📚 Affordability Metrics in City Rankings: How Tuition Fees and Living Costs Are Calculated | 城市排名的可负担性指标：学费与生活成本如何计算
Every year, dozens of "best student cities" rankings appear, and almost all of them claim to measure affordability. Yet two rankings can place the same city at number 3 and number 27, because affordability is not a price - it is a ratio between the cost of studying in a place and the resources a student can draw on there. This article explains, step by step, how tuition fees and living costs are actually converted into the index numbers that sit behind those league tables, and how you can interrogate the arithmetic yourself.
每年都会出现数十份"最佳留学城市"排名，而几乎所有排名都声称自己衡量了"可负担性"。然而同一座城市在不同榜单上可能一个排第 3、一个排第 27，原因在于：可负担性不是价格本身，而是"在某地读书的成本"与"学生可动用的资源"之间的比率。本文逐步拆解，学费与生活成本究竟是如何被换算成这些排行榜背后的指数数字的，以及你如何亲自审视这套算法。
1. What Affordability Actually Measures | 可负担性究竟衡量什么
Affordability is a relative concept. A city where tuition costs 30,000 currency units per year may be more affordable than a city where tuition costs 8,000, if graduates in the first city earn three times as much and rent absorbs a smaller share of income. Economists therefore treat affordability as a ratio - cost divided by capacity to pay - rather than as a level of prices.
可负担性是一个相对概念。某城市学费一年 30,000 货币单位，可能反而比学费只要 8,000 的城市更"可负担"，如果前者毕业生收入是后者的三倍、且房租占收入比重更低。因此经济学把可负担性视为一个比率——成本除以支付能力——而不是价格水平本身。
A useful distinction is between the sticker price and the net price. The sticker price is the published fee; the net price is what a student actually pays after scholarships, fee waivers, bursaries and subsidised housing are subtracted. Rankings that use sticker prices systematically understate affordability in countries with large public subsidy, such as Germany or the Nordic states.
一个有用的区分是"标价"与"净价"。标价是公布学费；净价是扣除奖学金、学费减免、助学金与宿舍补贴之后学生真正支付的金额。使用标价的排名会系统性低估公共补贴力度大的国家（如德国或北欧国家）的可负担性。
- Affordability = cost ÷ resources (a ratio, not a price) | 可负担性 = 成本 ÷ 资源（比率，不是价格）
- Sticker price vs net price: subsidies change the ranking | 标价与净价之别：补贴会改变排名
- Time frame matters: annual cost vs whole-programme cost | 时间口径很关键：年度成本与整个学位成本
2. Building Block One: Tuition Fees | 构成要素之一：学费
Tuition is the most visible cost component and also the most heterogeneous. Providers charge different rates to domestic and international students, band fees by subject (laboratory-based programmes such as medicine, engineering and studio art usually cost more than classroom-based ones), and either publish a per-year fee or an all-in programme fee. A ranking must decide which of these numbers to standardise on.
学费是最显眼的成本项，也是异质性最强的成本项。院校对本国学生与国际学生收费不同，按学科分档收费（医学、工程、工作室艺术等实验实践型专业通常高于课堂型专业），并且有的公布每年学费、有的公布整个学位总学费。排名必须决定以哪一种数字做标准化。
Four tuition traps appear again and again in published rankings. First, mixing a three-year bachelor programme with a four-year one without converting to cost per year. Second, using national average fees when actual fees vary enormously between institutions inside the same city. Third, ignoring the direction of the exchange rate on the date the data was collected. Fourth, confusing the tuition fee with the total cost of attendance, which is the fee plus living costs plus mandatory insurance and administrative charges.
已发表的排名中反复出现四类学费陷阱。第一，把三年制本科与四年制本科混在一起，却没有换算为每年成本。第二，使用全国平均学费，而同一城市内不同院校的实际学费差异巨大。第三，忽略数据采集当日汇率的方向。第四，把学费与"总就读成本"混淆——后者是学费加生活费加强制保险与行政费用。
| Tuition data problem | 学费数据问题 | Why it distorts the ranking | 为何会扭曲排名 |
| Domestic vs international rate | 本国费率与国际费率 | Differences can exceed 400% within one university | 同一所大学内差异可超过 400% |
| Subject bands | 学科分档 | Engineering may cost 1.5× an arts degree | 工科可能是文科的 1.5 倍 |
| Programme length | 学制长度 | 3-year vs 4-year changes total cost by one third | 三年制与四年制使总成本相差三分之一 |
| Net price after aid | 资助后净价 | Sticker rankings penalise generous systems | 只看标价会惩罚资助慷慨的体系 |
3. Building Block Two: The Student Living-Cost Basket | 构成要素之二：学生生活成本篮子
Living costs are measured through a basket: a fixed list of goods and services with quantities that represent what a typical student consumes in a year. The basket is then priced in each city. Rent normally dominates, taking between 40% and 60% of a student budget in large Western cities, which is why small errors in rental data swing the final ranking far more than errors in the price of bread.
生活成本通过"篮子"来衡量：一份固定清单，列出典型学生一年消费的商品与服务及其数量，然后在各城市对篮子定价。房租通常占主导——在西方大城市可占学生预算的 40% 至 60%——因此租金数据的小误差对最终排名的影响，远大于面包价格的误差。
A standard student basket usually contains the categories below. Note that some items, such as visa and health insurance charges, are not really "cost of living" in the ordinary sense but are unavoidable for international students, so rankings for international audiences include them.
标准学生篮子通常包含以下类别。需要注意的是，签证费与健康保险费严格说并不是通常意义的"生活成本"，但对国际学生而言无法回避，因此面向国际受众的排名会将其纳入。
- Rent (shared flat, studio, or university hall) | 房租（合租、单间或学校宿舍）
- Food (groceries plus a small allowance for eating out) | 饮食（食材加少量外食额度）
- Public transport and occasional intercity travel | 公共交通与偶尔的城际出行
- Utilities: electricity, heating, water, broadband, mobile data | 水电燃气、宽带、手机流量
- Books, printing, laboratory or software costs | 书籍、打印、实验或软件费用
- Mandatory health insurance and visa/administrative fees | 强制健康保险与签证/行政费用
- Clothing, personal care, leisure and sports | 衣物、个人护理、休闲与运动
4. Weighting the Basket | 篮子的加权
Once quantities are chosen, the cost of the basket is a weighted sum. Each item has a price pᵢ and a quantity qᵢ, and the total is the sum of price times quantity across all items. The quantities are the weights: they decide how much each price matters.
数量确定后，篮子的成本就是一个加权和。每一项有价格 pᵢ 与数量 qᵢ，总成本是所有项"价格 × 数量"之和。这些数量就是权重：它们决定每项价格有多重要。
Basket cost C = Σ (pᵢ × qᵢ)
Weights are usually derived from household expenditure surveys or student spending surveys. Two biases follow. Substitution bias arises because students switch away from items that become expensive - if rent rises sharply, students take on more flatmates rather than paying the higher rent, so a fixed basket overstates the true increase. Weight drift arises because student consumption patterns change over time, so a basket fixed in 2015 becomes less representative every year.
权重通常来自家庭支出调查或学生消费调查，随之产生两类偏差。"替代偏差"源于学生会回避变贵的项目——如果房租大涨，学生会多找室友而不是支付更高租金，因此固定篮子会高估真实涨幅。"权重漂移"源于学生消费结构随时间变化，2015 年固定的篮子会逐年失去代表性。
Laspeyres index = (Σ pₜq₀ ÷ Σ p₀q₀) × 100
The formula above is the classic Laspeyres index, which holds quantities at base-year levels q₀ and lets prices move to period t. It is the standard tool in cost-of-living work and is the reason official inflation data often looks slightly higher than the price changes students actually experience.
上式即经典的拉氏指数：数量固定在基年水平 q₀，价格变动到 t 期。这是生活成本研究的标准工具，也正是官方通胀数据常常略高于学生实际感受到的价格变化的原因。
5. Index Numbers: Base Year and Normalisation | 指数化：基年与标准化
Raw local-currency costs cannot be compared directly, so they are converted into index numbers. A base city or base year is set equal to 100, and every other observation is expressed relative to it. This makes the ranking readable but hides absolute values: an index of 175 tells you nothing about whether the underlying cost is 8,000 or 40,000.
原始的本币成本无法直接比较，因此要转换为指数。选定一个基城市或基年等于 100，其余观测值相对它表示。这让排名易于阅读，却隐藏了绝对水平：指数 175 并不能告诉你底层成本是 8,000 还是 40,000。
Cost index = (Cₜ ÷ C₀) × 100
Two further refinements matter. First, real versus nominal: if the data covers several years, nominal costs must be deflated by a price index to remove general inflation, otherwise a city that simply appears in a later year looks more expensive. Second, min-max normalisation, which rescales each indicator onto a common 0 to 1 range so that tuition and rent can be added together without one dominating purely because of its units.
还有两项关键处理。第一，名义与实际：若数据跨多年，名义成本必须用价格指数平减以剔除一般通胀，否则一个只是出现在较晚年份的城市会显得更贵。第二，极差标准化，把各指标重新缩放到共同的 0 至 1 区间，使学费与房租可以相加，而不会因单位差异导致某一项主导结果。
Normalised score = (x − xₘᵢₙ) ÷ (xₘₐₓ − xₘᵢₙ)
6. Purchasing Power Parity and Currency Conversion | 购买力平价与货币换算
To compare cities across countries, local currencies must be converted into a common unit. Using market exchange rates is simple but misleading, because market rates are driven by capital flows and traded goods, while students spend most of their money on non-traded services such as rent, haircuts and bus tickets. Purchasing power parity (PPP) conversion factors price a representative basket of goods in each country and derive the exchange rate that would equalise its cost.
跨国比较城市，必须把本币换算成统一单位。用市场汇率简单，但会产生误导：市场汇率由资本流动与可贸易品驱动，而学生支出的大头是非贸易服务，如房租、理发、公交票。购买力平价（PPP）换算因子先给各国一篮子代表性商品定价，再推导出使该篮子成本相等的汇率。
PPP-adjusted cost = Nominal cost in local currency ÷ PPP conversion factor
PPP has two limitations that rankings rarely disclose. It is a national average, not a city-level measure - within one country, the PPP-adjusted cost of living in the capital can be 40% higher than in a provincial town, yet both receive the same conversion factor. And PPP baskets reflect average household consumption, not student consumption, so a country where students rely heavily on subsidised canteens and dormitories will look artificially cheap.
PPP 有两个排名很少披露的局限。其一，它是全国平均值而非城市级指标——同一国家首都的 PPP 调整后生活成本可能比省会城市高 40%，但两者使用同一换算因子。其二，PPP 篮子是家庭平均消费而非学生消费，因此学生大量依赖补贴食堂与宿舍的国家会显得被人为压低。
7. Constructing the Composite Affordability Score | 构建综合可负担性得分
A composite index combines several indicators into one number. The designer chooses the indicators, the weights, and the direction of each indicator: costs enter negatively (higher is worse) while income or scholarship availability enters positively (higher is better).
综合指数把多个指标合成为一个数字。设计者需选择指标、权重以及每个指标的方向：成本类为负向（越高越差），收入或奖学金可得性为正向（越高越好）。
Composite score = Σ (wⱼ × xⱼ), where Σ wⱼ = 1
The most important single decision is the comparison denominator. Three common choices appear in published rankings: cost per year alone (a pure price measure), cost divided by average graduate salary (an affordability ratio), and cost divided by student income including permitted part-time work (a cash-flow measure). These produce genuinely different league tables, and a ranking that does not state its denominator cannot really be evaluated.
最关键的一项决策是比较的分母。已发表排名中常见三种选择：仅用年度成本（纯价格指标）；成本除以毕业生平均薪资（可负担比率）；成本除以含合法兼职收入的学生收入（现金流指标）。三者会产出真正不同的榜单，而未说明分母的排名实际上无法被评价。
Affordability ratio = (Annual cost ÷ Expected annual income) × 100
In practice, most city rankings report a ratio of total annual cost to expected gross starting salary, expressed as a percentage. A ratio of 65% means the first year of study costs about two thirds of a graduate's starting annual pay - a figure students can interpret immediately.
实践中，多数城市排名报告的是"年度总成本 ÷ 预期税前起薪"的比率，以百分比表示。比率 65% 意味着第一年就读成本约等于毕业生起薪的三分之二——这个数字学生可以立刻理解。
8. Student-Specific Adjustments: Subsidies and Concessions | 学生专属调整：补贴与优惠
Students are not ordinary consumers, and the adjustments below can change a city's relative position by more than any tuition difference. Failing to model them is the most common technical criticism of student city rankings.
学生不是普通消费者，下列调整对城市相对位置的影响可能超过任何学费差异。未能建模这些因素，是对学生城市排名最常见的技术批评。
- Tuition waivers, merit scholarships and need-based bursaries | 学费减免、优秀奖学金与助学金
- Subsidised university housing and canteens | 学校宿舍与食堂补贴
- Student transport passes and cultural discounts | 学生交通卡与文化优惠
- Free or subsidised health cover | 免费或补贴的医疗保障
- Rules on part-time work hours and minimum wage | 兼职工作时长规定与最低工资
- Post-study work rights, which affect the expected return on the investment | 毕业后工作权利，影响投资的预期回报
Consider two cities with identical sticker costs. In City A, 60% of students receive a 5,000-unit scholarship and rent for a university hall is capped at 300 per month. In City B, scholarships cover 10% of students and private rent averages 700 per month. The net costs differ by roughly 40% even though the published prices look the same, and only a net-price ranking captures this.
设想两座标价完全相同的城市。A 城 60% 的学生获得 5,000 单位奖学金，学校宿舍月租上限 300；B 城奖学金只覆盖 10% 的学生，私人房租平均每月 700。尽管公布价格看起来一样，净成本相差约 40%，而只有净价排名才能体现这一点。
9. Data Problems, Bias and Limitations | 数据问题、偏差与局限
Every ranking is a model, and every model has assumptions. The most frequent problems in affordability metrics are listed below; a critical reader should look for these before trusting any position change.
每份排名都是一个模型，而每个模型都有假设。以下列出可负担性指标中最常见的问题；审慎的读者在相信任何名次变化之前，应先检查这些点。
- Vintage mismatch: tuition from one year combined with rent from another | 年份错配：某年学费与另一年房租混用
- Aggregation error: national data applied to a specific city | 加总错误：用全国数据替代具体城市数据
- Sample bias: cost surveys answered mainly by students in central districts | 样本偏差：成本调查主要由市中心学生填写
- Housing mismatch: average city rent compared with a student's shared room | 住房口径不符：用城市平均租金对比学生的合租房
- Quality adjustment: a cheap city may be cheap because provision is poor | 质量调整：便宜的城市可能因为供给质量差而便宜
- Commercial interest: sponsors, recruitment targets and advertising revenue | 商业利益：赞助方、招生目标与广告收入
- Missing data imputed silently, with no uncertainty range published | 缺失数据被静默插补，且不公布不确定区间
Sensitivity analysis is the standard test. A well-built index is published with alternative weightings, and if a city's position moves by ten places when rent is given a 5 percentage point higher weight, the ranking is reporting the analyst's choices as much as it is reporting the city.
敏感性分析是标准检验。成熟的指数会公布替代权重下的结果：如果房租权重提高 5 个百分点，某城市的名次就移动十位，那么这份排名报告的更多是分析者的选择，而非城市的真实情况。
10. Worked Example: Comparing Three Cities | 实例演算：比较三座城市
The table below applies the method. All figures are annual and in the same currency unit. City A is the base city with an index of 100. Expected gross starting salary is used as the denominator for the affordability ratio.
下表演示整套方法。所有数字为年度值，使用同一货币单位。A 城为基城市，指数为 100。可负担比率以预期税前起薪为分母。
| Item | 项目 | City A | A 城 | City B | B 城 | City C | C 城 |
| Tuition | 学费 | 9,000 | 18,000 | 4,000 |
| Rent (12 months) | 房租（12 个月） | 7,200 | 10,800 | 5,400 |
| Other living costs | 其他生活成本 | 4,800 | 6,600 | 4,200 |
| Total annual cost | 年度总成本 | 21,000 | 35,400 | 13,600 |
| Cost index (A = 100) | 成本指数（A=100） | 100.0 | 168.6 | 64.8 |
| Expected starting salary | 预期起薪 | 32,000 | 55,000 | 20,000 |
| Affordability ratio | 可负担比率 | 65.6% | 64.4% | 68.0% |
| Scholarship | 奖学金 | 1,000 | 6,000 | 0 |
| Net affordability ratio | 净可负担比率 | 62.5% | 53.5% | 68.0% |
Three lessons follow. First, on raw cost City C wins easily, with an index of 64.8 against City B's 168.6. Second, once cost is divided by expected salary the gap almost disappears: 68.0% against 64.4%, so the cheapest city is not the most affordable one. Third, after scholarships are modelled, City B becomes clearly the most affordable at 53.5%. The ranking flips twice without a single price changing.
由此得到三点启示。第一，就原始成本而言 C 城轻松胜出，指数 64.8 对 B 城 168.6。第二，一旦用成本除以预期薪资，差距几乎消失：68.0% 对 64.4%，最便宜的城市并非最可负担的城市。第三，纳入奖学金后，B 城以 53.5% 明显成为最可负担者。价格一个都没变，排名却翻转了两次。
This is exactly why headline claims such as "the most affordable student city in the world" should always
Published by TutorHao | Exam Prep Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图