The joy economy and the power of markets

#blog

I've been working at a hostel lately and learning about how shifts work. While laundry and reception work isn't particularly brain-tingling, figuring out how the systems that support it has been quite stimulating.

I'm working in a team of 6 volunteers, doing 30 hours a week each, and was curious to figure out how the working hours varies with the number of volunteers.

By nutting this out with some math, I found a neat shift-rotation optimisation that gives everyone more time off with the same work (some might call it common-sense). But the banger realisation I had is that workforces can manage themselves using a free market where workers can exchange shifts, and that the workplace is a better example of The power of markets than most monetary economies we see today. If you despise words like "economy" and "capitalism", it's likely because Neoliberalism cripples us all by convincing us that free markets require minimal intervention (which is absolutely not what "free market" means in economics). It was a satisfying reminder that Economics is everywhere, and that contrary to what we see today, we can use economics to provide levels of fairness and efficiency that is impractically difficult to achieve with central-planning.

Anyway, enough ideological yapping. Let's put on our manager hats. If you want to skip the math, skip forward to the joy economy.

Modelling shifts with math

I'll provide my modelling process because I think being able to represent your logistics on paper like this is useful. It allows helpful equations and optimisations to emerge that you might not notice if you kept the model in your head.

Finding employee weekly hours with employee count

We know our input variables are

Employee count, which hovers between 5 and 6 people (not including the hostel owners, who occasionally work)
The shift duration, which is 5 hours for all shifts
The shift roster, which looks like

Shift start time	6:30am	9:30am	2:30pm	7:30pm
People	1	2	1	1

So there are five different shifts. We can model a day of the five shifts distributed to employees (let's say we have 5, and I'm not writing their names because they're just faceless numbers to me now) with little Tetris-looking blocks like so:

Let's use a simple rotation system to ensure everyone experiences the same shifts. We are assuming everyone can learn to do every shift, which is a fair assumption if you filter your job candidates or have a clear training program. We can handle employee preferences for shifts later, which we'll see can be resolved without any brain-power on our parts using the magic of markets.

Oh look! We can see that with five shifts and five people, the cycle repeats after five days. We can now calculate the employee work time per week by calculating:

\begin{aligned} \frac{employee hours worked}{week} = & hours worked by employee#1per cycle \times \frac{cycles}{week} \\ = & shifts worked by employee # 1 per cycle \\ \times \frac{hours}{shift} \times \frac{cycles}{week} \\ = & 5 shifts \times \frac{5 hours}{1 shift} \times \frac{7 cycles}{5 weeks} \\ = & 35 \frac{hours}{week} \end{aligned}

There are 7 cycles for every 5 weeks because a cycle lasts 5 days, so we need an extra $2 / 5$ days of the cycle to top it up to 7 days, meaning we have $1 + 2 / 5 = 7 / 5$ cycles per week.

Letting an equation emerge

Great! Let's find out how this changes with the number of employees by adding a 6th employee.

We can see the cycle now repeats after 6 days, meaning there are $7 / 6$ cycles per week, and employee $# 1$ has one day off per cycle (Saturday).

Notably, employee $# 1$ still works 5 shifts per cycle even though there is an extra day in the cycle. So with 6 employees:

\begin{aligned} \frac{employee hours worked}{week} = & shifts worked by employee # 1 per cycle \\ \times \frac{hours}{shift} \times \frac{cycles}{week} \\ = & 5 shifts worked by employee#1per cycle \times \frac{5 hours}{1 shift} \times \frac{7 cycles}{6 weeks} \\ = & 26.166 \dots \frac{hours}{week} \end{aligned}

We now have a formula for calculating the hours worked per employee

\begin{aligned} \frac{employee hours worked}{week} = & shifts worked by employee # 1 per cycle \\ \times \frac{hours}{shift} \times \frac{cycles}{week} \\ = & 5 shifts worked by employee#1per cycle \times \frac{5 hours}{1 shift} \\ \times \frac{7}{number of employees} \cdot \frac{cycles}{weeks} \\ = & \frac{7 \cdot 5 \cdot 5}{number of employees} \end{aligned}

So we can table this:

Employees	Hours worked per week
5	35
6	29.16
7	25
8	21.875
9	19.4
10	17.5

Optimisation: extra time off with rotation

Let's consider what happens with 6 employees when employee $# 1$ has their day off on Saturday. They finish their 7:30pm shift (5 hours) on Friday at 11:30pm, and then on Sunday, they start their shift at 6:30am. Yuck! They've only got Saturday off, they'll be tired after they finish on Friday (can't enjoy their night as much), and they have to go to bed early on Saturday night to be up on Sunday.

But what if we rotate shifts the other way?

Now employee $# 1$ has their day off on Tuesday, finishing their 6:30am shift on Monday and starting their next shift at 7:30pm on Wednesday. Now they can rest on Monday afternoon and stay up late on their day off knowing they can take it easy on Tuesday morning before work.

By simply rotating shifts in the opposite direction, we can give employees almost an extra day off and massively increase the utility of their time off.

The joy economy, and how capitalism kills your business

The hostel owners refuse to take on more than 6 volunteers, and I think this is to their detriment.

For the hostel I've just modelled the shifts of, I'm not actually working for money here; I'm doing a work trade in exchange for accommodation and some food. With 6 volunteers, we're working 30 hours a week and get one day off. In honesty, this is pretty lackluster; I've done previous worktrades which were only 15-20 hours per week, had two full days off, and provided all meals. I took this opportunity because it's out of the Schengen zone (freeing up later time in Schengen countries for other work trades), the cost of living is quite cheap here, it's surrounded by beautiful scenery for hike, and we needed a last-minute place to stay that ticked these boxes (worktrade season is hectic in European summer).

So naturally, given the work expectations with 6 volunteers are high relative to the standard, volunteers end up becoming quite resentful quickly. They leave as soon as a better opportunity arises (higher staff turnover and training costs), and they don't stay in the hostel when they're not working (because you naturally get roped into extra tasks as the work is spread so thinly).

This has an invisible but palpable effect on the vibe of the hostel; when volunteers are happy and social (at the beginning of their time, before they become jaded), lots of guests extend because they enjoy the social atmosphere, and the hostel makes more money. Guests are more likely to hang out at the hostel with volunteers, learn about guided tour offers, and buy drinks from the bar. All of these benefits bring in increased revenue for the hostel.

The costs of hosting extra volunteers is not high; it costs a bed per night (but the hostel has never been full, so this does not have a high opportunity cost), slightly extra in food costs (for the included dinner, which is often very light in portions and lacks satisfying protein), and perhaps extra in training costs. But with 7 volunteers, the work time goes down to 25 hours with 2 days off; with 8 volunteers, they work 22 hours with 3 days off (almost 4 days with our earlier rotation optimisation).

I strongly believe that this reduced workload, at little to no extra cost to the hostel, would present a massive improvement in social gravity and overall atmosphere, with monetary benefits shown in accomodation extensions, tours and extras.

But the point I want to make is that sometimes monetary benefits are not visible until you work with another economy. In this case, optimising for the economy of workers' time and joy creates more benefits that are invisible if you only consider how much money and time you'll spend on hosting and training staff as the owner. Not only that, but if you tunnel-vision on the monetary economy, you risk burning out your staff, creating boring or negative experiences for guests, and ultimately ruining your hostel's reputation. Not to mention that you'll come to view your volunteers as selfish, lazy and in need of micromanagement, which will turn any hardworking volunteer into exactly what you fear.

The joy economy is present and important in any service or hospitality-based job. Happy, relaxed baristas make for better conversation and coffees which make for higher customer retention and reviews, not to mention the employee retention. There is a runaway flow-on effect for joyful employees, just as there is a vicious destructive loop for unhappy employees. You might not notice you've killed your business until the tens of overworked baristas you've turned over have burnt enough coffees to kill your cafe rating, and at this point, you may never recover without dropping your prices to compete once more.

By focusing squarely on the monetary economy central to capitalism, you will miss improvements and deficiencies that leak through the joy economy. Even if you're a sociopathic manager, it's in your monetary interests to increase employee joy.

These ideas are echoed in product-based jobs like software engineering too; a fantastic related read is why employee churn is lethal to software companies.

Replacing your manager with a market

So we have a simple cycling roster that's fair in that everyone works the same shifts at some point, but it's not realistic. Employees have different preferences for shifts (early-risers vs night-owls), and might not have the skills for every shift (they might lack training or social skills for reception, for instance). How do we fix this?

Let employees freely trade their shifts, and train employees in every shift type. By ensuring all employees can (and must) work every shift by default, you prevent employees from monopolising easy shifts where they can quiet quit, and prevent less-skilled employees from being stuck with harder shifts they can't escape. This might not be practical with specialised labour, but in a rotating workforce for entry-level (quickly trainable) jobs, it indicates that having a fantastic education program is well-worth the effort.

The best part about this is that, in return for this training, as the shift roster creator you free yourself of the combinatorial cluster-fuck that is balancing everyone's preferences and skills. Employees will not give away universally preferable (ie easy) shifts, nor accept universally hard ones. Giving employees shift variety means they can find their niches; an introvert and extrovert will naturally exchange isolated laundry work for social reception work.

And this is the magic of economics; an economy is a market (a place where you can trade stuff), usually accompanied by a convenient medium of exchange (like money or time). The purpose of an economy is to distribute resources according to need or desire, and when it's free from monopolies, it can distribute resources more effectively than any group of people or computers.

You might think this is too much trust to place on the employees to manage themselves. Too often, a single manager in charge of the roster becomes the team's bane of existence, and it's no wonder that they often fail to meet people's needs when they are complex and numerous. But the hidden reality is that you're not trusting any one person when you use a market; you're trusting people, who are more than the sum of their parts.