Pages

Monday, July 11, 2022

post pandemic lottery

A number of industries have been dramatically impacted over the past 30 months because of the pandemic.  People who were originally strapped for cash when the pandemic began, were later flush with some of the extra money governments had slushing around the system.  Over time, markets of all types rose.  There was a froth of speculation in new-age market products and “investments” [eg SPACs, NFTs, cryptocurrency, non-profitable tech].  Some of these fever of course has since regressed, and too has the implosion of wealth that was leveraged by this corner of the investment community.

The lottery however didn’t sustain new highs.  In fact it lost ground, as lottery revenue appears to have stalled and by some measures is somewhat lower than they were pre-pandemic.  On both multi-state games of Mega Millions and Powerball.  To the degree lottery sales do not continue to grow [real, let alone nominal] with the rest of the economy, state education budgets relying on such income would become stressed.

In this article we explore the concept of lottery engineering, via more complicated winning probabilities, and other marketing.  And how the lottery, since the pandemic, has become an even worse value for those who play.  Last, how lottery odds compare to other speculations, such as casinos and sports, and new digital investments described above.

topic 1: growth and odds
We suspected in the previous article that the engineered worse winning odds was leading to greater excitement at chaotically larger jackpots. Resulting in a higher general flow of lottery revenue.  We now have 7 data points to look at, concerning the most recent game changes.

In the summary below, we show the change in winning probability, as well as the change in Mega Millions and Powerball lottery revenue per draw [and total lottery revenue in brackets].  In both cases, for the two years after the change versus the two years prior to the change:

January 2009 Powerball [+4 white balls, -3 red balls]
probability 1:195m (33% harder)
$40m v $52m [$54b v $56b]

January 2012 Powerball [unch white balls, -4 red balls]
probability 1:175m (10% easier)
$56m v $85m [$60b v $70b]

October 2013 Mega Millions [+19 white balls, -31 red balls]
probability 1:258m (47% harder)
$38m vs $50m [$68b v $71b]

January 2014 Powerball [unch white balls, unch red balls]
probability 1:175m (0% harder)
$85m v $85m [$69b v $71b]

October 2015 Powerball [+10 white balls, -9 red balls]
probability 1:292m (66% harder)
$64m v $95m [$71b v $80b]

October 2017 Mega Millions [-5 white balls, +10 red balls]
probability 1:303m (17% harder)
$38m vs $85m [$80b v $90b]

It is plain to see in the 7 examples above that there was never lower revenue subsequent to each iteration of game changes.  The lottery economy just rose.  Of course the general direction of the game complexity was mostly in the direction of complexity as well, and it's always easy for leaders to conflate the two.

Also, as a general matter, looking and counting the frequency of certain large jackpot [eg one billion dollars] sizes isn’t useful for a statistical exercise.  After all, we know what the general direction to be, but there is ergodicity in this system and its ultimate magnitude and specific timing could be random.

topic 2: enter covid-19
Prior to the pandemic, the lottery was able to employ marketing tricks in order to stimulate demand. Sometimes the buzz worked.  For example engineering more complicated winning odds, which has quietly led to ever increasing jackpots.  Along with the associated excitement of a "record" jackpot"!

But this system hasn’t changed in a while.  For exmaple, the game hasn’t changed its probabilistic chances [either Powerball or Mega Millions] in nealry five years.  We’ve not had a new state [ie #46] enter the national system since 2019.

It's worth noting that our previous article was widely cited [front page New York Times, Alabama Review, Washington Post] and interview [televisión española] and was used by senators to scuttle Alabama’s most recent vote in 2016 in join lottery.

Then coronavirus hit. The world forever changed. Behaviors changed.  Spending online rose.  Gambling online rose.  But physical spending on lottery didn’t rise.  What could explain it?

People stayed in, and turned in a different direction.  They haven’t returned to lottery.  The tradition of going to the local supermarket to buy a lottery ticket, let alone food, was discouraged in the pandemic depths of 2020.  Lottery revenue suddenly fell -3% in 2020 versus 2019: a slightly larger drop than during the global financial crisis.

As an instant reaction to this unexpected new world, lottery officials had to change their game.  But given their fear over the existential lottery threat, they didn’t opt this time to reverse their previous engineering of making the game more difficult.

The result was less revenue flow and a more difficult game, combined to make a severe business challenge for the lottery.  Making matters worse, they cannibalized their own market somewhat.  They instead adjusted the minimum jackpot for winning to something lower.  And permanently blamed coronavirus and “low interest rates” for this decision.  

Lottery payouts however don’t seem to have changed, yet [annual data comes with a lag].  An early indication that they will however will be in the coring out of the middle-prizes.

This is an illustration of median Mega Millions and Powerball jackpot changes over 5-quarter increments:



We used quartile charts here to illustrate the impact of random jackpot surges within a small sample size.  For example, we saw a $1.7 billion dollar combined jackpots in January 2021.  The next largest combined jackpots from 2019 onwards were in 2019 and earlier this year.  And both of them were near the high-$800 million range [just half of the $1.7 billion from early 2021!]

And peeling back one level down in jackpot sizes, we see in the 2.5 years pre-pandemic [which equal the first two periods of the four-period chart above] versus the latter 2.5 years post-pandemic:

Mega Millions went down from 7 -to 6- jackpots of at least $300m
Powerball went down from 7 -to 5- jackpots of at least $300m

Also bear in mind this trend towards smaller and fewer jackpots mildly overlaps with the tail-effect of the late-2017 Mega Millions engineering we discussed in the previous section. Is it likely we don't see another billion dollar jackpot for some time?

Meanwhile we have education budgets that are statistically misrepresenting [and likely fooling themselves] that everything is back to normal. The presentation of lottery revenue data shows things are relatively back only by looking at the statistics with the inclusion of the early-2021 jackpot outlier. 

If one instead assumes steady state is the vast amount of known median data [as we have done above], then we see that there has been no rebound and the post-pandemic data above [eg past 12 months] is lower than the pre-pandemic data [eg 2019].

By this point the economy has already recovered briskly. Stay-home and mask restrictions long lifted.  So too have been interest rates.  And yet the lottery still has not been able to recover their revenue as easily.  

And at these ssemingly crucial crossroads, the lottery continues to weaken under the surface.  Perhaps materially.  One example of detrimental changes was the lack of reversal to the Covid-19 era decision to cut minimums based on interest rates.

It’s difficult to see the actuarial math now working back in its favor given other market headwinds.  And as a result, the consumer's ticket value is perpetually weaker.

topic 3: the 1% actuarial change
Do small changes in the long-term rates actually matter?  Let's use a simple example of the 30-year interest rate i, which fell from 2.4% pre-coronavirus, to near 1.4% months later at the 2020 lows.

If one was “promised” a 30-year annuity-due lottery prize for example [in lieu of an equally-valued lump sum payment] of $500m, the actuarial model is straightforward [assuming no sudden deaths, credit issues, etc] at i=2.4%:


Where v=1/[1+i]=0.977, and d=1-v=0.023.  Or a X=500m/21.8, and an annual annuity of $23 million [versus $24 million if receiving a typical deferred annuity].

The issue here is continuing to offer a $23 million annuity when i falls to 1.4%, implies an impossibly larger jackpot size [and associated proportional revenue increase] of $570m in order to offset this collapse in government rate.

This of course wasn’t an issue previously as rates have fallen for many decades, including a 2019 1% drop in i from 3.4%, to the pre-coronavirus 2.4%.  The reason this wasn't a business isue before is, as we noted, the lottery revenue was generally increasing during this falling rate regime and could still sustain the prizes.

What the lottery did to now mask this issue for players is to simply [and quickly] lower the minimum jackpot sizes blaming it on i.  But now, interest rates have since increased not only from those 1.4% lows in the 2020 pandemic depths, but back through the 2.4% pre-coronavirus level, and all the way towards 3.4%.  

So a 2% interest range change!  And meanwhile the lottery has simply unbudged themselves at their reduced minimums.  Evidencing stress in maintaining the pre-coronavirus revenue levels to sustain the high hundreds of millions [or even a billion] dollar jackpot they've gotten the public accustomed to.

We have a lower value ticket now as a part of a stymied system.  Offering only a slower -and less exciting- runway to achieving any sort of medicore jackpot, assuming one eventually wins.

topic 4: risks in digital growth
At these crossroads, the lottery has had to be inventive in forging new growth avenues and opportunistically capitalizing on trends in innovation.  Unfortunately given its size, these other sources of income are not enough to make-up for the loss of the core business model.

As noted, only 1 new state has been added to the national system since our last article.  And the remaining states are mostly modest in their economic size.  Mobile operators such as the early pre-acquisition start-ups have sprung up to offer a new channel for states to play lottery [eg 8 cent charge per 10 tickets].

We also see a new entrant -Jackpocket- [covering 15 large states and recently valued at only $200 million] has made an introduction with sports tie-ins and perhaps sustainable independent branding.  We’ve also seen a flush of money used to facilitate sports gambling mobile apps getting state licenses.  

The gamification of sports gambling, combined with dynamic sports bets [which adjust games and odds, live during the game] makes it difficult to both understand the traditional odds [see my previous math on NCAA brackets] for anyone, including major bookies.  As we note below, this makes the lowered customer value of playing, even more terrible.

Add to this a 9% fee on initial deposits  for Jackpocket, and it is almost assuredly similar to the high fees that would reduce the payouts from the lottery [see comment above about distributions from mobile lottery operators].

But in this growth era across the economy, sometimes leveraged money circulates to give the appearance of growth.  For examle, along with online casinos, we’ve seen a tsunami of sports-betting advertising in major travel locations and at sporting venues.

The economics of these new gaming [gambling?] models are egregious to everyone.  From high "convenience" fees for the customer, and onerous for the technology companies [eg currently double-digit “fees” paid back to the lottery for casino games that should have higher payouts].  

Also the mobile experience provides greater and more frequent access to games [a frightening five times initial maximum spend on Jackpocket versus what high lottery spenders and already spending].  Since our 2016 lottery article, new surveys of consumer lottery spending suggest some randomness across income strata, though a typically high share coming from older citizens.  Still, as we noted in that article, the level of spending across the population is extraordinarily high and ill-advised.

The total lottery market share loss [versus the nominal economic growth of the economy post-coronavirus] isn’t fully made up for by the triple-digit revenue growth from these small mobile casinos and sports games.  But the numbers are close when you also include the other market based froth [described at the start of the article] of the past couple years.

topic 5: gaming odds with casinos and sports
Different types of speculation each come with its own chances of winning large or small prizes.  Casinos are slightly simpler than sports, though probability mathematics could be generally challenging for most people.  But casinos and sports probabilities are similar enough to be able to describe together, here.

We discussed one example above for a complex sports bet being a 1:20m chance of winning [eg perfectly selecting early March Madness brackets].  And note that you would need a 1:30m range chance of winning lottery game in order to have the same basic ticket value [given the reduced lottery payouts].

But let’s consider the fundamtntal two-team sporting game odds.  For example, who will win the World Cup, between two evenly matched countries?  Similar to who will win a national election.  If we assumed the sports games the and lottery payouts were the same [we know they are not, but this is for mathematical illustarion] then the expected value over the course of a single game or bet would be equal.  

However there is also the consideration of taxes on the winning, which would always tilt the “value” towards a loss.  Leaving that aside, we have the compounded nature similar to the disproportional gambler's ruin problem, famous among the most renowed probablists over a century ago while playing in Casino de Monte-Carlo.  

And other complications such as the fact that the original “evenly matched” odds one assumes is actually only a guess.  The actual chance could be more or less than this due to the individual game/team randomness, and also estimation errors in the the odds themselves for the unique game you [and only you] are interested in betting on.

Of course even in the most basic sports betting, speculators generally bet on lopsided odds match-ups.  Implying the compounded effect of gambling [chapter 6 of Statistics Topics] is even greater.  And returning to the more exotic game bets, the confidence intervals are generally wider due to the nature of artificial intelligence models being solved.  Instead of the preferred closed-form probability theory, on small and unique sample sizes.

The flawed overconfidence of the gamer would almost certainly be matched by larger and less efficient waging.  And everything above works for sports as well as casinos.  The most elementary casino game is roulette [with bare options such as a number or color], followed by Black Jack [with bare options such as a single draw, single player, and large shoes].

Last, there's generally a flawed mentality of gaming [or gambling entertainment] when it comes to casinos and sports, while there's more of a retirement mentality when it comes to the simpler lottery. Though one rarely realized.  Perhaps why more money is wasted on the lottery.  And why money has also flowed into other speculative [and perhaps fraudulent], new market-based products recently. For example the mirage of pumped-up cryptocurrencies being the get-rich quick scheme, which many in society could retire on.


supporting salil statistics:
follow via RSS or e-mail

No comments:

Post a Comment