歷年來，台灣有「螺絲王國」的美譽，出口金額一直位居全球第三大扣件(螺絲、螺帽)出口國，僅次於大陸、德國，2022年因為疫情的關係，各類產業各有消長，但螺絲業外銷金額，卻大幅成長，總出口量為166萬噸，首度衝破60億美元，達到63.55億美元，逼近2,000億台幣大關，創歷史上的新高；不過，自2023年起，來自國際訂單金額，卻急轉直下，總出口量為127萬噸，減少了23%；2024年更是五窮六絕，至少十餘家螺絲產業鏈相關企業退出市場，台灣螺絲業明確出現衰退的警訊；不只訂單量大幅萎縮，為了因應ESG、節能減碳(CBAM)、AI、人才招募困難、薪資上漲、貨運載重限制等所導致的成本增加、利潤更被壓縮了，更是雪上加霜。
鋼鐵業為工業之母，其下游應用產品包羅萬象，包括各類金屬製品、機械設備、運輸工具、模具、螺絲螺帽、鋼線鋼纜及工業設施及建築工程上所需之各種鋼材；棒線是扣件產品的原料，佔其成本比例很高；中鋼是國內最大之各式鋼板、線材的供應者，中鋼棒線(中高碳)的季盤價資料，是在前一季的第三個月份公布的；要跟中鋼購買原料，得先一次付清貨款後才能依中鋼的交期獲得原料，有資金耗用的問題要面對，棒線(中高碳)是可以長期存放並無材料變質的狀況，但仍屬於庫存，也代表有資金積壓上的問題，所以本研究希望，能夠透過找到適當的棒線(中高碳)季盤價價格的預測模型，以便讓扣件業的廠商，能夠了解未來新的一季中鋼棒線(中高碳)季盤價價格，可以做為企業採購與財務操作的參考之用。
本研究採用複迴歸(Multiple Regression)、針對棒線(中高碳)季盤價價格進行預測；因要取有較大的解釋力及較低的VIF值的組合，故採用了”逐步迴歸分析析(stepwise regression analysis)”策略所選擇的自變數，透過研究結果得到:月減二(季盤價生效當月往前推兩個月)-中鋼收盤股價-最小，在棒線(中高碳)價格之權重最高，再來是月減四(季盤價生效當月往前推四個月)-波羅的海乾散貨指數-最小、月減二(季盤價生效當月往前推兩個月)-62%品位鐵礦粉 CFR 期貨-最小、月減二(季盤價生效當月往前推兩個月)-波羅的海乾散貨指數-最大，這四個為棒線(中高碳)價格之主要正相關變量；以採用此四個自變數所構成的複迴歸模型，去計算季盤價，再與實際的棒線(中高碳)的季盤價資料做殘差分析，其殘差資料，可以看出分布大致對稱，近似常態分布，表示此模型是具有一定程度的可靠性，模型計算出來的結果可以提供扣件業的廠商做為參考。

Over the years, Taiwan has earned the reputation as the "Kingdom of Screws," consistently ranking as the world's third-largest exporter of fasteners (screws, nuts), trailing only China and Germany. In 2022, despite fluctuations across various industries due to the pandemic, Taiwan’s fastener export value surged significantly. Total export volume reached 1.66 million tons, with export revenue surpassing 6.355 billion USD for the first time, approaching the 200 billion TWD mark — a historic high.

However, starting in 2023, international order values sharply declined. Export volume dropped to 1.27 million tons, a decrease of 23%. The situation worsened in 2024, described as a period of severe downturn, with over ten fastener-related companies exiting the market, signaling a clear decline in Taiwan’s fastener industry.

The challenges extend beyond shrinking order volumes. Rising costs driven by ESG compliance, energy-saving and carbon reduction measures(such as CBAM), AI integration, talent recruitment difficulties, wage increases, and freight capacity constraints have further squeezed profit margins, exacerbating the industry's difficulties.

This study employs multiple regression to forecast the seasonal market prices of medium-high carbon steel bars and rods. To select explanatory variables with greater explanatory power and lower Variance Inflation Factor(VIF) values, the stepwise regression analysis method was adopted to identify the optimal set of independent variables.

The results indicate that the variable with the highest weight on the medium-high carbon steel bars and rods price is the two-month lagged minimum closing stock price of China Steel Corporation (i.e., the minimum closing price two months prior to the effective month of the seasonal market price). Following this are:the four-month lagged minimum Baltic Dry Index (BDI) (i.e., the minimum BDI four months prior to the effective month of the seasonal market price),the two-month lagged minimum 62% CFR futures price(i.e., the minimum 62% CFR future price two months prior to the effective month of the seasonal market price),the two-month lagged maximum Baltic Dry Index (BDI) (i.e., the maximum BDI two months prior to the effective month of the seasonal market price).

These four variables are the primary positive correlates influencing the price of medium-high carbon steel bars and rods.

Using these four independent variables, a multiple regression model was constructed to calculate the seasonal market prices. Residual analysis comparing the model’s predicted seasonal market prices with the actual prices shows that the residuals are approximately symmetrically distributed and nearly normally distributed. This indicates that the model possesses a reasonable degree of reliability.

The results generated by this model can thus serve as a valuable reference for manufacturers in the fastener industry.

[1] 王貞媛 (2016)，彈性生產，以台灣扣件產業為例，國立清華大學科技管理研究所碩士論文，新竹，台灣。
[2] 呂銀坤 (2012)，台灣螺絲產業為因應中國的低價競爭的策略研究-相關同業間的策略聯盟，國立中山大學管理學院高階經營碩士學程在職專班碩士論文，高雄，台灣。
[3] 李倩瑜 (2014)，應用基因演算法結合時間序列於台灣地區鋼鐵價格漲跌幅之預測，國立臺北科技大學工業工程與管理系碩士班碩士論文，台北，台灣。
[4] 林暉凱 (2017)，扣件業盤元價格預測，國立成功大學工學院工程管理碩士在職專班碩士論文，台南，台灣。
[5] 洪源在 (2003)，台灣區扣件業者海外投資因素之研究，國立成功大學管理學院高階管理EMBA碩士論文，台南，台灣。
[6] 馮世達 (2007)，台灣螺絲螺帽產業競爭力分析，國立成功大學資源工程學系碩士論文，台南，台灣。
[7] 聚亨公司網站 (2025)，棒線(中高碳)生產流程，https://www.tycons.com.tw/index.php?option=module&lang=cht&task=pageinfo&id=32&index=2，最後瀏覽日：2025/4/24。
[8] 劉守芬 (2019)，布蘭特原油價格、波羅地海海岬型指數、鐵礦砂現貨價格與鋼鐵價格指數間之關聯性研究，國立高雄科技大學國際企業系碩士論文，高雄，台灣。
[9] Lee, C.Y., Chou, B.J., & Huang, C.F. (2022), Data science and reinforcement learning for price forecasting and raw material procurement in petrochemical industry, Advanced Engineering Informatics, Volume 51,2022,101443,ISSN 1474-0346, https://doi.org/10.1016/j.aei.2021.101443
[10] Lee, C.Y., & Liang, C.L. (2018), Manufacturer's printing forecast, reprinting decision, and contract design in the educational publishing industry, Computers & Industrial Engineering, Volume 125, 2018, Pages 678-687, ISSN 0360-8352, https://doi.org/10.1016/j.cie.2018.05.049
[11] Lee, C.Y., & Yang, S.H. (2023), Graph Spatio-Temporal networks for manufacturing sales forecast and prevention policies in pandemic era, Computers & Industrial Engineering, Volume 182,2023,109413,ISSN 0360-8352, https://doi.org/10.1016/j.cie.2023.109413