This is an academic paper for Computational Economics Course at Davidson College with Dr. Shyam Gouri Suresh. Written December 10th 2021.
- Introduction
Climate change is a severe global problem that needs to be addressed immediately. Fashion – especially fast fashion – is a significant contributor to environmental degradation. Research by McKinsey & Co. indicates that the fashion industry was responsible for about 2.1 billion metric tons of greenhouse gas (GHG) emissions in 2018. This is about 4% of the worldwide total and almost the same quantity of GHGs per year as the entire economies of France, Germany, and the United Kingdom combined (Berg, A. & Magnus K. 2020). The need to shift to sustainable fashion is urgent, but how can this be achieved?
There are many stakeholders in the fashion sector, including brands, retailers, factories, farmers, policymakers, investors, and consumers. All influencers play a part in reducing fashion’s carbon footprint, making the industry more sustainable. In this paper, I seek to focus on the role of brands and consumers and their influence on the fashion cycle. The way brands market and make their products and consumer decisions impact how environmentally friendly the fashion industry can be. However, influencing consumer behavior toward sustainable fashion is not well understood or widely adopted.
Fast fashion harms the environment but is highly profitable. The pace at which brands launch new products influences how quickly trends become obsolete—some brands such as SHEIN, Zara, H&M release new products on a two-week cycle. A fast trend cycle sways consumers into overconsuming clothing and accessories to keep up with the latest styles. Frequent consumption of low-quality products is highly profitable but devasting in terms of carbon emissions associated with producing these goods. Slowing down the fashion cycle is critical to making the industry more sustainable. Some luxury brands such as Patagonia, Louis Vuitton, and Eileen Fisher have taken an ideological stance on sustainability and worked to create lines with less unnecessary products and focus on long-lasting items (Sun, Bellezza, & Paharia, 2021).
On the consumer end, people’s decision of when to buy clothes, what to buy, and how long to keep their products ultimately is dependent on consumer discretion. Consumers’ fashion decisions are functional and emotional, based on social influences, monetary limitations, preferences, and information about the durability of the products they purchase. High-end luxury goods are more expensive but can be more sustainable than mid-range products because they have a longer life cycle. Many consumers continue to buy multiple low-to-middle-end goods failing to consider the durability of their purchase or because of monetary constraints.
The fashion industry is a complex system of consumers and brands, each optimizing based on a set of competing priorities. Creating a sustainable fashion system depends on brands’ marketing decisions and people’s consumption patterns, which are highly influenced by socio-economic factors, and not solely practicality. Thus, I seek to model how dynamics between buyers and firms can create sustainable trends in fashion. Sound change is a slow and adaptive process. Brands continue to use unsustainable business practices so long as consumer demand is strong. Consumers continue to make unsustainable decisions as long as brands make convenient products and market their products on a fast cycle. The fashion sector is of particular interest for sustainability because styling decisions are so highly socially influenced. In this model, I seek to understand how consumers’ attention to the durability of their products affects sustainable behavior by brands and consumers.
- Model Description
Brands decide how long to space out new launches based on the sales generated in the previous launches. In my model, I assume that brands are profit-maximizing and that they solely base decisions on monetary outcomes and take no ethical stance on sustainability issues. So, brands update the period between launches based on consumer behavior. On the other hand, consumers decide when to buy or discard items in their closets based on critical parameters. Such parameters include discretionary income, the durability of the clothing, and the social influence of their peers.
Every round, if the brand’s cycle is over, then the brand will launch a new product. The only parameter they decide is how long the cycle will last for weeks and how durable their product is. A brand chooses between four choices: two weeks, four weeks, eight weeks, or twelve weeks. The duration is chosen randomly but weighted by the sales of previous decisions. I add a forgetting parameter where each round the weights are multiplied by a value close to one but slightly smaller (0.999). This forgetting parameter ensures that firms aren’t overly influenced by their early decisions and get stuck at inferior equilibrium, while also making sure that a firms weights do not grow infinitely large in the long run.
All the weights start at an initially high value to increase the amount of experimenting that each does. Figure 1 shows the effect of increasing the initial weight value on the overall variation in the duration decisions by the brands. As the initial weight gets larger, the brands are more experimental and vary their decisions over the course of a simulation more than when the initial weights are lower. Specifically, using an initial weight of 1000 increases variation in the decisions by firms by 6 weeks as opposed to using an initial weight of 1.
Figure 1: Average Variation in Firms’ Durability Decision at Different Initial Weights
On the consumer end, each agent is randomly assigned discretionary income of either $10, $50, $or $100. Agents get paid this amount every other round, so every two weeks. Agents’ balances get set to zero every fourth week, so it’s assumed by the end of the month, each agent has spent their discretionary income or put it into savings. Agents must wear something – and always need one garment. However, how long they choose to keep their attire is dependent on the utility they get from holding on to their current item compared to the utility from buying from one of the brands.
Each round, the agent will evaluate this decision by comparing the durability of the current goods being offered, the brand’s popularity, how many of the other agents wore the brand in the previous round, and the cost of the item relative to their current discretionary income balance. An agent will calculate the utility from buying a new piece of clothing from one of the three brands and keeping the same item which is simply how much durability the agent has left in the garment times a defined durability weight. The durability weight (Beta) can be thought of as attention that an agent pays to the quality of their garments or as a sustainability-minded parameter where agents place more value on holding on to their garments for longer because they know it’s a more sustainable decision. The utility calculations are below. The utility function is Cobbs-Douglas.
Equation (1) is the marginal utility of buying the latest item from one of the brands. Durability is the durability of the product offered by the brand in that period and it is weighted by one-fourth of the durability weight. This is because the agent cares about the durability of the product, but weighs the durability of a new good less than the durability of their current good. The popularity of the brand is the number of agents who in the previous round bought from that brand, the added one is to make sure if no one bought the brand that the utility would not go to zero, but will not affect the utility derived from the durability of the good. Equation (2) describes the marginal cost and is simply the price of the garment relative to the discretionary income of the agent. Equation (3) is the marginal utility of keeping the agent current good for one more period which is the current durable state of the good weighted by times the number of agents who did not buy a good in the previous round. The marginal cost of not buying a good is 0. Agents optimize by maximizing utility minus the marginal cost. In each round, agents make this evaluation and choose the optimal decision.
One of the outcomes of interest is how consumer behavior changes as they pay more attention to the durability of their goods. Do agents adopt sustainability purchasing practices such as buying fewer goods, buying more durable goods, or holding on to their goods for longer? The other dependent variable is the length of time between launches. Do brands learn to slow down the fashion cycle in response to consumer behavior? I tested these brand and consumer effects under different conditions for durability weights. I also tested the effect if halfway through the simulation there is a “sustainability movement” in which half of the agents increase their durability attention by three units.
- Results
Table 1: Description Of Statistics
As expected, the more weight that durability has for the agents, the longer they hold onto their goods. As shown in Figure 2, if all agents increase their attention by 1 unit, they leave on average 16,714 fewer weeks of durability unused over the course of 200 weeks. This is a significant drop in durability unused — more than 50 standard deviations of magnitude. Similarly, agents buy new products less frequently as the durability weight increases. If an agent increases their attention to durability by one unit, the number of sales decreased by 1776 — which is 10 times the standard deviation. This is expected because utility formulas are designed to capture attention to durability, so the utility of keeping a good and not buying increases exponentially as the durability weight increases.
Figure 2: Amount of Durability that Agents Leave Unused vs. Durability Attention
Figure 3: The Number of Purchases vs. Durability Attention
The variance between trials decreases significantly after the durability is above four for both the number of sales and the durability left unused, and stabilizes around zero weeks of durability left unused and 21,000 number of sales in 200 rounds. Thus, as agents’ attention to durability gets above a certain threshold, they become a near zero-waste society consistently. The high variation at low durability weights is most likely due to durability neglect, and because the price of the good makes a larger impact on the decision to buy a brand.
Figure 4: Average Duration Decision by Customers at Different Durability Attention Levels Across Different Income Groups
Consumers increase the level of durability they chose to buy as indicated by the positive slopes for all three income groups. The effect of increased durability weight is most significant for low-income agents. This is because this group is also the most cost-sensitive. Whereas, the highest income group is less financially limited and consistently buys the most durable goods regardless of attention to durability. For low-income groups, to move one standard deviation in average durability, the agent would need to increase their durability weight by 5 units.
Figure 5: The Likelihood that The Most Popular Brand Also Has the Highest Probability of Choosing the Longest Duration for Different Levels of Durability Attention
Figure 5 shows the likelihood that in a simulation the most popular brand in terms of the number of sales also has the highest probability of choosing the longest duration of clothing offered out of the three brands. Meaning out of the three most probable decisions by the three brands, the most popular brand has the longest or equal to the longest most probable durability among the three brands. As for durability increases, the probability of the most popular brand acting the most sustainably increases. However, the effect size is very small and only increases by 2.1 percentage points for every increase in durability attention. This small effect size demonstrates that even at higher levels of durability, the number of sales for brands that release less durable products on a faster cycle can remain popular. This is likely because people hold onto their goods for longer when they are more attentive to durability. Thus, the number of sales for durable products remains lower as people purchase these products less often.
Figure 6: Average Most Probable Duration Decision By Brands At Different Levels of Durability Attention
Figure 6 suggests that brands learn to lengthen their product cycles and increase durability as agents pay more attention to durability. The slope is positive, but there is high variability and the effect size only increases by 0.21 weeks on average for an increase in attention by 1 unit. This is not as drastic an effect as I expected, but indicates that sustainability practices may be harder for brands to catch up with even when consumers change their preferences and behaviors drastically.
Figure 7: Change in Most Proable Duration Decision By Brands After a Sustainability Movement at Different Initial Durability Weights
After a sustainability movement in which half of the agents increase their durability weight by 3, brands learn to slow down their product cycles by about 1 week. This is a very weak result. The starting durability weight seems to have a minimal effect on the change induration decision by brands. There is an effect from the sustainability movement which is interesting because only half of the agents actually adopted a greater sustainability sensitivity indicating that peer effects may play a role in the success of a sustainability movement.
- Conclusion
My results have implications for the success of the fashion industry to reduce emissions. Consumers are turning their attention to their own sustainable practices and the actions of brands. My results indicate that the number of sales may decrease as a result of this movement, brands who want to be successful should increase the durability of their goods and produce higher quality products on a slower cycle.
My results are consistent with empirical evidence that suggests that a rise in a new luxury trend amongst generation Z and millennials (Mohr, Fuxman, & Mahmoud, 2021). Similarly, my results are corroborated by Stahle and Muller’s study. Their findings suggest the rising significance of sustainability in the fashion sector. Stahle and Muller emphasize the role that brands play as intermediaries; our results indicate that brands will increase their sustainability practices in order to increase profits as people pay more attention to the durability of their products. It is in the best interest of retailers to become more sustainable.
In my model, the agents for the most part had homogenous sustainability and attention preferences with the expectation of when sustainability movement is introduced. However, McNeill and Moore find that consumers are likely to value certain different factors when making the decision to buy clothes. They conclude that consumers fit into three groups: ‘Self’ consumers, concerned with hedonistic needs, ‘Social’ consumers, concerned with social image, and ‘Sacrifice’ consumers who strive to reduce their impact on the world. A more realistic extension of this model would be to have three different groups with different weights on these factors.
References
Berg, A. & Magnus K. (2020). Fashion On Climate: How The Fashion Industry Can Urgently Act To Reduce Its Greenhouse Gas Emissions. Mckinsey & Co.
Mcneill, L., & Moore, R. (2015). Sustainable Fashion Consumption And The Fast Fashion Conundrum: Fashionable Consumers And Attitudes To Sustainability In Clothing Choice. International Journal Of Consumer Studies, 39(3), 212-222.
Mohr, I., Fuxman, L., & Mahmoud, A. B. (2021). A Triple-Trickle Theory For Sustainable Fashion Adoption: The Rise Of A Luxury Trend. Journal Of Fashion Marketing And Management: An International Journal.
Strähle, J., & Müller, V. (2017). Critical Aspects Of Sustainability In Fashion Retail. In Green Fashion Retail (Pp. 7-26). Springer, Singapore.
Sun, J. J., Bellezza, S., & Paharia, N. (2021). Buy Less, Buy Luxury: Understanding And Overcoming Product Durability Neglect For Sustainable Consumption. Journal Of Marketing, 85(3), 28-43.
