of public transit.

D.  Concluding Remarks

The price of gas is certainly high.  Although not quite a record (when general inflation is accounted for) it is close.  This has led to a number of proposals aimed at reducing those prices.  Particularly popular politically has been to cut fuel taxes for at least some period, with this championed both by President Biden (for federal fuel taxes) and in a number of states (where several have done this already for the state-level fuel taxes).  Many also blame OPEC for managing supplies in order to drive crude oil prices higher.  To address this, there have both been major sales out of the Strategic Petroleum Reserve (of one million barrels of crude a day), as well as diplomacy to try to get others to boost their supply of oil.

Under current market conditions, however, these initiatives should not be expected to reduce prices.  The issue right now is that refineries are the binding constraint.  They are producing as much of the refined products (fuels, etc.) as they can, but limits on their capacity keep them from producing more.  One sees this in the crack spread, which jumped up in late February immediately following the Russian invasion of Ukraine.  A substantial share of Russian refinery capacity became unusable, and this served as a tipping point in an already tight market.

Under such conditions, a lower price for crude oil will not lead to lower retail prices for fuels.  While it would benefit refiners (the crack spread would widen), the prices at the pump would not be affected unless refiners were somehow then able to raise their production.  Similarly, a cut in fuel taxes should not be expected to lead to lower fuel prices at the pump.  Rather, refiners would receive a windfall as they would receive a higher share of the retail price.  Refiners are already doing extremely well, with a crack spread in recent months that has been three to four times what it averaged between 2014 and 2021.  There is no need to make this even more generous.

To reduce retail prices, one should instead reduce demand.  One measure that would do this would be simply to stop charging fares on public transit.  Inducing only some of those now driving to use transit more often could have a significant impact on prices.  This is because the demand for fuels is not terribly responsive to price (consumers in the US do not cut back on their car use all that much when prices are higher), at the same time as the supply of fuels is limited by refinery capacity (so the supply of fuels cannot go up by much despite higher prices).  With both the demand and supply curves close to vertical, a small shift left or right in the curves can have a big impact on prices.

It would not cost all that much to end the collection of transit fares either.  Not only can it be done instantly (simply stop collecting), but the total public transit systems received in 2019 in fares paid (as well as in other revenues, such as from advertising) was only $16 billion.  One could easily cover this by increasing the federal taxes on fuels.  As noted above, a cut in fuel taxes would not lead to lower fuel prices.  For the same reason, an increase in fuel taxes (within limits) would not lead to higher fuel prices.  And just a $0.25 per gallon increase in federal fuel taxes would raise roughly $50 billion on an annualized basis.

It should be kept in mind that all this is based on current market conditions.  Those conditions can change, and change suddenly – as we saw in late February with the launch of the Russian invasion.  Thus, for example, while the crack spread is currently very high, this is in part a function of where crude oil prices are.  As of the week ending July 8, the price of West Texas intermediate was $103 per barrel.  With gas and diesel prices where they were then, the crack spread was $1.02 per gallon – far above the $0.36 per gallon it had averaged between 2014 and 2021.  But at a higher price for crude oil, the crack spread would fall.  At $131 per barrel (and with gas and diesel prices where they were as of the week ending July 8), the crack spread would be back at $0.36 per gallon.  And at $146 per barrel, the crack spread would be zero.  Presumably, if crude prices approached such a level refiners would cut back on production, leading to higher gas and diesel prices.  Crude oil prices would then be the binding factor, and efforts to lower those prices (e.g. by sales out of the Strategic Petroleum Reserve, or more OPEC production) could then matter.

The point of this blog post is that that is not where we are now.  Current conditions call for a different policy response.

 

=========================================================

Annex:  Supply and Demand Curves to Show the Impacts of the Options

For those of you familiar with simple supply and demand curves, it is easy to see the impacts of the policy options discussed verbally in the text above.

The supply curve of fuels from refineries slopes upward from a curve that is relatively shallow to something increasingly steep and ultimately to vertical.  At relatively low levels of production, where there is a good deal of excess capacity in the refineries, a small rise in prices for the fuels will elicit a strong supply response.  But as production approaches the maximum capacity of what the refineries can produce (in the near term, given existing plant), there can only be little and ultimately no more production no matter how high the price goes.

The demand curve is steep.  That is, if prices rise by some amount, the quantity of fuels demanded does not fall by all that much.  The price elasticity of demand is low.

Retail taxes per gallon of fuel add to the supply cost.  That is, in the figure above, the red curve (marked S2) is what the supplies would be at some lower (possibly zero) retail fuel tax per gallon sold, while the blue curve (marked S1) is what the supply would be at some higher tax rate.  The supply curve will shift upwards.  That is, for any given quantity of supply, a higher price will be needed for that amount to be supplied.

When the supply curve is relatively shallow and upward sloping, as in the lower left of the diagram, then a cut in the tax (from the blue curve to the red), with a demand curve such as D3, will lead to some increase in supply and a significantly lower price.  The price, in the diagram, would fall from P3 to P4.  This is the logic behind the proposals, such as have been made by President Biden, for a temporary cut in federal fuel taxes.

However, this is not where current market conditions are.  Rather, refineries are operating at close to their maximum capacity, and one is in an area where the supply curve is close to vertical.  When the supply curve is vertical, a reduction in fuel taxes will simply shift that vertical curve downwards, but with one vertical curve simply sitting on top of the other vertical curve.  While a reduction in the tax per gallon will increase how much the refiner receives, after taxes, it will not lead to a higher amount being supplied (refiners cannot produce any more) nor will it lead to a lower price for consumers.  The lower taxes will simply be reflected in higher profits for the refiners.

In terms of the supply and demand curves depicted above, one would be in an area such as that depicted with the demand curve D1 with a price of P1.  If the supply curve is shifted downwards due to the tax cut (from the blue curve S1 to the red curve S2), with nothing done to affect the demand curve, then the price remains at P1.

In contrast, if the market conditions are such that the demand curve is at D1 and the supply curve is close to vertical, yielding a price of P1, a relatively modest shift in the demand curve to the left, i.e. from D1 to D2, leads to a sizeable fall in the price – from P1 to P2.  The fall in the price is large because both the demand curve and the supply curve are steep, and indeed close to vertical for the supply curve.  In such conditions, modest changes in demand can have a big impact on the price.

A shift of the demand curve shows how much demand would change (at the given price) due to a change in some underlying factor other than price.  Inducing drivers to shift to public transit by ending the charging of fares on transit systems is one such example.  There are others, such as encouraging more work from home (so no commute at all is needed).  And should the economy fall into a recession (which I see as increasingly likely in 2023), there will also be a reduction in fuel demands.  But the latter is not a cause of lower prices that one should hope for.

 

The SpaceX Starship: Revolutionary, If It Works (Which It Probably Will – Eventually)

Source: Wikipedia – Super heavy-lift launch vehicle

A.  Introduction

SpaceX is currently developing a truly gigantic rocket it has named Starship.  It would be revolutionary.  Not only would it lift a heavier payload than the Saturn V, the rocket that took Apollo to the Moon and until now the heaviest lift launch vehicle that has successfully flown, but Starship is also being designed to be fully reusable.  Both the first stage and the orbiter would fly back to the launch pad, where they each would be caught in mid-air as they land by arms extending from the launch tower.  They would then be refueled and after minimal reinspection be able to take off again, within hours.  And each flight would cost only $2 million.

If all this works as intended.  And that remains to be seen.  But there are good reasons to believe it will, eventually.

It is certainly likely that there will be explosions or other causes of failure in the early orbital test flights now upcoming, and that even once operational the turn-around time will at first be a good deal longer than a few hours and the cost a good deal more than $2 million for a flight.  But even a cost that is ten times higher would still be incredibly cheap for such a lift capacity.  And the iterative process SpaceX follows of testing to failure, learning from the test, redesigning to address the problems found, and then testing to failure again, is a process that has allowed SpaceX to work through to a successful design.  It took some years of such tests before the first stage of the Falcon 9 rocket of SpaceX was successfully flown back to the launch site (or to a platform on a drone ship in the ocean) and landed, and then reused after some refurbishment.  But eventually, after a number of test failures, SpaceX worked out how to do this.  It is now routine.  And prior to SpaceX demonstrating this technology, the established view was that this would be basically impossible for an orbital launch vehicle.

Furthermore, in April 2021 NASA awarded SpaceX a close to $3 billion contract to build a variant of the Starship orbiter that would carry NASA astronauts from lunar orbit to the surface of the Moon and back.  It did this following a year of NASA engineers closely examining the SpaceX proposal (along with competing proposals from two others), reviewing the Starship plans with full access to all the technical information and to the development and testing plans.  NASA concluded the SpaceX Starship system could be relied upon to deliver on its proposal.  This is a tremendously important vote of confidence in the Starship system.

This post will first review the Starship system and what it promises to deliver.  It is really pretty astounding.  The big question is whether it will work, and that remains to be seen.  But the post will review the development process SpaceX is following, and contrast that with the sharply different process NASA is following with its Space Launch System (SLS) rocket.  The SLS, with its high cost and long delays, was discussed in some detail in an earlier post of this blog.  The contrasts in the approaches taken are stark.  The SLS will be a similarly sized rocket as Starship (a bit smaller), but has followed a very different design and development process.  Development began in 2011, with a design where the major components (the engines and the solid-fuel boosters strapped on the side) were the same as those used on the Space Shuttle or from other existing sources.  This should have saved both time and expense.  Yet despite this, there has yet to be a test flight of the SLS even though it is now more than ten years later.  The much delayed first flight is now scheduled (as I write this) for February 2022 (and had been set as November 2021 just a few months ago), and it is expected that recently discovered problems will delay this further.  Hopefully that first (and only planned) test flight will succeed.  If it does not, it is not at all clear what will happen to NASA’s plans to return to the Moon under the Artemis program.  As discussed in my earlier post, each SLS costs $2 billion to build, and under current production plans a second one will not be available for another two years in any case.

The Starship system that NASA has chosen to carry crew to the Lunar surface and back is also quite astounding.  This post will review what it would be able to do, and contrast this with the two competing proposals that NASA considered – the proposal from a team with Blue Origin (owned by Jeff Bezos) in the lead and one with the firm Dynetics in the lead.  The contrast is huge, where the SpaceX Starship proposal would deliver far more in several different dimensions (and at far lower cost).

The capabilities of the Starship also immediately raise the question of whether NASA should make use of it in a total revamp of its Artemis program to return astronauts to the Moon.  Instead of stuffing the crew into the relatively tiny Orion capsule for the trip from the Earth to lunar orbit, launched on an SLS that costs $2 billion per launch, why not use what would otherwise be an empty Lunar Starship for the journey?  The Lunar Starship will not only have all the life support systems needed by the crew for a multi-week journey, but the fully furbished habitable volume on the Lunar Starship is over 1,000 cubic meters, vs. just 9 cubic meters in the Orion capsule.  Certain technical issues would of course need to be worked out, but that should not be an insurmountable obstacle.

There likely will be obstacles, however, but political ones rather than technical ones.  This post will conclude with a discussion of those issues.  In my earlier blog post, I estimated that just for the period from when the programs started to what is planned by FY2023, Congress will have appropriated and NASA will have spent a total of $21.8 billion on the Orion capsule and a total of $32.4 billion on the SLS, for a total of $54.2 billion on them together.  It would be embarrassing, to say the least, to recognize that they turn out not to be needed, and that the SpaceX Starship system would not only be far less expensive but also far more capable.  And politicians do not appreciate being embarrassed.  While the politicians have asserted that they have been funding these programs to achieve certain space exploration goals, what appears to have driven the support of a number of the key Senators and Congressmen on the committees that set the NASA budget has been less the space exploration goals and more the resulting number of well-paid jobs in their constituencies.  Seen in this light, the high cost of the SLS / Orion systems is not a flaw but a feature.

But this is not sustainable.  How much the US government will spend on space exploration is limited, and treating it as a jobs program will mean national goals will either be long delayed or never met at all.  And while private companies such as SpaceX can help NASA achieve national goals (faster and at lower cost), they can only help if they are used to help.  Congress has often been opposed, with both Democrats and Republicans unfortunately aligned on this.  China is now moving fast in its space exploration program, and NASA may soon be left far behind if Congressional priorities do not change.

B.  The SpaceX Starship, and the SpaceX Approach to Development and Testing

SpaceX is currently developing, and actively testing, the extremely large rocket it has named Starship.  If it works, it will be revolutionary.  Starship will be huge – larger than the Saturn V as well as the SLS – and capable of delivering to low earth orbit a payload of 100,000 kg initially with this expected to grow to 150,000 kg or more as it is further developed.

The diagram at the top of this post shows Starship in comparison to other heavy-lift vehicles to give you a sense of its size.  It is huge.  And while the design is still evolving as its testing program proceeds (with the version of Starship shown in the diagram from a year or two ago), its basic dimensions will remain the same.  The Saturn V is well known, and is the heaviest-lift launch vehicle that has ever flown.  The SLS Block 1 will be the initial version of the SLS, with its first flight test now planned for 2022.  The SLS Block 2 Cargo is a planned follow-on SLS variant – still to be developed – that would have the heaviest lift capacity of the SLS series if it is ever built.   Finally, the N1 launch vehicle was developed, in secrecy at the time, in the USSR in the 1960s to carry its cosmonauts to the Moon.  There were four tests between 1969 and 1972.  Each failed with an explosion, and the especially spectacular explosion on the second test on July 3, 1969 – just 13 days before the launch of Apollo 11 – has been estimated to have been comparable in magnitude to that of the nuclear bomb dropped on Hiroshima. It was the biggest man-made non-nuclear explosion in history.

The SpaceX Starship will be a two-stage vehicle, with a first stage (named the Super Heavy) that will have a thrust at take-off that is more than double what it was on the Saturn V.  The first stage will initially use 29 of the newly-developed engines named Raptor, with a plan to increase this to 33 Raptor engines very soon.  And that number might rise to as many as 37.  The engines burn liquid methane with liquid oxygen.  The second stage will be powered by 6 Raptor engines (three optimized for burning in the vacuum of space and three for operation at ground level), with the spacecraft carrying the cargo and/or crew fully integrated into it.  Confusingly, this second stage/spacecraft has been named Starship, which is also the name of the whole rocket including the first stage Super Heavy booster.  To ease possible confusion, some refer to the second stage/spacecraft as the “Orbital Starship”, and I will as well.  The Orbital Starship would come in several variants, with vehicles just for cargo (Cargo Starship), for a human crew (Crew Starship), and to serve as a fuel depot in space (Fuel Depot Starship).  And as will be discussed in more detail below, NASA has extended a contract to SpaceX to build a variant that will take an astronaut crew to the Lunar surface and back (Lunar Starship).

Importantly, all components of Starship will be fully reusable, with the plan that there would be just minimal to no maintenance required between launches.  The Super Heavy booster would return to the launch site and relight some subset of its engines for a soft landing.  Indeed, as noted above, the plan is that it would come down precisely at the launch tower, where two extended arms on the tower would catch it as it (slowly) comes down.  No landing legs would be required.

The Orbital Starship would also be fully reusable.  It will have heat shield tiles that are chemically very similar to those used on the Space Shuttle, along with wing flaps and tail (see the drawing above) to guide the Orbital Starship as it re-enters the atmosphere and then lands.  A restart of a subset of the engines near the end will allow for a soft landing.  And as with the Super Heavy, the plan is for it to return to the launch tower to be caught as it comes down by the two large arms.  The plan (and hope) then is that the Starship and Super Heavy could be re-attached, refueled, and launched again within hours, with just minimal inspection required to ensure all was fine.

The payload will also be huge.  The immediate design goal is a payload to low Earth orbit of 100 tons.  This would be more than the SLS (95 tons in the initial, Block 1, version), although less than the Saturn V (140 tons).  But SpaceX plans to raise the payload capacity relatively soon to 150 tons through various means, and some have argued it might grow to even more.  As noted above, the number of Raptor engines on the first test flight will be 29, but will soon be increased to 33 and eventually possibly to 37.  Elon Musk (in a tweet on Twitter in July 2021) also mentioned the possibility that the number of Raptor engines on the Orbital Starship might be increased from six to nine at some point.  The additional three would all be the Vacuum Raptor variant.  This flexibility that has been built into the design of the Starship system – where adding core engines does not necessitate a complete revamping – is really quite remarkable.

The power of each engine will also soon be raised.  The Raptor 2 engine, which is already now starting to be produced, will have 230 or more tons of force – a big increase over the 185 tons of force in the first version of the engine that has been on the test flights thus far.  SpaceX also has a history of upgrading the performance of its rockets over time.  The initial version 1.0 of the Falcon 9 rocket could carry a payload to low Earth orbit of 9.0 tons.  But this rose to 22.8 tons in the Block 5 model that is now standard (22.8 tons when the boosters are expended, or 16.8 tons when the booster is recovered). That is, the Falcon 9 can now carry to orbit 2 1/2 times what it could when it first flew.

Will Starship work?  That remains to be seen, but testing is underway and there are reasons to be optimistic.  As I write this in December 2021, there have been seven test flights, all (so far) of the Starship upper stage, with simple up and down flights of up to 12.5 kilometers and a landing attempt on each.  There were a series of sometimes spectacular crashes (or as SpaceX calls them, RUDs, for Rapid Unscheduled Disassemblies), but the seventh (and hence final flight of the series) was fully successful, so they have moved on.  The plan now is to conduct a first test of the full Starship, with the Super Heavy plus the Orbital Starship launching from the SpaceX base in Boca Chica, Texas (on the Gulf of Mexico at the border with Mexico).  The first test would be of an almost orbital flight (more than 80% of what a full orbit would be) to land about 100 kilometers northwest of Kauai in the Hawaiian Islands.  The Super Heavy would return most of the way to the Boca Chica base but come down softly in a “landing” in the waters of the Gulf of Mexico around 20 kilometers off the coast on this first try.  The Orbiter Starship would come down through the atmosphere, testing its heat shield in particular, to “land” in the waters of the Pacific.

It will be an ambitious first test of the complete rocket, and the likelihood of everything working right the first time is low.  But this is in keeping with the SpaceX development and testing approach (to be further discussed below), which is to test early and often, and then iterate on the designs until they work.  While no date has been publicly announced for this first test (SpaceX is a private company and under no obligation to announce this), the indications are that the aim is for a first flight sometime early in 2022.  However, this will depend on SpaceX receiving necessary permits (related to environmental and safety issues) from the FAA, and it is not clear as I write this when this process will be completed.  While SpaceX ignored this requirement on one of the early test flights of Starship, it is now on notice from the FAA and it is doubtful they will ignore them again.

As noted before, if the design works it would be revolutionary.  It would be fully reusable, with both the first (Super Heavy) and second (Orbital Starship) stages returning not simply to a base but to the very launch tower where they would be caught in mid-air.  Probably the biggest question mark is whether the heat shield that covers one side of the Orbital Starship will prove to be durable and quickly reusable with no maintenance required.  As noted above, the chemical composition of the material is similar to that used on the heat shield for the Space Shuttle.  But on the Space Shuttle, extensive checks and maintenance of the heat shields proved to be necessary after each flight, with this a major contributor to the high cost of the Shuttle.  The original plan (and hope) was that a Space Shuttle could be turned around and flown again within two weeks of its landing from its preceding flight.  This proved impossible.  In the later years of the Shuttle program, each operational Shuttle was generally flying only once a year.

SpaceX has learned and applied important lessons from the experience with the Space Shuttle, and has addressed the heat shield tile issues in two important ways.  First, while most of the over 21,000 heat shield tiles on each Space Shuttle Orbiter had to have a unique shape due to the shape of the Shuttle, most (although not all) of the tiles on the Starship Orbiter will have a standard, hexagonal, shape.  This will make them far easier to replace when needed, as a new, custom-molded, tile will not (normally) need to be made each time.  And second, the tiles on the Starship Orbiter will be attached to built-in spikes on the body of the Starship, rather than attached with just a special glue.  It is hoped this will make them more durable.

If a fully and rapidly reusable design can be made to work, the cost of each flight of the Starship will be incredibly low.  Elon Musk has said that the fuel would cost only about $900,000 on each flight, and with the other operational costs the total would only be about $2 million per flight  This might well be optimistic (Elon Musk often is), but even if it is, say, ten times higher at $20 million per flight, then with a 100,000 kg payload the cost per kilogram will be roughly one-hundredth of the cost per kilogram of a launch on the similarly sized SLS.  And it will be roughly one-tenth of the cost per kilogram of launching on the current lowest-cost launcher – the Falcon Heavy.  And if it truly works out to be $2 million per launch, then the cost will be one-thousandth of what it would cost on the SLS.  This is all pretty incredible – if it works.

The development cost will also be far below what it has cost NASA to develop its similarly sized SLS.  As discussed in the earlier blog post, the cost of developing the SLS will have reached over $32 billion by FY2023.  While the cost of developing the Starship has not been published (SpaceX is a private company), and indeed as not is known by anyone what in the end it might be as development is still underway, Elon Musk has said in an interview that he expects it will come to about $5 billion to complete.  If that turns out to be the case, that would be less than one-sixth the cost of the SLS.  And while it is not clear whether the $5 billion includes the cost of including a section of the Starship to house the crew, if it does then for comparability the cost of developing the Starship with the crew quarters should be compared to the cost of SLS and Orion together  That will total over $54 billion – $32.4 billion for the SLS plus $21.8 billion for Orion.  That would be ten times higher than the cost of developing a crew capable Starship, if the $5 billion estimate for Starship turns out to be correct.

But the design remains to be proved.  While the first test flight will be important, the odds are high that there will be a failure at some point during that test.  As long as the Starship manages not to explode on the launch pad (with the damage that would cause to the launch pad) the flight should be considered a success.  Valuable information will have been obtained.  Particularly valuable, if it gets to that point, will be information on how well the heat shield holds up on re-entry.

The process followed by SpaceX is iterative.  A design is developed, a prototype is built, and it is then quickly tested – to the point of failure to see how much it can do.  Alterations are then made in response to what did not work in the test, with the revised design then soon tested again to take it to the next stage.  Failures are common and indeed expected.  Musk has noted that if failures did not occur, then you were not pushing it far enough.

Importantly also with this approach to development, manufacturing methods need to be developed to allow that frequent testing to the point of failiure to be at a cost per test that is not high. As Musk noted in a tweet in February 2020:

“Hardest problem by far is building the production system of something this big. …  Building many rockets allows for successive approximation. Progress in any given technology is simply # of iterations * [times] progress between iterations.”

This would then also drive design decisions.  For example, and directly counter to the conventional wisdom, the Starship hull and tanks are made of stainless steel.  Steel is seen as heavy – not good for a flying vehicle – and almost all rockets and spacecraft have been made of aluminum.  In the initial design, Starship was to be made of advanced carbon fiber.  Carbon fiber is light and can be made to be strong, but for several reasons the decision was made to switch to stainless steel.  Prominent among them was that stainless steel is relatively easy to work with – it can be bent or reshaped when needed for a design change – plus it is cheap.  As Musk noted in an interview in January 2019 (just after they announced Starship would be made of steel rather than carbon fiber), the type of stainless steel they need costs only about $3 per kg, vs. $135 per kg for carbon fiber.  Furthermore, there is about 35% wastage when one works with carbon fiber, so the true cost per ton for carbon fiber is over $200 per ton.  Any scrap from the stainless steel can be melted down and used.

There are also other advantages to steel, including its far higher melting point (which means less depends on the heat shield tiles, and they can then be made both simpler and lighter).  But the flexibility provided by using a relatively inexpensive and easily worked material is key in the SpaceX iterative development approach with its frequent testing and then redesign.

The decision to switch to stainless steel from the original plan to use carbon fiber also illustrates the flexibility in the overall approach followed by SpaceX.  Elon Musk’s first public announcement of his intention to build what evolved into Starship was made in November 2012.  The Raptor engines that would power the rocket were already being tested on NASA test stands in 2014.  The initial design, then at a 12-meter diameter, was revealed in September 2016.  This was then changed dramatically, to a 9-meter diameter design, in September 2017.  Through this point, the rocket would have been made of carbon fiber.  But then in December 2018, Musk revealed they had decided to switch to stainless steel for the reasons noted above.  And then just four months later, in April 2019, they were already conducting their first flight tests of what they called the Starhopper vehicle.  These were tests basically of the Raptor engine, the use of the engine to allow for a soft landing (and the controls required for this), and the stainless steel design.  The Starhooper test vehicle was tethered for its first “flight” on April 3, 2019 and it rose just one foot.  They then took it to one meter two days later on April 5.  The first untethered flight was in July 2019 to 20 meters, and the final test was in August 2019 when it rose to 150 meters and moved sideways for a short distance before returning to land softly on the ground.

There were then a series of seven tests of what became increasingly similar to the Orbital Starship starting just one year later, between August 2020 and May 2021.  The test vehicle had just one Raptor engine for the first two tests (rising just to 150 meters, and then landing) and then three Raptors in the subsequent tests (where it rose to a height of up to 12.5 km before returning to the pad to try to land).  After a few spectacular crashes and explosions, the Starship had a clean landing at the pad on the seventh and hence final flight.  There were modifications made following each test flight, fixing what went wrong, and they got it right by the seventh such flight.

The iterative SpaceX approach is in sharp contrast to that used by NASA in its development of the far more costly SLS.  As noted before, there has not yet been even one flight test of the SLS design, with the first now scheduled for February 2022 but most likely until sometime later due to recent issues being discovered.  Yet it has been under development since 2011, and arguably from before as the SLS design has many similarities with the Ares V rocket that at that time was under development (but then superseded by the SLS).  And the SLS design is based on existing rocket components, with the four main engines the same as those on the Space Shuttle (the Ares V would have had five of those engines, but otherwise the same).  The two solid-fuel side boosters are also taken from the Space Shuttle (and indeed make use of ones left over from the Space Shuttle program for the initial several SLS boosters to be built – but each with five segments rather than four).  And the second stage of the initial SLS is taken from that used on recent Delta IV rockets.

Drawing from existing rocket components is not necessarily wrong, as it should have led to lower costs and a shorter development period.  It is thus particularly difficult to understand why it has taken so long – with no flight test even after ten years – and why the cost has been so high.  In contrast to SpaceX, NASA has followed an approach where great care (and expense) is taken to try to ensure the full and final design is completely right, with few tests of the overall system.  Indeed just one test flight is planned for the SLS.  If this test fails, it is possible the total program will end.  It will depend on whether they can determine whether the cause of the failure can be found and relatively easily fixed.  A failure would in any case cause major delays of probably years.  A second SLS will not be available for two years (in the current schedule), and it could take longer depending on what they determine was the cause of the failure on the first test.  That is a lot riding on just one test, for a program expected to cost over $32 billion.

C.  Starship as a System to Land Astronauts on the Moon

1)  Introduction

The decision by NASA in April 2021 to select the SpaceX proposal to ferry astronauts from lunar orbit to the surface of the Moon and back was highly significant for several reasons.  First, and importantly, it was a clear vote of confidence by NASA management and its technical staff that the Starship system would not only work but would work soon.  NASA teams had thoroughly reviewed the proposal of SpaceX (as well as the two competing proposals, from Blue Origin and Dynetics) for more than a year, with full access to the technical teams at SpaceX.  As a critical, outside, reviewer seeking to find any holes that there might be in the SpaceX plans, NASA’s conclusion that the SpaceX proposal was doable is of great significance.

Second, the use of Starship for this purpose illustrates well the capabilities Starship will have, once it is operational, for not only such lunar missions but more broadly.  It is therefore of interest to examine in some detail what SpaceX is now being contracted to do, how it would work, and what Starship might then be used for, in particular in support of NASA’s goal to return to the Moon and establish a base there.  Contrasting the SpaceX proposal for what NASA is calling its “Human Landing System” (or HLS) with the two competing proposals that NASA considered is also of interest as it highlights how uniquely capable the Starship system is in comparison to the best that others can offer.

2)  Lunar Starship – The SpaceX Proposal for the HLS

NASA has contracted SpaceX for its Human Landing System, where it would serve as one component of the system NASA would use to return men (and as NASA repeatedly emphasizes in its PR materials, also women and a person of color) to the Moon.  Other major components of that system include the Orion spacecraft to carry crew from the Earth to lunar orbit, a “Lunar Gateway” that would be a modest-sized space station in permanent orbit around the Moon and serve as a waystation to transfer crew and cargo to the vehicles that would bring them to the lunar surface, and the SLS rocket that would carry the crew and possibly cargo to lunar orbit from the Earth.

NASA developed this basic plan to return to the Moon in the mid-2010s, with a goal of a first landing back on the Moon by 2028.  This time frame was then greatly compressed in 2019 by the Trump administration, with NASA charged with accomplishing this by 2024.  Vice President Mike Pence (in his capacity as head of the National Space Council) made the announcement, but stated this was “at the direction of the President”.  While the reason for the choice of 2024 was never officially stated, few failed to notice that 2024 would have been the final year of a second Trump term in office, had he not lost the election in 2020.

NASA then modified its plans in accordance with this new mandated schedule.  While the new schedule was never realistic, the revised, compressed, schedule did have real impacts on the mission designs and on contracts signed.  And while NASA has now finally acknowledged that the 2024 target will not be possible, it is still officially following the plans as drawn up during the Trump presidency.  The only change is the first landing on the Moon with a crew would not take place in 2024, as planned before, but in 2025 (which is still far from realistic – several more years should be added).  The new NASA Administrator Bill Nelson announced the 2025 date during a press conference in November 2021, and blamed the delay on the seven months lost due to the litigation by Blue Origin protesting the award of the HLS contract to SpaceX (which will be further discussed below).  Bill Nelson is a former Democratic Senator from Florida who has long been closely involved with NASA space programs, flew on the Space Shuttle in 1986 as a member of Congress, and was one of the key Senators (along with Republican Senator Shelby of Alabama and others) who wrote into legislation in 2010 the requirement that NASA develop the SLS (discussed further below).

Based on these still official NASA plans (albeit slightly delayed to 2025), NASA would organize the return to the Moon through a series of missions that it has labeled the Artemis program (where Artemis was the twin sister of Apollo in Greek mythology).  Specifically:

a)  Artemis I:  This would be the first, uncrewed, test flight of the SLS together with the Orion capsule, with a scheduled launch date (until recently) of February 2022.  It has been repeatedly delayed.  The original plan (in 2011) was that the SLS would have flown no later than 2017.  And while a new official launch date has not yet been set as I write this, no one expects that it will launch in February, this time due to new problems recently found in a malfunctioning computer for the SLS engines.

When the launch does take place, the SLS would send the unmanned Orion capsule first into Earth orbit and then to the Moon, passing just 60 miles above the lunar surface before entering into a high orbit of 38,000 miles above the Moon.  It would spend several days there and then return to Earth, passing again 60 miles from the lunar surface, and then testing its heat shield in the high-speed re-entry to Earth.  The entire mission would take several weeks.

b)  Artemis II:  This would be the first crewed launch of the SLS and Orion.  Until recently the official plan was for it to go in 2023, but with the acknowledgment that the first crewed landing on the Moon will not be before 2025, it has been pushed to May 2024.  This is still unrealistic.  And as the Office of the Inspector General of NASA noted in a recent assessment of the Artemis program, sending a crew on a mission around the Moon on the first flight of the Orion capsule with life support systems in place is an “operational and safety risk”.  The life support system has not been included in the Orion capsule being tested on the Artemis I mission to save on cost (and maybe time).

Artemis II would be a ten-day mission, with the SLS first launching the Orion into a high Earth orbit, during which the Orion and its life support systems would be monitored to see whether all is working properly.  After about two days it would then be launched to the Moon, but not into lunar orbit.  Rather, it would be sent on a fly-by trajectory where it would use lunar gravity to swing around the Moon, in a large figure-8, and then return directly to Earth.

c)  Artemis III:  This would be the first crewed landing on the Moon, with about one week to be spent on the lunar surface and the whole mission taking about two weeks.  The Orion would be launched on the SLS and would go to a very high lunar orbit (what they call a “Near Rectilinear Halo Orbit”, or NRHO).  The NRHO will have an apogee of 43,500 miles and a perigee of 1,900 miles, and just one orbit will take seven Earth days.

The original plan was that the Lunar Gateway would have been prepositioned in this orbit to serve as a staging area where the Orion would unload its crew and cargo, with the HLS then loaded with the crew and cargo from it to take them to the lunar surface.  But while there may be some components of the Lunar Gateway system ready in lunar orbit by then, there are skeptics on whether it will be ready by whenever Artemis III might be launched.  In that case, the transfers of crew and cargo would be done directly from the Orion to the HLS.  This of course calls into question why it is needed at all.  There might be a quite valid scientific justification for such a facility in orbit around the Moon, but so far the rationale given has been its role in the Artemis missions.

SpaceX won the contract for the HLS.  Under its proposal, a version of the Orbital Starship would be developed which would operate solely in the space environment to ferry crew and cargo back and forth to the lunar surface.  It would never re-enter the Earth’s atmosphere and hence would not need the heat shield.  Nor would there be a need for the large wing flaps and tail that on the Orbital Starship are required for aerodynamic control on re-entry to the Earth.  And while I have not seen this written in any of the limited plans that have been made publicly available (details on the Artemis plans that have been made public are surprisingly sketchy), it is not at all clear that the Lunar Starship would require all six of the Raptor engines that are on the Orbital Starship.

Three of those Raptor engines are optimized for operation in the vacuum of space (Vacuum Raptors), and three (Sea-Level Raptors) are optimized to operate in the atmosphere near sea-level for the return and soft-landing on Earth.  I have not been able to find in any reports whether all six are needed for the initial launch into Earth orbit, but with three designed for operation at sea-level, it is not clear that all six are.  Furthermore, the Lunar Starship itself would be lighter than an Orbiter Starship (due to no heat shield nor flaps) so less thrust would be needed.  It is therefore not clear whether the three Sea-Level Raptor engines would be needed, and if not this would also save a good deal of weight.  And with three Raptor engines sufficient for a landing on Earth, one would assume that three would more than suffice to land in the far lower gravity of the Moon.  Indeed, while full details have not been made public, the final phase of the landing of the Lunar Starship on the Moon would not use the Raptor engines at all, but rather a large number (possibly 24) of medium-sized thrusters positioned in the mid-body of the Lunar Starship for the final landing.  This is so that the Raptor engines at the bottom do not blow out an excessive amount of debris as they land.

Keep in mind also that once a vehicle is in orbit, the thrust required in order to, say, launch the craft on to a trans-lunar trajectory (TLI) from Earth orbit to the Moon (or vice versa) does not have to be all that strong.  The power required comes from the combination of thrust times the duration of the burn, so a set of engines with a lower overall thrust could achieve the velocity required by having the engines stay on for a longer period of time.  This is in contrast to a launch from the Earths’s surface, where a high thrust is needed to escape the pull of gravity.  Once in orbit, one does not need to rush this.  So again, it is not clear that one would need to keep all six Raptor engines on the Lunar Starship.  The only issue is whether the initial launch of the Lunar Starship into Earth requires all six Raptors, even though only three are optimized for operation in a vacuum.

The Lunar Starship would be launched into Earth orbit on the Super Heavy booster, as would be standard.  Some set of its own Raptor engines would also be fired for this.  It would then be refueled in Earth orbit before launching to the Moon.  This ability to refuel in orbit is a key ability of the Starship system, and is central to the flexibility that Starship allows in serving multiple objectives.  Prior to the launch of the Lunar Starship, a Fuel Depot Starship would have been launched into Earth orbit, where it would have received fuel carried and then transferred to it from multiple Orbital Starship launches.  The Fuel Depot Starship would essentially be a set of two large fuel tanks (one for liquid methane and one for liquid oxygen – the fuel plus oxidizer used by Starship) that has been well-insulated given the cryogenic temperatures the fuel and oxidizer need to be kept at.  And since the Fuel Depot Starship would be kept in orbit and not land back on Earth, there would be no need for a heat shield nor the wing flaps (nor for all six of the Raptor engines I assume) on it either.

A fully-fueled Starship can hold 1,200 tons of fuel (or technically, fuel plus oxidizer).  But each Starship launched from Earth would be able to carry 100 tons of cargo initially, with this expected to grow relatively soon to 150 tons.  Assuming the 150-ton capacity, Elon Musk noted that eight Starship flights to the Fuel Depot Starship would provide the 1,200 tons needed to fill the tanks on a Lunar Starship.  But how full the tanks would need to be will depend on how much the Lunar Starship would weigh (with no heat shield, wing flaps, nor possibly some of the Raptor engines), and how heavy of a payload would be taken to the lunar surface.  Musk concluded that only four flights to deliver fuel might be sufficient.

Much of this is still not clear – at least in what has been made public – and the final answer will depend on how heavy a payload the Orbital Starship will be able to carry to low Earth orbit (100 tons or 150 tons or something in between), how much lighter the Lunar Starship might be than the regular Orbital Starship, and how heavy the payload to the Moon would be.  And with differing payloads, the number of refueling flights needed might well differ from mission to mission.  None of this has as yet been spelled out, and likely is not yet fully known to anyone given the factors that are still uncertain.  But the key point is that the Starship system provides for flexibility where if additional flights are needed to lift the fuel required for refueling of the Lunar Starship in orbit, they can be carried out as needed.

Fully refueled, the Lunar Starship would then be sent to lunar orbit, to either the Lunar Gateway (if it has been built) or just to the similar orbit, to await the Orion capsule carrying the crew who would be launched on an SLS.  The Lunar Starship would likely already have most or all of the cargo required, but if there is anything additional carried on the Orion it would be transferred along with the crew.  The Lunar Starship would then carry the crew and cargo to the surface of the Moon and serve as a base for the crew during the time they spend on the lunar surface.  On the first mission (Artemis III) the current NASA plan, as noted above, is that this would be one week.  Also, the contract specifications NASA wrote for the HLS competition was that the initial mission to the Moon would be for two astronauts to land while two would remain in lunar orbit at either the Lunar Gateway (if available) or just in the Orion capsule.  After the initial mission, NASA’s plan is that the HLS should be upgradable so as to be able to carry four astronauts to the surface and back.  But given the extremely large capacity of the SpaceX Lunar Starship, it is likely that they will build in the capacity to carry a crew of four from the start.  And if NASA keeps to its plan to use Orion to carry astronauts to lunar orbit from the Earth, then that will be the ceiling as the Orion can carry no more than four (other than in an emergency).

Once the Lunar Starship carries the crew back to the Orion capsule waiting in lunar orbit, the Orion with the crew would return to the Earth.  The Lunar Starship, designed to be fully reusable for this role, would return on its own to Earth orbit where it could then be refueled by multiple regular Starship launches, stocked with any cargo desired, and then sent back to the Moon for the next NASA mission.  And the cycle could keep repeating.

3)  Comparison of the Three HLS Proposals:  SpaceX, Blue Origin, and Dynetics

After an initial request for proposals for the HLS in 2019, NASA provided funding in May 2020 to three teams to develop their proposals further.  The three chosen were Space X (with its Lunar Starship); a team that called themselves the “National Team” but which was led by Blue Origin and which is typically referred to as the Blue Origin proposal; and finally a team led by the not too well known aerospace firm Dynetics.  The HLS vehicle of Dynetics is named “ALPACA” (an acronym for Autonomous Logistics Platform for All-Moon Cargo Access).

The three proposals differed radically, both in the approaches taken and in their size.  Such diversity in approaches is certainly good and healthy, but at the same scale they would look like this:

Source: IEEE Spectrum, January 6, 2021

 

The mission plan using the Lunar Starship was described above.  The plans are quite different for the Blue Origin and Dynetics proposals.  A short description of each will be helpful before the three proposals are directly compared.

Blue Origin has worked in partnership with Lockheed Martin, Northrup Grumman, and Draper for its proposed lunar lander.  Its design is the most traditional, with a broad similarity to the design of the lunar lander (the LEM) of the Apollo program.  That is, there will be three separate components, with a descent stage (to be built by Blue Origin itself, and called the “Descent Element”), an ascent stage (to be built by Lockheed Martin, and called the “Ascent Element”), as well as a “Transfer Element” (to be built by Northrup Grumman) that is basically a rocket engine with fuel tanks that would take the proposed HLS from the high lunar NRHO orbit to a low lunar orbit where it would disconnect and the descent stage would take over.  Draper (or Draper Labs) would be responsible for the descent guidance system and flight avionics.

Together they call themselves the “National Team”, with Blue Origin in the lead.  They have named their proposed lunar lander the “Integrated Lander Vehicle” (or ILV).  The firms in the National Team basically come out of the traditional aerospace industry.  While some might consider Blue Origin (owned by Jeff Bezos) as different – as a private, entrepreneurial, company more akin to SpaceX – it really isn’t.  Much of the Blue Origin leadership has been drawn from traditional aerospace firms, and it has followed a development process more similar to that of traditional aerospace than that of SpaceX.  Its CEO, Bob Smith, came to Blue Origin from Honeywell Aerospace, and had previously held senior positions at the United Space Alliance (a joint venture of Lockheed Martin and Boeing that managed aspects of the Space Shuttle program for NASA) and before that at The Aerospace Corporation.

The three components of the Blue Origin ILV would be flown to the NRHO lunar orbit on three flights of the United Launch Alliance (ULA) Vulcan Centaur rocket that is now under development.  The United Launch Alliance is a 50/50 joint venture of Boeing and Lockheed, and the Vulcan Centaur would be a follow-on launcher that would replace ULA’s Atlas V and Delta IV boosters that are now being phased out.  The Vulcan Centaur will use rocket engines (named “BE-4”) being developed by Blue Origin, but due to repeated delays in delivering those engines, ULA has had to postpone the first test flight of the vehicle.  At one point it was supposed to have flown in 2020.  Most recently, the formal plan is for a test flight in 2022, but some have noted that with the continuing delays at Blue Origin, the first test flight might not be until 2023.

Alternatively, instead of flying the three components of the Blue Origin ILV to the NRHO on three of the still-to-fly Vulcan Centaur launch vehicles (with the three components of the ILV then assembled together into one unit there), it could be flown to lunar orbit already assembled on one flight of the still-to-fly SLS.  But additional SLS vehicles are not available, and would cost $2 billion each if they were.  A single Vulcan Centaur launch is expected to cost perhaps $120 to $150 million.

After receiving the NASA crew and cargo in the NRHO lunar orbit, the Blue Origin ILV would then be taken to a low orbit around the Moon with the Northrup Grumman Transfer Element, after which it would disconnect and the Blue Origin Descent Element would take over for the final descent to the surface.  Using the Transfer Element basically saves on the weight of the larger tanks (and associated hardware) that would otherwise be required if all of the trip from the NRHO to the lunar surface were powered by the Descent Element.  And as discussed below, it might be possible to reuse the Transfer Element on subsequent missions, although this would require that the fuel it needs be brought to the lunar NRHO orbit in some way.

The habitable space in the ILS would be a pressurized cabin on the Ascent Element, and the crew would live there for the time they spend on the lunar surface (about a week on the initial mission).  Cabin space would be tight, and only enough for a crew of two on the initial flights.  NASA’s original plans for the return to the Moon was for a crew of four, but reducing this to two for the initial two flights was one of the simplifications NASA introduced (albeit at a greater overall cost in the long run) when the Trump administration instructed NASA to accelerate its plans and get a crew to the Moon by 2024.  But the intention is for crews of four after the first two missions, and one of the criteria NASA indicated it would consider in the competition for the HLS contract was whether the proposed lander could be relatively easily scaled up to handle a crew of four.  One factor NASA cited when it decided not to accept the Blue Origin proposal for the HLS was that such a scaling-up to accommodate a crew of four would be difficult with its design.  A substantially larger cabin in the Ascent Element would be needed, which would weigh more as well as take more space, which would require not only a substantially redesigned Descent Element to handle the extra weight but also more powerful Ascent Element engines to return the crew to the NRHO lunar orbit.

Once the Ascent Element returns the crew to the NRHO orbit, the crew along with any cargo (lunar rocks) being brought back would transfer to the waiting Orion (or first to the Lunar Gateway, if it is there yet), and then return to Earth on the Orion.

What is not fully clear is what will then happen in terms of reusability of (portions of) the Blue Origin HLS.  The Descent Element can clearly only be used once, as it would remain on the surface of the Moon.  The Ascent Element might in principle be usable again (if designed for this), and possibly also the Transfer Element (if it is designed to hold sufficient fuel to bring it back to the NRHO orbit following its use for bringing the HLS to a low lunar orbit).  However, each would then need to be refueled if they were to be used again, and probably two Vulcan Centaur launches would be required to bring those fuels (which are also different for each – liquid hydrogen and liquid oxygen for the Ascent Element and hypergolic fuels that ignite on contact for the Transfer Element).  it is not clear whether much, if anything, would be saved by developing the tankers to bring those fuels to the NRHO and then transferring them to the Ascent and Transfer Elements.  The tankers would then be thrown away as they could not be returned to Earth, and it is not clear whether much would be saved over just building new copies of the Ascent and Transfer Elements (along with the Descent Element) and then sending them fueled to the NRHO.

Dynetics has given the name ALPACA to its proposed HLS.  Dynetics is a medium-sized aerospace firm, based in Huntsville, Alabama, that was acquired by the defense contractor Leidos just a few months before NASA announced in April 2020 that Dynetics would be one of the three HLS proposals it would fund for further development.  While Dynetics would work with a number of subcontractors (the two most important being Sierra Nevada  Corporation – an aerospace firm – and Draper once again for the avionics), the responsibility for the design is fully with Dynetics.

The Dynetics design is quite radically different from any that have been considered before.  As seen in the artist’s rendering above, the Dynetics ALPACA would be basically a crew cabin with rocket engines and their tanks on the two sides.  Three launches on a Vulcan Centaur would be required to bring it to the NRHO lunar orbit – one for the empty ALPACA and two for the fuel and oxidizer (liquid methane and liquid oxygen).  The fuels would then need to be transferred to the ALPACA in the NRHO lunar orbit, and NASA cited this as a concern when it reviewed the Dynetics proposal.  The technology still has to be developed.  While there would also be in-space transfers of the cryogenic fuel and oxidizer in the SpaceX Starship proposal – and in far greater volumes – this would be done in Earth orbit for the Starship instead of lunar orbit for the Dynetics ALPACA.  If a problem arose, it could be more easily addressed if still in Earth orbit.  Furthermore, development of this refueling ability is central to the Starship system, and hence SpaceX is devoting a good deal of attention to ensure it will be able to do this.  Dynetics and its partners would not require this ability for anything other than their ALPACA.

Once fueled and ready, the Orion would be launched on the SLS, rendevous with the ALPACA in the NRHO lunar orbit for the transfer of the crew and cargo, and the ALPACA would then take the crew to the surface.  NASA found an important issue during its review of the Dynetics proposal, however.  Given the weight of ALPACA, the thrust of its engines, and the fuel that would be carried, NASA concluded that there would be insufficient fuel to keep ALPACA from crashing into the lunar surface, even with no crew or cargo at all.  That is, its payload capacity would be negative.  Dynetics conceded that this was indeed an issue with its current design, but was confident that they would be able to lighten ALPACA sufficiently so that it would be able to carry the crew and cargo and not crash.  NASA, however, was skeptical.  Such spacecraft normally increase in weight as they are further developed, as issues are found requiring modifications to resolve those issues.  It is rare that there would be a reduction in their weight.  And NASA’s bid specification was that the HLS had to be able to bring to the lunar surface a minimum of 865 kg (of crew and cargo, including the space suits needed to venture outside), with a preferred goal of 965 kg.

Assuming it could land, the ALPACA proposal had the significant positive in its design in that the crew cabin would be very close to the ground.  In the Lunar Starship proposal, the habitable crew cabin would be far above the ground, and an elevator arrangement would be used to carry the crew back and forth from the surface.  In the Blue Origin ILV, the crew cabin would also be far above the ground (although not as high up as on the Lunar Starship), but the crew would need to climb down and up a 12 meter (39 foot) ladder in their lunar spacesuits to reach the surface.  That would be cumbersome and probably tiring despite the low lunar gravity (as the lunar spacesuits would be heavy and bulky), and possibly catastrophic should anyone slip and fall.

At the end of the stay, the Dynetics ALPACA would then fly the crew back to the NRHO orbit to link up with the Orion (or Lunar Gateway if it is there), where the crew and any returning cargo would be transferred for the return to Earth on the Orion.  The ALPACA could then wait there, and if to be used on a subsequent mission, could be refueled with two launches of a Vulcan Centaur carrying fuel to it as would have been done on the initial mission.

The designs and mission plans are therefore radically different in the three proposals.  The proposed costs, as well as the payload and other capacities of the resulting bids, also differed dramatically, as summarized here:

Lunar Starship

Integrated Landing Vehicle (ILV)

ALPACA

Lead Firm

SpaceX

Blue Origin

Dynetics

Other Primary Firms

none

Lockheed Martin Northrup Grumman Draper

Sierra_Nevada  Draper

Max Payload to Lunar Surface

100 tons

850 kg

negative

Habitable Volume

1,000 m3

12.4 m3

14.5 m3

Floor Area

325 m2

4.7 m2

5 m2

Max Crew Size

Many (well more than 4 if desired)

   2 at first; redesign    for 4

2 at first; later 4

NASA Development Award, April 2020

$139.6 m

$479.7 m

$239.7 m

Final Bid Price, April 2021

$2.94 b

$6.00 b

$9.08 b

Depending on the number of refueling flights of Starship to Earth orbit, the Lunar Starship could conceivably carry as much as 100 metric tons of cargo to the lunar surface.  This dwarfs the estimated 850 kg that the Blue Origin ILV could bring (which is close to, but slightly below, the minimum NASA specification that it should be able to bring 865 kg to the surface).  And as noted above, NASA does not believe the Dynetics ALPACA would be able to carry even itself to the surface without crashing into it.

The habitable volume of the Lunar Starship would also be enormous, at 1,000 cubic meters.  This is larger than the entire habitable volume of the International Space Station (which is 916 cubic meters), and the ISS had to be slowly assembled in space over a period of 13 years.  In sharp contrast, the habitable volume of the Blue Origin ILV would be just 12.4 cubic meters, which it says would be enough for a crew of two, but acknowledges would not be enough for a crew of four.  Hence the need for an extensive redesign of the Blue Origin ILV if it were to be used for subsequent Artemis missions when they would need to accommodate a crew of four.  And the Dynetics ALPACA would have a volume of 14.5 cubic meters – only a bit more than on the Blue Origin ILV – which it argues would be sufficient for a crew of four.  Keep in mind that the habitable space for the crew is not simply for their landing on the lunar surface, but for their stay there of about one week on the first Artemis mission with this increasing to a planned 30 days on the following Artemis mission and even more later.  Such accommodations would be tight, especially for a crew of four.

Resting on the lunar surface, where there is gravity and not simply the weightlessness of a vehicle in orbit, the floor area will also matter.  For the Lunar Starship, the floor area (on several different floor levels) would come to 325 square meters (3,500 square feet).  That would be the floor area of a very substantial house in the US.  In contrast, the floor area on the Blue Origin ILV would be just 4.7 square meters, and only slightly more at 5 square meters on the Dynetics ALPACA.  Five square meters for a crew of two for a week is not much, and especially not for a crew of four for a month or more on the lunar surface.

NASA announced in April 2020 it would provide funding to the three competitors for them to develop more concretely their proposals.  But the dollar amounts provided were not equal.  NASA was relatively quite generous with providing almost $480 million to the Blue Origin team, but just half this (almost $240 million) to Dynetics for its ALPACA proposal.  And to SpaceX the grant was just below $140 million.  One cannot argue that NASA was unfairly favoring SpaceX in its grants for the further development of the proposals.

Finally, the price SpaceX bid for the contract – at $2.94 billion – was less than half the $6.00 billion price Blue Origin offered, and less than one-third the price of $9.08 billion Dynetics said it would need.  The price includes the development of the vehicle, a test flight where it would go through the full mission sequence (including landing on and then returning from the lunar surface) but without a crew, and then the Artemis III mission itself with a crew of two.

Based on all this, it should not be difficult to see why NASA chose the SpaceX proposal.  Basic feasibility is of course key, and NASA certainly paid close attention to this.  It found, for example, that by its calculations, the Dynetics ALPACA would not be able to land on the Moon with any payload at all.

But after its detailed review of the SpaceX plans during the year it was funding the more concrete development of those plans, and with full access to all the SpaceX technical teams and the work they had done, NASA engineers concluded the SpaceX proposal was feasible.  As noted before, this in itself was a tremendous vote of confidence in the still-to-fly Starship.  And NASA also concluded that SpaceX would be able to deliver on these plans quite soon.  While I personally have strong doubts that this will be possible by 2024 (or even 2025), this time frame is of interest as it implies that NASA engineers believe the basic Starship system will be flying successfully very soon, i.e. by sometime in 2022.  It implies that they have concluded that the current Starship design is basically doable, and that a major redesign will not be necessary for it to be successful.

Despite the clear superiority of the SpaceX proposal over those of Blue Origin and Dynetics, the latter companies did not see it that way.  Soon after NASA’s decision was announced in April 2021, both Blue Origin and Dynetics appealed to the US Government Accountability Office (GAO), arguing that NASA had not properly followed government procurement procedures.  Such appeals are not uncommon, and sometimes lead to reversals.  But the GAO concluded, in a report issued on July 30, that NASA had followed the proper procedures and was justified in making the award to SpaceX.

While Dynetics accepted this and moved on, Blue Origin decided then to file a court case protesting NASA’s decision.  This was frustrating to many, as NASA could not grant the award to SpaceX nor work with SpaceX on the contract while an appeal was underway.  Both sides (NASA and Blue Origin) agreed, however, to expedited court procedures where the judge would make a decision (without a jury) and would do so by early November.  The judge’s decision was then announced on November 4, rejecting Blue Origin’s claims.  The full decision was released on November 18, after Blue Origin had been given the chance to make redactions of commercially confidential information in that judicial opinion.

4)  The Possibilities with Lunar Starship

The mission profiles discussed above follow what NASA has set out for its Artemis plans to return to the Moon.  That is, NASA set how the HLS chosen would be used, and the missions are based on use of an Orion to take the astronauts to the high NRHO lunar orbit and an SLS to launch them there (along with, possibly, use of the Lunar Gateway in the NRHO orbit to serve as a staging area).  The Lunar Starship would then take the crew from the lunar orbit to the surface of the Moon and back.  The Starship would have been launched into Earth orbit, refueled there, then flown to the NRHO orbit, and following the delivery of the crew back at the NRHO would be flown back to Earth orbit for refueling after which it could be used again for the next lunar mission.

But if a Lunar Starship can do all this, there is the obvious question of why one needs the Orion, the SLS, and the Lunar Gateway, at all.  The Lunar Starship will have all the life support systems, seats, and other equipment required to carry the astronauts to a lunar landing and then to support them there for extended periods (one week on the first mission, a month on the second, and more later).  It will also have lots of space and an ability to carry a cargo load far in excess (more than 100 times as large) of what NASA set in its minimum requirements.  Forcing the four astronauts to squeeze into the Orion capsule (with a habitable volume of just 9 cubic meters) for the multi-day flight to lunar orbit, while the Lunar Starship (with a habitable volume of 1,000 cubic meters) would fly there empty, seems absurd.  It would be similar to forcing a group of four to squeeze into a Volkswagon Beetle for a drive across the US, while a luxury bus drove the same route, but empty, only to be used for a final short segment at the end of the journey.

I fully acknowledge that there will be technical issues to work through.  Fuel loads would have to be calculated and would need to suffice.  But the flexibility in the Starship system, where additional in-orbit refuelings could be provided if there is a need, combined with a scaling back, if need be, of the payload to be lifted (a 100-ton capacity means there is a lot of room for adjustment), means that if a Lunar Starship in the NASA HLS role is feasible, then it should certainly be for this expanded role as well.

The mission plan would then be that once there is a fully fueled Lunar Starship in Earth orbit (with whatever number of refueling flights to the Fuel Depot Starship that might require, with the Lunar Starship then fueled from what has been stored in the Fuel Depot Starship), the NASA crew of four (or even a substantially higher number, as there would be plenty of room) would be flown to transfer to the vehicle while it is still in Earth orbit.  They could be flown there on a regular Orbital Starship flight or even on one or more flights of a Falcon 9 with the now well-tested Crew Dragon capsule.  They would then be flown on the Lunar Starship directly to the lunar surface, or possibly first to a lunar orbit and then to the surface.

Furthermore, given its vast size the Lunar Starship could in effect be an instant base on the Moon.  There would be no need to bring to the Moon via a series of separate flights all the components that would otherwise be needed to build such a base.  NASA’s long-term Artemis plans had envisaged this following the first two Artemis landing missions (at least in the plans set before Lunar Starship was chosen).  A permanent, habitable, structure (“Artemis Base Camp”) would have been built on the lunar surface to house the crews, with this slowly expanded in size as additions are brought on subsequent missions.  With the available space in the Lunar Starship, there would be no need.  And the Lunar Starship would also have the space needed, as well as the cargo capacity, for the lunar rovers that NASA has planned.

The Lunar Starship as a lunar base would also have the rather unique capability of being packed up, lifting off, and then flown to a new location, if there is any desire to do this.  There might be value in exploring various sites around the Moon – in a search for frozen waters supplies, for example.  One would just have to ensure that adequate reserves of fuel were brought along.  And if, as some believe likely, significant amounts of frozen carbon dioxide along with frozen water are found on the Moon (most likely near the South Pole, which is why this has been chosen for the early missions and the planned Artemis Base Camp), then it would be possible to produce the liquid methane fuel the Starship uses along with the liquid oxygen, and refuel the Lunar Starship while it is there.

If there is a need for additional cargo, one could also design a version of the Lunar Starship to carry cargo only.  It would fly there on its own (which is now a standard technology – indeed under the NASA HLS contract the Lunar Starship would be required to complete an entire unmanned flight as it would under Artemix III, including landing on the Moon, as a test of the entire system before any crew is put aboard).  If it was not then to be returned to the Earth and thus not need the fuel to do so, but rather remain on the surface as a permanent addition to a growing lunar base, such a cargo flight could bring a load of as much as 200 tons.  This is huge.  The weight (mass really) of the entire International Space Station is, after all the additions over the years, now 420 tons.

And the possibilities the Starship system would open up are not just limited to the Moon.  Very recently, a group of prominent planetary scientists issued a White Paper arguing that with the availability of Starship, the traditional approach taken to planetary exploration missions should be radically re-thought.  With the capacity of the Starship, along with its low cost, planetary missions could become not simply far more frequent but also simplified in design.  Rather than spend a good deal of money and a good deal of time in refining the spacecraft to reduce its weight by a few ounces or its dimensions to fit into the limited capacity of the existing rockets that are used, one could simply use off-the-shelf equipment when Starship is used to send it on its way.  Finally, there is of course the missions to Mars for which the Starship system has been designed.  But all this goes beyond what I intended to cover in this already long post.

Finally, flying the lunar missions on Starship rather than with the Orion, SLS, and associated systems would be far cheaper.  As an extremely rough calculation:  Assume that each Starship launch will cost $20 million (ten times the $2 million Elon Musk has said it will ultimately cost) and that 10 Starship launches would be required for each mission (for refueling in orbit – note that the Lunar Starship and the Fuel Depot Starship would remain in space following their initial launch and would be reused), for a total cost of $200 million for the launches.  For simplicity, assume all the other costs of the launches and then for carrying out the mission would be similar and hence the total cost would be $400 million to carry out the mission.  While crude, this estimate almost certainly errs on the high side.  The actual cost is likely to be lower, but take $400 million for the purposes here.

In contrast, the Office of the Inspector General (OIG) of NASA in a recent (November 15, 2021) report assessing NASA’s Management of the Artemis Missions estimated that the cost of a single SLS launch with an Orion capsule for an Artemis mission would come to $4.1 billion.  The SLS itself (none of which is reusable – all is thrown away in each launch) would cost $2.2 billion.  This is similar to, but 10% more than, the estimated $2.0 billion cost per launch for the SLS made by the White House Office of Management and Budget (OMB) and released in a letter to Congress in 2019.

The NASA OIG estimated that the cost of an Orion crew vehicle, including the cost of the Service Module being built by the European Space Agency, would total $1.3 billion, of which $300 million would be for the Service Module.  In my earlier blog post, I used the contracted amount NASA would pay for the third through the fifth Orion vehicles ($900 million each), but the earlier ones would be higher and the $1.0 billion estimate of the OIG is consistent.  Interestingly, I had thought in my earlier blog post that the European Space Agency would be covering the cost of the $300 million Service Module they are making for the Orion, but the OIG report clarifies that this will not really be the case.  While the Europeans will be paying the contractor that will build the module, NASA will then not charge the Europeans $300 million (per Orion) for costs incurred on behalf of the Europeans at the International Space Station.  That is, this is a barter agreement, so NASA is in fact paying that $300 million cost, but indirectly.  By including it in the ISS budget, where it is not broken out in what is made publicly available, one cannot determine the true cost of the complete Orion vehicle without access to unpublished information.

Finally, the NASA OIG also includes in the cost of an SLS launch the pro-rated share of what it costs to maintain the ground facilities (the launch pad, etc.) needed for an SLS launch.  Since there will be only (at most) one SLS launch per year, this will be the annual budget for maintaining and then using (once) that capacity, which the NASA OIG estimates to be $568 million per launch.

The NASA OIG therefore estimates the total cost per launch of the SLS with Orion will be $4.1 billion.  This would be ten times the generously estimated cost of doing the same via Starship, and for far less capacity.

D.  Politics

Why then keep funding the SLS and Orion?  In the near term, it might make sense.  After all, despite all that has already been done to test and develop Starship, the testing is not yet complete and orbital testing is still to start.  And while NASA technical staff concluded it should ultimately work, it might not.  The key tests will be in the next year, however, and we should know by the end of 2022 (or before) whether Starship will work as planned.  There will certainly be failures at first in the upcoming orbital tests.  As discussed above, this is to be expected in the iterative development process used by SpaceX.  But with iterative improvements, with solutions found to the problems uncovered, it will eventually work.  We will only know that for sure, however, when it does.

Whether Starship will work should be known, however, by no later than the end of 2022.  Provided Starship has demonstrated that it will indeed work, should one then expect to see NASA (with Congress) decide to close down the SLS and Orion funding and switch over to the far more capable and far less costly Starship system?  The answer to that is:  not likely.

First of all, it will be politically embarrassing to admit that it was a mistake.  By the end of FY23, over $54 billion will have been spent ($32 billion for the SLS and $22 billion for the Orion).  The fundamental error came when Congress in 2010 forced the Obama administration to start development immediately of a heavy-lift launcher that became the SLS.  The Obama administration had recommended that NASA should first examine and test certain new technologies (in particular in-orbit refueling ability), with the results then used to determine how best then to design such a launcher.  Congress, and in particular the Senate, insisted (and wrote into law in the 2010 NASA authorizing legislation) that the heavy-lift launcher be designed immediately and furthermore that it make use of the key components of the Space Shuttle (e.g. the engines) in that design.  By mandating this, those in their states and districts who had been employed building such components for the Space Shuttle would remain employed.  But with such requirements written into law, some wags have concluded the SLS should not stand for Space Launch System but rather Senate Launch System.

Republican Senator Shelby of Alabama (home of the NASA Marshall Space Flight Center in Huntsville – the center with the primary responsibility for the SLS) was a key advocate, as was Republican Senator Kay Bailey Hutchison of Texas (home of the NASA Johnson Space Center in Houston – responsible for manned space flight operations).  Senators from Mississippi (home of the Stennis Space Center, where large rocket engines are tested), Louisiana (home of the Michoud Assembly Facility), and Utah (home of the company that builds the solid rocket boosters that are strapped on to the first stage) were also important advocates.  As was one Democratic Senator – Bill Nelson of Florida (home of the Kennedy Space Center), who was the chair of the subcommittee that drafted the NASA Authorization Act of 2010 when Democrats controlled the Senate.

Former Senator Bill Nelson is now NASA Administrator.  Canceling SLS and Orion would be an embarrassment for him, even if Starship demonstrates its full ability to carry out the nation’s lunar exploration plans – and do so far more effectively and at far lower cost.  As NASA Administrator he has already signaled that NASA intends to continue to use the SLS, and for possibly 30 years or more.  In October, NASA issued a “Request for Information” to the aerospace industry, requesting proposals from firms willing to produce, operate, and maintain the SLS rocket system until the 2050s, with the stated intention that this should reduce the “baseline per flight cost” of an SLS launch by “50% or more”.  The NASA Request for Information did not say, however, what the current baseline cost was from which the 50% would be taken, and refused to respond to reporter queries on what NASA estimates that cost to be.  But as noted above, the White House Office of Management and Budget indicated in a letter to Congress in 2019 that the cost was $2.0 billion per launch.  A more recent estimate from the NASA OIG puts that cost at $2.2 billion.

But NASA itself has consistently refused to reveal what the per launch costs of the SLS will be.  This is despite repeated recommendations over the years by oversight bodies that it should.  For example, the Government Accountability Office (GAO), which is formally part of the legislative branch reporting to Congress, recommended NASA provide such an accounting in a 2017 report, when it was already clear that the SLS would be delayed and well over the original planned cost.  More recently, the NASA OIG report of November 15, assessing the management of the Artemis program, recommended that NASA management provide such figures.  But NASA management has continued to refuse.  The NASA OIG report made nine recommendations, and NASA management accepted seven (two partially).  But it rejected the two recommendations on the issue of being transparent on costs.

Congress has, however, not forced NASA to reveal those costs.  While the overall budget of NASA is known year by year, as well as the budgets of major sub-aggregates of certain individual programs as NASA has organized them, one cannot work out from this the overall cost of a major program such as Artemis (with responsibilities for portions of the program spread across much of the agency) nor work out a separation between the cost of developing a new vehicle or system such as the SLS and the cost then of using that system on individual missions.  There is a lack of transparency, although entities such as the GAO or the NASA OIG (as well as journalists and other outsiders) may try to come up with estimates.

Senators and Members of Congress may be quite comfortable, however, with this lack of transparency.  It then does not draw attention to the overall costs (and how high those costs might have grown to) even though they still wish to see (and to publicize) how much is being spent in their particular districts.  And NASA is glad to provide this.  For examples, see here, here, here, here, here, and here.  Spending more is seen not as a flaw but as a feature.

The SpaceX Starship, if it works, may have a cost of just $2 million per launch.  Even if the cost turns out to be ten times that, it will still be dramatically less costly than an SLS alone, and even more so of an SLS plus Orion system.  Furthermore, and importantly, it would be far more capable.  It would enable far more to be achieved in meeting the stated national objectives of the space exploration program than would be possible with the SLS plus Orion system.  It would be quite revolutionary.  But it will only be used for this purpose if the American political system allows it.

There Have Been Real Consequences From Not Taking Covid Seriously

A.  Introduction

Earlier posts on this blog have documented that vaccination rates against Covid-19 have been systematically lower in accordance with the share of a state’s vote for Trump in the 2020 election, and that mask-wearing to protect the individuals and those around them have also been systematically lower.  The higher the share voting for Trump in a state, the lower the share vaccinated and the lower the share wearing masks.

Those choices have had consequences.  As shown in the charts above, it should not then be surprising that states with a higher share of their vote for Trump have seen, on average, a higher number of cases of Covid-19 (per 10,000 of population) as well as a higher number of deaths.  The relationship is statistically a very strong one.  While many factors affect the likelihood of being infected with Covid-19 and of dying from it (including factors such as urban density, extent of travel, health status of the population, adequacy of the health care system, and more), political identification by itself appears to be a strong and independent factor.

In what is literally a life and death issue, one would have thought that rational self-interest would have dominated.  It has not.  Following a review of the data, this post will discuss some possible reasons why.

B.  The Relationship Between the Incidence of Covid-19 Cases and Deaths and the Share Voting for Trump

The figures at the top of this post plot the relationship between the number of cases of Covid-19 in a state (per 10,000 of population), or the number of deaths (also per 10,000), and the share in the state who voted for Trump in 2020.  The Covid data come from the CDC.  It was downloaded October 26, but since case and death counts from the states may not be fully reported to the CDC for up to a week to ten days, I used October 15 as the end date for the analysis here.  “Cases” are confirmed cases, and “deaths” are deaths as a consequence of Covid-19, both as defined in the CDC guidance for how these should be recorded.

For the start date I used July 1, 2020.  This came at the end of the first wave of Covid-19 cases and deaths.  Cases and deaths in this first wave were excluded for two reasons.  One is that the first wave arrived suddenly in mid-March and with an intensity that surprised many.  The nation was unprepared, with little done to prepare for the disease that was spreading around the world as Trump was claiming it was all under control, that it was “going to disappear”, and that it would soon “go away”.  Also, the CDC had bungled the initial testing (where testing was more readily accessible in parts of Africa than in the US in the key initial months), so the full extent of the developing problem was not clear until it hit.  The response, and the then only possible response, was to quickly institute lockdowns, and this was soon done in all 50 states.  The lockdowns were effective, albeit costly, and by late April the approach had succeeded in starting to bring down the daily number of new cases.  Case numbers continued to fall in May and into June.

But starting in early May, disparate decisions were taken across the different states on how fast to lift the lockdown measures.  Some opened up early and with little guidance on or advocacy for the wearing of masks, while others opened up more cautiously.  But with the opening up, and the refusal by a significant share of the population to wear masks and to follow social distancing recommendations, the daily number of new cases stopped falling and by around mid-June began to rise again.  The daily number of deaths followed a similar pattern but with a lag of about two weeks, and so began to rise around the end of June. Thus July 1 can be taken as a turning point – the end of the first wave and the start of the second.  While differences across the states had already started to develop from early May (when decisions were taken on how rapidly to open up), the consequences of the varying approaches only became clear as the second wave started to build.  On average across the nation, this was around July 1.

The second reason to exclude this first wave is that the quality of the data for that initial period was poor.  The Trump administration was slow in launching and then ramping up testing, with testing limited even well into April to those who showed obvious symptoms or who had been in close contact with someone with a confirmed case.  Thus many cases were missed.  While testing has been far from perfect throughout this pandemic, it was much worse in the earlier months than it was later.  For this reason as well, excluding the estimates from the earlier months will provide a better measure of how successful or not the different states were as they responded to the pandemic in their different ways after the initial lockdowns.

Excluding the first wave leads to the exclusion of 6% of confirmed cases and 18% of deaths from the overall totals as of October 15, 2021.  Most thus remained.  Note also the disparity in these figures.  That the official figures recorded that just 6% of the confirmed cases in the US (as of October 15) were in this initial, first wave, period, while this same period recorded 18% of deaths, strongly suggests that cases were significantly undercounted in that first wave.

The charts then show the incidence of total confirmed cases of Covid-19, or deaths from it, per 10,000 of population, over the period from July 1, 2020, to October 15, 2021, with this plotted against the share of the vote that Trump received in that state in 2020.  The relationship is a strong one:  The higher the share of the state vote for Trump, the higher the incidence of Covid-19 cases and of deaths.  Taking averages, the average number of confirmed cases over this period per 10,000 in the states won by Trump was 1,461 (i.e. 14.6% of their population) vs. 1,113 in the states won by Biden.  That is, there were on average 31% more cases in the states won by Trump.  The number of deaths from Covid-19 came to 21.2 per 10,000 in the states won by Trump vs. 15.3 in the states won by Biden, or 38% more in the states won by Trump.

But averaging across all the states won by Trump or by Biden is not terribly meaningful as there will be a mix of voters in every state.  Furthermore, there were a number of states where the vote was close to 50/50.

It is thus more meaningful to examine the trend across the different states, as a function of the share voting for Trump.  This trend is provided in the regression line shown in each chart, where simple, linear, ordinary least squares regression was used.  The statistical relationship found was very strong, and especially so for the regression for the number of cases of Covid-19.  The R-squared (a measure of how much of the variation in the values is accounted for by the regression line alone) was extremely high for such a cross-state sample as here – at 0.63 for the number of Covid-19 cases and a still high 0.36 for the number of Covid-19 deaths.  (R-squared values can vary between 1.0, in which case the regression line explains 100% of the variation across states, and 0.0, in which case the regression line explains none of the variation.)

The higher correlation (the higher R-squared) observed in the relationship for the number of cases than in the relationship for the number of deaths is what one would expect.  To die from the disease, one must first have caught it.  Hence this will depend on the number of cases in the state.  But deaths from it will then depend on additional factors such as the age structure of the population, general health conditions (obesity rates, for example), as well as the availability and quality of health care services (hospitals, for example).  These factors will vary by state, and hence add additional variation to that found for the number of confirmed cases.

The slope of the regression line is an estimate of how many additional cases of (or deaths from) Covid-19 to expect (per 10,000) for each 1% point higher share of the vote for Trump.  For each additional 1% point in the share of the vote for Trump in a state, there were on average 23.8 more cases (per 10,000 of population) of Covid-19 during the period examined, and on average 0.36 more deaths (per 10,000).  The t-statistics for these slope coefficients were both extremely high, at 9.1 for the number of cases and 5.2 for the number of deaths.  A t-statistic of 2.0 or higher is generally taken to be an indicator that the relationship found is statistically significant (as it implies that in 95% of the cases, the slope is something different from zero – a slope of zero would imply no relationship).  A t-statistic of 3.5 would raise that significance to 99.9%.  The t-statistics here of 9.1 and 5.2 are both far above even that mark.

One can also use the regression lines to address the question of what the impact would have been on Covid-19 cases and deaths if everyone behaved as Biden voters did (or as Trump voters did).  The regression lines look at how the incidence of cases or deaths change based on each additional percentage point in the vote for Trump.  If one extrapolates this to the extreme case of zero votes for Trump (and hence a “pure” Biden vote), one can estimate what cases and deaths would have been if all behaved as Biden voters did.

This is a straight line, i.e. linear, extrapolation of the effects, and the limitations from this assumption will be addressed in a moment.  But using linearity, the effects are easily calculated by simply inserting zeroes for the Trump share of the vote into the regression equations, so that one is left with the constants of +96.94 for the number of cases (per 10,000 of population) and -0.69 for the number of deaths.  That is, there would have been a predicted 97 (per 10,000) cases of Covid-19 over this period in the US rather than the actual figure of 1,261 (per 10,000).  This is 92% lower.  And the number of deaths would have been essentially zero (and indeed would have reached zero with still some share voting for Trump – based on the regression equation coefficients it would have been at the 2% point share for the Trump vote).

Are these results plausible?  Would cases and deaths have fallen by so much if all of the population had behaved (in terms of wearing masks, social distancing, getting vaccinated once vaccines became available, and other such behaviors) as the Biden voters did?  The answer is yes.  Indeed, the linear extrapolation is conservative, as infectious diseases such as Covid-19 spread exponentially.

If in some state each infected person infects, on average, two further people, the number infected will double in each time period for the disease.  This is exponential growth, with a reproduction rate of two in this example – a doubling in each period.  For Covid-19, the time period from when a person is infected to when that person may, on average, spread it to another, is a week and a half.  A person becomes infectious (can spread it to others) about one week after they became infected with the disease, and then can infect others for about a week (with the average then at the half-way point of that week).  Thus 100 cases of active infections in some region would double to 200 in that time period of a week and a half, then to 400 in the next time period, and so on.  If, in contrast, responsible behavior (such as vaccinations and mask-wearing) reduces the reproduction rate to one-half rather than two, then 100 cases will lead to 50 in the next time period, to 25 in the next, and so on down to zero.

In any given state there is a mix of Biden voters and Trump voters.  While there are many factors that matter, if these two identities reflect, on average, differing shares of people that do or do not choose to be vaccinated, wear masks, and so on, then the average reproduction rate will vary depending on the relative shares of such voters.  That average reproduction rate will be lower in states with a higher share of Biden voters, and for a sufficiently high share of Biden voters (a sufficiently low share of Trump voters), there will be an exponential decline in new infections from Covid-19.  The linear extrapolation based on the regression equations would thus be a conservative estimate of the number of cases to expect when most of the population behaves as the Biden voters have.

There are, of course, many factors that enter into whether a person is infected by someone with Covid-19, and whether they then die from the infection they got from someone.  But the charts and the regression results suggest that the share of the population in a state voting for Biden or for Trump is, by itself, strongly correlated with how likely that was.  Why?

C.  Personal Behavior and Political Identity

The fact of, and then the consequences from, this political divide for infection by Covid should not be a surprise to anyone.  As noted before, Trump voters are far less likely to be vaccinated or to wear masks to protect themselves and others from this highly infectious, and deadly, disease.  This then translates into higher infection rates, and the higher infection rates then to higher deaths.

One sees this unwillingness to be vaccinated also in surveys.  The most recent of the regular surveys by the Kaiser Family Foundation (published on October 28) found that 90% of Democrats had received at least the first dose of the Covid vaccine, while only 61% of Republicans had.  Furthermore, 31% of Republicans declared they would “definitely not” be vaccinated, while just 2% of Democrats held that view.  Gallup surveys have found similar results, with a survey from mid-September finding that 92% of Democrats had received at least the first dose of the Covid vaccine, but that only 56% of Republicans had.  And 40% of Republicans in that survey said they are not planning on being vaccinated ever, while only 3% of Democrats said that.

Not surprisingly, one then sees this reflected in state politics.  Republican governors (such as Abbott of Texas and DeSantis of Florida) have gone so far as to issue executive orders to block private companies from protecting their staff and their customers from this disease, and even to prohibit local school boards from taking measures to protect schoolchildren.

The direct result is that the virus that causes Covid-19 has continued to spread.  An infectious disease such as Covid-19 will only persist as long as it is being spread on to others.  It cannot survive on its own.  The issue, then, is not just that someone refusing to wear a mask or to be vaccinated is highly likely to catch the disease, but that that person is likely to spread it to others.  While Republican governors such as Abbott and DeSantis have said this is a matter of “personal freedom”, it is not that at all.  No one is free to do harm to others.  It is the same reason why there are laws against drunk driving.  Drunk drivers are more likely to cause crashes (not all of the time, but often), and those crashes will harm others, up to and including killing others.  Spreading Covid-19 is similar, up to and including that those who become infected may die from it.

For whatever causal reason, the facts themselves are clear.  But why has a significant share of the population chosen to behave this way?  This is now more speculative, and goes into an area that I openly acknowledge is not my area of expertise.  With that proviso, some speculation.

It is clear that political identity has played a central role, where Trump from the start treated the then developing pandemic as an issue where you were either with him – and his assertion that he had it all under control – or against him.  This started with Trump’s assertion in an interview on January 22, 2020 (from Davos, Switzerland) that he had no worries, that “we have it totally under control”, and that “It’s going to be just fine”.  This claim continued through February (as cases were growing in the US), where on February 27 he said “It’s going to disappear.  One day it’s like a miracle.  It will disappear.”  And in campaign rallies in February, he claimed to his cheering supporters that he had been doing a superb job in stopping the virus and that any charge to the contrary was simply a “hoax” coming from the Democrats.

Thus, from the start, Trump made the issue a political one.  If you were a true supporter of Trump you could not treat the disease as something of concern – Trump had taken care of it.  Any assertion that the developing pandemic was in fact serious, and needed to be addressed, was a “hoax” perpetrated by the Democrats.

Trump then continued to assert all would soon be well, saying on March 10 that “it will go away”, on April 29 that “This is going away.  It’s gonna go.  It’s gonna leave.  It’s gonna be gone.”, on May 11 that “we have prevailed”, on June 17 that “It’s fading away.”, and on July 19 that “It’s going to disappear”.  But more than 600,000 Americans have died since July 19, 2020, not far short of the 651,000 Americans who have died in battle in all of America’s wars since 1775.  From the start of the pandemic, more than 750,000 Americans have now died.

Trump’s politicization of Covid-19 was then amplified when, at the April 3 press conference in which he announced the CDC recommendation that everyone should wear face masks when going out, he immediately then added that he would not himself wear a face mask.  Face masks are highly effective in hindering the spread from person to person of the virus that causes Covid-19, and until vaccines became available, were the best way to hinder that spread.  But wearing a face mask is also highly visible.  For those who saw themselves as supporters of Trump, and believed what he said (that the virus was going away, that he had it under control, and that any concerns over this were merely a hoax promoted by the Democrats), then it was not surprising that many would follow Trump’s highly public example and not wear a mask either.  Some even went so far as to shoot, and kill, store personnel when told they should wear a face mask inside some store.

It is not surprising that such views would then carry over to vaccination.  Having rationalized not wearing a mask, it is easy to rationalize a refusal to be vaccinated.  And rationalizations could easily be found just by watching Fox News.  In the six months from April through September this year, for example, Fox News chose to air a claim undermining vaccination on all but two of those more than 180 broadcast days.  Many were also exposed to claims that can only be described as truly bizarre, such as that the vaccination will be secretly inserting a microchip into your body for the government to track you, with Bill Gates behind it all; or that it will make you magnetic with this managed through 5G telecom towers; or that it will re-write your body’s DNA; and more.

One can therefore easily come up with rationalizations not to be vaccinated, of varying degrees of plausibility, if you are predisposed against it.  But many of those providing such rationalizations must have realized that their rationalizations often did not make much sense.  Rather, their decisions appear to have been driven more by a visceral or emotional reaction (vaccinations just “feel” wrong) than as an outcome of a rational process.  That is, the decision not to be vaccinated was made first, based on emotions or feelings, with the rationalizations then arrived at later to justify a decision that had already been made.  (Such a process is in accord with the “social intuitionist” model of Jonathan Haidt, where decisions are made first, in a visceral reaction based on emotion, while rationalizations then come later to justify that decision.)

In the case of Covid-19, those decisions on vaccination (and earlier on wearing masks) were made in accordance with political identity – a perceived loyalty to Trump – rather than in recognition of the very real risks that would follow if one contracted Covid-19.  Wearing a mask or accepting a vaccination would simply be “wrong” and disloyal.

I have found it astonishing how strong this emotional reaction has apparently become.  Covid-19 is new (it did not even exist just two years ago), it is deadly (where on average about 1.5% of those infected have died – with a much higher fatality rate than this average for those who are older or who have other health issues), and may have serious long-term ill effects even for those who do not die from it.  Yet this visceral reaction appears to have been so powerful that many supporters of Trump still refuse to be vaccinated, despite the risk of genuine life and death consequences.

I should hasten to add that not all voters for Trump have refused to be vaccinated.  Indeed, according to the surveys, about 60% (a majority) have as of October.  There are also highly vocal partisans on the left who have refused to be vaccinated.  Their reasons are likely very different from that of the typical Trump voter, but the underlying cause appears still to be intuitive – the feeling that such vaccinations are simply “wrong”.  But the issue is that the relative shares of the two groups have been very different:  A far higher share of those who voted for Trump have refused vaccination than is the case for those who voted for Biden.  The consequences are as shown in the charts at the top of this post.

As noted before, the cause for this relationship cannot be known with certainty, and what I have presented here should be viewed as speculative on my part.  There may well be other explanations.  For example, a related but somewhat different explanation would be that a common third factor explains both the tendency of some to vote for Trump and also to be resistant to vaccinations.  Those in this group may put faith in conspiracy theories (including, but not limited to, terrible consequences from being vaccinated), distrust authority, proudly but stubbornly insist on doing the opposite of whatever is recommended, and for such reasons not only refuse to be vaccinated but also vote for Trump.

Whatever the explanation, the results have been tragic.  This has also been a lesson in how strongly some will keep to a held position, even as they have seen prominent figures, and sometimes friends or even family members, come down with this disease.  When an issue becomes one of identity, it appears that even with such tragic consequences there will be many who steadfastly refuse to change.